WIRELESS SENSING AND ACQUISITION SYSTEM FOR NON DESTRUCTIVE

EVALUATION AND STRUCTURAL HEALTH MONITORING

By

Deepak Kumar

A DISSERTATION

Michigan State University

in partial fulfillment of the requirements

Submitted to

for the degree of

Electrical Engineering – Doctor of Philosophy

2021

ABSTRACT

WIRELESS SENSING AND ACQUISITION SYSTEM FOR NON DESTRUCTIVE

EVALUATION AND STRUCTURAL HEALTH MONITORING

By

Deepak Kumar

Wireless sensing and communication techniques are superior and preferred over tethered counter-
parts. Wireless system has already replaced or replacing commonly used data transmission methods
for applications such as voice communication, internet, sensing, monitoring, etc. Recently, real-
time infrastructure or asset monitoring for safety critical application is gaining attention, due to its
ease of better visibility and predictive maintenance. A few of those applications belong to industries
such as aerospace, defense, oil and gas, nuclear, etc., which require top of the line materials and
an extra level of attention to safety. Every material or product used for such applications need to
go through extensive testing and evaluation to mitigate any risk of failure. The next generation
nondestructive evaluation system would provide sensing and characterization capabilities from raw
materials to the final product for manufacturing on factory grounds as well as monitoring products’
desired operation in their respective in-service conditions. However, the current evaluation and
internet-of-things (IoT)-based monitoring techniques have limited capabilities and cannot meet
the sensitivity, range, versatility, and scalability requirements for these applications while being
economical and maintenance-free. In this work, the current challenges associated with traditional
NDE and SHM methods have been described and an alternative approach is proposed using new
wireless sensing mechanisms.

First, a wireless near-field high-Q sensing resonator is described, which has a high sensitivity
to dielectric changes in composites and polymers. Second, a battery-free wireless communication
system with enhanced range is demonstrated that overcomes high power requirements and radio
frequency (RF) clutter issues. Third, passive (resistive and capacitive) sensors are integrated with
the battery-free RFID platform and the sensor information is transmitted via a hybrid data frame
with digital ID and analog sensor data. Fourth, the hybrid data frame is exploited further, an active

piezoelectric transducer pair is used to actuate an acoustic pulse and measure time-of-flight (ToF)
via the battery-free RFID tag, without using a power-hungry analog to digital conversion approach.
The RFID-based wireless sensing system is versatile, scalable, and economically superior to battery-
based IoT nodes and it can last for decades in field conditions without any routine maintenance that
truly opens the pathway for embedded sensing in next generation NDE systems.

This dissertation is dedicated to my beloved family. . .

iv

ACKNOWLEDGEMENTS

My last few years at MSU have been very enjoyable, exciting, enlightening, and enriching. I have
a lot to be thankful for and I would like to express my gratitude to all people whom I have worked
with and who have helped me learn and grow during this PhD.

First and foremost, I would like to express my deepest gratitude to my advisor, Dr. Yiming
Deng, who in many ways has molded my professional life. He has given his time and patience to
make me what I am today and for which I am forever indebted. Next, I am thankful to the rest of
my committee members Dr. Premjeet Chahal, Dr. Satish Udpa, Dr. Lalita Udpa, and Dr.Venkatesh
Kodur for their valuable suggestions and feedback through my PhD work. I am especially thankful
to Dr. Chahal for guiding me during my turbulent times and encouraging me both professionally as
well as personally. His inspirational ideas to tackle difficult engineering problems, support during
failure, and encouragement to retry was expectional.

I would like to thank the US Department of Transportation (DOT), Gas Technology Institute,
Chicago (GTI) and Consolidated Edison, New York for sponsoring my research and making all this
work possible, especially Arthur Buff, Ernest Lever and Vijay Srinivasan. I sincerely appreciate
the MSU’s College of Engineering and Department of Electrical and Computer Engineering for
multiple fellowships and travel grants. Special thanks to Mr. Brain Wright, Mr. Gregg Mulder,
Mr. Cameron Crump, and Mrs. Roxanne Peacock for all technical and logistical help.

Having spent most of my time in and around the lab, I’m very thankful to previous members
Dr. Chaofeng Ye, Dr. Anders Rosell, and Dr. Zhiyi Su for their valuable advices and welcoming
when I joined the lab. I was extremely lucky to find great friends who made this PhD a better
experince; Saran, Saikat, Suhas, Anmol, and Paul. I would like to mention other colleagues and
friends Xiaodong Shi, Mohand Alzuhiri, Dr. Oleksii Karpenko, Dr. Anton Efremov, Dr. Saptarshi
Mukherjee, Dr. Portia Banerjee, Dr. Vivek Rathod, Rajendra Prasath, Eric Tarkleson, Subrata
Mukherjee, Srijan Datta, Guillermo Huanes-Alvan, Fares Alharbi, Xuhui Huang, Zi Li, Bharath
Shenoy, Ciaron Hamilton and everyone else in NDEL who offered a great opportunity to learn from

v

the vastly different expertise and research directions.

Finally, I would like to thank my parents, Kuldeep and Kavita, and little sister, Bharti, for
constant sources of love, support, and encouragement throughout my life. Special shout-out to
Satya, Kamal, Kirnesh, and Karan who never fails to inspire and entertain me. I am also thankful
to Cara for all her continuous support during my PhD journey. I dedicate this thesis to them.

vi

TABLE OF CONTENTS

.
.

.
.

.
.

.
.

.
.

.
.

.
.

CHAPTER 1

LIST OF TABLES .
LIST OF FIGURES .

x
. xii

.
.
INTRODUCTION .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1 Motivation for NDE and SHM applications
. . . . . . . . . . . . . . . . . . . . .
1.2 Near-field Dielectric Inspection Methods for NDE . . . . . . . . . . . . . . . . . .
1.2.1 Capacitive Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3 Far-field Inspection Methods for SHM . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
Passive Resistive and Capacitive Sensor Tag . . . . . . . . . . .
1.3.1.1
1.3.1.2 Active Ultrasonic Sensor Tag . . . . . . . . . . . . . . . . . . .

1
3
5
6
8
8
8
9
. 10
1.4.1 Wireless High-Q resonators for NDE . . . . . . . . . . . . . . . . . . . . . 11
1.4.2 Battery-free Wireless Sensor Tags for SHM . . . . . . . . . . . . . . . . . 11
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1.4 Scope of wireless systems in NDE and SHM . . . . . . . . . . . . . . . . . . . .

1.3.1 Battery based Sensor Tags

1.5 Thesis Organization .

.

.

CHAPTER 2 WIRELESS NEAR-FIELD HIGH-Q SENSING RESONATOR FOR DI-

.

.
.

.
.

.
.

. .

2.2.1

Introduction .

2.3.1 Dielectric Reference Measurements

2.1
.
2.2 Theory and Design .

ELECTRIC EVALUATION OF COMPOSITES AND POLYMERS . . . . . 14
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Single Element Probe @ 5 MHz . . . . . . . . . . . . . . . . . . . . . . . 17
2.3 Measurement and Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . 19
. . . . . . . . . . . . . . . . . . . . . 19
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Polyethylene Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
. . . . . . . . . . . . . . . . . . . . . . . . . . . 29
. 32

.
2.4.1 CFRP Imaging .
2.4.2 GFRP Imaging .
2.4.3

2.5 Array Design and Measurement
2.6 Summary .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.4 Dielectric Imaging .

.

.

.

.

.

.

.

.

.

CHAPTER 3 WIRELESS FAR-FIELD HARMONIC RFID COMMUNICATION PLAT-

.
.

.
.

.
.

. .

. .

. .

.
. .

Introduction .

3.1
3.2 Design .

FORM FOR ENHANCED RANGE . . . . . . . . . . . . . . . . . . . . . . 33
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2.1 Harmonic RFID Interrogator . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2.2 Harmonic RFID Tag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.2.2.1 RFID IC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.2.2.2 RFID Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3 Results .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.4 RF Analysis of Harmonic Communication . . . . . . . . . . . . . . . . . . . . . . 47

. .

.

.

.

.

.

.

.

vii

3.5 Discussion, Challenges and Limitations
3.6 Summary .

. . . . . . . . . . . . . . . . . . . . . . . 51
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

.

.

.

.

.

.

.

CHAPTER 4 PASSIVE SENSOR INTEGRATION WITH ADC-FREE RFID TAG FOR

.

. .

.
.
.
.
. .

Introduction .

4.2.1 Design . .

4.1
.
4.2 Analog RF Sensing .
. .
3D Printed Pressure Sensor

.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.

ULTRA-LOW POWER OPERATION . . . . . . . . . . . . . . . . . . . . . 53
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
. 57
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.2.1.1
. . . . . . . . . . . . . . . . . . . . 58
4.2.1.2 Harmonic RF Sensor Tag . . . . . . . . . . . . . . . . . . . . . 61
4.2.1.3 Harmonic RF Interrogator . . . . . . . . . . . . . . . . . . . . . 63
4.2.2 Results
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.2.3 Discussion and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . 68
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.3.1.1 Hybrid Single-Frequency RFID Tag Architecture . . . . . . . . . 71
4.3.1.2 Hybrid Harmonic-Frequency RFID Tag Architecture . . . . . . . 74
4.3.1.3 Hybrid Multi-Frequency Interrogator
. . . . . . . . . . . . . . . 78
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.3.2.1 Wireless Demonstration . . . . . . . . . . . . . . . . . . . . . . 80
. 89

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4.3 Hybrid RFID Sensing .

4.4 Summary .

4.3.1 Design . .

4.3.2 Results

.
. .

. .

. .

. .

. .

. .

. .

. .

. .

.
.

.

.

.

.

.

.

.
.

.
.

.
.

. .

. .

. .

.
. .

CHAPTER 5 ACTIVE SENSOR INTEGRATION WITH ADC-FREE RFID TAG FOR

Introduction .

5.3 Results .

5.1
5.2 Design .

5.2.1 Active Sensing Architecture
5.2.2
5.2.3

ULTRA-LOW POWER OPERATION . . . . . . . . . . . . . . . . . . . . . 90
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
. . . . . . . . . . . . . . . . . . . . . . . . . 92
Integrating Acoustic Actuation Transducer . . . . . . . . . . . . . . . . . . 92
Integrating Acoustic Detection Transducer . . . . . . . . . . . . . . . . . . 94
5.2.3.1 Acoustic continuous wave detection . . . . . . . . . . . . . . . . 95
5.2.3.2 Acoustic pulse detection . . . . . . . . . . . . . . . . . . . . . . 95
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
5.3.1 Wireless Acoustic Actuation and Detection Demonstration . . . . . . . . . 96
5.3.1.1 Detection of Acoustic Time-of-Flight in Steel . . . . . . . . . . . 96
. 97
CHAPTER 6 CONCLUSION AND FUTURE WORK . . . . . . . . . . . . . . . . . . . 99
.
. 99
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Imaging using an array . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
6.2.1
6.2.2 Miniaturization .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
6.2.3 Ultrasonic Material Characterization . . . . . . . . . . . . . . . . . . . . . 101
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6.1 Summary .
.
6.2 Future Work . .

APPENDICES .

5.4 Summary .

.
.
. .

. .

. .

. .

. .

.
.

.
.

.
.

.

.

.

.

.

.

.

.

.

.

.

.

.

viii

APPENDIX A

APPENDIX B

BIBLIOGRAPHY .

.

.

AN RF BACKSCATTERER SYSTEM FOR REAL-TIME MULTI-
SENSING APPLICATIONS . . . . . . . . . . . . . . . . . . . . . 103
IDENTIFICATION AND SENSING SYSTEM (ISS): AN OPEN
SOURCE RFID PLATFORM . . . . . . . . . . . . . . . . . . . . . 120
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
.

.

.

.

.

.

ix

LIST OF TABLES

Table 2.1: Comparison of commonly used NDE Methods for composites and polymers . . . 16

Table 2.2: Dimensions of the fabricated sensor probe . . . . . . . . . . . . . . . . . . . . . 18

Table 2.3: Dimensions of the polyamide sample

. . . . . . . . . . . . . . . . . . . . . . . 23

Table 2.4: Dimensions of the aluminum sample . . . . . . . . . . . . . . . . . . . . . . . . 25

Table 2.5: Dimensions of the CFRP sample . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Table 2.6: Dimensions of the GFRP sample . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Table 2.7: Dimensions of the HDPE sample . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Table 2.8: Dimensions of the array probe . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

Table 3.1: Dimensions of the Hairpin bandpass filter . . . . . . . . . . . . . . . . . . . . . 39

Table 3.2: Dimensions of the five stage bandpass filter

. . . . . . . . . . . . . . . . . . . . 41

Table 3.3: Dimensions of the hRFID antenna . . . . . . . . . . . . . . . . . . . . . . . . . 43

Table 4.1: Commercially available RFID ICs with sensitivity data . . . . . . . . . . . . . . 56

Table 4.2: Dimensions of the 3D-printed pressure sensor . . . . . . . . . . . . . . . . . . . 60

Table 4.3: Dimensions of the analog pressure sensor tag . . . . . . . . . . . . . . . . . . . 62

Table 4.4: Data frame format for Digital ID with Analog Sensor Data Reply . . . . . . . . . 70

Table 4.5: Data frame format for Analog Sensor Data Reply . . . . . . . . . . . . . . . . . 70

Table 4.6: Dimensions of the designed 915 MHz antenna

. . . . . . . . . . . . . . . . . . 74

Table 4.7: Dimensions of the designed 434 MHz antenna

. . . . . . . . . . . . . . . . . . 77

Table 4.8: Dimensions of the designed 868 MHz antenna

. . . . . . . . . . . . . . . . . . 78

Table 5.1: Data frame format for Digital ID with Acoustic Transducer Data Reply . . . . . . 92

Table A.1: Dimensions of the 915 MHz antenna . . . . . . . . . . . . . . . . . . . . . . . . 106

x

Table B.1: Data frame format for Digital ID Reply . . . . . . . . . . . . . . . . . . . . . . 122

Table B.2: Data frame format for Digital ID with Sensor Data Reply . . . . . . . . . . . . . 122

Table B.3: 4-bit codes for different sensing parameters . . . . . . . . . . . . . . . . . . . . 122

Table B.4: 8-bit codes for different sensor types . . . . . . . . . . . . . . . . . . . . . . .

. 123

xi

LIST OF FIGURES

Figure 1.1:

Illustration of application for NDE and SHM at different levels

. . . . . . . . .

Figure 1.2: Example illustration of NDE and SHM for same product with different stakeholders

Figure 1.3: Schematic of conventional capacitive imaging; (A) Using impedance ana-

lyzer, (B) Differential bridge circuit, (C) LC resonator . . . . . . . . . . . . . .

Figure 1.4: Schematic of conventional capacitive probe interaction with dielectric sample

under test with an example imaging output plot . . . . . . . . . . . . . . . . . .

Figure 1.5: Battery-based passive sensor nodes acquiring data from digital or analog

resistive or capacitive sensor

. . . . . . . . . . . . . . . . . . . . . . . . . . .

2

3

7

7

9

Figure 1.6: Battery-based active sensor node with acoustic signal actuating and detecting

capabilities for monitoring mechanical properties

. . . . . . . . . . . . . . . . 10

Figure 2.1: Schematic of the LC tank along with the fabricated sensor probe

. . . . . . .

. 18

Figure 2.2: Measurement setup for the single probe . . . . . . . . . . . . . . . . . . . . .

. 18

Figure 2.3: Measured S11 (magnitude and phase) of the single LC sensor probe . . . . . .

. 19

Figure 2.4: Frequency response of the standard dielectric materials (Teflon = 2.1, Plexi-

glass = 3.1) with known relative permittivity . . . . . . . . . . . . . . . . . . . 20

Figure 2.5: Frequency response of the composites and polymers with different dielectric

constants .

.

.

.

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Figure 2.6: Frequency response of the conductive materials with different conductivity . . . 21

Figure 2.7: Frequency response (magnitude and phase) of GFRP with defect and no defect . 22

Figure 2.8: Change in resonance frequency for increasing depth of sub-surface defects in

polyamide target .

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Figure 2.9: Schematic of the dielectric imaging setup . . . . . . . . . . . . . . . . . . . .

. 24

Figure 2.10: Measurement setup for dielectric imaging with the single probe mounted on

a X-Y raster scanner .

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

xii

Figure 2.11: (A) Image of the aluminum target with dimensions, (B) Imaging results for

profiling of defects using single probe

. . . . . . . . . . . . . . . . . . . . . . 25

Figure 2.12: (A) Image of the CFRP target, (B) Imaging results for profiling of defects

using single probe . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Figure 2.13: (A) Image of the GFRP target, (B) Imaging results for profiling of defects

using single probe . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Figure 2.14: (A) Image of the HDPE target, (B) Imaging results for profiling subsurface

defects in HDPE target

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Figure 2.15: Schematic of the designed array probe along with the dimensions . . . . . . . . 30

Figure 2.16: Measured frequency response of the 1 x 4 array probe under no loading conditions 30

Figure 2.17: Measured results for GFRP with surface defects for (A) All four elements of

the array, (B) Focused frequency Shift in fourth element of the array . . . . . . 31

Figure 3.1: Range limiting phase noise issue with conventional RFID system . . . . . . .

. 34

Figure 3.2: Hypothesis for harmonic communication towards range enhancement . . . . . . 35

Figure 3.3: The hRFID communication system . . . . . . . . . . . . . . . . . . . . . . .

. 35

Figure 3.4: The hRFID communication platform using conventional UHF system . . . . .

. 38

Figure 3.5: Hairpin bandpass filter, F1 (2.6-3.4 GHz): (A) Schematic with dimensions,

and (B) Simulated and measured S-parameters . . . . . . . . . . . . . . . . . . 39

Figure 3.6: Five stage bandpass filter, F2 (1.7-1.9 GHz): (A) Schematic with dimensions,

and (B) Simulated and measured S-parameters . . . . . . . . . . . . . . . . . . 40

Figure 3.7:

(A) Schematic of the designed hRFID antenna along with its dimensions,
and (B) Measured frequency response of RFID IC and hRFID antenna with
conjugate matching points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Figure 3.8: Measured backscattered signal visualized using a spectrum analyzer, with a
transmitted power of +8 dBm and a read range of 0.35 m: Received signal
at the fundamental and the third harmonic frequency with and without the
RFID tag in the interrogator’s field of view . . . . . . . . . . . . . . . . . . . . 44

xiii

Figure 3.9: Measured backscattered signal visualized using a spectrum analyzer, with a
transmitted power of +8 dBm and a read range of 0.35 m: Reference signal at
the second harmonic frequency generated using a frequency doubler and the
down converted signal at the fundamental frequency after mixing . . . . . . . . 46

Figure 3.10: Received third harmonic signal from the designed hRFID tag. . . . . . . . . . . 47

Figure 3.11: Received backscattered power at the interrogator, with transmitted EIRP of
36 dBm, from the designed hRFID tag for both the fundamental frequency
and the harmonic frequency operation.

. . . . . . . . . . . . . . . . . . . . . . 48

Figure 3.12: Power budget analysis of backscattering based communication system: (A)
Detection range using fundamental and harmonic bands with no tag losses,
and (B) Detection range using fundamental and harmonic bands with 5 and
10 dB losses, respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Figure 3.13: Effects of the IC sensitivity on the RFID read range with fundamental and

harmonic communication. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Figure 4.1: Example battery-free wireless sensing system architecture for the pipeline

infrastructure .

. .

. .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

Figure 4.2: Air cavity-based fully 3D printed pressure sensor

. . . . . . . . . . . . . . . . 58

Figure 4.3: Exploded view of 3D printed pressure sensor with three main components . . . 59

Figure 4.4: Fully assembled 3D printed pressure sensor with component annotation and

dimensions .

.

.

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Figure 4.5:

(A) Pressurizing chamber (pipe) with integrated pressure sensor, (B) Schematic
of the measurement setup, and (C) Fabricated pressure sensor . . . . . . . . . . 60

Figure 4.6: Measured impedance at 2 GHz for pressure range from 0 to 20 psi with a

direct wired connection using vector network analyzer (VNA) . . . . . . . . . . 61

Figure 4.7: Schematic of the pressure sensing harmonic RF tag along with its dimensions.
Signal received using 2 GHz antenna forwarded to the hybrid coupler for phase
shift according to the integrated pressure sensor and the second harmonic
signal with double shifted phase, generated using diode frequency doubler, is
transmitted out using 4 GHz patch antenna . . . . . . . . . . . . . . . . . . . . 62

Figure 4.8: Simulated and measured frequency response; (A) Fundamental (2 GHz) and

(B) Harmonic antenna (4 GHz) . . . . . . . . . . . . . . . . . . . . . . . . . . 63

xiv

Figure 4.9: Schematic of the harmonic RF Interrogator consist of a VNA for RF source
and phase measurement unit, transmitting and receiving antennas, and mis-
cellaneous RF peripheral circuits for communicating with the harmonic RF
sensor tag .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

.

.

.

.

.

Figure 4.10: Simulated frequency response of the designed phase shifting hybrid coupler
with a variable reactance from 6.82  to 36.73 : (A) No power loss is
observed at 2 GHz due to changing reactance, (B) A total phase shift of 28.55
◦ is observed at 2 GHz due to changing reactance

. . . . . . . . . . . . . . . . 64

Figure 4.11: Measured phase change due to variable reactance introduced with direct
current (DC) biasing the varactor diode in a wired configuration (Ɗ =
18.28
) using a single frequency approach and wireless configuration (Ɗ =
36.78
) using the harmonic frequency approach . . . . . . . . . . . . . . . . . 66

Figure 4.12: Pressurized polypropylene pipeline setup with the fabricated harmonic RF

sensor tag for real-time wireless communication . . . . . . . . . . . . . . . . . 67

Figure 4.13: Phase change is wirelessly measured using harmonic RF interrogator due to
applied pressure ranging from 0 to 20 psi. The linearity of change in phase
due to pressure is represented using coefficient of determination (R2 = 0.9915) . 68

Figure 4.14: Schematic of the hybrid single-frequency RFID tag . . . . . . . . . . . . . . . 71

Figure 4.15: (A) Schematic of a lumped hybrid coupler with integrated reactance sensors

for modulation, (B) Simulated and measured response of the coupler at 915 MHz 71

Figure 4.16: Measured change in phase due to shift in capacitance of the varactor diode

using VNA, at 915 MHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

Figure 4.17: (A) Measurement setup for testing backreflection at 915 MHz, (B) Measured

backreflected signal with digital ID and analog sensor data.

. . . . . . . . . . . 73

Figure 4.18: Meandered Dipole antenna for 915 MHz . . . . . . . . . . . . . . . . . . . .

. 73

Figure 4.19: (A) Simulated and Measured frequency response of the 915 MHz antenna,

(B) Simulated radiation pattern of the antenna . . . . . . . . . . . . . . . . . . 74

Figure 4.20: Schematic of the hybrid harmonic-frequency RFID tag . . . . . . . . . . . . . 74

Figure 4.21: (A) Schematic of a lumped hybrid coupler with integrated reactance sensors

for modulation, (B) Simulated and measured response of the coupler at 434 MHz 75

Figure 4.22: Measured change in phase due to shift in capacitance of the varactor diode

using VNA, at 434 MHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

xv

Figure 4.23: (A) Measurement setup for testing backreflection at 868 MHz, (B) Measured

backreflected signal with digital ID and analog sensor data.

. . . . . . . . . . . 76

Figure 4.24: Meandered Dipole antenna for 434 MHz . . . . . . . . . . . . . . . . . . . .

. 77

Figure 4.25: (A) Simulated and Measured frequency response of the 434 MHz antenna,

(B) Simulated radiation pattern of the antenna . . . . . . . . . . . . . . . . . . 77

Figure 4.26: Meandered Dipole antenna for 868 MHz . . . . . . . . . . . . . . . . . . . .

. 78

Figure 4.27: (A) Simulated and Measured frequency response of the 868 MHz antenna,

(B) Simulated radiation pattern of the antenna . . . . . . . . . . . . . . . . . . 78

Figure 4.28: Block diagram of the hybrid multi-frequency interrogator

. . . . . . . . . . .

. 79

Figure 4.29: Wireless setup for hybrid single-frequency RFID communication . . . . . . . . 80

Figure 4.30: Demodulated hybrid data received at 915 MHz wirelessly from designed

RFID tag .

.

.

.

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Figure 4.31: Snapshot of the hybrid single-frequency communication with digital information 81

Figure 4.32: Analog sensor data recorded using oscilloscope from 915 MHz hybrid RFID tag 82

Figure 4.33: Analog sensor data recorded using oscilloscope from 915 MHz hybrid RFID

tag; O.0V input (Red and 1.0V input (Blue)) . . . . . . . . . . . . . . . . . . . 83

Figure 4.34: Analog sensor data recorded using oscilloscope from 915 MHz hybrid RFID

tag; (A) 0.2V, (B) 0.4V, (C) 0.6V, (D) 0.8V . . . . . . . . . . . . . . . . . . . . 83

Figure 4.35: Difference in reference (RAD) and sensor (SAD) data for different sensor

voltage across Varactor

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

Figure 4.36: Wireless setup for hybrid harmonic-frequency RFID communication . . . . .

. 85

Figure 4.37: Demodulated hybrid data received at 868 MHz wirelessly from designed

RFID tag .

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

.

.

.

.

.

.

.

.

.

.

Figure 4.38: Snapshot of the hybrid harmonic-frequency communication with digital in-

formation .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

Figure 4.39: Analog sensor data at harmonic signal recorded using oscilloscope . . . . . . . 86

Figure 4.40: Analog sensor data at harmonic signal recorded using oscilloscope; 0.0V

input (Red and 1.0V input (Blue)) . . . . . . . . . . . . . . . . . . . . . . . . . 87

xvi

Figure 4.41: Analog sensor data recorded at 868 MHz using oscilloscope; (A) 0.2V, (B)

0.4V, (C) 0.6V, (D) 0.8V . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

Figure 4.42: Difference in RAD and SAD of 868 MHz signal for different sensor voltage

across varactor . .

. .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Figure 5.1: Schematic of active sensor integrated hybrid harmonic RFID tag . . . . . . . . 93

Figure 5.2:

(A) Acoustic actuation and detection wired setup, (B) PZT actuation pulse
signal generated by microcontroller and detected signal profile with averaging . 93

Figure 5.3: Schematic of the PZT integration with non-linear transmission line . . . . . . . 94

Figure 5.4:

(A) Test setup to check the analog modulation capabilities of RFID tag, (B)
Harmonic output of the NLTL is modulated with 200 kHz acoustic signal
detected by PZT .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

.

.

.

Figure 5.5:

(A) Test setup to check the acoustic pulse modulation capabilities of RFID
tag, (B) Demodulated acoustic signal detected by PZT . . . . . . . . . . . . . . 96

Figure 5.6:

(A) Wireless hybrid harmonic RFID tag with integrated PZT setup for acoustic
actuation and detection, (B) Picture of the practical setup . . . . . . . . . . . . 97

Figure 5.7: Wireless acquired and demodulated harmonic signal for acoustic signal . . . . . 98

Figure A.1: Schematic of a passive wireless system with multi-sensing RF tag and Interrogator105

Figure A.2: UHF RFID Antenna: (A) Schematic of the dipole antenna and dimensions,

(B) Simulated and measured frequency response . . . . . . . . . . . . . . . . . 106

Figure A.3: (A) Schematic of RF-to-DC converter, (B) Input RF power vs Output DC power 107

Figure A.4: (A) Input voltage vs Input current at 100 kHz oscillation frequency, (B)

Oscillation frequency in kHz vs required Input power in microwatts . . . . . . . 109

Figure A.5: Circuit diagram of the designed multi-sensing module . . . . . . . . . . . . . . 110

Figure A.6: Sensing information modulation: (A) Schematic of the modulation scheme
and (B) Modulated sensor information of 100 kHz and 150 kHz on 908 MHz
carrier frequency .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

.

.

.

Figure A.7: Schematic of the RF interrogator radiating +36 dBm EIRP and visualizing the

back reflected sensor information using a spectrum analyzer and an oscilloscope 112

xvii

Figure A.8: Wired validation: Received side band power of a carrier signal due to first

and second sensor with varying input power

. . . . . . . . . . . . . . . . . . . 114

Figure A.9: Frequency change of temperature and luminescence sensors (A) with linear

fit and (B) with second order polynomial fit . . . . . . . . . . . . . . . . . . . . 115

Figure A.10:The prototype of RF Tag with designed antenna, integrated sensors, and

assembled PCB .

. .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

Figure A.11:Wireless demonstration with 9 dBm transmitted power at 908 MHz: Received

RF power at multisensor tag vs distance and backscattered . . . . . . . . . . . . 117

Figure A.12:Wireless demonstration with 9 dBm transmitted power at 908 MHz: (A)
Frequency spectrum of the demodulated sensor data, and (B) Received sensor
information in time domain with a switching frequency of 6 kHz (∼ 160 
s)

. . 118

Figure B.1: Schematic of the digital single frequency tag for receiving and transmitting

at 915 MHz .

.

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

Figure B.2: (A) Schematic of the designed energy harvester, (B) Measured frequency

response of the energy harvester with input power sweep . . . . . . . . . . . . . 123

Figure B.3: (A) Measurement setup for ASK modulation at 915 MHz, (B) Measured back

reflected RF signal with modulated ID . . . . . . . . . . . . . . . . . . . . . . 124

Figure B.4: Schematic of the digital harmonic tag for receiving at 434MHz and transmit-

ting at 868MHz .

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

Figure B.5: Schematic of the 434-868 MHz digital RFID tag . . . . . . . . . . . . . . . .

. 125

Figure B.6: (A) Measurement setup for ASK modulation at 434 MHz, (B) Measured back

reflected RF signal with modulated ID . . . . . . . . . . . . . . . . . . . . . . 126

Figure B.7: Block diagram of the ISS interrogator . . . . . . . . . . . . . . . . . . . . . .

. 127

Figure B.8: (A) Modulated “Query” command on 915 MHz signal, (B) Demodulated

received reply “RN16” from 915 MHz . . . . . . . . . . . . . . . . . . . . . . 128

Figure B.9: Digital RFID data received at 915 MHz with +7 dBm transmitted power; (A)

with 5 inch interrogation distance, (B) with 10 inch interrogation distance

. . . 129

Figure B.10: Snapshots of tag and interrogator communication in digital single-frequency
mode; (A) Decoded output with 0.0V at tag, (B) Decoded output with 0.2V
at tag, (C) Decoded output with 0.4V at tag, (D) Decoded output with 0.6V at
tag, (E) Decoded output with 0.8V at tag, (F) Decoded output with 1.0V at tag . 130

xviii

Figure B.11: Received digital sensor data at 915 MHz . . . . . . . . . . . . . . . . . . . .

. 131

Figure B.12: Digital RFID data received at 868 MHz with +7 dBm transmitted power with

10 inch interrogation distance . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

Figure B.13: Snapshots of tag and interrogator communication in digital harmonic-frequency
mode; (A) Decoded output with 0.0V at tag, (B) Decoded output with 0.2V
at tag, (C) Decoded output with 0.4V at tag, (D) Decoded output with 0.6V at
tag, (E) Decoded output with 0.8V at tag, (F) Decoded output with 1.0V at tag . 132

Figure B.14: Received digital sensor data at 868 MHz . . . . . . . . . . . . . . . . . . . .

. 133

xix

CHAPTER 1

INTRODUCTION

In the age of artificial intelligence (AI), data is key and the size reflects its worth. The number
of sensors and sensing resolution always go hand in hand, therefore, a futuristic world with AI
will require integrating more sensors than ever. High resolution sensing is the path forward for
mitigating even the minimal risk of any failure. The collected data from installed sensors not only
reflect safety parameters but also help predicting necessary maintenance to maximize the operating
life of the infrastructure. These upraised requirements has pushed the inspection industry to bring
forth the best sensing and data transmission technologies at a faster pace.

The scope of infrastructure safety is very wide and commonly divided into two main categories;
Non Destructive Evaluation (NDE) and Structural Health Monitoring (SHM). The difference is
NDE and SHM is very narrow but precise, because both mechanisms are put in place for ensuring
and increasing safety. NDE works at a more fundamental level with material properties, whereas,
the SHM is more concerned about the reliable operation of the designed product for a specific
application. Fig.1.1 illustrates the two sides of infrastructure that require either NDE or SHM
safety mechanism.

Fundamentally, bulk materials like metals, composites, and polymers are used in various
different shapes or forms to create any product. More than 90% of artificially made common use
objects belong to one of the material class mentioned here. Therefore, selecting a bulk material is
the first step in every product design. Bulk materials come with different properties like conductive
or insulating, malleable or ductile, etc.
that needs to be quantified for classifying materials
into different grades for easier selection and use. The material property quantification methods
are usually part of NDE techniques, which can test the material for its electrical or mechanical
properties. For example, metals can be classified according to their respective conductivity,
stiffness, density, etc., and polymers can have different dielectric permittivity, porosity, etc. These
electrical and mechanical properties are interdependent that indirectly gives a measure of each

1

Figure 1.1: Illustration of application for NDE and SHM at different levels

other [1]. Researchers have developed numerous techniques to evaluate material properties at both
micro level and bulk level [2–8]. Usually the resolution is inversely proportional to technique’s
complexity, therefore, a resolution trade-off can be made according to the application, after-all
complexities arise from necessities.

The next step after material selection is product development and operation. The newly devel-
oped product hangs on both sides of the evaluation and monitoring methods. To test the quality of
the newly developed product, a second evaluation test of material properties can be a good measure
to ensure safety and reliability as well as a short operation or dry run under monitoring scenario is
also an acceptable method. For example, a medium density polyethylene pipe is manufactured in a
plant, a quality assured polyethylene with known material properties would be molded into a new
pipe shape, the molding process can alter the properties of the material due to applied heat and cur-
ing pressure. Measuring the density or dielectric permittivity again after the mold is a good quality
assurance step but the pipe can also be checked with a quick pressure test with monitoring sensors.
If the pipe quality comes out well, it can used in a refinery or chemical plant to transport oil and
gas, where its operating condition also needs to be monitored continuously for avoiding any failure.

2

Therefore, at the operation stage, SHM kicks-in with its capability of measuring environmental/
global properties; like pressure, temperature, humidity that contributes to its optimal operation.

Fig.1.2, shows the polyethylene pipe as an example, where the manufacturer is interested in
parameters like permittivity and density while user at a refinery plant is interested in pressure and
temperature.

Figure 1.2: Example illustration of NDE and SHM for same product with different stakeholders

1.1 Motivation for NDE and SHM applications

An efficient evaluation and monitoring system requires a comprehensive map of the structural
integrity of the target object. Real-time monitoring techniques are best suited for providing
instantaneous mapping of the structural stability and the nature of the targeted objects for enhancing
predictive maintenance strategies. In general, a target object such as a pipeline is constructed using
a combination of fundamental building blocks such as metals, composites, and polymers, whose
properties define its structural integrity and mechanical stability.

Structural health monitoring (SHM) of massive infrastructures is necessary due to the rapidly
growing urban infrastructure as it has huge impact on human safety and well-being [9, 10]. Struc-

3

turally deficient infrastructures over time increase the possibility of catastrophic failures leading to
loss of human life and economy. There has been several thousand deaths reported around the globe
due to infrastructure collapse or failure in the last decade [11–14]. For example, in the United
States alone, PHMSA and NTSB reported numerous pipeline failures in last decade leading to
billions of dollars in monetary loss [15–17]. Although the exponential growth in the number of
massive structures has mandated regulatory authorities to enforce rigorous maintenance protocols
to mitigate risks for improved safety, the failure incidents are not uncommon. Hence, there is al-
ways a growing need to develop NDE and SHM technologies for efficient condition monitoring and
evaluation techniques for predictive maintenance. Implementing efficient predictive maintenance
strategies enables saving lives and prevents huge economical losses. Moreover, SHM is not limited
to just civil infrastructure but can be extended for providing predictive maintenance strategies in
multiple domains such as transportation, aerospace, environment, etc.

The common properties that are evaluated for structural integrity are stress, strain, permittivity
(
), permeability (
), and conductivity [18]. These intrinsic properties depend directly on the
combination of materials used in manufacturing the object. Any variation in these properties from
the standard values will enhance the visibility of the structural integrity of the object. For example,
the periodic evaluation of a material’s stress and strain is required to detect the occurrence of cracks
or other anomalies in walls of a bridge or a pipeline, allowing precautionary safety measures to be
enforced in real-time. Apart from the intrinsic properties, extrinsic properties such as humidity,
temperature, vibration, moisture, and pressure are also commonly monitored. These properties are
important to predict the damage caused due to external environmental factors on the structure. For
example, excessive pressure on a pipeline may compromise the structural integrity over time. The
combination of the intrinsic and extrinsic properties provides a comprehensive map of the structural
integrity of the target object.

In order to promote best predictive maintenance strategies, multiple properties should be mon-
itored in real-time for providing instantaneous feedback on the structural integrity of the target
object [19–21]. A number of NDE and SHM techniques are available in the literature for monitor-

4

ing both the intrinsic and extrinsic properties of the target [22–27].

In general, for electrical probing of the properties of the targeted object, different monitoring and
evaluation methods track the three fundamental electrical parameters, resistance (R), capacitance
(C), or inductance (L) to estimate the intrinsic or extrinsic properties. For example, by employing
microwave based method, the defect of the target material is mapped by tracking the change in
impedance across the microwave sensor and indirectly estimating the effective dielectric constant of
the target [28–30]. Overall, the monitoring and evaluation of different materials and their intrinsic
as well as extrinsic properties fall under major sensing methods that directly or indirectly depends
on the change in the three fundamental electrical parameters, Resistance (R), inductance (L), or
capacitance (C).

Dielectric imaging is an important technique for evaluating composites and polymers. Elec-
tromagnetic methods are well suited for this task as RF waves behave according to the different
dielectric properties of the medium. Far-field and near-field imaging techniques have been widely
studied for dielectric characterization [31–34]. The far-field imaging resolution is related to the
wavelength, therefore, a 300 GHz system is needed for sub-mm resolution. Due to high frequency
requirements, the system becomes very complex and the limited availability of RF parts makes it
difficult to design such system. The alternative is near-field imaging using fringing fields that can
achieve sub-mm resolution with a raster scanning mechanism.

The conventional methods for both near-field NDE inspection and far-field SHM inspection are

discussed below.

1.2 Near-field Dielectric Inspection Methods for NDE

The capacitive and microwave imaging techniques are commonly known to evaluate the di-
electric properties of composites and polymers. Both techniques operate in near-field region that
detects a perturbation in the electric field due to permittivity change. Capacitive and microwave
imaging methods can evaluate dielectric materials like CFRP and GFRP for common problems like
air bubbles, contamination, delamination, and epoxy saturation [31,35].

5

The microwave system operates at a respectively higher frequency compared to capacitive and
it uses radiative probes for detection. Due to the radiative nature of microwave probes, making the
multidimensional array for simultaneous data acquisition is very difficult. The isolation between
array elements is poor, which reduces the sensitivity of the imaging system. The near-field
microwave imaging in S or C-band comes across the issue of large aperture window of the probe,
which corresponds to a lower spatial resolution due to convolution over a larger volume. The
higher frequency operation in X or K-band has complex design and limited component availability
issues. The alternative capacitive imaging is better in terms of isolation and circuit design ease.
The procedure of conventional capacitive imaging is explained below.

1.2.1 Capacitive Imaging

The capacitive imaging conventionally operate in shortwave frequency range (3–30 MHz) that use
a parallel plate capacitive probe for generating fringing fields in its near-by region. The fringing
field die off very quickly with the distance that helps with making array probes without isolation
issue. The reading mechanism is the primary factor for determining the sensitivity of the probe.
In literature, researchers have used direct probing using impedance analyzers for detecting change
in capacitance due to the dielectric medium [36]. The direct capacitance probing is too slow for
imaging applications. Other capacitive imaging system includes differential reading using bridge
circuit, which output change in voltage only due to the change in capacitance. The differential
method is fast for imaging but not sensitive due to impedance mismatches, which reflect back most
of the input power.

For matching with a capacitive imaging probe, researchers have used series or parallel inductor
matching to create an LC resonator. The resonance-based methods are superior over other described
methods. However, the Q-factor for wired LC resonator is not very high [37]. In wired configuration
the loaded quality factor (
 
) is measured at resonance frequency ( 

) as shown in Equation.1.1.


 
 =








6

(1.1)

where 

 is the bandwidth either at −3 dB or −10 dB.
Fig.1.3, shows the conventional capacitive imaging circuits for dielectric evaluations.

Figure 1.3: Schematic of conventional capacitive imaging; (A) Using impedance analyzer, (B)
Differential bridge circuit, (C) LC resonator

The capacitive probe radiates fringing-field that interact with the dielectric sample under test
and perturb according to the effective relative permittivity. The measurement setup is shown in
Fig.1.4.

Figure 1.4: Schematic of conventional capacitive probe interaction with dielectric sample under
test with an example imaging output plot

The conventional methods do not exhibit high-Q or high sensitivity, therefore, there is a scope
of improvement for capacitive imaging in the directions of achieving high-Q, sub-mm resolution,
and simultaneous actuation of array elements and detection.

7

1.3 Far-field Inspection Methods for SHM

Structurally deficient infrastructure increases the probability of catastrophic failure over time,
leading to a significant impact on human safety. The demand for developing efficient structural
health monitoring (SHM) techniques for promoting effective predictive maintenance strategies is
necessary to prevent deaths and substantial economical losses. The evaluation should be nonde-
structive in nature and should provide instantaneous and accurate feedback on the nature of the
structure.

1.3.1 Battery based Sensor Tags

In literature, a number of battery-based SHM platforms are proposed for monitoring and evaluating
different structural properties using a plethora of sensors [38,39]. Common battery-based platforms
that are used for wireless SHM have an intercommunication link between sensor nodes for real-time
data monitoring. Few example platforms are Zigbee, LoraWAN, SigFox, BLE, etc. These platforms
are designed for a long-range reliable communication of data from the sensor nodes. Although
battery-based SHM platforms are fully evolved, a few challenges still exist in realizing an economical
system, such as requirement of an on-board battery at each node, limited lifetime, and requires
periodical maintenance. The best alternative solution to the current active sensing platforms
is to employ infinite lifetime battery-less passive sensing platforms which limits maintenance
requirements and reduces the overall cost.

1.3.1.1 Passive Resistive and Capacitive Sensor Tag

The passive (resistance and capacitance) sensors are commonly used for all IoT based infrastruc-
tures. The sensor nodes consist of a microcontroller, a wireless communication module, sensing
circuit, and a battery to power everything. The microcontroller runs a logic in-loop, which fully
wake up the microcontoller and acquires data from the sensing circuitry either using digital bus
like SPI, I2C, UART, etc. or using ADC to convert the voltage potential into a digital number. The

8

acquired data is stored in local microcontroller’s memory that eventually transmitted out to either
the cloud or control station using the wireless network. The architecture shown in Fig.1.5.

Figure 1.5: Battery-based passive sensor nodes acquiring data from digital or analog resistive or
capacitive sensor

The commonly used products in our daily life, for example, smartphones, smartwatches, in-
telligent lighting, garage opener, pet trackers, thermometers, etc. follows the same architecture as
described. Some of these devices last for months before requirement of battery replacement or
charging and some need a daily boost, which depends on the amount of activity and use case.

1.3.1.2 Active Ultrasonic Sensor Tag

The active sensors like ultrasonic transducers require an actuation signal for generating an acoustic
wave, which travel through a medium under test and change its properties like traveling speed,
dispersion, etc. according to the material properties. The traveled signal can be picked up at a
distance which holds the time-of-flight information that can be used to back calculate the velocity,
young’s modulus, and other mechanical properties [40].

The architecture of such a node includes a microcontroller, wireless module, ultrasonic pulser-
receiver module, and a battery as shown in Fig.1.6. The small changes in time-of-flight are picked
up by the pulser-receiver module that corresponds to changing material behaviour with traveling
acoustic waves. Active ultrasonic sensor tags are commonly used in automotive applications for
driver assistance systems.

9

Figure 1.6: Battery-based active sensor node with acoustic signal actuating and detecting capabili-
ties for monitoring mechanical properties

1.4 Scope of wireless systems in NDE and SHM

A number of sensors are necessary to monitor different parameters of interest. These sensors
should be efficient and should be able to provide the necessary information in real-time. There are
two methods for probing these sensors, wired and wireless. The conventional wired techniques
involves employing physical wires or probes to connect and obtain data from different sensor
nodes. A major limitation of wired approach is that it is tedious to run wires on the external
surface of the target object from the cluster of sensor nodes making this approach expensive as
well as impractical for monitoring large infrastructures such as pipelines, bridges, aircrafts, etc.
For example, multiple sensor nodes are connected to the pipeline which detects different electrical
parameters.

To overcome the impracticality of wired systems to probe large infrastructures wireless
probing of the sensors is an excellent choice to make the system reliable. The sensor nodes should
be able to communicate the sensing information wirelessly to the interrogator or among themselves
in an efficient manner to provide an economical solution for evaluating the structural integrity of
large objects.

Wireless techniques offer a significant advantage over wired techniques due to the ability to

10

non-destructively extract data from sensors or transducers that are not in the direct line of sight or
located in places that are not easily accessible. With the advent of the internet, and the advancement
in wireless communication techniques, a number of conventional systems have been transformed
to a "smart system" in which the operation of the system is controlled remotely. The conversion
of conventional system to a cloud based remote controlled system commonly known as Internet-
of-things (IoT), have revolutionized the world [41]. Applying IoT to create a "Smart" monitoring
system for a huge infrastructure by enabling real-time and reliable wireless data communication
among the cluster of sensor nodes enhances the visibility of the structure’s health [42]. There are
two common techniques available in the literature for implementing an IoT based SHM system are
either battery-based or battery-less [43,44].

1.4.1 Wireless High-Q resonators for NDE

The wireless high-Q resonance probe would be ideal for imaging dielectric composites and poly-
mers [45]. The high-Q directly correlates with the sensitivity and small changes due to cracks,
delamination, foreign objects, etc. can be easily detected [5, 46]. The wireless mechanism allows
to read the resonance of an LC tank without directly loading it with the source impedance and make
the probe more immune to higher order modes and spurious noise signal in the system. With a
high Q-factor, the change in reflection coefficient due to defect will be significantly higher than a
mismatched sensor probe.

1.4.2 Battery-free Wireless Sensor Tags for SHM

A number of battery-free sensing solutions have been proposed in the literature as an alternative
low cost system for SHM. One such system is the passive Radio Frequency Identification (RFID)
with integrated sensors that are suited to monitor different SHM parameters such as temperature,
humidity, and moisture [47, 48]. To collect sufficient information to determine the nature and
map the structural integrity of the target, multiple parameter monitoring becomes essential. In the
literature, two types of passive multisensing RFID platforms are available, one is a CMOS based

11

RFID sensing platform, and the other is a passive wireless identification and sensing platform
(WISP) [49–51]. These platforms are mostly limited to monitoring extrinsic parameters and very
few intrinsic parameters that do not allow a comprehensive structural integrity mapping of the
target.
In this report, a few passive sensing techniques are proposed that allows to obtain an
improved comprehensive map for structural integrity of any large infrastructure.

Comprehensive mapping of the structural integrity of the target requires a host of discreetly
positioned sensors or transducers monitoring both intrinsic and extrinsic parameters. To realize
an efficient and an economic system, the overall power consumption should be reduced and this
can be achieved by adapting passive battery-free sensors or transducers for monitoring different
parameters. Furthermore, adapting efficient wireless techniques to retrieve data from the passive
sensors or transducers allows remote monitoring of the target. A high performing system requires
an efficient retrieval of data, which is possible by developing new routines using conventional
modulation techniques which are noise immune, power-efficient, and are capable of modulating
multiple sensors or transducer information onto the re-transmitted signal [52–55].

1.5 Thesis Organization

The dissertation thesis is organized as follows:

• Chapter 2 presents the technique to make a wireless near-field high-Q sensing resonator for

dielectric evaluation of composites and polymers.

• Chapter 3 introduces the harmonic RFID to enhance the communication range by eliminating

the RF clutter.

• Chapter 4 extends the capabilities of RFID-based battery-free communication system by
integrating the passive sensors and transmitting a hybrid data packet without increasing any
power requirement.

• Chapter 5 exploits the concept of hybrid data packets even further by actuating and detecting

the acoustic signal using a battery-free RFID tag.

12

• Chapter 6 finally concludes the thesis and presents possible future work.

13

CHAPTER 2

WIRELESS NEAR-FIELD HIGH-Q SENSING RESONATOR FOR DIELECTRIC

EVALUATION OF COMPOSITES AND POLYMERS

2.1 Introduction

This chapter introduces a wireless near-field sensing system for evaluating the composite and
polymer-based dielectric materials. Composites and polymers are widely used in numerous indus-
tries like aerospace, renewable energy, automotive, defense, etc. for making low cost, light weight,
and long lasting products.

Composites and polymers have an ever increasing demand, so does the requirement of evaluating
techniques for avoiding any potential operational issues. Quality analysis of these materials is
critical during manufacturing to prevent higher maintenance and repair costs. These materials are
produced using various methods under different environmental conditions, where the output yield
and quality can vary significantly. Some of the common manufacturing issues of the composites
include surface and subsurface defects such as delamination, voids, cracks, fiber breakage, improper
bonding and misalignment, etc. [56]. While polymers show problems like density differences,
bonding, permeation, and brittling issues, etc.

There is a growing demand for developing rapid, low cost, non-destructive evaluation (NDE)
techniques to identify and quantify these issues in composites and polymers to prevent any future
economic and structural losses.

In recent years, the advancement in sensing technology has fueled the development of different
probing techniques to evaluate the structural integrity of the target under test. In literature, many
novel sensing modalities are available analyzing different aspects of the composite and polymer
evaluation from direct contact surface mounted probes to non-contact, non-invasive probes. Some
of the common NDE techniques are vibration analysis, microwave imaging, thermography, shearog-
raphy, X-ray imaging, acoustic emission technique, and ultrasonic scanning [56–60]. Most of these

14

methods suffer from common limitations such as rigorous data analysis and processing, complex
scanning, high power transducers, not real-time or robust, lower defect sensitivity, increased cost
and surface mount with hard-wired connections. For example, the acoustic methods require physi-
cal contact with the materials by surface mounting of sensors using a low attenuation flexible glue,
a large number of sensors are required to cover the complete material area makes it impossible to
perform real-time scanning on the manufacturing line [59]. Moreover, the thermography methods
are sensitive to spatial and temporal temperature variations leading to unreliable sensitivity and
depends on the resolution of the expensive thermal camera for imaging [61–63]. The technique also
requires a preparation of the surface before implementation due to hot and cold spot. Microwave
imaging are the alternative EM methods for evaluating defects in composite and polymer structures.
Single element near-field microwave probes can test the dielectric materials, but the high frequency
operation makes the circuitry more complicated and expensive. Design of multi-element near-field
array probe is intricate due to element size and coupling effects. Additionally, sensitivity of the
direct-wired probes depends on the length of the cable, bending effects of the cable, cable losses
and connectors. The Q-factor of these sensors are low due to contact loading leading to dampening
effects. Table 2.1 lists some of the common NDE methods for composite and polymers. The
comparison shows the fundamental limits and the cost of implementation and operation of these
methods. In order to overcome these limitations, a low cost, low frequency, real-time, non-contact,
and non-invasive wireless sensing technique is needed. The growth of hand-held wireless devices
enables the realization of an economical wireless sensing system.

The surface and subsurface manufacturing defects of the composites and polymers can be
detected by monitoring the change in material properties such as the relative permittivity, tensile
strength, deflection temperature, etc. The defects in composite lead to a change in effective dielectric
constant across the surface of the material.
In literature, dielectric characterization of different
targets such as solids, liquids, and gases has been well studied using wireless near-field and far-
field sensor probes [70–73]. Inductor-Capacitor (LC) based wireless resonant tanks are commonly
used dielectric probes in which the inductance or capacitance is perturbed by the target under test

15

Table 2.1: Comparison of commonly used NDE Methods for composites and polymers

Method

Probe/ Feed Type

Contact/ Wired

(Capital/

Cost
Operation)
Medium/ Low

Near-Surface/ Wired

Low/ Low

Near-Surface/ Wired

Medium/ Low

Operation/

Distance
High Power
Distance Operation

Medium/ Low

Medium/ Low

Operation/

Distance
High Power
Via Couplant/ Wired

High/ High

Medium/ Low

Contact/ Wired

Distance Operation

Medium/
Medium
High/ Very Low

Fundamental Limitations

Limited to ferromagnetic ma-
terials
Limited to Conductors

High Frequency Circuitry, El-
ement Coupling
High Frequency Circuitry,
Heavy Data Processing
Sensor Embedding in Struc-
ture
No Good Contrast, Radiation
Hazard
Poor SNR due to high Atten-
uation
Heavy Data Processing

Surface preparation

Cur-

Mi-

Flux

Magnetic
Leakage [64]
Eddy
rent [65]
Near-field Mi-
crowave [60]
Far-filed
crowave [66]
Frequency Selec-
tive Surfaces [67]
X-Ray [59]

(Ultra-

C-Scans
sonic) [68]
Ultrasonic
Guided Wave [69]
Thermography
[61]
Wireless Near-
field (This Work)

Near-Surface/ Wireless

Low/ Very Low

Limited to dielectrics

leading to shift in the resonance frequency. There is a direct correlation between the sensitivity to
the Q-factor of the sensor that warrants high-Q sensor design for improved sensitivity [74].

The goal of this work is to develop a novel wireless LC resonant tank-based imaging sensor for
detecting dielectric defects in real-time for composites and polymers. The imaging probe requires
to have high-Q, sub-mm resolution and simultaneous actuation and detection capability in
array configuration.

Two different sensor designs are studied in this work, a single capacitive sensor probe and an
array probe. The single probe consists of a parallel plate capacitor with strong fringing fields along
the edges and a coupled inductor coil forming an LC resonant tank. The inductor coil also allows
wireless magnetic coupling with a secondary pick-up coil for resonance interrogation. The resonant
frequency of the sensor probe varies due to the capacitive loading of the fringing fields by the target

16

under test.

The second approach is the array probe that consists of four independent LC resonant tanks with
resonance frequencies closer to each other but at independent resonant frequencies. The resonant
peaks have sufficient frequency spacing that prevents peak overlap during target sensing. The four
tanks are fabricated side by side in a 1x4-array fashion and this array is interrogated using a single
external pick-up coil. The array allows two-dimensional simultaneous scanning of the targets under
test.

In this work, the capabilities of the proposed sensors are demonstrated by detecting sub-surface
and surface defects along different composites and polymers such as glass fiber reinforced polymer
(GFRP), and high density polyethylene (HDPE).

2.2 Theory and Design

2.2.1 Single Element Probe @ 5 MHz

A single capacitive sensor probe is investigated to track changes in the relative permittivity of the
target under test. The sensor consists of an LC tank, which has an unshielded parallel capacitor
fabricated on a Roger’s 4350 board with the dielectric constant of 3.66 and a thickness of 1.52 mm.
Fig. 2.1 shows the schematic and the actual developed sensor probe with dimensions in Table 2.2.
The strip lines with solder pads are designed to couple the capacitor plates with a commercial
inductor coil (100 

) to form the LC tank. The strip lines on either side of the Roger’s board are
separated to avoid an increase in the effective capacitance, but offer additional series inductance.
The length and width of the strip lines dictate the additional series inductance. Capacitive loading of
the target material changes the effective impedance of the LC tank leading to shift in the resonance
frequency.

The parallel plate capacitor is designed to have strong fringing fields along the edges of the
capacitor that are used for target monitoring. In order to enhance the fringing field at the probe
end, the width of the capacitive sensor is kept more than the width of the strip lines. The fringing
fields can be tailored according to the defect types of the target under test.

17

Figure 2.1: Schematic of the LC tank along with the fabricated sensor probe

Table 2.2: Dimensions of the fabricated sensor probe

Parameter

Dimensions (mm)

Parameter

Dimensions (mm)

d
2

a
1
e
0.6

b
8
f
4

c
29
g
1

The default response (in Air) of the RF sensing probe is measured using an external pick up
coil for interrogation of the LC tank connected to a Vector Networks Analyzer (E5070B). The
measurement setup is shown in Fig. 2.2.

Figure 2.2: Measurement setup for the single probe

18

The interrogator pick-up coil is designed such that the self-resonance of the inductor is way
beyond the resonance frequency of the sensor probe. The pick-up coil is optimally positioned
above the sensing resonator probe using a 3D printed cylindrical cap. The coil is wound around
the cylindrical cap to interrogate the sensor probe. The reflection coefficient is recorded using the
VNA. The measured resonance frequency of the single probe LC tank is at 5.34 MHz. Fig. 2.3
shows the measured S11 demonstrating the high Q-factor (∼3800) of the sensor probe.

Figure 2.3: Measured S11 (magnitude and phase) of the single LC sensor probe

2.3 Measurement and Characterization

2.3.1 Dielectric Reference Measurements

The dielectric reference measurements are performed to demonstrate the change in resonance
frequency due to the change in target properties. In this experiment, well-characterized and standard
dielectrics were chosen as targets for demonstrating the sensitivity of the sensor probe. Teflon and
Plexiglass with the dielectric constant (

) of 2.1 and 3.1, respectively, were chosen [75, 76]. The
targets were placed in the fringing field region of the LC tank and the resonance frequency was
recorded. Equation 2.1 gives the formula for the resonance frequency ( 

) of an LC tank.

19



 =

1
√

2






(2.1)

The resonance frequency is inversely proportional to the total capacitance of the sensor, which
is directly proportional to the effective dielectric constant. The first set of experiments shows the
change in resonance frequency for different dielectric standards in Fig. 2.4. All measurements were
taken in triplicates and averaged out, with 1 mm lift-off distance from the dielectric sample. It
can be inferred from the graph that, as the dielectric constant increases, the resonance frequency
decreases and hence the peak shifts towards the left (lower frequency) as expected.

Figure 2.4: Frequency response of the standard dielectric materials (Teflon = 2.1, Plexiglass = 3.1)
with known relative permittivity

The single element probe detects less than 5% change in dielectric constant and this was verified
by measuring, second set of experiment, two samples of high frequency laminates with precisely
defined relative permittivity, Rogers’s 4003 substrate (thickness = 1.52 mm, dielectric = 3.55) and
Rogers’s 4350 substrate (thickness = 1.52 mm, dielectric = 3.66). The frequency shift was ∼ 0.45
kHz between the two samples, which were measured using VNA with 1 Hz resolution.

The third set of experiment was performed on composites and polymers such as polyamide (PA-
11), high density polyethylene (HDPE), and glass fiber reinforced polymer (GFRP). The dielectric
properties of the targets are 2.8-3.5, 2.3-2.5, and 3.5-3.4, respectively [77–79]. Fig. 2.5 shows the

20

plot for change in resonance frequency for different composites and polymers. It can be inferred
from the plot that, the developed sensor has good sensitivity towards a small change (

=0.2) in
dielectric constant.

Figure 2.5: Frequency response of the composites and polymers with different dielectric constants

The fourth set of experiments were performed to analyze the performance of the sensor
probe with conductive targets. Aluminum and Carbon fiber reinforced polymer (CFRP) with a
conductivity (
) of 1
7
/
 and 1000 
/
, respectively, were chosen as targets. Fig. 2.6 shows
the change in resonance frequency for different conductor targets.

Figure 2.6: Frequency response of the conductive materials with different conductivity

The fifth set of experiment shows GFRP with surface defects, which was placed as a target for

21

testing the sensitivity of the single probe. The surface circular defect had 4 mm radius and a scan
across the surface was performed. Each measurement was repeated five times and averaged out.
Fig. 2.7 shows the change in resonance frequency and phase for the change in effective dielectric
constant due to presence of surface defects on glass fiber reinforced polymer.

Figure 2.7: Frequency response (magnitude and phase) of GFRP with defect and no defect

The sixth set demonstrates the sensitivity of the single probe in detecting subsurface defects.
Polyamide with different depths of subsurface defects ranging from 3 to 6 mm was chosen as a
target, dimensions of the target are shown in Table 2.3. A horizontal scan was performed across
the surface of the target. The effective dielectric constant of the surface decreases with increase
in the depth of the defects, which in turn decreases the effective capacitance of the probe thereby
increasing the resonance frequency. The measurements were performed in triplicates and the data
was averaged out. Fig. 2.8 shows the change in resonance frequency with reference to the resonance
frequency of the single probe without any defects for different depths. The resonance frequency
change increases with increase in depth as expected. The sensitivity of the probe depends on the
strength of the fringing fields and the depth of penetration along the parallel plate capacitor. Based
on the sensing application and location of the subsurface defects, the fringing fields of the parallel
plate capacitor can be tailored by modifying the design to accommodate greater penetration depth
with increased sensitivity.

22

Figure 2.8: Change in resonance frequency for increasing depth of sub-surface defects in polyamide
target

Table 2.3: Dimensions of the polyamide sample

Parameter

Dimensions (mm)

a
160

b
8

c
3

d
4

e
5

f
6

g
20

h
24

2.4 Dielectric Imaging

The sensitivity of the single probe is verified using an imaging setup. The probe is mounted on
an X-Y raster scanner and the target is placed at a distance of 1 mm from the probe. An RF source
operated at the resonance frequency of the single probe (5.34 MHz and 0 dBm input power) is
connected to a power splitter which generates two signals, the reference signal and the measurement
signal. The measurement signal is fed into the input of the directional coupler and the coupler is
connected to the pick-up coil of the probe. The reflected signal (magnitude and phase) from the
pick-up coil is fed to one of the input ports of the lock-in amplifier from the directional coupler. The
reference signal from the power splitter is fed into the lock-in amplifier for synchronization. The
lock-in amplifier generates voltage signals for the difference in reference and reflected signal. The
output of the lock-in amplifier is fed into a National Instrument Data Acquisition card (PCIe-6341)
which samples and digitizes the data using an imaging software and the output is plotted on a
computer. The output is the change in voltage signal as a function of position on the target. A low
frequency -3 dB directional coupler was designed for this purpose. The schematic of the imaging

23

setup is outlined in Fig. 2.9 and the measurement setup is shown in Fig. 2.10.

Figure 2.9: Schematic of the dielectric imaging setup

Figure 2.10: Measurement setup for dielectric imaging with the single probe mounted on a X-Y
raster scanner

Aluminum was selected as a first demo target material with surface defects. The dimensions of
the sample were 180 mm x 220 mm x 3.5 mm and had twelve different sizes of surface defects as
shown in Fig. 2.11A and Table 2.4. The defects were also in the shapes of basic geometries such as
triangle, circle and squares to analyze the sub-mm resolution of the developed probe. The output

24

from NI-DAQ was plotted in terms of complex voltages to understand the nature of the defects.
Fig. 2.11B shows the changes in output voltage of the reflected signal for different surface defects
on aluminum. It can be inferred clearly from Fig. 2.11 that the probe was able to recognize different
surface defects with different sizes and shapes.

Figure 2.11: (A) Image of the aluminum target with dimensions, (B) Imaging results for profiling
of defects using single probe

Table 2.4: Dimensions of the aluminum sample

Parameter Dimensions (mm) Parameter Dimensions (mm)

30
20
5
32
20
8
5

a
b
c
d
e
f
g

180
220
48
23
10
3
45

h
i
j
k
l
m
n

The calibration (aluminum) sample shows promising results and demonstrates the sub-mm
resolution capabilities. Dielectric sensor probes shift its resonance according to the effective
capacitance and work well with the new custom designed imaging setup. Next step is to check the
capabilities of the probe over composite and polymer sample that represents the real world issues.

25

Three different samples are selected for demonstrating the dielectric imaging capabilities; CFRP,
GFRP and HDPE. The imaging results are shown in the sections below.

2.4.1 CFRP Imaging

First, CFRP sample is selected, which is provided by Con-Edison for detecting multiple defects like
air gap, contamination, epoxy saturation, mold, and delamination. The CFRP is also wrapped on a
steel sample, mimicking the corrosion protection layer and acting as new pressure layer. Fig. 2.12A
and Table 2.5 shows the sample image and dimensions. Fig. 2.12B shows the imaging results that
clearly detect the embedded defects.

Figure 2.12: (A) Image of the CFRP target, (B) Imaging results for profiling of defects using single
probe

Table 2.5: Dimensions of the CFRP sample

Parameter Dimensions (mm)

a
b

200
240

26

2.4.2 GFRP Imaging

GFRP was chosen as the second dielectric target for verifying the RF probe capabilities. The
dimensions of the target are 128 mm x 126 mm x 3 mm and had multiple defects such as impact
damage at the center and fiber breakage (through hole defects) along the four edges of the target
as shown in Fig. 2.13A and Table 2.6. Fig. 2.13B shows the changes in imaging results clearly
distinguishing the defects.

Figure 2.13: (A) Image of the GFRP target, (B) Imaging results for profiling of defects using single
probe

Table 2.6: Dimensions of the GFRP sample

Parameter Dimensions (mm)

n
o
p

128
126
8

27

2.4.3 Polyethylene Imaging

Polyethylene was chosen as the third target to demonstrate the sensitivity of the probe in detecting
subsurface defects. The dimensions of the target are 60 mm x 40 mm x 8 mm and had two different
depths of subsurface defects in the shape of a triangle at a depth of 3 mm from the target surface as
shown in Fig. 2.14A with Table 2.7 and the imaging verification results are shown in Fig. 2.14B.
The results show that the developed probe had stronger fringing fields with higher penetration depth
to clearly differentiate the subsurface defects. The penetration depth can be tailored based on the
intended application for the detection of deeper subsurface defects.

Figure 2.14: (A) Image of the HDPE target, (B) Imaging results for profiling subsurface defects in
HDPE target

Table 2.7: Dimensions of the HDPE sample

Parameter Dimensions(mm)

q
r
s
t

60
140
25.4
38

The imaging verification step demonstrates the capability of the single probe as an RF sensor
for detecting surface and sub-surface defects on conductive, composites and polymers materials.
The single probe showed very minimal effects when measured using either side of the elongated
capacitor while performing the image scans. The orientation of the probe effects the edge response

28

due to the field distribution along one directions, which can be deconvoluted by signal processing or
running multiple scans in different directions. The sensitivity of the probe depends on the strength
of the electric-field of the parallel plate capacitor and on the penetration depth of the fringing fields.
The developed probe has a sub-mm resolution and can differentiate a small change in dielectric
constant (Ɗ

 (cid:62) 0.2), which can be further enhanced by tailoring the fringing fields of the parallel
plate capacitor. The spatial resolution is dictated by the scanning step size and the probe size. The
probe size depends upon the separation between electrodes (1.5 mm) that can easily detect any crack
with size 1 mm or larger. Moreover, the probe is shown to detect sub-surface defects 3 mm from
the surface and this could be further improved by tailoring the penetration depth of the fringing
fields by further optimizing the probe design. To reduce deflection time and enhance performance,
the single probes can be placed in a 2D array format and can be used to simultaneously scan a
large surface are while maintaining the resolution. The distance between the sensor probe and the
target should be fixed throughout the scan and during successive scans between different samples.
A standard set of dielectrics should be used as test targets before the actual measurements to verify
the frequency shift dynamics to prevent false positives.

2.5 Array Design and Measurement

A 1x4 array probe was designed with four individual single probe sensing elements on the same
substrate. The resonance frequencies of the four probes were spaced apart such that it does not
interfere or overlap with each other. The optimization of the resonance frequency was performed
by varying the capacitance or inductance of the individual LC tanks. The schematic of the array
probe along with the dimensions are shown in Fig. 2.15 and Table 2.8.

A single pick up coil was used to interrogate the array and the measured result of the 1x4 array
is shown in Fig. 2.16. The resonance frequencies with high Q factor for the four LC tanks were at
2.41, 3.51, 4.56 and 5.46 MHz, respectively.

GFRP sample was selected to demonstrate the sensitivity of the array probe. A sample of
dimensions 128x126 mm with surface defects was placed 1 mm from the fourth element of the

29

Figure 2.15: Schematic of the designed array probe along with the dimensions

Table 2.8: Dimensions of the array probe

Parameter Dimensions (mm) Parameter Dimensions (mm) Resonance freq (MHz)

2.41
3.51
4.56
5.46

a
b
c
d
e

1
8
29
2
0.6

f
g
h
i
j

4
1
0.25
2.25
4

Figure 2.16: Measured frequency response of the 1 x 4 array probe under no loading conditions

30

array probe to illustrate 2D scanning. Fig. 2.17A shows the measured result of the array probe
with the change in resonance frequency for all four frequencies due to the GFRP sample. Since the
defect was placed in the near-field region of the fourth element, it can be inferred from the graph
that the resonance frequency shifts only for that particular element and remains the same for the first
three elements. Fig. 2.17B shows the change in resonance frequency for the fourth array element
due to surface defects in GFRP material. The LC array can be further improved by making an n
x n and can be used for simultaneously scanning a large area of the target. The standard dielectric
laminates Roger’s 4003 and Roger’s 4350 are tested that gives a frequency shift of ∼ 0.5 kHz.

Figure 2.17: Measured results for GFRP with surface defects for (A) All four elements of the array,
(B) Focused frequency Shift in fourth element of the array

The array probe allows a second degree of freedom in measuring defects. A major factor that
influences the accuracy of the measurements is the isolation between successive frequency peaks
in the array. The number of elements in the array are limited by the available frequency band and
the isolation factor. The challenge building an efficient array lies in incorporating multiple sensor
elements within a short frequency spectrum and at the same time maintaining a high-Q factor for
each of them. Another limiting factor for the number of elements is the frequency shift between
successive sample scans. Ideally, the shift should not exceed the isolation factor.

31

2.6 Summary

This work demonstrated a low-cost passive wireless sensor for evaluating structural integrity
of composites and polymers. The sensor probe can be easily adapted to scan different kinds of
materials for surface and subsurface defects, including material loss, fiber breakage, and impact
damage as presented. The electric-field of the capacitive probe can be tailored to detect defects at
various depths depending on the application.

The 1 x 4 array probe can be extended to a faster 2D scanning by increasing the number of array
elements (n x n). The developed sensor probe can detect less than 5% change in dielectric constant
making it a highly sensitive probe for practical applications that can analyze the structural integrity
of different targets. The array sensor probe can be readily used in manufacturing lines for quality
control applications.

32

CHAPTER 3

WIRELESS FAR-FIELD HARMONIC RFID COMMUNICATION PLATFORM FOR

ENHANCED RANGE

3.1 Introduction

Radio frequency identification (RFID) system is a well established infrastructure used across
multiple avenues such as logistics, security, automobiles, health-care, toll tags, animal tracking,
quality control, and applications based on internet-of-things (IoT) [80–84]. A typical RFID system
consists of RFID tags as markers for tracking and identification, and an interrogator to read the
tags [80]. A passive UHF RFID tag consists of an antenna and an application specific integrated
chip that can be powered up using the received RF signal. The interrogator transmits the query
signal which is received by the tag, the signal powers up the chip which then re-transmits the
modulated signal back to the receiver with the identification information [85]. Although the
current system is well-established, there exists a number of drawbacks such as self-jamming and
multi-path interference that limits realizing the full potential of the RFID system [86,87].

Typically, a passive UHF RFID system operates in a single frequency band, 902-928 MHz in
the United States. The isolation of transmitter and receiver paths is vital to prevent self-jamming
of the RF signal. Most RFID readers use either a bistatic or a monostatic antenna configuration
to transmit and receive the information. Due to single frequency operation, both configurations
suffer from low transmitter-to-receiver isolation. This signal leakage weakens the sensitivity and
limits the performance of the RFID system [88]. The low isolation (leakage) limitation of the
interrogator increases the phase noise and raises the power spectral density higher than the side
bands of the weak reflected signal, increasing the bit-rate error between the RFID tag and the
interrogator. Moreover, another limiting factor that reduces the overall system performance is the
multi-path interference due to reflections from surrounding structures, as well as interference from
other tags operating in the same frequency band. The problem is depicted in Fig.3.1.

33

Figure 3.1: Range limiting phase noise issue with conventional RFID system

Over the last decade, a number of techniques have been proposed that cancels self-jamming and
limits the multi-path interference. For example, a sample of the transmitted signal with opposite
phase is added to the receiver to cancel out the self-interference [89, 90]. Moreover, adaptive self-
jamming cancellation algorithms have been proposed that automatically optimize the parameters
chosen for the cancelling signal and alter the cancelling signal based on the dynamic changes in the
tag environment such as reflections from surrounding objects when the tag is in motion to suppress
self-jamming [91, 92]. However, these techniques do not provide a robust, reliable solution to the
limitations of the system and hence in recent years, harmonic RFIDs and software defined radio
(SDR) based interrogators are proposed as an alternative solution for conventional single frequency
RFID systems [86, 93, 94]. The primary idea for using clutter free communication method is to
enhance the detection range as shown in Fig.3.2. The communication distance is divided into 3
different regions. Region 1 represents the range of the conventional RFID system where the tag
turns on using harvested power and the reflected signal is strong enough to be detected by the
interrogator. Region 2 represents where RFID tag turns on using the harvested power due to its
low sensitivity but do not able to reflect the strong signal that can be detected by the interrogator
due to high phase noise levels. Region 3 represents where the tags do not able to harvest enough
received RF power so it would not reflect any signal back. Among these 3 regions, using harmonic
communication, the range can be extended up to the end of Region 2, which offers a clutter free

34

operating mechanism.

Figure 3.2: Hypothesis for harmonic communication towards range enhancement

A harmonic radar concept is illustrated in Fig.3.3. The harmonic radar illuminates the tag
at the fundamental frequency ( 
0), and the non-linear elements in the tag generates higher order
harmonics (
 
0) and re-radiates it back to the interrogator.

Figure 3.3: The hRFID communication system

In literature, a number of papers have been published on the concept of using harmonic radar for
RFID applications. These tags exploit the non-linearity in active components to generate harmonics
separating the uplink and the downlink frequency thereby completely eliminating self-jamming.
Moreover, these tags do not require a more sophisticated RF isolator to prevent leakage or cross-

35

talk at the front-end, reducing the system complexity. The tags range from chipless harmonic
passive tags (battery-free) to harmonic tags integrated with RFID digital components [86, 95].
The nonlinearity is achieved by using a single or a series of active devices, such as a Schottky
or a Varactor diode [96].
In order to adapt the harmonic radar concept with the current RFID
technology, two possible options are commonly explored. First option is to redesign both the RFID
tag and the interrogator to generate harmonics and read harmonics, respectively. Second option
is a cost-effective one to use the current infrastructure to generate and read harmonics from the
commercial RFID tags.

The commercially available UHF RFID tags are passive in nature and are constrained by the
availability of real-estate for additional sophisticated electronics such as on-board filters to suppress
the nonlinearities generated from the diodes and MOSFETs of the RFID chip. The concept of
harmonic generation from commercially available UHF RFID tags was explored in [97] and the
authors concluded that the current RFID tags can generate and scatter harmonics depending on
three parameters. First, the specific tag antenna design, second, the IC sensitivity or the minimum
RF power required for the interrogator query signal to activate the chip, and third, the power and
bandwidth of the interrogation signal. An RFID test platform for characterizing the nonlinearities
generated by the conventional tags is presented in [98]. A method to read the harmonic backscattered
information by manipulating the physical layer of the interrogator is presented in [99,100], where an
oscilloscope is used to detect the backscattered RFID signal at the harmonic frequency. Recently,
the integration of UHF passive RFID tag with harmonic energy harvesting function is explored,
wherein the third harmonic power generated by the RFID ICs is utilized to harvest energy and
improve the read range of the tag [101, 102] as well as for activating sensors coupled to the
tag [103].
In order to exploit the full potential of the harmonic signals generated by the RFID
IC, not just for harvesting or sensing but also as carrier of information, harmonic communication
platform using the existing single frequency UHF RFID system is necessary.

In this work, a method to utilize the existing UHF RFID platform for harmonic communication
is explored for the first time by designing an interface RF harmonic layer between the commercially

36

available RFID interrogators and the RFID tags. The advantages of the harmonic system are
realized fully in a conventional UHF RFID system by utilizing the proposed architecture without
any internal design change to the interrogator as shown in Fig.3.3. A new harmonic hRFID tag
antenna design is proposed for efficiently radiating the harmonic signal back to the interrogator.
The proposed system utilizes the advantages of harmonics in terms of improved signal to noise
ratio, higher clutter rejection, and enhanced read range. An analysis of harmonic communication
is also performed demonstrating the relationship between the read range, the RFID IC sensitivity,
and the received signal power. The proposed system can be used for interrogating commercially
available existing UHF RFID tags or custom antenna-designed hRFID tags. The harmonic solution
with increased read range can be easily adapted for any current UHF RFID application.

3.2 Design

The hRFID system shown in Fig.3.3 consists of two main units: A) hRFID Interrogator, and B)

hRFID Tag.

3.2.1 Harmonic RFID Interrogator

The hRFID communication platform has significant advantages than the conventional single band
RFID system. However, the current RFID infrastructure is built upon the conventional UHF RFID
platform and millions of interrogators are in place. Replacing all interrogators with hRFID system is
impractical and uneconomical. Hence, a cheaper alternate solution is proposed by using the existing
conventional RFID system for harmonic communication by integrating an additional RF front-end
to the interrogator for receiving the backscattered tag information at the harmonic frequencies.

The additional RF front-end will act as an intermediate communication layer between the
conventional RFID interrogator and the hRFID tag. The harmonic RF layer receives the query
signal from the interrogator at the fundamental frequency ( 
0) and radiates it using an UHF
antenna operating at the fundamental frequency band. The hRFID tags that are present within
the field of radiation will receive the query signal and respond back to the interrogator with

37

the identification information at the harmonic frequency. The harmonic RF layer receives the
backscattered information using an harmonic antenna that operates at the harmonic frequency band
and down converts it to the fundamental frequency. The downconverted backscattered signal is fed
back into the interrogator which demodulates and extracts the digital ID information. The proposed
method overcomes the challenges associated with realizing a good conventional RFID system such
as cancelling self-jamming and limiting multipath interference. The detailed block diagram of the
harmonic RF layer integrated with the conventional RFID system is shown in Fig.3.4.

Figure 3.4: The hRFID communication platform using conventional UHF system

The conventional Impinj Speedway R420 RFID interrogator system is used for communicating
with the hRFID tags by integrating the designed harmonic RF layer. The harmonic RF layer
isolates the transmitting and receiving RF signals at the fundamental frequency using three 3-port
RF circulators (MECA CS-0.90), each with a 20 dB isolation factor. The circulators are connected
in series in which the forward propagating signal is fed into the successive circulators without any
loss and the third port of each circulator is matched with a 50-ohm load allowing a 60 dB collective
isolation and a reflection coefficient of -28 dB. The purpose of the series circulators is to allow the
interrogator’s signal to reach the UHF transmitting antenna and at the same time, it prevents the
reflected RF signal at the fundamental frequency from the tag to reach the interrogator. A reference

38

fundamental signal is extracted from the circulator’s output port before forwarding it to the UHF
antenna using a directional coupler (MC ZABDC20-25H-S+) with a 22 dB coupling factor. The
Laird’s circularly polarized patch antenna with 6 dBi gain is used for radiating the fundamental
frequency signal to the tag.

Figure 3.5: Hairpin bandpass filter, F1 (2.6-3.4 GHz): (A) Schematic with dimensions, and (B)
Simulated and measured S-parameters

Table 3.1: Dimensions of the Hairpin bandpass filter

Parameter

Dimensions (mm)

a
8.0

b
3.32

c
1.0

d

0.575

e
3.92

f

15.0

The third harmonic reflected signal from the hRFID tag is received using a Vivaldi antenna.
The choice of third harmonic signal for reflecting information from the tag is explained in the next
section and also in refs [97, 98]. In this work, the harmonic system is designed to query the RFID

39

tag at the fundamental frequency and read the backreflected information at the third harmonic
frequency. Since the Vivaldi antennas are wide band, it is necessary to filter out the fundamental
frequency signal and allow only the third harmonic signal to be downconverted in order to extract
the information. To achieve this, a hairpin filter in a bandpass configuration is designed and
fabricated on Rogers’s 4350 substrate (thickness = 1.52 mm, dielectric = 3.66, loss tangent = 0.003)
using a standard photolithography process. The schematic of the designed hairpin filter along with
its dimensions, simulated, and measured results are shown in Table.3.1 and Fig.3.5.

Figure 3.6: Five stage bandpass filter, F2 (1.7-1.9 GHz): (A) Schematic with dimensions, and (B)
Simulated and measured S-parameters

Fig.3.5 shows that the designed filter provides a 45dB isolation of the reflected signal from the
tag at the fundamental frequency (915 MHz), and allows only the third harmonic frequency (2.745
GHz). The received third harmonic signal is amplified using a Mini-Circuit’s LNAs (MC ZX60-

40

Table 3.2: Dimensions of the five stage bandpass filter

Parameter

Dimensions (mm)

Parameter

a
11.6

f

Dimensions (mm)

12.44

b
3.27
g

12.56

c

52.36

h
9.30

d
47.0

i

45.71

e

j

12.56

0.35

53LNB-S+ and ZX60-43+) with a total gain of 36 dB. To feed the information into the conventional
UHF RFID interrogator, the amplified third harmonic frequency signal should be downconverted
to the fundamental frequency using an RF mixer. For this purpose, the second harmonic signal
at 1.830 GHz is required at the LO port of the mixer. To obtain the second harmonic signal, the
reference 915 MHz signal from the output port of the directional coupler is fed into a frequency
multiplier (MC FK-3000+) for doubling. The output of the frequency multiplier is amplified using
an LNA (MC ZX60-2531MA-S+) with a gain of 34 dB and is fed into a five-stage bandpass filter.
The purpose of the filter is to allow just the second harmonic frequency (1.830 GHz) into the mixer.
The filter is designed with Roger’s 4350 of thickness 1.52 mm and the schematic with dimensions
and its frequency response are shown in Fig.3.6.

The second harmonic signal is fed into the LO-port of the RF mixer, and the third harmonic
signal reflected from the tag is fed into the RF-port of the RF mixer for down conversion. After
down conversion, the output fundamental frequency signal with the RFID information is received
at the IF-port of the RF mixer (MC ZX05-42MH-S+). The output of the mixer is fed back into the
conventional RFID interrogator through a circulator.

The designed harmonic RF layer allows uplink at the fundamental frequency and downlink at
the third harmonic frequency. The RF layer operates within the bounds of the conventional RFID
system at the same time providing the advantages of a harmonic platform.

3.2.2 Harmonic RFID Tag

The conventional single band RFID tags designed by various companies (for example, IMS, Alien,
Impinj, etc) are commercially available. The fundamental architecture of a tag consists of an UHF

41

antenna and an RFID IC. The IC performs logical controls and responds to specific query signals
from the interrogator. The IC uses a MOSFET for modulating the ID information on a backscattered
RF signal by switching between "on" and "off" states. The digital part of the IC requires a DC power
for operation and is generated using an RF-to-DC converter with rectifying diodes in a charge pump
configuration [86]. The internal circuitry of the IC with non-linear components including diodes
and MOSFETs collectively generates the third harmonic signal. The details of the third harmonic
generation from commercial RFID tags are available in the literature [97, 98]. The conventional
off-the-shelf RFID tags (COTS-RT) are designed to operate in a single frequency band (902-928
MHz), and do not radiate the internally generated third harmonic signals efficiently. This section
shows the design of an RFID tag antenna for efficient third harmonic radiation. The radiated third
harmonic signal has very low signal strength, but for the case of harmonics, the noise level is also
extremely low, which in turn increases the SNR of the received signal.

3.2.2.1 RFID IC

The RFID chip chosen for integration with the designed hRFID tag antenna is the Higgs-3 from
Alien Technology [104]. According to the data sheet, the chip has a -15 dBm read sensitivity
and is capable of operating in the frequency band of 860-960 MHz. Due to the requirement of a
smaller footprint and low sensitivity level, RF filters are not integrated as a part of any commercially
available RFID chips. This leads to an internal generation of harmonics that are not suppressed
and these contains the modulated identification information as well. This third harmonic signal can
be radiated back to the interrogator using a dual band antenna design. The Alien’s Higgs-3 IC has
a characteristic input impedance of 39.5-j60  and 15+j0  at fundamental (915 MHz) and third
harmonic frequency (2745 MHz), respectively.

3.2.2.2 RFID Antenna

To efficiently radiate the third harmonic signal, a new antenna design is required that operates in
two different frequency bands, the fundamental band from 902-928 MHz, and the third harmonic

42

band from 2706-2784 MHz. Commercially available RFID tags commonly use a single meandered
dipole configuration for an antenna operating at a single UHF band. Whereas, the hRFID antenna
consists of two antennas operating at the fundamental and the third harmonic frequency bands.

Figure 3.7: (A) Schematic of the designed hRFID antenna along with its dimensions, and (B)
Measured frequency response of RFID IC and hRFID antenna with conjugate matching points

Table 3.3: Dimensions of the hRFID antenna

Parameter

Dimensions (mm)

Parameter

Dimensions (mm)

a
8.0
f

13.5

b
4.0
g
5.0

c
35.0
h
2.0

d
1.0
i
1.5

e
7.0
j

32.0

ANSYS HFSS® (2019, ANSYS, Inc., Canonsburg, PA, USA) were used to design the dual

43

band antenna. The schematic of the designed antenna along with its dimensions is shown in
Fig.3.7A. The antenna is designed on Roger’s 4350 substrate and fabricated using conventional
photolithography. The measured frequency response of the designed hRFID antenna with conjugate
matching is shown in Fig.3.7B.

3.3 Results

For the demonstration of a harmonic RFID platform using the conventional UHF system, three

sets of experiments are performed.

Figure 3.8: Measured backscattered signal visualized using a spectrum analyzer, with a transmitted
power of +8 dBm and a read range of 0.35 m: Received signal at the fundamental and the third
harmonic frequency with and without the RFID tag in the interrogator’s field of view

First set of experiments demonstrates the third harmonic generation by conventional RFID
tag (on Higgs-3 chip), down conversion and detection. The second set demonstrates the efficient
harmonic radiation from the designed hRFID tag. Third set of experiments demonstrates the

44

improved read range of the designed hRFID tag using harmonic communication. The hRFID
interrogator platform explained in the previous section is assembled, and the return power from
the RFID tag at the fundamental and third harmonic frequencies are measured via a circulator
coupled to a patch antenna and a Vivaldi antenna, respectively. The reflected signal is acquired
using a spectrum analyzer (HP8562A). The RFID interrogator is configured to transmit +8 dBm
at a single UHF frequency (926 MHz) for analysis purposes. The spectrum analyzer is set at the
fundamental (926 MHz) and at the harmonic (2778 MHz) frequencies with a span of 500 kHz and
a bandwidth of 10 kHz. Fig.3.8 row 1 shows the received RF signal without any RFID tag. It can
be inferred from the figure that at the fundamental frequency, the leaked interrogator signal with
query information is received, showing the isolation limitation of any single band communication
system. The leakage increases the noise level causing self-jamming of the fundamental reflected
signal. Whereas, at the harmonic frequency, the noise level is dictated by the thermal factors and
is below -100 dBm.

Next, the conventional RFID tag is placed at a distance of 0.35 m from the interrogating
antennas, and the reflected signal is received with the same setup. Fig.3.8 row 2 shows the received
backscattered signal at the fundamental and at the harmonic frequencies. The increase in power of
the side bands at the fundamental frequency is clearly visible representing the RFID tag information
in the frequency domain. The RFID tag also reflects back the information at the third harmonic with
a lower strength in comparison to the fundamental frequency but with a higher SNR. The RFID
modified interrogator with the harmonic RF layer receives the third harmonic signal and down
converts it. The RF mixer requires a second harmonic RF signal for reference, which is generated
by doubling the interrogator’s fundamental frequency signal, shown in Fig.3.9 row 1. The mixing
of the amplified third harmonic and reference signal gives an output at the fundamental frequency,
which is the downconverted signal with information on the sidebands detected by the interrogator,
as shown in Fig.3.9 row 2.

The conventional RFID tag is replaced with the designed hRFID tag. The designed hRFID tag
generates the harmonic signal 7 dB higher than the conventional RFID tag with the same input

45

Figure 3.9: Measured backscattered signal visualized using a spectrum analyzer, with a transmitted
power of +8 dBm and a read range of 0.35 m: Reference signal at the second harmonic frequency
generated using a frequency doubler and the down converted signal at the fundamental frequency
after mixing

power and interrogation distance as shown in Fig.3.10.

The designed hRFID tag is tested and a comparison between the fundamental and the harmonic
communication for a range of up to 6.5 m is demonstrated. The interrogator is set to transmit
+30 dBm of power using a 6 dBi gain patch antenna. The designed hRFID tag receives the query
signal and reflects back the information at fundamental as well as at the harmonic frequency bands.
The fundamental received signal measured using the conventional setup and the harmonic signal
measured using the modified receiver are shown in Fig.3.11. The fundamental signal is probed
using an Impinj interrogator without using the harmonic setup. The downconverted harmonic signal
is the mixer’s output at the fundamental frequency, generated from the received third harmonic

46

Figure 3.10: Received third harmonic signal from the designed hRFID tag.

signal. The interrogator’s detection limit using the conventional RFID communication is at ∼5 m,
whereas using the harmonic communication is at ∼6 m. The RFID chip stops responding to the
query signal at a certain distance (∼6.5 m) due to the limited IC sensitivity (-15 dBm) and it also
poses the limit on the tag detection distance. As the read range approaches 5.5 m for operating the
tag in the fundamental frequency, the received signal falls below the noise level due to self-leakage
as well as interference and hence the tag is not detected by the interrogator.

The designed harmonic interrogator is tested with multiple (two) hRFID tags. They have
different antenna designs, matched at both fundamental and third harmonic frequency bands. The
Impinj interrogator’s average read rate (total detected tags per sec) for a single tag is in the range
of 60-65 and it changed to 71-75 when multiple tags are present in the field of view.

3.4 RF Analysis of Harmonic Communication

The harmonic communication platform is superior to the conventional UHF system due to the
capability of eliminating self-jamming and limiting multipath interference. This improvement is
due to the lower reference noise level at the harmonic frequency in comparison with the fundamental
frequency. In a single frequency band operation, the power leakage from the RF coupler or the

47

Figure 3.11: Received backscattered power at the interrogator, with transmitted EIRP of 36 dBm,
from the designed hRFID tag for both the fundamental frequency and the harmonic frequency
operation.

circulator increases the noise level significantly. In literature, the noise level for a conventional
RFID system is reported anywhere between -80 dBm to -60 dBm, with varying bandwidth and
data rates [105]. Whereas, the noise level at harmonic frequency is dictated by just the thermal
factors and is observed to be below -100 dBm, as observed in Fig.3.7A. A system analysis study for
harmonic RFID communication with -70 dBm and -100 dBm noise levels at the fundamental and
harmonic frequencies, respectively, is presented. The study is primarily focused on the read range
of the passive RFID systems. Two different factors are considered in estimating the maximum
operating range of an RFID tag: 1) Backscattering efficiency and 2) the IC sensitivity.

The FCC allows to transmit a maximum of 4 W (+36 dBm) EIRP at the fundamental frequency
in the ISM band. The received power at the tag located at a distance from 0 to 100 m spacing from
the interrogator, and the backscattered signal strength at the interrogator can be calculated using
Friis’s transmission equation. The modified Friis’s transmission is presented in Eq.3.1, which can
calculate the maximum possible read range for a passive RFID tag operating with a single frequency
or with a harmonic frequency for both the uplink and downlink communication.

48

arg min




[
 
 − (


 + 
 




 

 + 

 


+20


10(
/4







) − 
 + 20


10(
/4





 

))]




 + 

 






 + 
 




 



(3.1)

where, 
 is the maximum distance between the interrogator and the tag (read range) that needs
to be estimated. 


 is the transmitted power by the interrogator at the fundamental frequency,

 





 is the gain of the interrogator’s transmitting antenna, 

 





 is the gain of the interrogator’s
receiving antenna, 





 is the frequency of the transmitting signal by interrogator, 



 

 is the
frequency of reflected signal by tag, 
 



 

 is the gain
of the tag’s transmitter antenna, 
 is the speed of light, 
 is the backscattering efficiency or losses
due to modulation and reflection, 
 
 is the noise level at the receiving frequency.


 

 is the gain of the tag’s receiver antenna, 

 



Figure 3.12: Power budget analysis of backscattering based communication system: (A) Detection
range using fundamental and harmonic bands with no tag losses, and (B) Detection range using
fundamental and harmonic bands with 5 and 10 dB losses, respectively.




 and 

 



Eq.3.1 is visualized in Fig.3.12 with +30 dBm 


, 6 dBi 
 



 

 and


 



 

, 

 ( 





), and reflection ( 



 

) at three different frequencies 

, 2 

, and 3 

. 
 

for receiving at 

 is -70 dBm, and for receiving at 2 

 and 3 

 is -100 dBm. Fig.3.12A shows
the maximum possible communication distance of 15.4m, 64.12m and 52.3m for the fundamental,
second harmonic, and third harmonic frequencies, respectively, without any RFID tag losses (
= 0).
With the tag losses (
) of -5 dB [106] at fundamental and -10 dB [107] at both harmonic frequencies




, 0 dBi 
 



49

gives a maximum communication distance of 11.4 m, 35.6 m, and 29.0 m for the fundamental,
second harmonic and third harmonic frequencies, respectively, as shown in Fig.3.12B.

The study mentioned above assumes that the tag can be activated at any received power, but
the RFID tags have a minimum power requirement that is defined by the IC sensitivity. The IC
sensitivity plays an important role in determining the maximum communication distance. The
commonly used RFID IC’s can be activated with -12 to -15 dBm received power. Fig.3.13 shows
that a minimum activation power (-12 dBm) can be received up to a distance of 6.5 m with
above-mentioned transmitter and receiver parameters. The backscattered signal from 6.5 m with a
lossy tag maintains the power level above the noise level at all receiving frequency configurations.
The minimum IC sensitivity limit, for which the received signal stays above the noise level, can
also be back-calculated for both the fundamental and harmonic frequencies using Eq.3.1. For a
fundamental frequency of 915 MHz, -17 dBm is the calculated limit for IC sensitivity. Using the
calculated limit, the maximum read range of the tag is estimated to be 11.6 m. The IC sensitivity
limit defines the maximum read range for the respective frequency of operation. If the IC sensitivity
goes below the defined limit, the read range will not increase further because the reflected signal
will be below the noise level at the receiver end. For receiving at the second (1.830 GHz) or the
third (2.745 GHz) harmonic frequency, the maximum read range can be further increased due to a
lower noise level. For the second and third harmonic frequencies with a fixed tag loss of 10 dB,
the maximum read range is 36 m with an IC sensitivity of -27 dBm, and 29 m with IC sensitivity
of -25 dBm, respectively.

The RF analysis of the harmonic communication shows a path to achieve higher SNR with
improved clutter rejection, and increased read range. The study reflects the importance of the
choice of an RFID IC, and the parameters of the tag in terms of sensitivity, power loss, and noise
level with their effects on the read range.

50

Figure 3.13: Effects of the IC sensitivity on the RFID read range with fundamental and harmonic
communication.

3.5 Discussion, Challenges and Limitations

The designed hRFID platform performs better than the conventional RFID system as demon-
strated in the previous sections. The key improvement of the hRFID system is the 30 dB reduction
in noise level achieved by using the harmonic frequency for communication. Furthermore, the
designed harmonic RF front-end for the interrogator can be miniaturized for an easy plug-in that
can be directly plugged into the current RFID interrogators for realizing the hRFID platform making
it an economical solution.

The harmonic generation efficiency of the RFID ICs changes with different manufacturers,
which can be further studied. The harmonic radiation from the passive RFID tag comply with the
FCC regulations, as harmonics are considered as spurious signals and allowed if it stays below a
certain power level (hRFID radiates in the order of 10−11 W) [108]. The harmonic communication
overcomes the bottleneck of a conventional RFID with lower noise level that allows an opportunity
for further fine tuning the IC sensitivity. The demonstrated harmonic communication platform

51

using the conventional UHF RFID infrastructure shows the advantages of self-jamming cancellation
leading to a higher SNR as well as the availability of three times the bandwidth for communication
(2706-2784 MHz). In order to fully realize the potential of increased bandwidth, a change in RFID
protocol is required that allows the RFID tags to use wider channels to transmit the information.
Overall, the proposed platform is beneficial for any RFID applications that requires improved read
range, especially in a cluttered environment.

3.6 Summary

In this work, a hRFID platform is demonstrated using a conventional UHF RFID system for
transmitting the information using the inherently generated third harmonic from the RFID IC.
An RF harmonic interface layer between the conventional RFID interrogator and the RFID tag
is designed. This layer is used as a plug in to any conventional UHF interrogator to transmit
the query signal at the fundamental frequency and receive the backscattered information at the
third harmonic frequency. Harmonic communication is established using a conventional RFID
tag that generates spurious harmonic signals. Moreover, the third harmonic radiation efficiency
is increased by designing a new harmonic tag antenna coupled to a conventional RFID IC. An
increased read range is achieved along with the elimination of self-jamming and limiting of multi-
path interference without any internal design changes to the conventional interrogator. A harmonic
analysis is performed showing the relationship between the RFID IC’s sensitivity, tag losses, noise
level of the interrogator, and the read range. The demonstrated system is best suited for any RFID
applications that require increased data rate with improved read range.

52

CHAPTER 4

PASSIVE SENSOR INTEGRATION WITH ADC-FREE RFID TAG FOR ULTRA-LOW

POWER OPERATION

4.1 Introduction

A low power long-range communication system is desired in rapidly modernizing society as it
has a huge impact on human safety and economy in multiple sectors such as civil, environment,
transportation, etc. [9,10]. Real-time sensing of different parameters of any object provides timely
feedback that ensures enhancing predictive maintenance strategies to minimize human interaction
and promote safety and reliability. To make any system intelligent, a number of sensors are
required to provide continuous and detailed information on the nature of the system starting from
an autonomous vehicle to structural health monitoring [18, 109]. Moreover, the cluster of sensors
should be efficient and be able to communicate the sensing information with each other or to
the controller in order to perform a quick and accurate corrective action. Conventional wired
communication approach towards collecting sensor data from a large number of nodes is neither
compatible nor economical and hence a practical wireless approach such as Internet-of-Things
(IoT) based platforms are preferred [110]. IoT has revolutionized the world by converting many
applications into a "Smart" operation with integrated sensors and devices. Applying IoT to create a
"Smart" monitoring system by enabling real-time and reliable wireless data communication among
the cluster of sensor nodes enhances the visibility [42]. One of the recent challenges is to develop
a low-cost, battery-free, long-range, and real-time IoT sensing system. The impact of the system
will benefit areas such as structural health monitoring (SHM).

An example application of pipelines is selected for making the case for battery-free wireless
sensor tags and demonstrating the working prototype. Pipelines are the safest and economically
viable arterial networks for transporting natural gases and oils across the globe. The pipeline
network expands at a faster rate due to the projected growth in population and increased rate of

53

urbanization, creating a demand of natural gas for domestic and industrial use [111]. The natural
gases are gathered, transported, and distributed at various pressure levels via metals, composites,
and plastic pipelines. Any material subjected to a high or low cyclic pressure over a prolonged
period of time induces a stress on the material. This stress over time compromises the safety and
reliability of the pipeline and may lead to a catastrophic failure [112–114]. In order to prevent
such failures, there is a growing need for the development of an economical, real-time, scalable,
structural integrity monitoring, and sensing system for the pipeline infrastructure implemented in
refineries, chemical plants and manufacturing facilities.

In literature, a number of structural health monitoring (SHM) techniques exist to evaluate
the integrity of pipes, aiding in improving the safety and lifespan of the pipeline infrastructure
[115,116]. The different indicators that are monitored for elucidating the pipeline health are cyclic
pressure, surface temperature, humidity, and vibration [115,116]. Among these indicators, pressure
is the most important and found to be the primary cause of damage to the pipelines [117].

A number of techniques have been proposed to monitor the pipeline pressure; such as piezo-
based resistive sensors, micro-electromechanical systems (MEMS)-based capacitive sensors, and
fiber Bragg grating (FBG)-based negative pressure sensors [117–119]. The fabrication process of
all the sensors is very complex, which either requires a hefty cleanroom process or costly tools for
precision. For example, the MEMS-based capacitive sensors require expensive substrate material,
chemical etching, wire bonding, and separate housing [119]. Additive manufacturing is a viable
alternative of this process, which has been previously proven for electronic system and component
design [120]. Three-dimensional printing technique was previously used for designing multiple
components in a pressure sensor [121–123]. However, a fully 3D-printed pressure sensor is still
a feat to achieve. Most of the published research is focused on various other applications and gas
pipeline is not the primary focus, which leads to over complexity, limited communication range,
and scalability issues.

One of the major limitations of these sensor units is that a direct wired connection is required
for data acquisition. The direct probing or wired connection limits the application of sensors and

54

poses a problem for inaccessible environments, like under ground buried pipelines or large chemical
plants with a requirement of thousands of sensing nodes.

A number of wireless sensing approaches have been proposed as an alternate solution to
overcome the disadvantages of the wired sensing methods [124, 125]. The wireless methods
monitor the pressure in the pipeline by deploying a large number of active sensor nodes. The
integrated nodes on the surface of the pipes transmit the data wirelessly to the interrogator. The
disadvantage of active sensors is the need for an on-board battery, charging and/or replacement
requirements increase the implementation and maintenance cost of the system [124–126]. Battery-
free, low-cost, real-time sensing is ideal to provide a robust and economical system with longer life
span and lower maintenance [127,128].

Inductor-capacitor (LC) resonator-based passive sensor tags have been proposed in the literature
for monitoring the pressure based on capacitive loading, which in turn shifts the resonance frequency
[129]. LC resonant sensors operate in near-field configuration and have a limited interrogation
distance [5].

A number of other IoT based sensor systems are also proposed for SHM applications, majority
of them operate using digital sensor nodes that require an external power source or battery for its
operation, leading to a power hungry solution [38,39]. The digital sensor nodes require an analog-
to-digital converter (ADC) and a microcontroller for reading the digital data and transmitting it back
to the interrogator. For large-scale deployment, these sensor nodes are not suited for continuous
monitoring due to their limited life time, increased overall cost, and additional maintenance of
the system. Similar problem arises for the cold food supply chain, where a low cost long lifetime
sensing platform is required [41]. To overcome the limitations of current active systems, a number
of battery free passive sensing solutions have been proposed in the literature [47, 130, 131]. One
such system is the battery-free Radio Frequency Identification (RFID) with integrated sensors that
are suited to monitor different parameters such as temperature, humidity, and moisture [47, 48].
To collect sufficient information to determine the nature of the problem with the structure being
monitored, multiple parameter monitoring becomes essential. One significant challenge in IoT

55

based sensing systems is to design power efficient sensor nodes that can detect and transmit the
information back to the interrogator without any battery. A table of commercially available RFID
technologies with sensing capability and sensitivity data is shown in Table 4.1.

Table 4.1: Commercially available RFID ICs with sensitivity data

RFID IC

Sensor Ports Sensitivity (dBm)

SL900A [ [49]]
PhaseIV Engr

1
2
3 Andy100 [ [132]]
4 Magnus 3 [ [133]]
5
6

NXP UCODE
Higgs 4 [ [134]]

2
1
1
1
0
0

0
-8
-2
-9.9
-21
-20

Battery-free sensing RFID platforms reported earlier are either a single CMOS IC based or
multiple discrete components integrated on PCB substrate [49–51]. CMOS-based RFID platform
consists of a low power on-chip sensing capability that allows a compact RFID tag design with
digital ID and sensing at a shorter read range [49]. Interestingly, CMOS based digital circuitry for
identification and logic control is highly efficient and requires a very low operational power [134].
On the other hand, discrete component-based platforms are equipped with a microcontroller, ADC,
external memory, and multiple sensors powered by energy harvested from the interrogation signal
and reflect back the digital data by modulating backscattered signal [50, 51, 135]. Although both
CMOS-based and discrete component-based platforms are battery-free, the integrated power hungry
ADC limits the real-time monitoring as well as the read range of the RFID sensing system.

The discussed far-field sensing platforms operate at a single frequency (902-928 MHz), which
are prone to clutter due to multi-path interference and have a lower signal-to-noise (SNR) due
to self-jamming, as shown in the previous chapter and open literature [127]. To overcome these
limitations, the harmonic communication approach is ideal, that eliminates self-jamming, reduces
multi-path interference, lowers down phase noise, and increases the SNR. Harmonic-based sensors
have been previously reported in the literature for sensing temperature, and pH [136,137] for liquid
monitoring.

56

4.2 Analog RF Sensing

A passive harmonic RF tag with an integrated pressure sensor is designed for the real-time mon-
itoring of the pipeline infrastructure [138]. The harmonic RF system for gas pipeline infrastructure
consists of an RF tag coupled with a pressure sensor and an RF interrogator.

Figure 4.1: Example battery-free wireless sensing system architecture for the pipeline infrastructure

The harmonic RF sensor tag has a receiver (2 GHz) and transmitter (4 GHz) antenna, a harmonic
doubler, and an integrated pressure sensor with the hybrid coupler. The integrated pressure sensor
is designed with a cavity based parallel plate reactive impedance element/sensor that changes its
reactance with the change in applied pressure from 0–20 psi. The designed sensor is fabricated
using the additive manufacturing technique, which uses a thin liquid crystalline polymer (LCP)
substrate as its diaphragm. The hybrid coupler translates the change in reactance to the change
in the received 2 GHz signal’s phase, with the power efficient modulation technique, which is
backscattered at a harmonic frequency of 4 GHz. A schematic of the analog sensing system over
the pipeline infrastructure is shown in Figure 4.1.

57

4.2.1 Design

The design of analog passive wireless sensing harmonic system includes: (A) 3D printed pressure
sensor, (B) passive harmonic RF sensor tag, and (C) harmonic RF interrogator.

4.2.1.1

3D Printed Pressure Sensor

The designed pressure sensor is an air cavity based parallel plate capacitor, where one electrode
has a fixed position and second electrode is deformable due to the pressure as shown in Figure 4.2.
The air cavity in between the electrode acts as dielectric medium and holds a single atmospheric
pressure or reference pressure. The external pressure is applied through the tunnel directly onto the
diaphragm with copper electrode.

Figure 4.2: Air cavity-based fully 3D printed pressure sensor

The pressure sensor is fabricated in three parts, as shown in Figure 4.3. The first component has
the air cavity with the fixed electrode and electrical access holes for rigid coaxial cable. The copper
is plated into the cavity’s bottom of the 3D printed part. The second component is the deformable
electrode, which is realized using 18 µm-thick LCP film with copper plating. The LCP film is fixed
on top of the cavity, such that it makes an electrical connection with the outer shell of the rigid
coaxial cable. The third component is a pressure tunnel that directs the pressure onto the electrodes
and holds the LCP film in its place. The top part has four alignment legs, which perfectly fits with
the bottom component. An extra layer of polymer resin is coated on top of the fully assembled

58

pressure sensor and cured under ultraviolet (UV) light, for eliminating any pressure leaks into or
from the air cavity and acts as a seal of the pressure sensor. The outer resin layer also holds the
three separately fabricated components together and replace any use of extra adhesive.

Figure 4.3: Exploded view of 3D printed pressure sensor with three main components

Figure 4.4 shows a fully assembled view of the fabricated pressure sensor, where the three
pressure sensor parts fit perfectly and allow a coaxial cable to make an electrical connection to the
electrodes for reactance measurement.

Figure 4.4: Fully assembled 3D printed pressure sensor with component annotation and dimensions

The total volume of the fabricated sensor is 10 × 13 × 6 mm3. The diameter of the pressure
tunnel is 6.6 mm and the depth of the pressure cavity is 1 mm. The change in pressure deforms

59

the LCP electrode and reduces the separation of the two electrodes, therefore a change in electrical
response, which can be measured using an impedance analyzing equipment.

Table 4.2: Dimensions of the 3D-printed pressure sensor

Parameter

Dimensions (mm)

a
10.0

b
6.0

c
13.0

d
6.6

e
1.0

A sealed cavity-based sensor with a deformable diaphragm in a parallel plate configuration is
used to measure the pressure, ranging from 0 to 20 psi. The pressure sensor is integrated inside
a polypropylene plastic pipe and probed using an SMA connector from the outside, as shown in
Figure 4.5A. A vector network analyzer (VNA) is used in a direct wired configuration to measure
the impedance of the pressure sensor at 2 GHz, as shown in Figure 4.5B.

Figure 4.5: (A) Pressurizing chamber (pipe) with integrated pressure sensor, (B) Schematic of the
measurement setup, and (C) Fabricated pressure sensor

The measured impedance is plotted on a normalized Smith chart for varying pressure, as shown
in Figure 4.6. The reactive impedance changes from 10.23–17.80  for the applied pressure. A
small series resistance is observed in the acquired data, which is due to the parasitics associated with
the high frequency and losses in the cable. The direct wired measurements support the hypothesis of
significant change in reactive impedance due to the applied pressure difference between the sealed

60

cavity and pipeline pressure. Moreover, it provides an insight into the parasitic real impedance of
the sensor and provide data for simulating the ideal phase shift for a given range of impedance.

Figure 4.6: Measured impedance at 2 GHz for pressure range from 0 to 20 psi with a direct wired
connection using vector network analyzer (VNA)

4.2.1.2 Harmonic RF Sensor Tag

The passive harmonic RF tag uses a reactance-based phase shifting mechanism for wirelessly
measuring the change in the applied pressure. The hybrid coupler is used as a phase shifter, which
is a four-port reciprocal device with input, output, isolation, and coupled ports. The phase shifter is
designed and simulated using ANSYS HFSS® (2019, ANSYS, Inc., Canonsburg, PA, USA) on an
FR-4 substrate with a dielectric constant (

) of 4.4, thickness of 1.52 mm and a loss tangent (tan

) of 0.02. The phase shifter operates at 2 GHz with a 3 dB coupling factor and a 20 dB isolation.
The phase shifter’s input is connected to a 2 GHz patch antenna for receiving the continuous
RF wave from the interrogator. The received RF signal is forwarded to the output and the coupled
ports of the coupler, to which a reactive impedance pressure sensor is connected. The RF signal
at the output and the coupled ports experience a total internal reflection due to the termination of
a reactive sensor element. The internally reflected signal is received at the isolated port without

61

any internal power loss, but with a shifted phase according to the reactance. The isolated port is
connected to a non-linear device (Schottky diode), which generates a second harmonic signal at
4 GHz. The Schottky diode (BAT 15-03) doubles the shifted phase along with a frequency. The
phase modulated 4 GHz output of the frequency doubler is transmitted back using a patch antenna.
The schematic of the designed harmonic RF tag is shown in Figure 4.7 along with the dimensions.

Figure 4.7: Schematic of the pressure sensing harmonic RF tag along with its dimensions. Signal
received using 2 GHz antenna forwarded to the hybrid coupler for phase shift according to the
integrated pressure sensor and the second harmonic signal with double shifted phase, generated
using diode frequency doubler, is transmitted out using 4 GHz patch antenna

Table 4.3: Dimensions of the analog pressure sensor tag

Parameter

Dimensions (mm)

Parameter

Dimensions (mm)

a
46.0
h

17.64

b
35.0

i

18.47

c
2.0
j

30.0

d

46.14

k
18.0

e
2.86

l
2.0

f

4.92
m
22.0

g
20.5
n

17.32

The fundamental and harmonic antennae are placed in cross-polarization for increasing the SNR
and minimizing the interference at the interrogator. The simulated and the measured frequency
responses for both the fundamental and the harmonic patch antenna are shown in Figure 4.8. The
measured response closely matches the simulated response, and the observed minimal differences
are due to fabrication tolerances.

62

Figure 4.8: Simulated and measured frequency response; (A) Fundamental (2 GHz) and (B)
Harmonic antenna (4 GHz)

4.2.1.3 Harmonic RF Interrogator

The harmonic RF interrogator is designed to communicate with the passive phase shifting-based
RF tag. The interrogator generates an RF signal using a VNA with a signal strength of 13 dBm at 2
GHz and it is transmitted towards the harmonic RF tag using a Vivaldi antenna. The harmonic RF
tag receives the fundamental frequency and backscatters the pressure sensor information modulated
at 4 GHz harmonic signal. The harmonic interrogator receives the modulated signal using a Vivaldi
antenna in a cross-polarized configuration.

Figure 4.9: Schematic of the harmonic RF Interrogator consist of a VNA for RF source and phase
measurement unit, transmitting and receiving antennas, and miscellaneous RF peripheral circuits
for communicating with the harmonic RF sensor tag

The received harmonic signal is amplified using LNA’s (ZX60-53LNB-S+ and ZX60-43-S+)

63

with a total amplification of 34 dB. The amplified 4 GHz signal is fed into the RF port of the
mixer for down converting it to a 2 GHz signal. A reference signal is extracted from the 2 GHz
transmission signal using an RF splitter and fed into the LO port of the mixer. The down converted
output signal at 2 GHz with the modulated phase information is acquired using VNA at the IF Port
of the mixer. The schematic of the harmonic RF interrogator communicating with the harmonic
RF sensor tag is shown in Figure 4.9.

4.2.2 Results

Initially, a simulation-based validation study is performed in order to test the phase shifting mech-
anism using a reactive element. A four-port phase shifter is designed in Keysight’s ADS RF
simulation software (2017, Keysight Technologies, Santa Rosa, CA, USA) with a coupled reactive
impedance element. The range of reactive impedance is referenced from the previous direct wired
measurements shown in Figure 4.6.

Figure 4.10: Simulated frequency response of the designed phase shifting hybrid coupler with a
variable reactance from 6.82  to 36.73 : (A) No power loss is observed at 2 GHz due to changing
reactance, (B) A total phase shift of 28.55 ◦ is observed at 2 GHz due to changing reactance

The reactive elements are connected to the two of the four ports of hybrid coupler, while using
two leftover ports as input and output of 2 GHz signal. The S-parameter study is used in ADS for
validating the phase shifting mechanism. The simulated transmission coefficient (S21) of the phase
shifter is shown in Figure 4.10. The phase of the transmitted signal at 2 GHz is changed from

64

−176.21◦ to 155.2◦ (Ɗ = 28.55◦) with a change in coupled reactance from 6.82–36.73 . The
simulated phase shift due to reactive impedance is significant for practical implementation and can
be easily detected using a standard phase measurement device. Moreover, the transmitted power
at 2 GHz remains unchanged, which shows no loss at the operating frequency for the entire range
of the changing reactive impedance. Pressure information is modulated on the 2 GHz signal in the
form of a phase with no power loss is a significant feat for passive RF tag technology, where a small
amount of energy saving can lead to a huge improvement in range or signal-to-noise ratio. In this
paper, the no power loss advantage is exploited for better harmonic generation using the Schottky
diode, where the frequency conversion losses are inversely proportional to the input power of the
signal [94].

The proposed technique is validated by performing two experiments. First, the phase shifter is
designed in ANSYS HFSS® (2019, ANSYS, Inc., Canonsburg, PA, USA) for 2 GHz and fabricated
on a FR-4 board for validating the simulation results. The fabricated phase shifter’s response,
without a transmitter or a receiver antenna, is measured using the VNA at 2 GHz. A power
combiner is used to connect a single reactive element (Varactor diode) to the two ports of the phase
shifter. The reactive impedance of the Varactor diode is varied by applying a bias voltage from 0 V
to 1.5 V. The change in phase, due to the change in reactance of the transmission coefficient (S21)
at 2 GHz in direct wired configuration, is shown in Figure 4.11. The 1.5 V change in bias potential
leads to a change of 18.28◦ in phase. The fabricated phase shifter performed in a similar way to
the simulated results. Next, the receiver and transmitter patch antennae are connected to the phase
shifter with a harmonic doubler and the wireless phase measurement experiment is performed.
The harmonic RF sensor tag is placed at a separation of 8 inches from the interrogator antenna for
wireless communication. The wirelessly acquired phase change due to the applied bias voltage is
shown in Figure 4.11. A phase shift of 36.78◦ is observed for the wireless configuration, which
is approximately double the single frequency wired configuration. In Figure 4.11, both responses
are the average plots of three repetitions and error bars represent the standard deviation between
all readings. A constant phase offset with reference to 0◦ is added in both the wired and wirelessly

65

acquired data.

Figure 4.11: Measured phase change due to variable reactance introduced with direct current (DC)
biasing the varactor diode in a wired configuration (Ɗ = 18.28
) using a single frequency approach
and wireless configuration (Ɗ = 36.78
) using the harmonic frequency approach

Second, a pressurized pipeline setup is designed to test the harmonic RF tag with an integrated
pressure sensor. A section of the polypropylene pipe, 24 inches in length and 2 inches in diameter,
is sealed using threaded polypropylene caps.

An 1/8 inch precision flow valve is connected to both the ends of the pipe for controlling the
pressure. A high pressure nitrogen cylinder is used to pressurize the pipe through a compressed
yor-lok fitting coupled to a precision flow valve. Due to the safety of the research lab and personnel,
natural gas pressure pipe application is demonstrated using nitrogen gas. A 20 psi safety valve is
connected that sets the limit on the maximum applied pressure. The designed pressure sensor can
also work at higher pressure ranges, but due to safety, the upper limit is set to 20 psi for this work.
The working principle of the designed sensor is demonstrated for a limited pressure range from 0
to 20 psi. The integrated pressure sensor inside the pipeline is connected to the designed harmonic
RF tag, as shown in Figure 4.12. The harmonic RF tag is capable of receiving RF signal at 2 GHz
and modulate the phase of output signal according to the applied pressure.

66

Figure 4.12: Pressurized polypropylene pipeline setup with the fabricated harmonic RF sensor tag
for real-time wireless communication

In our experiment, the applied pressure range is from 0 to 20 psi. Due to applied pressure, the
harmonic RF interrogator, placed at 18 inches’ distance from the RF tag, measured a phase shift
of 16.6◦ in the received signal at 4 GHz. The wirelessly measured phase is shown in Figure 4.13,
where the measurements are acquired at an interval of 5 psi. All measurements are performed
thrice for checking the repeatability and standard deviation at each acquisition point, as shown in
Figure 4.13. The plotted average data of three repetitions show a linearity of 0.9915 (R2), which is
comparable to existing commercial pressure sensors [139].

The designed harmonic RF sensor tag has a sensitivity of detecting a minimum change of 1 
in reactive impedance at 2 GHz and is compatible with other types of reactive impedance based
sensors for measuring higher pressure or other physical parameters such as temperature, moisture,
etc. The interrogation range of the system can be further enhanced by radiating the maximum
allowable power of 4 W, whereas in this work, only 19 mW was used. A maximum communication
range of 15 ft can be achieved with 36 dBm (4W) transmitted power at 2 GHz, 20 dB doubler diode
conversion loss and −90 dBm receiver sensitivity at 4 GHz.

The 3D printed pressure sensor eliminated the complex fabrication requirement in cleanroom,

67

Figure 4.13: Phase change is wirelessly measured using harmonic RF interrogator due to applied
pressure ranging from 0 to 20 psi. The linearity of change in phase due to pressure is represented
using coefficient of determination (R2 = 0.9915)

while providing similar capabilities, such as piezo-based resistive sensors, MEMS-based capacitive
sensor, or FBG-based pressure sensors [117–119]. The passive wireless feature successfully
communicates the pressure information from the pipe to interrogator, while eliminating the battery
requirement and lowering down the maintenance cost [127].

4.2.3 Discussion and Limitations

To avoid the limitations of the power-hungry ADC, we used the alternative solution of analog
modulation to send the sensor information directly over the backscattered signal. But the phase
modulation techniques are affected by noise in the propagation environment. Furthermore, in
case of direct phase modulation of the sensor information, the distance between the tag and the
interrogator should be calibrated in order to avoid the ambiguities towards change in distance.
An alternative approach to analog sensing for multisensor integration and detection is shown in
Appendix A. The battery-free multisensor system uses frequency modulation technique to reflect
back sensor data, which is immune to multipath effect or spatial placement. But it also suffers

68

from other limitations like only a single RF tag can be in the field of view for reliable information
transfer.

Therefore, to avoid ambiguity issues with analog RF sensing and digitization issues with
digital RF sensing, a new platform called Hybrid RFID sensing is proposed. The hybrid RFID
platform use power efficient ID communication mechanism from conventional RFID infrastructure
and combine with power efficient analog sensing to create a novel solution for sensor integration
problem in battery free tags.

4.3 Hybrid RFID Sensing

The conventional single-frequency RFID tag IC’s without sensing capabilities are very sensitive
[134]. The minimum power required to activate the IC is only -20 dBm (10uW), which translates
into longer communication range of about 50-ft. On the other hand, the digital RFID tags with
sensor integration capabilities need at least -6 dBm [49, 132]. The higher power requirement
reduces the range significantly to only few feet. The external digital sensor consumes more power
due to operations like analog-to-digital conversion. The efficient ADC’s inside ultra-low power
microcontroller can take up to 0.6 mA of current at 2.2V input and requires additional energy for
overhead computation like register update, calibration and readout [140]. The minimum power
required to acquire data using inbuilt ADC is approximately 1.32mW, which is significant without
even considering the additional power required to actuate sensor. To harvest enough power using a
typical energy harvester circuit with 20-25% efficiency, a higher input power is required, which is -6
dBm in existing digital sensor integrated RFID tags. The alternative technologies of analog sensing
do not put additional constraints over power requirement, therefore superior over ADC mechanism.
But analog sensing has it’s own limitations, which can be tackled in this section by integrating it
with digital ID tags. Therefore the power requirements stay as low as ID only mechanism but the
tag capabilities will include sensing as well.

The hybrid RFID sensing system is a combined solution using digital and analog signals. It
utilize the power efficient digital ID mechanism from conventional systems and analog sensing

69

from latest work shown in previous section. It takes best parts from both digital and analog worlds.
Hybrid RFID sensing system does not affect the communication range and reliably transmits the
sensor’s information. The analog sensor data can be received using an hybrid interrogator and
digitized with high precision ADCs. The data frame for hybrid RFID tag is shown in Table.4.4.

Table 4.4: Data frame format for Digital ID with Analog Sensor Data Reply

P EPC ST SP CRC-16 ASD

where, P is preamble, EPC is Electronic Product Code, ST is Sensor Type, SP is Sensing
Parameter, CRC is Cyclic Redundancy Check, and ASD is Analog Sensor Data. The ASD is
further divided into two components, as shown in Table.4.5. The data frame has both digital bit
and analog sensor information, therefore, it is called an hybrid method for RFID sensing.

Table 4.5: Data frame format for Analog Sensor Data Reply

RAD TG SAD

where RAD is the reference analog sensor data, TG is the time gap, and SAD is the sensor
analog data. RAD is used as a reference measurement for eliminating any ambiguity and precise
detection of the change in phase of the RF signal. The design of hybrid RFID sensing is presented
below, which use the digital RFID work shown in Appendix B as building block for providing logic
and ID to the analog tag.

4.3.1 Design

The hybrid sensing system is designed for both single and harmonic frequency operation. The
digital end with energy harvester, microcontroller, modulator, and non-linear transmission line
stays the same as explained in Appendix B. The novel analog sensing hybrid coupler is integrated
with digital tags in unique way for both frequency operations. The architecture of the designed
RFID tags are shown below.

70

4.3.1.1 Hybrid Single-Frequency RFID Tag Architecture

The hybrid single-frequency RFID tag has three main components; a digital RFID system, a hybrid
coupler, and a 915 MHz antenna. The digital RFID system is used as explained in section B.2.1.
The schematic of the hybrid single-frequency RFID tag is shown in Fig.4.14.

Figure 4.14: Schematic of the hybrid single-frequency RFID tag

A new hybrid coupler is designed to operate at 915 MHz using lumped components for sensing

the change in impedance without any requirement of additional power.

Figure 4.15: (A) Schematic of a lumped hybrid coupler with integrated reactance sensors for
modulation, (B) Simulated and measured response of the coupler at 915 MHz

71

The high frequency (2 GHz) hybrid coupler for power efficient phase modulation was shown in
previous section. Based on the same fundamental principles the miniaturized 915 MHz coupler is
simulated in Keysight’s ADS RF simulation software (2017, Keysight Technologies, Santa Rosa,
CA, USA) for the correct component values and modulation check. The designed coupler with
schematic and frequency response is shown in Fig.4.15. The value of inductors (
1 = 8.9

 and

2 = 2.55

) and capacitors (
 = 5.6

) are tuned according to the additional PCB parasitics. An
ultra-low power RF switch is used to connect and disconnect in between the reference impedance
and sensor impedance elements.

A Varactor diode is attached as an impedance sensor, which changes its capacitance due to the
applied voltage across its terminals. The change in capacitance will shift the phase of the input
signal accordingly. The change in phase due to the applied voltage is measured using VNA. The
acquired phase data is shown in Fig.4.16.

Figure 4.16: Measured change in phase due to shift in capacitance of the varactor diode using
VNA, at 915 MHz

The phase-shifting hybrid coupler changed its phase from 76.53
 to 87.80
 for applied voltage of
0.0
 and 1.0
, respectively. The coupler can be integrated with the tag as shown in its architecture
schematic. An antenna will be connected at port 1 and a impedance mismatching modulation
switch at port 2, which will reflect back the input signal with shifted phase in ’OFF’ state. The
phase will shift twice due to forward and backward propagation via the coupler.

72

A hybrid single-frequency RFID tag is tested with antenna using direct wired connections as
shown in Fig.4.17A. The measured reflected signal is acquired using an oscilloscope and the plot is
shown in Fig.4.17B. The sensor’s capacitance information is hidden in phase of the sensor analog
data (SAD), which can be extracted by demodulation using heterodyne mixer. The reference analog
data (RAD) can be used to compare the phase shift of the SAD, which eliminates the effects of any
multipath reflection and tag movement.

Figure 4.17: (A) Measurement setup for testing backreflection at 915 MHz, (B) Measured backre-
flected signal with digital ID and analog sensor data.

A 915 MHz meandered dipole antenna is designed in ANSYS HFSS® (2019, ANSYS, Inc.,
Canonsburg, PA, USA) to receive and transmit the information from the ISS tag. The designed
antenna is shown in Fig.4.18 with dimensions in Table.4.6 and frequency response is shown in
Fig.4.19. The simulated parameters match closely with the measurement, while the simulated gain
is 0 dBi.

Figure 4.18: Meandered Dipole antenna for 915 MHz

73

Table 4.6: Dimensions of the designed 915 MHz antenna

Parameter

Dimensions (mm)

a
11.6

b
10.0

c
10.0

d
6.0

e
1.0

Figure 4.19:
Simulated radiation pattern of the antenna

(A) Simulated and Measured frequency response of the 915 MHz antenna, (B)

4.3.1.2 Hybrid Harmonic-Frequency RFID Tag Architecture

The hybrid harmonic-frequency RFID tag also has three main components; a digital RFID system,
a hybrid coupler operating at 434 MHz, and a 434 & 868 MHz antenna pair. The digital RFID
system is used as explained in section B.2.2. The schematic of the hybrid harmonic-frequency
RFID tag is shown in Fig.4.20.

Figure 4.20: Schematic of the hybrid harmonic-frequency RFID tag

74

A new hybrid coupler is designed to operate at 434 MHz using lumped components for sensing
the change in impedance without any requirement of additional power. The miniaturized 434 MHz
coupler is simulated in Keysight’s ADS RF simulation software (2017, Keysight Technologies,
Santa Rosa, CA, USA) for the correct component values and modulation check. The designed
coupler with schematic and frequency response is shown in Fig.4.21. The value of inductors
(
1 = 18.5

 and 
2 = 13

) and capacitors (
 = 10.5

) are also tuned according to the
additional PCB parasitics due to the trace lines and connectors. An ultra-low power RF switch is
used to connect and disconnect in between the reference impedance and sensor impedance elements
as in previous section.

Figure 4.21: (A) Schematic of a lumped hybrid coupler with integrated reactance sensors for
modulation, (B) Simulated and measured response of the coupler at 434 MHz

A Varactor diode is attached as an impedance sensor, which changes its capacitance due to the
applied voltage across its terminals. The change in capacitance will shift the phase of the input
signal accordingly. The change in phase due to the applied voltage is measured using VNA. The
acquired phase data is shown in Fig.4.22.

The phase-shifting hybrid coupler changed its phase from −86.34
 to −75.25
 for applied

75

Figure 4.22: Measured change in phase due to shift in capacitance of the varactor diode using
VNA, at 434 MHz

voltage of 0.0
 and 1.0
, respectively. The coupler can be integrated with the tag as shown in its
architecture schematic. An antenna will be connected at port 1 and a harmonic generating NLTL
at port 2, which will transmit back the 868 MHz signal with shifted phase. The phase will shift
twice due to frequency doubling with harmonic generation.

Figure 4.23: (A) Measurement setup for testing backreflection at 868 MHz, (B) Measured backre-
flected signal with digital ID and analog sensor data.

A hybrid harmonic-frequency RFID tag is tested without antenna using direct wired connections
as shown in Fig.4.23A. The measured reflected signal was acquired using oscilloscope and plotted
in Fig.4.23B. The modulation depth (>90%) of the backreflected signal is greater than the single
frequency operation, which corresponds to better SNR and higher detection probability. The

76

digital part can be demodulated and decoded as it is shown in section B.3. The analog data is
read separately with additional elements using a hybrid multi-frequency interrogator. The phase
information from RAD and SAD region can be compared for finding a reference change in phase
due to the sensor’s impedance change.

Figure 4.24: Meandered Dipole antenna for 434 MHz

Figure 4.25:
Simulated radiation pattern of the antenna

(A) Simulated and Measured frequency response of the 434 MHz antenna, (B)

Table 4.7: Dimensions of the designed 434 MHz antenna

Parameter

Dimensions (mm)

a
7.6

b
24.0

c
8.0

d
4.0

e
6.0

A 434 MHz meandered dipole antenna is designed in ANSYS HFSS® (2019, ANSYS, Inc.,
Canonsburg, PA, USA) to receive the information from the interrogator. The designed antenna is
shown in Fig.4.24 with dimensions in Table.4.7 and frequency response is shown in Fig.4.25. The
simulated parameters match closely with the measurements.

77

A 868 MHz meandered dipole antenna is designed in ANSYS HFSS® (2019, ANSYS, Inc.,
Canonsburg, PA, USA) to transmit the information to the interrogator. The designed antenna is
shown in Fig.4.26 with dimensions in Table.4.8 and frequency response is shown in Fig.4.27. The
simulated parameters match closely with the measurement.

Figure 4.26: Meandered Dipole antenna for 868 MHz

Table 4.8: Dimensions of the designed 868 MHz antenna

Parameter

Dimensions (mm)

a
21.6

b
10.0

c
18.0

d
14.0

e
9.0

Figure 4.27:
Simulated radiation pattern of the antenna

(A) Simulated and Measured frequency response of the 868 MHz antenna, (B)

4.3.1.3 Hybrid Multi-Frequency Interrogator

A new custom interrogator is designed to communicate with the hybrid tags operating at a single or
harmonic frequency. The basic functionality of an interrogator is to transmit and receive RF signals,

78

which needs to be modulated on the way-out and demodulated on the way-in. Both the modulator
and demodulator are part of the RF front-end, with additional peripherals like power amplifier,
LNA, filters, heterodyning mixer, etc. The block diagram of the designed interrogator with multi-
frequency operation and hybrid reception is shown in Fig.4.28. The fundamental functionalities
are similar to ISS interrogator explained in section B.2.3. A few changes are made to accommodate
the analog sensor data reception. A SPDT switch is added at the IF PORT of the heterodyning
mixer to select in the path in for digital and analog data. The communication starts with the digital
data that is decoded by a microprocessor. After receiving all digital bits, the microprocessor will
switch the read port towards an oscilloscope for analog data acquisition. An additional low pass
filter is added for eliminating high frequency components and the final signal will have the residual
DC bias, which corresponds to the in-phase component of the received signal. The change in phase
or amplitude of the back-reflected signal from the tag will change the DC output of heterodyne
receiver.

Figure 4.28: Block diagram of the hybrid multi-frequency interrogator

79

4.3.2 Results

4.3.2.1 Wireless Demonstration

Hybrid Single frequency RFID Tag

A wireless communication setup is created for detecting a hybrid RFID tag operating at 915
MHz. The interrogator at 915 MHz is turned on with an output power of +7 dBm, which can be
received by the tag placed at 10 inches distance. The block diagram of the setup and the picture of
setup is shown in Fig.4.29.

Figure 4.29: Wireless setup for hybrid single-frequency RFID communication

The transmitted signal from interrogator is received by the hybrid RFID tag, which harvests
the RF energy and turn on the microcontroller. The logic inside the microcontroller decodes the
commands and respond by modulating the backreflected RF signal with a digital ID, followed with
the analog data packet by simply turning on the switch of the sensor integrated coupler to introduce
an additional phase shift. The backreflected signal with digital ID and analog sensor data is received
by the interrogator and demodulated using mixer. The output of the mixer is shown in Fig.4.30.
The digital bits are clearly visible with preamble, EPC, sensor type, sensing parameter, and CRC.

80

The continuous wave for RAD (175

) and SAD(175

) can also be distinguished easily with a
15

 time gap in between.

Figure 4.30: Demodulated hybrid data received at 915 MHz wirelessly from designed RFID tag

Figure 4.31: Snapshot of the hybrid single-frequency communication with digital information

The received data contains both digital and analog information. The interrogator decoded the
digital data using microprocessor and switched the receiver to oscilloscope for recording the analog

81

data. The digital information is sent to the computer and a snapshot of the received data is shown in
Fig.4.31. While the recorded analog data is shown in Fig.4.32. A Varactor diode is connected with
the coupler for imitating a capacitive sensor. An external 0V is applied across the Varactor, which
represents the sensor analog data. Due to 0V as input, no change is observed in the RAD and SAD
level that represents different phases of the received signal. Therefore, no capacitance change is
represented by the difference of RAD’s and SAD’s average voltage level, which is approximately
zero in this case.

Figure 4.32: Analog sensor data recorded using oscilloscope from 915 MHz hybrid RFID tag

A second measurement is made with 1.0V applied across varactor. The received digital data is
same for the same tag but the received analog data is changed due to the change in capacitance (or
phase). Both analog received data with 0.0V and 1.0V input are shown in Fig.4.33. Both plots are
shown together for observing the 3.3

 change in the received signal.

The explained process is repeated for a range of input voltages from 0.0V to 1.0V with a step
size of 0.2V. The digital information is same for the different sensing values. The received analog
data for all other four sensor values is shown in Fig.4.34.

The average voltage potential for 175

 data is calculated for both RAD and SAD. The procedure
for all 6 different voltage potential is recorded and the difference in reference (RAD) and sensor

82

Figure 4.33: Analog sensor data recorded using oscilloscope from 915 MHz hybrid RFID tag;
O.0V input (Red and 1.0V input (Blue))

Figure 4.34: Analog sensor data recorded using oscilloscope from 915 MHz hybrid RFID tag; (A)
0.2V, (B) 0.4V, (C) 0.6V, (D) 0.8V

83

potential (SAD) is shown in Fig.4.35.

Figure 4.35: Difference in reference (RAD) and sensor (SAD) data for different sensor voltage
across Varactor

A interrogator successfully decoded the 915 MHz hybrid RFID tag’s data. The small change
in phase due to the small voltage potential can be detected. The acquired data is linear and show
good correlation with the wired acquisition using VNA.

Hybrid Harmonic RFID Tag
A wireless communication setup is created for detecting a hybrid RFID tag operating at 434
MHz for downlink and 868 MHz for uplink. The interrogator with two harmonic antenna’s is turned
on with 434 MHz output frequency and +7 dBm output power, which can be received by the tag
placed at 10 inches distance. The block diagram and the picture of setup are shown in Fig.4.36.

The transmitted signal at 434 MHz from interrogator is received by a hybrid harmonic RFID
tag, which harvests the RF energy and turn on the microcontroller. The microcontroller holds the
digital logic to operate according to the received commands from the interrogator and respond
by modulating the backreflected harmonic RF signal at 868 MHz with digital ID and analog
sensor data. The backreflected signal with hybrid data packet is received by the interrogator and

84

Figure 4.36: Wireless setup for hybrid harmonic-frequency RFID communication

demodulated using an RF mixer. The output of the mixer is shown in Fig.4.37. The digital bits are
distinctly visible with preamble, EPC, sensor type, sensing parameter, and CRC. The continuous
wave for RAD (175

) and SAD (175

) is also successfully received with a 15

 time gap in
between.

Figure 4.37: Demodulated hybrid data received at 868 MHz wirelessly from designed RFID tag

The received data contains both digital and analog information. The interrogator decoded the
digital data using microprocessor and switched the receiver to the oscilloscope for recording the
analog data. The digital information is sent to the computer and a snapshot of the received data is

85

shown in Fig.4.38. While the recorded analog data is shown in Fig.4.39.

Figure 4.38: Snapshot of the hybrid harmonic-frequency communication with digital information

Figure 4.39: Analog sensor data at harmonic signal recorded using oscilloscope

A Varactor diode is connected to the hybrid coupler as a capacitive sensor. An external 0V
is applied across the Varactor, which represents the sensor analog data. Due to 0V as input, a
very small change is observed in the RAD and SAD level that represents different phases of the
received signal. Therefore, minimal capacitance change is represented by the difference of RAD’s
and SAD’s average voltage level, which is approximately zero in this case.

86

Figure 4.40: Analog sensor data at harmonic signal recorded using oscilloscope; 0.0V input (Red
and 1.0V input (Blue))

Figure 4.41: Analog sensor data recorded at 868 MHz using oscilloscope; (A) 0.2V, (B) 0.4V, (C)
0.6V, (D) 0.8V

87

A second measurement is made with 1.0
 applied across Varactor. The received digital data is
same for the same tag but the received analog data is changed due to the change in capacitance (or
phase). Both analog received data with 0.0
 and 1.0
 input are shown in Fig.4.40. Both plots are
shown together for observing the change in the received signal.

The explained process is repeated for a range of input voltages from 0.0
 to 1.0
 with a step
size of 0.2
. The digital information is same for the different sensing values. The received analog
data for all other four sensor values is shown in Fig.4.41.

The average voltage potential for 175

 data is calculated for both RAD and SAD. The difference

in RAD and SAD of acquired data is shown in Fig.4.42.

Figure 4.42: Difference in RAD and SAD of 868 MHz signal for different sensor voltage across
varactor

A interrogator successfully decoded the harmonic-frequency tag’s data. The small change in

phase due to the small voltage potential can be detected.

88

4.4 Summary

The hybrid RFID system is implemented to reduce power requirement for battery-free sensing.
A hybrid coupler based sensing mechanism is demonstrated that can shift the phase of the RF
signal according to the reactance of the sensor. An additively manufactured capacitive pressure
sensor is designed and used to show the capabilities of the power efficient phase modulation system
and wirelessly transmit the data. The elimination of ADC using hybrid coupler based solution is
optimal for power efficiency, but introduce other problems like; incompatibility with multiple tag
communication, sensitive to movement, etc. Therefore, the analog data is combined with a digital
data frame to allow multiple tags detection and data acquisition. The hybrid data frame requires less
power than its ADC based alternatives and also work with single or harmonic frequency operation
mode. The hybrid harmonic has better SNR than its single frequency counterpart, which allows
reliable phase extraction. Moreover, the differential reading between RAD and SAD makes the
data acquisition more robust against noise prone environment and multiple path reflections.

89

CHAPTER 5

ACTIVE SENSOR INTEGRATION WITH ADC-FREE RFID TAG FOR ULTRA-LOW

POWER OPERATION

5.1 Introduction

Lack of maintenance and proper protection were found to be the major cause of mechanical
failure. In order to continue a safe and reliable operation, some proactive measures are required that
can detect the failure before it occurs and sound an alarm for maintenance. Ultrasonic technique is
one such method for evaluating the mechanical properties of materials by analyzing the vibration
response.

Researchers are developing the vibration sensing technology for years that sets a standard for
measurement techniques that are best performed in lab settings or field environment. The ultrasonic
pulser-receiver systems are very popular for detecting time-of-flight and tracking amplitude changes
of received signal, which has traveled through the material under test. The vibrations are actuated
and detected by mounting a transducer, which change or acquires the lateral displacement, velocity,
and acceleration [141, 142]. It include elements like PZT, PVDF, etc., which can be considered
active sensors because it requires an external energy source for generating signals.

In NDE, the acoustic system plays a vital role for evaluating various materials from metals to
composites and polymers. The acoustic equipment’s used for NDE are very standard table-top,
bulky, and high power systems, which is easy to handle only in lab conditions and ideal for precise
measurements. On the other hand, the acoustic system used for SHM purpose has a very different
application area. The SHM requires both bulky equipment’s for long range and deep penetration
applications and a small IoT based system for monitoring tiny cracks, material loss, permeation,
etc. The current SHM system in place is required to avoid any potential down time by acquiring
continuous data and performing predictive maintenance.

Our focus for this research is specifically on small IoT based platforms. The small sensing

90

devices for SHM in large facilities like refineries and chemical plants are gaining momentum due to
their 24x7 monitoring capabilities while providing more data around the plant for multiple systems
simultaneously. The profound application and large market size has pushed a few key companies like
Texas Instruments to make specially designed integrated circuits for acoustic pulser and receivers
to combine it with the IoT based communication platform to acquire real-time data [143, 144].
However,these current IoT based maintenance systems in place are battery powered and consume
more than 4

 of power due to high speed ADC, which also requires additional maintenance and
has a very limited life cycle of a few years.

In this work, we propose a battery-free RFID tag based platform for actuating and detecting
acoustic signals without any ADC requirement. The primary advantage of such system is its
multidecade life time with no maintenance. Moreover, the costs for battery-free RFID tags are
significantly lower than battery-based counter parts. The lower sensor tag cost pave an economical
path to deployment at large scale and collect high resolution data around any facility.

The design of the proposed solution is discussed first with the active-sensing architecture and
its main components and finally a wireless demonstration is presented to show a working prototype.

5.2 Design

The active sensors like PZT’s are conventionally used to actuate and detect acoustic waves. The
PZT transducer simply needs to be connected to a continuous wave or pulse source for actuation
and either an ADC or an oscilloscope for detection. In cases when the received signal strength is
too low for the oscilloscope, an amplifier can be used to boost the signal. For battery-free RFID
tags, a waveform source for actuation and an amplifier for detection are not among the viable
options. Therefore, a new technique is required to feed and read the transducer while reducing all
power requirements. The architecture of proposed battery-free sensor tag with integrated PZT is
explained below.

91

5.2.1 Active Sensing Architecture

The fundamental architecture for the PZT integration is inspired from the hybrid harmonic-
frequency RFID tag, explained in the previous chapter. The primary reason for selecting the
harmonic frequency is due to its superior SNR performance and on-board nonlinear frequency
mixing element (NLTL). The basic functionality of the tag remains the same, where it receives a
continuous wave at 434 MHz and turn on the microcontroller. The microcontroller reflects back
the digital ID by switching the modulator with a GPIO pin. Similarly, another GPIO is used to
generate a square pulse of 2.5

, which is connected to a PZT. The PZT actuates according to the
input pulse and couples the mechanical vibrations into the attached steel material. The acoustic
waves travel down in the material and can be picked-up using a second PZT. The received signal at
PZT is mixed with the wirelessly received 434 MHz RF signal using NLTL, which also generates
the 2nd harmonic signal at 868 MHz. The output of NLTL will have the second harmonic signal
with modulated PZT data that can be transmitted back to the interrogator for demodulation and
detection. The schematic of the designed active sensor integrated hybrid harmonic RFID tag is
shown in Fig.5.1. The data frame for acoustic actuation and detection with RFID communication
is shown in Table.5.1.

Table 5.1: Data frame format for Digital ID with Acoustic Transducer Data Reply

P EPC ST SP CRC-16 AP DP

where AP is actuation pulse and DP is detected pulse. The digital part in the hybrid acoustic

frame is same as explained in hybrid harmonic RFID tag.

5.2.2

Integrating Acoustic Actuation Transducer

The battery-free RFID tag do not have a specific waveform generator for acoustic actuation,
therefore, a microcontroller is used to generate a short pulse of 1.8
 that can have a minimum width
of 2.5

 or 400
 

. Generating a short pulse is not a power hungry operation, the microcontroller
uses the same GPIOs to generate the backreflected signal at a similar data rate. A PZT (Piezo

92

Figure 5.1: Schematic of active sensor integrated hybrid harmonic RFID tag

Ceraminc Plate 7x8x0.2mm) from Steminc were used to test the acoustic generation in steel plate.
The square pulse profile generated by microcontroller is shown in Fig.5.2B. The signal is launched
into a steel sample plate and detected by a second PZT placed 70

 away, the setup is shown in
Fig.5.2A. The signal received by second PZT is recorded using an oscilloscope, shown in Fig.5.2B.

Figure 5.2: (A) Acoustic actuation and detection wired setup, (B) PZT actuation pulse signal
generated by microcontroller and detected signal profile with averaging

The microcontroller was used as a pulse source for PZT and the actuated pulse successfully
coupled to the steel plate. The PZT will only be actuated when the RFID tag responds to the

93

interrogator with a digital ID, therefore each pulse actuation and detection can be specified to a
predefined target. For the current wired experiment, the observed time of flight to travel 70

 in
steel is 25.5

.

5.2.3

Integrating Acoustic Detection Transducer

The wired detection shown in the previous section for testing purpose needs to be made wireless
as well. However, the battery-free RFID tag cannot afford any high data acquisition rate ADC.
Therefore, the acoustic signal needs to be analog modulated with backreflected signal for decoding
at the interrogator’s end. The signal generated by the PZT has very low strength, so biasing a
Varactor diode connected to a hybrid coupler will not be effective due to the low sensitivity. An
alternative method for modulating low power signals is using a mixing component, which is present
on-board in the form of NLTL. The PZT is connected at the input port of the NLTL to the 434 MHz
receiving antenna, as shown in Fig.5.3.

Figure 5.3: Schematic of the PZT integration with non-linear transmission line

A wired setup is created for testing the analog modulation of PZT signal on the harmonic
signal generated by the NLTL. The PZT transducer integration is tested with two types of actuation
signals; continuous wave and pulses.

94

5.2.3.1 Acoustic continuous wave detection

A continuous wave of 200 kHz with 1.8 Vpp amplitude is applied at the actuation transducer,
which travels through the steel plate and can be detected at the second receiving transducer. The
block diagram of the test setup is shown in Fig.5.4A. The acoustic signal received by the PZT is
modulated on the harmonic signal, which can be transmitted back to the interrogator. The analog
modulation of the harmonic signal is shown in Fig.5.4B. The modulated side bands are clearly
visible at 200 kHz offset, with very low mixing losses.

Figure 5.4: (A) Test setup to check the analog modulation capabilities of RFID tag, (B) Harmonic
output of the NLTL is modulated with 200 kHz acoustic signal detected by PZT

5.2.3.2 Acoustic pulse detection

The continuous wave detection setup was created for sanity check, but it does not represent the
practical scenario for acoustic-based NDE and SHM. The practical applications use pulse for
evaluating materials and structures. Therefore, a second test setup is created for modulating and
detecting the pulse signal. A 2.5

 wide pulse of 1.8 Vpp is generated using a waveform generator
that is picked up by the receiving PZT and modulated on the 868 MHz harmonic signal. The
harmonic signal with acoustic signal pulses as side bands cannot be captured using a spectrum
analyzer, therefore, an additional setup for demodulation using RF mixer is used. The designed

95

test setup is shown in Fig.5.5A. The output of the NLTL is amplified by 25 dB and fed into the
mixer for demodulation, with an external RF source of 868 MHz. The demodulated acoustic pulse
is shown in Fig.5.5B with same ToF as previously measured by the oscilloscope.

Figure 5.5: (A) Test setup to check the acoustic pulse modulation capabilities of RFID tag, (B)
Demodulated acoustic signal detected by PZT

The NLTL shows the capabilities of detecting small acoustic signals and modulating back on
the harmonic signals. The next step is to combine the acoustic actuation and detection setup with
the wireless antennas for demonstrating a complete battery-free solution for the first time.

5.3 Results

5.3.1 Wireless Acoustic Actuation and Detection Demonstration

The wireless RFID tag as shown in Fig.5.1 is assembled with necessary tag software changes to
accommodate acoustic pulse generation and data transmission. The wireless setup with battery-free
RFID tag, sample steel plate and interrogator is shown in Fig.5.6.

5.3.1.1 Detection of Acoustic Time-of-Flight in Steel

The interrogator is started at 434 MHz output signal with an amplitude of +7 dBm. The 434
MHz signal wirelessly charge the battery-free RFID tag in field-of-view, which can also respond to

96

Figure 5.6: (A) Wireless hybrid harmonic RFID tag with integrated PZT setup for acoustic actuation
and detection, (B) Picture of the practical setup

interrogator’s commands accordingly. After the tag gets acknowledged, it backreflects the digital ID
information as we have seen previously for hybrid harmonic-frequency communication. Moreover,
for the PZT transducers, the microcontroller generates an actuation pulse of 2.5

 at the end of
digital ID transmission and switch on the backreflection for additional 175

 for the PZT to receive
and modulate the acoustic signal. The interrogator’s 868 MHz harmonic antenna receives the
harmonic modulated signal with digital ID and analog acoustic data. The digital information is
extracted by the interrogator’s microprocessor and acoustic data is directed towards the oscilloscope
as shown in the previous chapter. The wirelessly received and demodulated acoustic signal is shown
in Fig.5.7.

The wirelessly measured time-of-flight for acoustic signal is 25.5

 as measured before in the
wired configuration. The detection results also proves the viability of such a system that can operate
without battery and able to perform such complex tasks of detecting vibrations.

5.4 Summary

The acoustic actuation and detection demonstrates the capability of battery-free hybrid harmonic
RFID tag and multifrequency interrogation system. The PZT integrated hybrid RFID tag can

97

Figure 5.7: Wireless acquired and demodulated harmonic signal for acoustic signal

generate the actuation pulse without any heavy power requirements. The receiving PZT and
modulating NLTL are very sensitive towards very small amplitude of vibrations. The 434 MHz and
868 MHz operation allows real-time battery free communication and avails enough bandwidth for
detecting acoustic pulses. Additionally, the harmonic operation offers a high SNR that increases
the reliability of the communication system.

98

CHAPTER 6

CONCLUSION AND FUTURE WORK

6.1 Summary

The primary challenges of current NDE and SHM methods are identified, which includes lower
sensitivity towards dielectric evaluation, limited range due to RF clutter, lack of versatility in
sensor integration, scalability and maintenance issues for monitoring applications. Each challenge
is discussed in detail for various applications with multiple available techniques. Alternative
solutions are proposed to overcome each challenge by using (a) wireless high-Q resonator sensing
for dielectric evaluation, (b) battery-free RFID system for eliminating maintenance requirements
and harmonic communication for enhanced range, (c) hybrid data packet for sensor versatility while
achieving power efficiency and eliminating ADC.

The high-Q resonator is achieved by wireless coupling in near-field. The capacitive sensor in
LC resonator is an unshielded parallel plate configuration that change its capacitance according to
the near by dielectric properties. The designed probe was tested on both composite and polymer
samples that detected surface and subsurface defects at sub-mm resolution. A 1 × 4 array probe is
designed with the same fundamentals concept to resonate at 4 different frequencies for simultaneous
actuation and data read.

A harmonic-based RFID system is designed to extend the communication range of conventional
RFID by eliminating the RF clutter/self-jamming issues. The harmonic system provides better SNR
and opens a window of opportunity to make more sensitive RFID ICs.

A pressure sensor is designed using additive manufacturing techniques to demonstrate the power
efficient wireless sensing capability. The power efficient analog sensing is merged with digital ID
for eliminating high power requirements. The analog data is successfully extracted from the hybrid
data packet with good linearity.

Ultrasonic methods are very popular and effective as NDE and SHM technologies. But it has

99

never been considered as a power efficient technique. This work demonstrated a PZT actuation
using a battery-free RFID tag and acoustic detection without the use of an ADC by wirelessly
transmitting the information to the interrogator. The hybrid harmonic RFID tag with digital ID and
analog PZT data has opened new possibilities for embedded sensing.

The wireless sensing for both NDE and SHM can take on the current challenges. The hybrid
harmonic RFID with passive and active sensor integration can all together solve the excessive
maintenance problem of battery-based IoT infrastructure. Moreover, the lower cost of a battery-
free system allows large-scale deployment for high-resolution data acquisition to allow AI based
predictive maintenance.

6.2 Future Work

In this thesis, various novel sensing and acquisition concepts have been presented that can be
extended further for increasing the technology readiness level. The key areas that require focus are
(a) imaging using an array, (b) miniaturization, and (c) ultrasonic material characterization.

6.2.1

Imaging using an array

A novel concept of resonant array design is shown, where all elements can be simultaneously
excited and recorded. This work shows the response of array probe, but lacks the imaging results.
Therefore, a faster imaging scan using the designed array is the logical next step. The commercially
available systems like Eddyfi Ectane can be used for actuation and detection as it has an inbuilt
capability of exciting at multiple frequencies at the same time.

6.2.2 Miniaturization

The battery-free RFID sensing technology using hybrid data packet is fundamentally power efficient
and range enhancing. However, the current implementation using distributed components is big
and limits its application. The miniaturization can be started with antennas, which takes most of the
area. Antennas can also be printed on flexible substrate for easy handling, embedding process, and

100

light weight. Second, the energy harvester and microcontroller section can be significantly reduced
using a more efficient single chip RF harvester and low pin count, extremely low power logic
controller. Overall reduction in tag size and component count will also reduce power requirement
and make the hybrid RFID tag more sensitive.

Miniaturization of the interrogator can also help with the portability of the whole system.
Integrated RF sources, super-heterodyne mixing, and onboard high precision ADC can take this
technology to a whole new level.

6.2.3 Ultrasonic Material Characterization

The battery-free technology for wireless actuation and detection of ultrasonic signals via steel has
a lot of potential for industrial and academic applications. The time-of-flight data can be used
to calculate the velocity of an ultrasonic wave through the medium, which has a relation with
mechanical properties like young’s modulus and density. Moreover, the recovered pulse also holds
the harmonic information and can be extended towards non-linear ultrasonic.

101

APPENDICES

102

APPENDIX A

AN RF BACKSCATTERER SYSTEM FOR REAL-TIME MULTI-SENSING

APPLICATIONS

A.1 Introduction

Both amplitude and phase modulation techniques are affected by noise in the propagation
environment. Furthermore, in case of direct phase modulation of the sensor information, the
distance between the tag and the interrogator should be calibrated in order to avoid the ambiguity
towards the change in received signal strength. The best suited energy efficient technique for analog
modulation with higher immunity to noise is the frequency modulation technique. In literature,
only one work has been demonstrated using the analog frequency modulation for representing
RFID sensor data [145]. In this work, a Colpitts oscillator is used to generate different single tone
low frequency signals that are modulated onto the backscattered signal based on the change in
capacitance of the resonator. Additionally, the work shows that the rectification circuit using two
zero-bias Schottky diodes was able to generate enough DC power to drive a 150 kHz oscillator
with a received power of -22 dBm, whereas, the state-of-the-art CMOS based rectifiers requires at
least 10 stages of rectification to generate DC power (10 
W) from -16 dBm received power [146].
Furthermore, a linear relationship is shown between RF power in log scale and output DC voltage
in linear scale but the relationship stands valid only in log-log scale [147]. Hence, there is a need
for developing an efficient high-performing passive RFID multi-sensing system with improved read
range for IoT based sensing platforms.

In this chapter, a passive wireless multisensing system is presented with a power efficient
analog frequency modulation scheme for long-range real-time communication. The proposed
design exploits the available 300 kHz bandwidth of a single UHF RFID channel to modulate the
multiple sensing information for efficiently transmitting the sensor data, eliminating the need for
power hungry analog-to-digital conversion. The query signal from the interrogator activates the

103

tag and the sensor information is backscattered as a function of frequency. As an example, a
multisensing tag with integrated temperature and luminescence sensors are shown with a read
range of 10 feet and a sensitivity of -6 dBm. Additionally, a study for integrating the digital ID
along with the analog sensor information is presented, in which the digital identification bits and
analog sensor information are confluenced into a single hybrid data packet. The presented system
is compatible for communicating multiple sensor information passively over a long range for IoT
based applications.

A.2 Principle of Operation

The passive multisensor system consists of a custom designed UHF RFID interrogator and a
multisensor tag. The communication between the reader and the multisensor tag is initiated by
the interrogator by sending a continuous single frequency RF query signal within the range of
902-928 MHz. The multisensing tag consists of four units, a receiving antenna, a rectifying unit,
a multisensing module with an oscillator, and a modulation unit. The multisensor tag antenna
receives the RF query signal and feeds it to the rectifier through a matching circuitry that generates
the required DC bias to activate the frequency oscillator.

The oscillator generates specific single tone frequency signals to represent each of the sensing
parameters (for example, temperature, luminescence, humidity, etc) and a switch is used to select
between different sensors with a specific periodicity. A low frequency source or an oscillator is used
to generate a unique low frequency signal within a limited RFID channel bandwidth, corresponding
to each sensing parameter value for a single sensor element. A number of sensors can be integrated
to the same RF tag based on the total available bandwidth. Two example sensors are integrated
with the proposed platform to demonstrate a test prototype. Two sensors, a temperature sensor and
a light sensor, the first sensor; temperature (-25◦C to +25◦C) is assigned a bandwidth of 25 kHz
within the frequency range of 75-100 kHz, with every 0.5 kHz representing a degree in temperature
change. The second sensor, luminescence (0 fc to 20 fc) is assigned a similar 25 kHz bandwidth
within the frequency range of 125-150 kHz. The available single channel bandwidth for UHF RFID

104

communication is 300 kHz, and with an example bandwidth requirement of 25 kHz per sensor, up
to 6 different sensors can be integrated with the RF tag simultaneously. The low frequency signal
generated by the oscillator is modulated on to a high frequency carrier backscattered by the tag. The
backscattering modulation is performed using a MOSFET switch and the interrogator receives the
modulated back-scattered signal and demodulates the information to extract the sensing data. Now,
another switch is required to select one sensor at a time as the sensor information is transmitted
sequentially. The period of sensor switching is fixed at 6 kHz (∼160 
s). The received signal at the
interrogator will contain the sensor information from the first sensor within the first time period,
followed by the second sensor within the second period, and the sequence repeats periodically.
The schematic of the designed tag with multisensing module, modulator, tag antenna is shown in
Fig.A.1.

Figure A.1: Schematic of a passive wireless system with multi-sensing RF tag and Interrogator

A.3 Design

The passive wireless multiple sensing system is composed of an RF tag and an RF interrogator
as shown in Fig.A.1. The design details of each fundamental block of the multisensor RF tag and
RF interrogator are discussed in this section.

105

A.3.1 Antenna Design

An UHF RFID tag antenna is designed to transmit and receive the information from the RF
interrogator. A smaller footprint antenna is designed for the RF tag to operate in a frequency
band from 902-928 MHz by meandering the conventional dipole antenna. ANSYS HFSS® (2019,
ANSYS, Inc., Canonsburg, PA, USA) is used to simulate the design of the antenna with Roger’s
4350 (

 = 3.66, and tan
 = 0.003) dielectric substrate with thickness of 1.52 mm. The antenna is
fabricated using the conventional photolithography process. The schematic of the antenna along
with its dimensions and its frequency response is shown in Fig. A.2A. The measured frequency
response matches closely with the simulated frequency response, as shown in shown in Fig. A.2B.

Figure A.2: UHF RFID Antenna:
Simulated and measured frequency response

(A) Schematic of the dipole antenna and dimensions, (B)

Table A.1: Dimensions of the 915 MHz antenna

Parameter

Dimensions (mm)

Parameter

Dimensions (mm)

a
20.0
e
1.0

b
3.75

f
8.0

c
12.0
g
16.0

d
4.5
h

19.75

106

A.3.2 RF to DC converter

The antenna receives the RF signal from the interrogator as a single frequency continuous wave
at the chosen UHF RFID frequency of 908 MHz. The received signal is backscattered with the
modulated sensor information generated by the oscillator coupled with the multisensing module.
The oscillator requires a DC supply to operate, which is harvested from the received RF signal from
the interrogator using an RF-to-DC converter (rectifier). The implemented converter is a five-stage
Dickinson charge pump rectifier that can convert a range of input RF power (0 dBm to -6 dBm)
to DC output voltage. A Zener diode is connected in parallel to the output of the charge pump
circuit to limit the maximum rectifier’s up converted DC voltage to 5 V. A high efficiency regulator
(TPS79718) from Texas Instruments is used to regulate the up converted voltage signal to 1.8 V in
order to supply the required DC power for the oscillator.

The designed rectifier with voltage regulator is tested with a range of received RF power from 0
dBm to -7 dBm. The output voltage of 1.8 V is connected to a variable load resistance to calculate
the total available DC power at the output of RF-to-DC converter unit. The corresponding output
DC power range from 350 
W to 10 
W for the input RF power from 0 dBm to -7 dBm as shown
in Fig.A.3.

Figure A.3: (A) Schematic of RF-to-DC converter, (B) Input RF power vs Output DC power

The minimum power required to operate the frequency oscillator at 150 kHz is 30 
W, as
explained in the next section, which can be harvested from a minimum input RF power of -6 dBm,

107

making it the minimum operating limit or the sensitivity level of the designed multisensing RF tag.

A.3.3 Multi-sensing Module

The multisensing module generates a wide range of frequencies according to the number of sensors
coupled to the low frequency oscillator through a single-pole-multiple-throw (SPMT) switch.
The oscillation frequency range is defined by the available channel bandwidth of the RFID system.
Within the available bandwidth, a number of subbands are allocated to each sensor element coupled
to the oscillator. The different sensor elements change the oscillation frequency according to the
change in their respective physical parameter (for example, temperature, luminescence, humidity,
etc). Although a SPDT switch (TS5A2053) was used for the prototype as the prototype uses two
example sensors, multiple sensors can be connected by changing the switch. The switch couples
the oscillator to different sensing elements by selecting a single sensor at a given point of time. The
toggle frequency should be fast enough to sweep through all the coupled sensors in minimum time
while maintaining the integrity of the output frequency signal from the oscillator, which is defined
by the oscillator’s stability time. In this work the toggle frequency is set to 6 kHz for selecting
in between temperature and luminescence sensors, which is found experimentally to get the stable
oscillations in minimum time. The toggle frequency of 6 kHz corresponds to a single sensor read
time of approximately 160 
s.

The multisensing frequency source is realized using an ultra low power oscillator IC (LTC6906)
from Linear Technology with an operating voltage range from 1.8 V to 5.5 V. In the proposed passive
wireless multiple sensing application, the minimum power consumption is paramount, for which
an experiment is performed to generate a fixed 100 KHz signal while feeding the oscillator with a
range of input voltages from 1.8- 5 V, and the corresponding current consumption plot is shown in
Fig. A.4A. A linear relationship exists between the current and the voltage, and a minimum current
is drawn for the minimum input required voltage of 1.8 V.

A ultra low power oscillator IC that generates a single tone frequency based on the value of the
resistance across it is chosen due to the wide availability of a variety of resistive sensors. The selected

108

Figure A.4: (A) Input voltage vs Input current at 100 kHz oscillation frequency, (B) Oscillation
frequency in kHz vs required Input power in microwatts

oscillator can generate frequencies from 10 kHz to 1 MHz according to the coupled resistance.
The oscillator’s power consumption increases with the increase in frequency, an experiment is
performed to extract the power requirements to generate a range of frequency signals from 75- 325
kHz. A variable resistor is used to change the oscillation frequency, and the required minimum
input voltage of 1.8 V is supplied to the oscillator. The oscillator’s required DC power is calculated
by measuring the current consumption at 1.8 V for different oscillation frequencies. The plot of
oscillation frequency vs required DC power is shown in Fig. A.4B.

The oscillator is capable of generating a frequency as low as 10 kHz, where the DC power
requirements would be even lower. But in this paper, the frequency operation lower than 75 kHz
is avoided intentionally due to the presence of phase noise skirt in the UHF carrier frequency
signal. Hence, the lower frequency signals may get affected by the higher noise level and introduce
difficulties in detection by the interrogator. The lower frequency modulation can be used with the
other carrier signals with a very low phase noise level. The upper bound on modulation frequency
is set by the channel bandwidth of the UHF RFID protocol, which is 300 kHz or ±150 kHz on both
size of the carrier wave. According to the plot shown in Fig. A.4B, the maximum power required
to generate 150 kHz is 30
W. The circuit diagram of the multisensing module with modulator is

109

shown in Fig.A.5.

Figure A.5: Circuit diagram of the designed multi-sensing module

A.3.4 Modulator

The frequency signal generated by the oscillator according to the different sensor values is modulated
to a high frequency carrier signal for information transmission. A 25 kHz band is assigned to each
sensor with a guard band, also 25 kHz, in between each sensor bands, as shown in Fig. A.6A. The
frequency band for the first sensor is from 75 kHz to 100 kHz, and the second sensor is from 125
kHz to 150 kHz. The UHF RFID band is used for the communication, a frequency of 908 MHz
is selected to modulate, for instance, 100 kHz and 150 KHz signals representing first and second
sensors, respectively.

A MOSFET (BF904) from NXP semiconductors is used to modulate or mix the low frequency
information signal with the high frequency carrier signal. First, the multisensor RF tag is tested in
a wired setup. A 10 dB directional coupler is used to feed the 908 MHz signal across the drain and
source pin of the MOSFET, and the low frequency signal from the oscillator is fed into the gate
pin. The back-scattered signal with the sensor modulation is measured at the coupled port of the
directional coupler as shown in Fig. A.6B. The received frequency signal of 100 kHz and 150 kHz

110

belongs to the preallocated sub-bands, which represents a physical parameter with a specific value.
For validation purpose, the sub-bands are generated using variable resistors.

Figure A.6: Sensing information modulation: (A) Schematic of the modulation scheme and (B)
Modulated sensor information of 100 kHz and 150 kHz on 908 MHz carrier frequency

111

A.3.5 RF Interrogator

The wireless communication with a multisensor RF tag requires a custom designed RF interrogator,
which transmits and receives the UHF RF signal. The interrogator consists of an RF source that
can generate UHF continuous wave, which is radiated using a commercial patch antenna with 6
dBi gain. The multisensor RF tag receives a single frequency signal from the tag antenna and
backscatters the information in the form of modulated RF signal, which is received by using the
same transceiver antenna. A circulator is connected at the front end of the interrogator to isolate
the forward and backward propagating signal to and from the transceiver antenna. The received
signal is demodulated and the sensor response is obtained. A RF mixer was used for demodulation
by feeding a reference UHF signal into LO PORT and the back-scattered modulated signal into
the RF port after 36 dB amplification using a LNA. The demodulated signal is acquired at the
IF PORT of the mixer. The schematic of the RF interrogator is shown in Fig.A.7, where the RF
interrogator transmit and receives information to and from the multisensor RF tag. The complete
RF interrogator was assembled using commercial RF parts from Minicircuits.

Figure A.7: Schematic of the RF interrogator radiating +36 dBm EIRP and visualizing the back
reflected sensor information using a spectrum analyzer and an oscilloscope

112

A.4 Results

For demonstrating the real-time passive wireless multisensor platform, the experiments are
performed in two different configurations: Wired for validation and wireless for demonstration.
The wired validation is performed to check the effects of input RF power on side band power, a
measure of resistance required to generate the frequency from 75 kHz to 150 kHz, and the rela-
tionship of temperature and luminescence sensor with the resistance. The wireless demonstration
of temperature and luminescence shows the successful communication from RF tag to interrogator
by visualizing the demodulated sensor information using spectrum analyzer and oscilloscope in
frequency domain and time domain, respectively.

A.4.1 Wired Validation

The performance of multisensor RF tag is tested by establishing a direct wired connection through
the antenna port. Similar setup as described earlier, a 10 dB directional coupler is used to feed
the UHF RF power into the multisensor RF tag and the reflected signal is acquired at the coupled
port using a spectrum analyzer. First experiment is performed to test the minimum input RF power
required to activate the multisensor RF tag and the effects of variable input RF power on on the
reflected modulated side band power. The oscillator is set to generate 100 kHz and 150 kHz
frequency signal, which represents the information from the first and second sensor, respectively.
The minimum RF power required to activate the multisensor RF tag is found out to be -6 dBm
at which the oscillator generates a frequency up to 150 kHz and successfully modulates it on the
carrier signal. The back reflected power at the modulated side bands do not vary with the input RF
power range from 0 to -6 dBm, as shown in Fig. A.8, as the output voltage level of the oscillator is
fixed at 1.8 V
 
 according to the input DC supply.

Second set of experiments are performed for validating the resistance required to generate the
frequency range in the available bandwidth from 75-150 kHz. A variable resistor is used to change
the oscillation frequency. The decreasing resistance range of 440 k to 340 k is required for 75

113

Figure A.8: Wired validation: Received side band power of a carrier signal due to first and second
sensor with varying input power

kHz to 100 kHz and 260 k to 220 k for 125 kHz to 150 kHz. Another advantage of using such a
high resistance reduces the power consumption significantly. A guard band from 341 k to 259 k
(101 kHz to 124 kHz) is adopted for easier classification (differentiation) and detection of different
sensor type and data, respectively.

The resistance based temperature and luminescence sensors are selected for demonstrating the
real-time passive wireless multi-sensing capability. A commercial temperature sensor IC [148] is
selected as first sensor element that is able to detect a range from 25
C to 60
C, and a commercial
light-dependent resistor (LDR) [149] is selected as the second sensor element that is able to detect
the luminescence from 0-20 fc. The temperature sensor IC changes its resistance from 100 k to
8 k for the given temperature range, which is tailored to 440 k to 348 k by adding a series
resistance of 340 k in order to generate the respective oscillation frequencies with in the range
from 75-100 kHz. Similarly, for LDR, the resistance range is tailored to generate the resistance
range from 260 k to 220 k for a given luminescence range while generating the frequency from
125-150 kHz. The temperature and luminescence relation with frequency is shown in Fig. A.9.
As shown in Fig. A.9, the linear sensitivities are respectively 0.53 (kHz/
C) and 0.862 (kHz/fc)

114

for the temperature and light sensor. The R-square value for the linear fits is 0.84 and 0.464 for
temperature and light sensors, respectively. However, the R-square value improves to 0.987 and
0.711, respectively with second-order polynomial fit.

Figure A.9: Frequency change of temperature and luminescence sensors (A) with linear fit and (B)
with second order polynomial fit

A.4.2 Wireless Demonstration

A wireless communication in between the RF interrogator and the multisensing RF tag is demon-
strated by transmitting an RF power of +9 dBm using a 6 dBi gain patch antenna. The operating
frequency is 908 MHz and multiple sensor tag measurements are acquired with 6 to 14 inches of
separation from the interrogator’s antenna. Prototype of the passive RF tag with PCB assembly
and antenna is shown in Fig.A.10.

The wirelessly received power at the RF tag with varying read range is experimentally measured
and compared in Fig. A.11 with the ideal received power using Friis’s transmission equation. The
antenna’s interference from nearby objects, multi-path propagation, and noise in the system effects
the carrier’s received power and introduce a deviation of ±1.5 dBm from ideal measurements. The
multisensor RF tag, backscatter frequency modulated sensor information, and the received side

115

Figure A.10: The prototype of RF Tag with designed antenna, integrated sensors, and assembled
PCB

band power for both sensors are shown in Fig.A.11. The received side band power decreases with
the increase in interrogator-tag separation. The RF tag receives the required minimum RF power
of -6 dBm at a distance of 12 inches, which is the communication range with 9 dBm radiated
power. An additional experiment is performed to test the maximum read range while transmitting
the maximum allowable EIRP of 36 dBm, with the sensitivity of -6 dBm, the multisensor RF tag
is able to communicate the sensor information up to 10 ft.

The temperature and luminescence sensors reflect back the information signal in the form
of modulated side band frequencies. The received frequency for temperature and luminescence
sensor are 82.34 kHz and 149.59 kHz, which corresponds to 27.8
C and 13 fc. The temperature and
lighting conditions of the experimenting lab are verified using a commercial temperature and light
measuring instrument from URCERI. The frequency spectrum of the received demodulated signal
is shown in Fig. A.12A. The frequency signal (sensor data) peaks are clearly visible, the frequency
modulation of the sensor data is immune to any multi-path interference or noise in the propagation.
The time domain representation of the demodulated RF signal is shown in Fig. A.12B, where the
information signals from the first and second sensors are repeating in time with a time period of

116

Figure A.11: Wireless demonstration with 9 dBm transmitted power at 908 MHz: Received RF
power at multisensor tag vs distance and backscattered

approximately 160 
s. The temperature and luminescence signals with 82.34 kHz and 149.59 kHz
frequencies, respectively, are shown in zoomed section. The demodulated signal is amplified using
a 40 dB low frequency amplifier before acquiring the data using an oscilloscope.

The sensors used in the system are analog in nature. Hence, the minimum resolution of the
sensors depends on the measurement instrument used at the interrogator. For a standard oscilloscope
with 1% measurement accuracy for bandwidth up-to 200 kHz, the minimum resolution will be 2.6

C for temperature sensor and 1.2 fc for the photosensor.

A.5 Discussion, Challenges and Limitations

A passive multiple sensing platform is designed, developed and demonstrated using a power
efficient analog frequency modulation. The sensing system operates in UHF RFID band from 902-
928 MHz and back-scatters the multiple sensor information in the form of frequency modulation.
The ultra low power frequency oscillator is connected on the board, which changes its oscillation
frequency according to sensor information. In this work, a resistance based oscillator is used for
multisensor module. The oscillator can be designed with a combination of RLC elements where

117

Figure A.12: Wireless demonstration with 9 dBm transmitted power at 908 MHz: (A) Frequency
spectrum of the demodulated sensor data, and (B) Received sensor information in time domain
with a switching frequency of 6 kHz (∼ 160 
s)

the resonance can be changed by varying the resistance, capacitance, or inductance. The operating
frequency range is limited by the total available bandwidth in which a number of subbands are
allocated to different sensor types. To integrate a large number of sensors within a single channel
bandwidth, the size of subband and guard band can be reduced. For operating at a lower frequency,
the phase noise issues can be eliminated using harmonic communication method [87, 150]. The
system’s reliability depends on the voltage level of the oscillator, switch, and sensors. As long as
the tag obtains its minimum power, the regulator will maintain its constant voltage level ensuring

118

proper functioning of all components. The interrogation distance of the system is limited by the
DC power requirement of the multi-sensor RF tag that can be improved by implementing a more
efficient RF-to-DC conversion mechanism.

There can be interference among the sensors when used together in the system. However,
the interference is primarily due to the leakage through the switch selecting one sensor at a time
between the both. According to the datasheet, the SPDT switch has isolation of 68 dB for switching
clock operating below 10 MHz, which means the desired sensor signal level will be 68 dB above
the undesired sensor signal. To further reduce the interference, a higher isolation switch should be
used. Other challenge of the passive multiple sensing system is to communicate with multiple RF
tags simultaneously and extract the multisensor information from each tag. It requires an additional
information layer to classify different multisensor RFID nodes and extract the information. The
multiple tag communication protocol is a well studied and a number of efficient circuitries are
available in the literature [49, 134], but digital sensing is very power inefficient due to ADC
requirements.

A.6 Summary

A passive multisensing RF system is shown for real-time communication with the multiple
sensors coupled to an RF tag. The demonstrated multisensor RF tag is able to communicate with an
RF interrogator up-to a read range of 10 ft. The RF tag uses an oscillator to generate a unique single
frequency signal corresponding to the temperature and luminescence sensing values, and transmit
the information by frequency modulating a 908 MHz RFID carrier signal. Each of the sensors’
information is modulated within the available channel bandwidth of 300 kHz at the pre-allocated
frequency range from 75-100 KHz and 125-150 kHz. The back-scattered frequency modulated
signal is immune to noise or interference due to multi-path propagation, and power efficient as it
eliminates the requirement of any analog to digital conversion. The demonstrated analog frequency
modulation based passive RF system shows an efficient, real-time and long-range multi-sensor
information transmission method for IoT based applications.

119

APPENDIX B

IDENTIFICATION AND SENSING SYSTEM (ISS): AN OPEN SOURCE RFID

PLATFORM

B.1 Introduction

In this chapter, we present ISS; an RFID platform for both single band (902-928 MHz) and
harmonic (433-434 MHz) band operation. The platform operates on modified EPC Gen2 RFID
protocol to accommodate identification and sensing together. The ISS supports communication
with digital sensor tags, which makes it ideal for passive impedance (resistance, capacitance, and
inductance) measurement. The working ISS is demonstrated by wirelessly communicating with
a digital sensor. The custom-designed interrogator and tags for ISS wirelessly communicate with
single or multiple sensor tags in its field of view. Additionally, a performance evaluation study is
shown for single vs harmonic frequency operation.

B.2 Design

The design section of ISS comprised of three unique hardware elements; two types of RFID
tags and an RFID interrogator. The first designed RFID tag uses a conventional single-frequency
RFID communication mechanism, which is required to study well known phase noise issues.
Second, RFID tag uses a harmonic communication mechanism to overcome the limitations of the
conventional system. At last, a multifrequency interrogator is designed to communicate with both
single-frequency and harmonic-frequency RFID tags. All three elements are discussed in detail
below.

B.2.1 Digital Single-frequency RFID Tag

The design of a digital single-frequency RFID tag has been widely studied. The hardware imple-
mentation of this work is highly inspired from the Wireless Identification and Sensing Platform

120

(WISP), which acts as a base for this new development work [86, 135]. The software providing
logic to ISS is developed from scratch to accommodate new requirements, like; custom interroga-
tor, variable data rate, timing, etc. Therefore, we called it modified EPC Gen2 RFID protocol.
The hardware implementation and software stack for tag are discussed in following sections. The
schematic of the designed single frequency RFID tag is shown in Fig.B.1.

Figure B.1: Schematic of the digital single frequency tag for receiving and transmitting at 915 MHz

B.2.1.1 Modified EPC Gen2 RFID tag protocol implementation

The RFID tag protocol is divided into two main parts; receiver and transmitter. The receiver part
can decode the digital commands sent from the interrogator and respond accordingly. The general
EPC Gen2 protocol for RFID requires a tag to implement 11 mandatory commands for commercial
use. However, we only implemented 6 (Query, QueryAdjust, QueryRep, ACK, NAK, Req_RN)
for the research and development ease, which are enough to have a reliable communication with
multiple RFID tags in the field of view. The RFID tag holds one of the states (Ready, Arbitrate,
Reply, Acknowledged), which is switched according to the received command. When a tag will
reach "Reply" state, it prepares for transmitting identification or sensing information.

The transmitter part can either operate in ID only or ID with sensor configuration. If operating
in ID-only mode, the tag calculates CRC-16 for the given EPC and appends together all components

121

required according to the data frame shown in Table.B.1.

Table B.1: Data frame format for Digital ID Reply

P EPC CRC-16

Where P is Preamble, EPC is Electronic Product Code (40-bits), and CRC-16 is Packet Cyclic

Redundancy Check (16-bits).

On the other hand, if the tag has an integrated sensor, it uses ADC to convert the sensor data

into digital bits and append it together according to the data frame shown in Table.B.2.

Table B.2: Data frame format for Digital ID with Sensor Data Reply

P EPC ST SP SD CRC-16

Where, P, EPC, and CRC-16 are same as above. Additionally, ST is Sensor Type (8-bits), SP is

Sensing Parameter (4-bits), SD is Sensor Data (12-bits).

The preamble and product code are used as defined in EPC Gen2 protocol [151]. An example
EPC: 0×E200A10090 is used in this paper for demonstration. The implementation list of sensing
parameters and sensing type is shown in Table B.3 and Table B.4. The codes shown in this table
will be used throughout this chapter for consistency.

Table B.3: 4-bit codes for different sensing parameters

Sensing Parameters

Resistive
Capacitive
Inductive
Voltage

4-bit Code

0000
0001
0010
0011

According to this proposed format, 16 different types of sensing parameters and 256 different

sensors can be coded as prior knowledge, so each sensor tag can be characterized correctly.

The binary sequence for the data frame is encoded using FM0 (Miller method) technique before
transmission for reliable data transfer as described in EPC Gen2 RFID protocol. The default

122

Table B.4: 8-bit codes for different sensor types

Sensor Type
Temperature

Pressure

pH

Humidity

8-bit Code
00000000
00000001
00000010
00000011

reply rate is 60 kbps with encoded bits, which can go up-to 106 kbps. At higher data rate, the
power requirements for the computation also increase. The power harvesting mechanism and ASK
modulation mechanism for transferring bits is explained below.

B.2.1.2

915 MHz Energy Harvester

The single frequency operation works in a frequency range from 902-928 MHz. The schematic of
the designed energy harvester is shown in Fig.B.2A. The on-board energy harvester circuit converts
the RF signal into DC signal for executing the protocol stored in the microcontroller. The input
impedance of the energy harvester is a function of input power, therefore, a sweep from -2 to -10
dBm is performed to ensure good matching with wide range of input power levels. The measured
frequency response of the designed energy harvester with power sweep is shown in Fig.B.2B.

Figure B.2: (A) Schematic of the designed energy harvester, (B) Measured frequency response of
the energy harvester with input power sweep

123

B.2.1.3 ASK Modulator

An ASK modulator is used to backscatter the RF signal from the tag with the ID and sensor data
bits. An ultra low power SPST RF switch (HMC550) is used to introduce an impedance mismatch
that translates into an amplitude shift of the backreflected signal. A simple setup with RF circulator
and designed RFID tag shown in Fig.B.3A for checking the modulation signal. The port 1 of the
circulator is connected to RF source that feeds the RFID tag at port 2 and the reflected signal from
the tag is read by an oscilloscope from port 3. The acquired back reflected signal is shown in
Fig.B.3B.

Figure B.3: (A) Measurement setup for ASK modulation at 915 MHz, (B) Measured back reflected
RF signal with modulated ID

B.2.2 Digital Harmonic-frequency RFID Tag

The digital harmonic-frequency RFID tag is designed to overcome the phase noise issue of the
single-frequency operation as shown in Chapter 3. The tag protocol for harmonic RFID is the
same as the single frequency RFID system. The major difference is in the operating frequency and
generation of harmonic signals. A new energy harvester is designed to operate at 434 MHz, while
the switch for modulation is placed on the output path before the nonlinear elements for generating
harmonic signals. The schematic of the digital harmonic-frequency RFID tag is shown in Fig.B.4.

124

Figure B.4: Schematic of the digital harmonic tag for receiving at 434MHz and transmitting at
868MHz

B.2.2.1

434 MHz Energy Harvester

The energy harvester at 434 MHz is tested for impedance matching at various power levels. The
frequency response for the power sweep is shown in Fig.B.5. The energy harvester is fed from port
1 of RF splitter, which receives the signal from antenna connected to the input port.

Figure B.5: Schematic of the 434-868 MHz digital RFID tag

125

B.2.2.2 OOK Modulator

The OOK switch (HMC550) is used to modulate the digital data with backreflected harmonic
signals. The splitter at tag input feeds the RF switch from port 2. The output of the switch
is connected to the nonlinear element (NLTL) for harmonic generation. The output of NLTL is
connected to the oscilloscope for checking the modulation, the acquired signal is shown in Fig.B.6B.

Figure B.6: (A) Measurement setup for ASK modulation at 434 MHz, (B) Measured back reflected
RF signal with modulated ID

B.2.3 Multi-frequency RFID Interrogator

Most of the sensor integrated RFID development work is conventionally performed around the
existing EPC gen2 protocol and off-the-self interrogator, which was never meant to be designed
for sensing applications. The multiple bottle neck problems (self-jamming, digital only) of the
conventional interrogators cannot be eliminated using next generation tag design. For pushing the
boundaries in passive RF sensing and communication, a new interrogator needs to be designed as
well. We are proposing an architecture for a multifrequency interrogator.

A new custom interrogator is designed to communicate with the ISS’s digital single and
harmonic frequency RFID tags. The basic functionality of an interrogator is to transmit and receive
RF signals, which needs to be modulated on the way-out and demodulated on the way-in. Both

126

the modulator and demodulator are part of the RF front-end, with additional peripherals like power
amplifier, LNA, filters, heterodyning mixer, etc. The block diagram of the designed interrogator
with multi-frequency operation and hybrid reception is shown in Fig.B.7.

Figure B.7: Block diagram of the ISS interrogator

A continuous RF source of 915 MHz and 434 MHz is required. The 915 MHz and 434 MHz RF
sources are fed into an OOK modulator, implemented using SPST RF switch (name switch). The
modulator receives digital command bits from the on-board processing unit, to power all available
digital RFID tags in the field of view. The processing unit also selects the single or harmonic mode
of operation. The modulated signal from RF switch is passed down to a +30 dB power amplifier
(MPA-10-40) that can be transmitted using the interrogator’s respective antenna (915 MHz or 434
MHz). The tag in field-of-view reply to the interrogator’s command accordingly, which is received
using either a 915 MHz antenna or 868 MHz antenna, depends on the mode of operation. The
received signal is amplified +25 dB using LNA (MAR-8ASM+) and fed into the RF PORT of an
RF mixer. The RF mixer is part of the super heterodyning, required to precisely decouple the
modulated information from the carrier wave. A secondary RF source (either 915 MHz or 868
MHz) is used to provide the signal at the LO PORT of the mixer. The digital information is further

127

extracted using a comparator and 1-bit translator, so that the processing unit can understand the
signal and apply respective logic.

The interrogator transmits ‘Query’ command for the tags available in field-of-view, to respond.
If an RF sensor tag (single or harmonic) responds with a ’RN16’ to ’Query’ command, the
interrogator sends an ’ACK’ command for asking for its ’EPC’. If a tag receives the correct ’ACK’,
it responds with either digital ID-only or sensor integrated EPC. The interrogator checks the received
EPC and decodes the digital data. The digital sensor data can be directly stored in the interrogator’s
memory for later processing. The transmitted ’Query’ command at 915 MHz and the demodulated
received reply ’RN16’ are shown in Fig.B.8.

Figure B.8: (A) Modulated “Query” command on 915 MHz signal, (B) Demodulated received
reply “RN16” from 915 MHz

The digital processes used for the interrogator is implemented on Raspberry Pi 3’s board
with ARM microprocessor. A bare metal code is developed for modified EPC Gen2 interrogator
protocol, that can send 6 mandatory commands (Query, QueryRep, QueryAdj, ACK, NAK and
ReqRN). The code also controls the operating mode at single or harmonic frequency and set the
correct switches for receiving 915 MHz or 868 MHz signal.

B.3 Wireless Communication Demonstration

First, the interrogator is turned on with an output power of +7 dBm at 915 MHz.The digital
single-frequency RFID tag is placed 10 inches from the interrogator and communicated wirelessly

128

for both ID and sensor information. A variable voltage source is used to demonstrate the tag
capabilities to digitize the sensor’s signal and combine with the EPC data frame. The voltage
change can imitate a resistive type temperature or pressure sensor that can be connected to a voltage
divider circuit for acquisition.

Figure B.9: Digital RFID data received at 915 MHz with +7 dBm transmitted power; (A) with 5
inch interrogation distance, (B) with 10 inch interrogation distance

The raw received EPC with sensor data is plotted in Fig.B.9. The two plots show the change in
signal at the interrogator’s receiver due to change in tag’s distance. Fig.B.9A shows larger change
in amplitude of the received signal due to higher received power at 5 inch separation.

129

The received signal is converted into bits using a comparator and the bit level is shifted to 3.3V
for proper detection using the interrogator’s microprocessor. The received data bits are decoded
according to the width of 0’s and 1’s. Six different voltage values are applied to the tag from 0V to
1.0V at a step size of 0.2V. The interrogator decoded the values and gave a hex output as shown in
Fig.B.10.

Figure B.10: Snapshots of tag and interrogator communication in digital single-frequency mode;
(A) Decoded output with 0.0V at tag, (B) Decoded output with 0.2V at tag, (C) Decoded output
with 0.4V at tag, (D) Decoded output with 0.6V at tag, (E) Decoded output with 0.8V at tag, (F)
Decoded output with 1.0V at tag

The received hex values are plotted according to the known input voltage. The graph is shown

in Fig.B.11.

The designed digital RFID tag successfully worked in single-frequency communication mode.
The interrogator received the RFID tag signal and demodulated for extracting sensor’s information.
The digital sensor information is acquired and checked against the input parameter (voltage).

130

Figure B.11: Received digital sensor data at 915 MHz

Figure B.12: Digital RFID data received at 868 MHz with +7 dBm transmitted power with 10 inch
interrogation distance

131

Second, the interrogator is turned on with an output power of +7 dBm at 434 MHz.The
digital harmonic-frequency RFID tag is placed 10 inches from the interrogator and communicated
wirelessly. The variable voltage source is attached for imitating the sensor change, similar to
previous experiments. The received signal at 868 MHz is demodulated using an RF mixer and the
output is shown in Fig.B.12.

The received signal is converted into bits using a comparator and bit level shifter as before.
The experiment with six different voltage values at the tag from 0V to 1.0V at a step size of 0.2V
is repeated for harmonic-frequncy RFID tag. The interrogator decoded the received values at 868
MHz and gave a hex output as shown in Fig.B.13.

Figure B.13: Snapshots of tag and interrogator communication in digital harmonic-frequency
mode; (A) Decoded output with 0.0V at tag, (B) Decoded output with 0.2V at tag, (C) Decoded
output with 0.4V at tag, (D) Decoded output with 0.6V at tag, (E) Decoded output with 0.8V at
tag, (F) Decoded output with 1.0V at tag

The received hex values are plotted according to the known input voltage. The graph is shown

132

in Fig.B.14.

Figure B.14: Received digital sensor data at 868 MHz

The designed digital RFID tag successfully worked in harmonic-frequency communication
mode. The interrogator received the RFID tag signal and demodulated for extracting sensor’s
information.

B.4 Summary

The battery-free wireless RFID communication using single and harmonic frequency operation
is shown. The receiver with an energy harvester can power the microcontroller and demodulate
the digital signal. The software for RFID tag works reliably and able to decode commands from
the interrogator and respond accordingly. The tag software control the modulation switch for
impedance mismatch and able to reflect back the ID and sensor information. ADC is used to
digitize the sensor’s information, which is combined with the tag’s EPC and back-reflected to the
interrogator.

The interrogator is able to transmit and receive digital signals at multiple frequencies. The
receiver of the interrogator acquires the backscattered data from the digital RFID tags and decodes

133

ID and sensor information.

The multifrequency interrogator and tag implementation for digital battery-free communication
is achieved. The well-known shortcomings of the conventional single frequency RFID communi-
cation are mitigated using harmonic operation.

134

BIBLIOGRAPHY

135

BIBLIOGRAPHY

[1]

[2]

[3]

and Y. Deng,

S. Mukherjee, L. Udpa,

X. Shi, V. T. Rathod,
“Multi-
modality strain estimation using a rapid near-field microwave imaging system for
dielectric materials,” Measurement, vol. 151, p. 107243, 2020. [Online]. Available:
https://www.sciencedirect.com/science/article/pii/S026322411931108X
Z. Li, B. B. Shenoy, L. Udpa, and Y. Deng, “Magnetic Barkhausen Noise Technique for
Early Stage Fatigue Prediction in Martensitic Stainless-Steel Samples (invited),” Journal of
Nondestructive Evaluation, Diagnostics and Prognostics of Engineering Systems, pp. 1–18,
04 2021. [Online]. Available: https://doi.org/10.1115/1.4050842
S. Datta, X. Shi, S. Mukherjee, Y. Deng, and L. Udpa, “Model-Based Study
of a Metamaterial Lens for Nondestructive Evaluation of Composites,” Journal of
Nondestructive Evaluation, Diagnostics and Prognostics of Engineering Systems, vol. 3,
no. 4, 05 2020, 041001. [Online]. Available: https://doi.org/10.1115/1.4047027

[6]

[5]

[4] M. Rautela, S. Gopalakrishnan, K. Gopalakrishnan, and Y. Deng, “Ultrasonic guided waves
based identification of elastic properties using 1D-convolutional neural networks,” in 2020
IEEE International Conference on Prognostics and Health Management (ICPHM), 2020,
pp. 1–7.
D. Kumar, S. Karuppuswami, Y. Deng, and P. Chahal, “A wireless shortwave
near-field probe for monitoring structural
integrity of dielectric composites and
polymers,” NDT & E International, vol. 96, pp. 9–17, 2018. [Online]. Available:
https://www.sciencedirect.com/science/article/pii/S0963869517306886
G. Piscitelli, A. C. Esposito, L. Faella, R. Prakash, and A. Tamburrino, “Monotonicity
principle for nonlinear electrical conductivity tomography,” Inverse Problems, dec 2020.
[Online]. Available: https://doi.org/10.1088/1361-6420/abd29a
A. Tamburrino, G. Piscitelli, and Z. Zhou, “Monotonicity principle for magnetic induction
tomography,” 2020.
S. Karuppuswami, E. Rothwell, P. Chahal, and M. Havrilla, “A triaxial applicator for the
measurement of the electromagnetic properties of materials,” Sensors, vol. 18, no. 2, 2018.
[Online]. Available: https://www.mdpi.com/1424-8220/18/2/383
G. Webb, P. Vardanega, and C. Middleton, “Categories of SHM deployments: technologies
and capabilities,” Journal of Bridge Engineering, vol. 20, no. 11, p. 04014118, 2014.

[7]

[8]

[9]

[10] G. Song, C. Wang,

and B. Wang,

structures,” Applied Sciences,

civil
http://www.mdpi.com/2076-3417/7/8/789

“Structural health monitoring (SHM) of
[Online]. Available:

no. 8,

2017.

vol. 7,

[11] A. News.

[Online]. Available:
6-dead-10-injured-after-bridge-collapse-at-florida-intl-univ/3220344/

(2018) Miami bridge collapse.

https://6abc.com/

136

[12] R.

Insider.

(2018) Railway bridge collapse.

[Online]. Available:

https://www.

romania-insider.com/floods-romania-june-2018/

[13] T13. (2018) Cancura bridge collapse. [Online]. Available:
nacional/colapso-puente-cancura-deja-muerto-y-seis-heridos

http://www.t13.cl/noticia/

[14] T.

telegraph.

(2018) Calcutta bridge collapse.

[Online]. Available:

telegraphindia.com/india/no-room-for-repairs-complains-mamata/cid/1666742

https://www.

[15] PHMSA.

[Online]. Available:
data-and-statistics/pipeline/pipeline-incident-20-year-trends

(2018) Pipeline failures.

https://www.phmsa.dot.gov/

[16] NTSB. (2018) Pipeline failures. [Online]. Available: https://www.ntsb.gov/investigations/

AccidentReports/Pages/pipeline.aspx

[17] H. News. (2017) Natural gas leak. [Online]. Available: https://www.hcn.org/issues/49.22/

infographic-a-map-of-leaking-natural-gas-pipelines-across-the-nation

[18] C. R. Farrar and K. Worden, “An introduction to structural health monitoring,” Philosophical
Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, vol.
365, no. 1851, pp. 303–315, 2006.

[19] D. Kumar, S. Mondal, A. Kaur, Y. Deng, and P. Chahal, “An RF backscatterer system
for real-time multi-sensing applications,” Sensors and Actuators A: Physical, vol. 303,
p. 111685, 2020. [Online]. Available: https://www.sciencedirect.com/science/article/pii/
S0924424719307162

[20] D. Kumar and Y. Deng, “An efficient multisensing system for SHM of oil and gas
pipelines,” Review of Progress in Quantitative Nondestructive Evaluation, vol. 46, 2019.
[Online]. Available: https://www.iastatedigitalpress.com/qnde/article/id/8571

[21] C. Kralovec and M. Schagerl, “Review of structural health monitoring methods regarding a
multi-sensor approach for damage assessment of metal and composite structures,” Sensors,
vol. 20, no. 3, 2020. [Online]. Available: https://www.mdpi.com/1424-8220/20/3/826

[22] H. Ahmed, H. M. La, and N. Gucunski, “Review of non-destructive civil infrastructure
evaluation for bridges: State-of-the-art robotic platforms, sensors and algorithms,” Sensors,
vol. 20, no. 14, 2020. [Online]. Available: https://www.mdpi.com/1424-8220/20/14/3954

[23]

J. Zheng, Y. Zhang, D. Hou, Y. Qin, W. Guo, C. Zhang, and J. Shi, “A review of
nondestructive examination technology for polyethylene pipe in nuclear power plant,”
Frontiers of Mechanical Engineering, vol. 13, no. 4, 2018.
[Online]. Available:
https://doi.org/10.1007/s11465-018-0515-9

[24] S. Abbas, F. Li, and J. Qiu, “A review on shm techniques and current challenges
for characteristic investigation of damage in composite material components of aviation
industry,” Materials Performance and Characterization, vol. 7, no. 1, pp. 224–258, 2018.
[Online]. Available: https://doi.org/10.1520/MPC20170167

137

[25] W. Ostachowicz, R. Soman, and P. Malinowski, “Optimization of sensor placement for
structural health monitoring: A review,” Structural Health Monitoring, vol. 18, no. 3, pp.
963–988, 2019. [Online]. Available: https://doi.org/10.1177/1475921719825601

[26] H. Rocha, C. Semprimoschnig, and J. P. Nunes, “Sensors for process and structural
health monitoring of aerospace composites: A review,” Engineering Structures, vol. 237,
p. 112231, 2021. [Online]. Available: https://www.sciencedirect.com/science/article/pii/
S0141029621003813

[27] Z. H. Warsi, S. M. Irshad, F. Khan, M. A. Shahbaz, M. Junaid, and S. U. Amin, “Sensors
for structural health monitoring: A review,” in 2019 Second International Conference on
Latest trends in Electrical Engineering and Computing Technologies (INTELLECT), 2019,
pp. 1–6.

[28] S. Mukherjee, L. Udpa, S. Udpa, E. J. Rothwell, and Y. Deng, “A time reversal-based
microwave imaging system for detection of breast tumors,” IEEE Transactions on Microwave
Theory and Techniques, vol. 67, no. 5, pp. 2062–2075, 2019.

[29] R. Jacobs, M. Alzuhiri, M. Golkowski, and Y. Deng, “Low-power microwave induced
thermoacoustic imaging: Experimental study and hybrid fem modeling,” Progress In Elec-
tromagnetics Research, vol. 91, pp. 265–277, 2019.

[30] S. Mondal, S. Karuppuswami, R. Steinhorst, and P. Chahal, “A wearable passive pH sensor for
health monitoring,” in 2019 IEEE 69th Electronic Components and Technology Conference
(ECTC), 2019, pp. 1240–1245.

[31] S. Mukherjee, X. Shi, L. Udpa, S. Udpa, Y. Deng, and P. Chahal, “Design of a split-ring
resonator sensor for near-field microwave imaging,” IEEE Sensors Journal, vol. 18, no. 17,
pp. 7066–7076, 2018.

[32] C. Liu, M. Al Qaseer, and R. Zoughi, “Wideband double-ridged TEM horn for nondestructive
evaluation and imaging applications,” in 2020 Antenna Measurement Techniques Association
Symposium (AMTA), 2020, pp. 1–4.

[33] K. Brinker, M. Dvorsky, M. T. Al Qaseer, and R. Zoughi, “Review of advances in microwave
and millimetre-wave NDT&E: principles and applications,” Philosophical Transactions of
the Royal Society A, vol. 378, no. 2182, p. 20190585, 2020.

[34] M. T. Ghasr, Y. Gao, and R. Zoughi, “Handheld microwave imaging system for the inspection
of non-metallic structures,” Review of Progress in Quantitative Nondestructive Evaluation,
2019.
I. Mehdipour, R. Zoughi, and K. H. Khayat, “Feasibility of using near-field microwave
reflectometry for monitoring autogenous crack healing in cementitious materials,”
Cement and Concrete Composites, vol. 85, pp. 161–173, 2018. [Online]. Available:
https://www.sciencedirect.com/science/article/pii/S0958946517305565

[35]

[36] R. Kumar, J. H. Lang, T. Hamer, and D. L. Trumper, “Towards high-bandwidth capacitive

imaging,” in 2016 IEEE SENSORS, 2016, pp. 1–3.

138

[37] A. Jacques, “Q factor measurements on L-C circuits,” Amateur Radio in Space
(AMSAT), p. 11, 2012. [Online]. Available: http://ve2azx.net/technical/Q-FactorMeas_on_
LC_Circuits.pdf

[38] S. G. Taylor, E. Y. Raby, K. M. Farinholt, G. Park, and M. D. Todd, “Active-sensing
platform for structural health monitoring: Development and deployment,” Structural
Health Monitoring,
no. 4,
[Online]. Available:
https://doi.org/10.1177/1475921716642171

pp. 413–422,

vol. 15,

2016.

[39] N. A. Tanner, J. R. Wait, C. R. Farrar, and H. Sohn, “Structural health monitoring using
modular wireless sensors,” Journal of Intelligent Material Systems and Structures, vol. 14,
no. 1, pp. 43–56, 2003. [Online]. Available: https://doi.org/10.1177/1045389X03014001005

[40] E. Yasar and Y. Erdogan, “Correlating sound velocity with the density, compressive strength
and young’s modulus of carbonate rocks,” International Journal of Rock Mechanics and
Mining Sciences, vol. 41, no. 5, pp. 871–875, 2004.

[41] S. Mondal, K. P. Wijewardena, S. Karuppuswami, N. Kriti, D. Kumar, and P. Chahal,
“Blockchain inspired rfid-based information architecture for food supply chain,” IEEE In-
ternet of Things Journal, vol. 6, no. 3, pp. 5803–5813, 2019.

[42] S. Cho, C.-B. Yun, J. Lynch, A. Zimmerman, B. Spencer, and T. Nagayama, “Smart wireless
sensor technology for structural health monitoring of civil structures,” Steel Structures, vol. 8,
pp. 267–275, 01 2004.

[43] S. Karuppuswami, S. Mondal, D. Kumar, and P. Chahal, “A compact wireless passive breath
analyzer for health monitoring,” in 2020 IEEE 70th Electronic Components and Technology
Conference (ECTC), 2020, pp. 1010–1015.

[44] P. Chahal, D. Kumar, S. Mondal, and S. Karuppuswami, “Systems and methods for a

multiband sensing platform,” May 28 2020, uS Patent App. 16/695,758.

[45] R. Nopper, R. Niekrawietz, and L. Reindl, “Wireless readout of passive LC sensors,” IEEE

Transactions on Instrumentation and Measurement, vol. 59, no. 9, pp. 2450–2457, 2010.

[46] S. Karuppuswami, L. L. Matta, E. C. Alocilja, and P. Chahal, “A wireless RFID compatible
sensor tag using gold nanoparticle markers for pathogen detection in the liquid food supply
chain,” IEEE Sensors Letters, vol. 2, no. 2, pp. 1–4, 2018.

[47] A. Deivasigamani, A. Daliri, C. H. Wang, and S. John, “A review of passive wireless sensors

for structural health monitoring,” Modern Applied Science, vol. 7, no. 2, p. 57, 2013.

[48] H. Shen, L. Li, and Y. Zhou, “Fully integrated passive UHF RFID tag with temperature sensor
for environment monitoring,” in ASIC, 2007. ASICON’07. 7th International Conference on.
IEEE, 2007, pp. 360–363.

[49] SL900A EPC Class 3 Sensory Tag Chip - For Automatic Data Logging, AMS, 10 2018, rev.

1.12.

139

[50] A. P. Sample, D. J. Yeager, P. S. Powledge, A. V. Mamishev, J. R. Smith et al., “De-
sign of an RFID-based battery-free programmable sensing platform,” IEEE transactions on
instrumentation and measurement, vol. 57, no. 11, p. 2608, 2008.

[51] Y. Su, A. Wickramasinghe, and D. C. Ranasinghe, “Investigating sensor data retrieval
schemes for multi-sensor passive rfid tags,” in 2015 IEEE International Conference on
RFID (RFID), April 2015, pp. 158–165.

[52] S. Mondal, D. Kumar, and P. Chahal, “A continuous-mode single-antenna harmonic rfid
tag,” IEEE Microwave and Wireless Components Letters, vol. 30, no. 4, pp. 441–444, 2020.
[53] S. Karuppuswami, S. Mondal, D. Kumar, and P. Chahal, “Rfid coupled passive digital
ammonia sensor for quality control of packaged food,” IEEE Sensors Journal, vol. 20, no. 9,
pp. 4679–4687, 2020.

[54] S. Mondal, S. Karuppuswami, D. Kumar, A. Kaur, and P. Chahal, “A miniaturized dual
band antenna for harmonic rfid tag,” in International Symposium on Microelectronics, vol.
2019, no. 1.
International Microelectronics Assembly and Packaging Society, 2019, pp.
000 033–000 036.

[56]

[55] S. Mondal, D. Kumar, M. I. Ghazali, P. Chahal, L. Udpa, and Y. Deng, “Monitoring and
localization of buried plastic natural gas pipes using passive rf tags,” in AIP Conference
Proceedings, vol. 1949, no. 1. AIP Publishing LLC, 2018, p. 020020.
I. Amenabar, A. Mendikute, A. López-Arraiza, M. Lizaranzu, and J. Aurrekoetxea,
“Comparison and analysis of non-destructive testing techniques suitable for delamination
inspection in wind turbine blades,” Composites Part B: Engineering, vol. 42, no. 5, pp.
1298 – 1305, 2011. [Online]. Available: http://www.sciencedirect.com/science/article/pii/
S1359836811000539

[57] P. A. Joosse, M. J. Blanch, A. G. Dutton, D. A. Kouroussis, T. P. Philippidis, and P. S.
Vionis, “Acoustic emission monitoring of small wind turbine blades,” Journal of Solar
Energy Engineering, vol. 124, no. 4, pp. 446–454, 2002.

[58] R. A. Raišutis, E. Jasi¯unien˙e, and E. Žukauskas, “Ultrasonic NDT of wind turbine blades

using guided waves,” Ultragarsas, vol. 63, no. 1, pp. 7–11, 2008.

[59] R. Raišutis, E. Jasi¯unien˙e, R. Šliteris, and A. Vladišauskas, “The review of non-destructive
testing techniques suitable for inspection of the wind turbine blades,” Ultragarsas" Ultra-
sound", vol. 63, no. 2, pp. 26–30, 2008.

[60] Z. Li, C. Soutis, A. Haigh, R. Sloan, A. Gibson, and N. Karimian, “Microwave imaging for
delamination detection in T-joints of wind turbine composite blades,” in 2016 46th European
Microwave Conference (EuMC), Oct 2016, pp. 1235–1238.

[61] D. Palumbo, R. Tamborrino, U. Galietti, P. Aversa, A. Tatì, and V. Luprano, “Ultrasonic
analysis and lock-in thermography for debonding evaluation of composite adhesive joints,”
NDT E International, vol. 78, no. Supplement C, pp. 1 – 9, 2016. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0963869515000985

140

[62] C. Meola, S. Boccardi, G. Carlomagno, N. Boffa, F. Ricci, G. Simeoli, and P. Russo, “Impact
damaging of composites through online monitoring and non-destructive evaluation with
infrared thermography,” NDT E International, vol. 85, no. Supplement C, pp. 34 – 42, 2017.
[Online]. Available: http://www.sciencedirect.com/science/article/pii/S0963869516301323

[63] C. C. Ciang,

J.-R. Lee,

wind turbine system:
Science and Technology, vol. 19, no. 12, p. 122001, 2008.
http://stacks.iop.org/0957-0233/19/i=12/a=122001

“Structural health monitoring for a
a review of damage detection methods,” Measurement
[Online]. Available:

and H.-J. Bang,

[64] G. S. Park and E. S. Park, “Improvement of the sensor system in magnetic flux leakage-
type nondestructive testing (NDT),” IEEE Transactions on Magnetics, vol. 38, no. 2, pp.
1277–1280, Mar 2002.

[65] N. V. Nair, V. R. Melapudi, H. R. Jimenez, X. Liu, Y. Deng, Z. Zeng, L. Udpa, T. J. Moran,
and S. S. Udpa, “A gmr-based eddy current system for nde of aircraft structures,” IEEE
Transactions on Magnetics, vol. 42, no. 10, pp. 3312–3314, Oct 2006.

[66] Y. Xiong, Z. Liu, C. Sun, and X. Zhang, “Two-dimensional imaging by far-field superlens at
visible wavelengths,” Nano Letters, vol. 7, no. 11, pp. 3360–3365, 2007, pMID: 17918904.
[Online]. Available: http://dx.doi.org/10.1021/nl0716449

[67] D. F. Pieper and K. M. Donnell, “Application of frequency selective surfaces for inspec-
tion of layered structures,” in 2015 IEEE International Instrumentation and Measurement
Technology Conference (I2MTC) Proceedings, May 2015, pp. 1204–1209.
“A hollow fibre reinforced polymer composite
J. W. Pang and I. P. Bond,
encompassing self-healing and enhanced damage visibility,” Composites Science
and Technology, vol. 65, no. 11, pp. 1791 – 1799, 2005.
[Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0266353805000783

[68]

[69] F. Yan, j. Roger L. Royer, and J. L. Rose, “Ultrasonic guided wave imaging techniques in
structural health monitoring,” Journal of Intelligent Material Systems and Structures, vol. 21,
no. 3, pp. 377–384, 2010.

[70] S. Karuppuswami, A. Kaur, M. I. M. Ghazali, and P. Chahal, “RFID compatible sensor
tags for remote liquid sample interrogation,” in 2016 IEEE 66th Electronic Components and
Technology Conference (ECTC), May 2016, pp. 2401–2407.

[71] S. Karuppuswami, N. Wiwatcharagoses, A. Kaur, and P. Chahal, “Capillary condensation
based wireless volatile molecular sensor,” in 2017 IEEE 67th Electronic Components and
Technology Conference (ECTC), May 2017, pp. 1455–1460.

[72] S. Karuppuswami, A. Kaur, H. Arangali, and P. P. Chahal, “A hybrid magnetoelastic wireless
sensor for detection of food adulteration,” IEEE Sensors Journal, vol. 17, no. 6, pp. 1706–
1714, March 2017.

141

[73] S. Karuppuswami, H. Arangali, and P. Chahal, “A hybrid electrical-mechanical wireless
magnetoelastic sensor for liquid sample measurements,” in 2016 IEEE 66th Electronic
Components and Technology Conference (ECTC), May 2016, pp. 2535–2540.

[74] N. Wiwatcharagoses, K. Y. Park, and P. Chahal, “Metamaterial-inspired sensor on porous
substrate for detection of volatile organic compounds in air,” in 2016 IEEE 66th Electronic
Components and Technology Conference (ECTC), May 2016, pp. 1557–1562.

[75] W. B. Weir, “Automatic measurement of complex dielectric constant and permeability at

microwave frequencies,” Proceedings of the IEEE, vol. 62, no. 1, pp. 33–36, Jan 1974.

[76] R. Luebbers, “Lossy dielectrics in fdtd,” IEEE Transactions on Antennas and Propagation,

vol. 41, no. 11, pp. 1586–1588, Nov 1993.

[77] S. Chisca, I. Sava, V. E. Musteata, and M. Bruma, “Dielectric and conduction properties of
polyimide films,” in CAS 2011 Proceedings (2011 International Semiconductor Conference),
vol. 2, Oct 2011, pp. 253–256.

[78] B. Riddle, J. Baker-Jarvis, and J. Krupka, “Complex permittivity measurements of common
plastics over variable temperatures,” IEEE Transactions on Microwave theory and techniques,
vol. 51, no. 3, pp. 727–733, 2003.

[79] K. Naito, Y. Kagawa, S. Utsuno, T. Naganuma, and K. Kurihara, “Dielectric properties
of eight-harness-stain fabric glass fiber reinforced polyimide matrix composite in the THz
frequency range,” NDTE International, vol. 42, no. 5, pp. 441 – 445, 2009.
J. Landt, “The history of RFID,” IEEE potentials, vol. 24, no. 4, pp. 8–11, 2005.

[80]
[81] F. Wu, F. Kuo, and L.-W. Liu, “The application of RFID on drug safety of inpatient nursing
healthcare,” in Proceedings of the 7th international conference on Electronic commerce.
ACM, 2005, pp. 85–92.

[82] B. Yan and G. Huang, “Application of RFID and internet of things in monitoring and anti-
counterfeiting for products,” in Business and Information Management, 2008. ISBIM’08.
International Seminar on, vol. 1.

IEEE, 2008, pp. 392–395.

[83] M. Buettner, R. Prasad, M. Philipose, and D. Wetherall, “Recognizing daily activities with
RFID-based sensors,” in Proceedings of the 11th international conference on Ubiquitous
computing. ACM, 2009, pp. 51–60.

[84] D. Delen, R. Sharda, and B. C. Hardgrave, “The promise of RFID-based sensors in the

perishables supply chain,” IEEE Wireless Communications, vol. 18, no. 2, 2011.

[85] K. S. Rao, P. V. Nikitin, and S. F. Lam, “Antenna design for UHF RFID tags: A review and
a practical application,” IEEE Transactions on antennas and propagation, vol. 53, no. 12,
pp. 3870–3876, 2005.

[86] Y. Ma, X. Hui, and E. C. Kan, “Harmonic-wisp: A passive broadband harmonic RFID
IEEE, 2016,

platform,” in Microwave Symposium (IMS), 2016 IEEE MTT-S International.
pp. 1–4.

142

[87] S. Mondal and P. Chahal, “A passive harmonic RFID tag and interrogator development,”

IEEE Journal of Radio Frequency Identification, vol. 3, no. 2, pp. 98–107, June 2019.

[88] A. Boaventura, J. Santos, A. Oliveira, and N. B. Carvalho, “Perfect isolation: Dealing with
self-jamming in passive RFID systems,” IEEE Microwave Magazine, vol. 17, no. 11, pp.
20–39, 2016.

[89] M. Sajid and S. Chiu, “Canceling self-jammer and interfering signals in an RFID system,”

Aug. 16 2011, US Patent 8,000,674.

[90] S. Tedjini, Y. Duroc, G. A. Vera, C. Loussert, and M. Recouly, “RFID communication

system,” Sep. 27 2016, US Patent 9,455,769.

[91]

I. Mayordomo and J. Bernhard, “Implementation of an adaptive leakage cancellation control
for passive uhf RFID readers,” in RFID (RFID), 2011 IEEE International Conference on.
IEEE, 2011, pp. 121–127.

[92] R. Rimini, P. D. Heidmann, and J. P. Burke, “Method and apparatus for adaptive non-linear

self-jamming interference cancellation,” Jun. 3 2014, US Patent 8,744,377.

[93] M. I. M. Ghazali, S. Karuppuswami, and P. Chahal, “3-D printed embedded passive harmonic
sensor tag as markers for buried assets localization,” IEEE Sensors Letters, vol. 3, no. 4, pp.
1–4, April 2019.

[94] S. Mondal, D. Kumar, and P. Chahal, “A wireless passive pH sensor with clutter rejection

scheme,” IEEE Sensors Journal, vol. 19, no. 9, pp. 3399–3407, May 2019.

[95] H. Huang, P.-Y. Chen, C.-H. Hung, R. Gharpurey, and D. Akinwande, “Frequency hopped
wireless passive sensing system with harmonic transponder antenna sensor,” in Microwave
Symposium (IMS), 2015 IEEE MTT-S International.

IEEE, 2015, pp. 1–4.

[96] K. Rasilainen, J. Ilvonen, A. Lehtovuori, J.-M. Hannula, and V. Viikari, “On design and
evaluation of harmonic transponders,” IEEE Transactions on Antennas and Propagation,
vol. 63, no. 1, pp. 15–23, 2015.

[97] P. V. Nikitin and K. Rao, “Harmonic scattering from passive UHF RFID tags,” in Antennas
IEEE, 2009,

and Propagation Society International Symposium, 2009. APSURSI’09. IEEE.
pp. 1–4.

[98] G. A. Vera, Y. Duroc, and S. Tedjini, “RFID test platform: Nonlinear characterization,”
IEEE Transactions on Instrumentation and Measurement, vol. 63, no. 9, pp. 2299–2305,
2014.

[99] G. Vera, Y. Duroc, and S. Tedjini, “Analysis of harmonics in UHF RFID signals,” IEEE

Transactions on Microwave Theory and Techniques, vol. 61, no. 6, pp. 2481–2490, 2013.

[100] G. A. Vera, Y. Duroc, and S. Tedjini, “Tag-to-reader harmonic link in passive UHF RFID,”
IEEE, 2014, pp. 1–4.

in Microwave Symposium (IMS), 2014 IEEE MTT-S International.

143

[101] G. A. Vera, S. D. Nawale, Y. Duroc, and S. Tedjini, “Read range enhancement by harmonic
energy harvesting in passive UHF RFID,” IEEE Microwave and Wireless Components Let-
ters, vol. 25, no. 9, pp. 627–629, 2015.

[102] D. Allane, G. A. Vera, Y. Duroc, R. Touhami, and S. Tedjini, “Harmonic power harvest-
ing system for passive RFID sensor tags,” IEEE Transactions on Microwave Theory and
Techniques, vol. 64, no. 7, pp. 2347–2356, 2016.

[103] D. Allane, Y. Duroc, G. A. Vera, R. Touhami, and S. Tedjini, “On energy harvesting for

augmented tags,” Comptes Rendus Physique, vol. 18, no. 2, pp. 86–97, 2017.

[104] Higgs® 3, ALC-360-SOT, Alien Technology, LLC, 11 2016.
[105] N. Pillin, C. Dehollain, and M. J. Declercq, “Read range limitation in if-based far-field
RFID using ASK backscatter modulation,” in 2009 Ph.D. Research in Microelectronics and
Electronics, July 2009, pp. 348–351.

[106] D. Kuester and Z. Popovic, “How good is your tag?: RFID backscatter metrics and mea-

surements,” IEEE Microwave Magazine, vol. 14, no. 5, pp. 47–55, July 2013.

[107] S. Mondal, M. I. M. Ghazali, S. Karuppuswami, A. Kaur, and P. Chahal, “A nonlinear
transmission line based harmonic RF tag,” in 2017 IEEE 67th Electronic Components and
Technology Conference (ECTC), May 2017, pp. 2237–2242.

[108] G. Andia, Y. Duroc, and S. Tedjini, Non-Linearities in Passive RFID Systems: Third Har-

monic Concept and Applications.

John Wiley & Sons, 2018.

[109] H. Cho, Y.-W. Seo, B. V. Kumar, and R. R. Rajkumar, “A multi-sensor fusion system
for moving object detection and tracking in urban driving environments,” in 2014 IEEE
International Conference on Robotics and Automation (ICRA).
IEEE, 2014, pp. 1836–
1843.

[110] C. Tang, H. F. Rashvand, G. Y. Tian, P. Hu, A. I. Sunny, and H. Wang, “Structural health
monitoring with WSNs,” Wireless Sensor Systems for Extreme Environments: Space, Un-
derwater, Underground, and Industrial, pp. 383–408, 2017.

[111] US-EIA, Outlook for future emissons, 2018. [Online]. Available: https://www.eia.gov/

energyexplained/index.php?page=environment_outlook_for_emissions

[112] S. K. Rajah and S. Nasr, “Longitudinal stresses in pressure pipeline design—A critical
review,” Pipelines 2008: Pipeline Asset Management: Maximizing Performance of our
Pipeline Infrastructure, pp. 1–12, 2008.

[113] X. Wu, H. Lu, and S. Wu, “Stress analysis of parallel oil and gas steel pipelines in inclined

tunnels,” SpringerPlus, vol. 4, no. 1, pp. 1–25, 2015.

[114] J. Makar, R. Rogge, and S. McDonald, “Circumferential failures in grey cast iron pipes,” in
Proceedings of the AWWA Infrastructure Conference, Chicago, IL, USA. Citeseer, 2002,
pp. 10–13.

144

[115] Z. Liu and Y. Kleiner, “State-of-the-art review of technologies for pipe structural health

monitoring,” IEEE Sensors Journal, vol. 12, no. 6, pp. 1987–1992, June 2012.

[116] A. M. Obeid, F. Karray, M. W. Jmal, M. Abid, S. M. Qasim, and M. S. BenSaleh, “Towards
realisation of wireless sensor network-based water pipeline monitoring systems: a compre-
hensive review of techniques and platforms,” IET Science, Measurement Technology, vol. 10,
no. 5, pp. 420–426, 2016.

[117] J. Weber, J. Shi, and C. Zhang, “Embedded capacitance sensor array for structural health
monitoring of pipeline systems,” ASME Pressure Vessels and Piping Conference, vol. 5,
2017.

[118] H. Qingmin, R. Liang, J. Wenling, Z. Pinghua, and S. Gangbing, “An improved negative
pressure wave method for natural gas pipeline leak location using fbg based strain sensor
and wavelet transform,” Mathematical Problems in Engineering, vol. 2013, p. 8, 2013.

[119] M. Jang and K.-S. Yun, “Mems capacitive pressure sensor monolithically integrated with
cmos readout circuit by using post CMOS processes,” Micro and Nano Systems Letters,
vol. 5, no. 1, p. 4, 2017.

[120] M. I. M. Ghazali, Additive Manufacturing for Electronic Systems (AMES). Michigan State

University. Electrical Engineering, 2019.

[121] C. Hong, Y. Zhang, and L. Borana, “Design, fabrication and testing of a 3d printed fbg

pressure sensor,” IEEE Access, vol. 7, pp. 38 577–38 583, 2019.

[122] J. Park,

J.-K. Kim, D.-S. Kim, A. Shanmugasundaram, S. A. Park, S. Kang,
integrated
Sensors and
[Online]. Available:

S.-H. Kim, M. H. Jeong, and D.-W. Lee, “Wireless pressure sensor
with a 3d printed polymer
Actuators B: Chemical, vol. 280, pp. 201 – 209, 2019.
http://www.sciencedirect.com/science/article/pii/S0925400518317696

stent

for

smart health monitoring,”

[123] J. Yoon, B. Choi, Y. Lee, J. Han, J. Lee, J. Park, Y. Kim, D. M. Kim, D. H. Kim, M. Kang,
S. Kim, and S. Choi, “Monolithically 3d-printed pressure sensors for application in electronic
skin and healthcare monitoring,” in 2017 IEEE International Electron Devices Meeting
(IEDM), 2017, pp. 40.2.1–40.2.4.

[124] M. Cattani, C. A. Boano, D. Steffelbauer, S. Kaltenbacher, M. Günther, K. Römer, D. Fuchs-
Hanusch, and M. Horn, “Adige: An efficient smart water network based on long-range
wireless technology,” in Proceedings of the 3rd International Workshop on Cyber-Physical
Systems for Smart Water Networks, ser. CySWATER ’17. New York, NY, USA: ACM,
2017, pp. 3–6. [Online]. Available: http://doi.acm.org/10.1145/3055366.3055367

[125] A. M. Sadeghioon, N. Metje, D. Chapman, and C. Anthony, “Water pipeline failure
detection using distributed relative pressure and temperature measurements and anomaly
detection algorithms,” Urban Water Journal, vol. 15, no. 4, pp. 287–295, 2018. [Online].
Available: https://doi.org/10.1080/1573062X.2018.1424213

145

[126] A. B. Noel, A. Abdaoui, T. Elfouly, M. H. Ahmed, A. Badawy, and M. S. Shehata, “Struc-
tural health monitoring using wireless sensor networks: A comprehensive survey,” IEEE
Communications Surveys Tutorials, vol. 19, no. 3, pp. 1403–1423, thirdquarter 2017.

[127] J.-H. Kim, G. Sharma, N. Boudriga, S. Iyengar, and N. Prabakar, “Autonomous pipeline
monitoring and maintenance system: A RFID-based approach,” EURASIP Journal on
Wireless Communications and Networking, vol. 2015, no. 1, p. 262, Dec 2015. [Online].
Available: https://doi.org/10.1186/s13638-015-0495-y

[128] S. Mondal, K. Wijewardena, S. Karrapuswami, D. Kumar, P. Chahal, M. Ghannam, and
M. Cuddihy, “A wireless battery-less seat sensor for autonomous vehicles,” in 2020 IEEE
70th Electronic Components and Technology Conference (ECTC), 2020, pp. 2289–2294.

[129] L. Lin, M. Ma, F. Zhang, F. Liu, Z. Liu, and Y. Li, “Fabrications and performance
of wireless LC pressure sensors through ltcc technology,” Sensors, vol. 18, no. 2, 2018.
[Online]. Available: http://www.mdpi.com/1424-8220/18/2/340

[130] M. Boada, A. Lazaro, R. Villarino, and D. Girbau, “Battery-less NFC sensor for pH moni-

toring,” IEEE Access, vol. 7, pp. 33 226–33 239, 2019.

[131] K. Opasjumruskit, T. Thanthipwan, O. Sathusen, P. Sirinamarattana, P. Gadmanee,
E. Pootarapan, N. Wongkomet, A. Thanachayanont, and M. Thamsirianunt, “Self-powered
wireless temperature sensors exploit RFID technology,” IEEE Pervasive computing, no. 1,
pp. 54–61, 2006.

[132] EPC C1 G2 compliant UHF RFID tag with power harvesting and SPI communication for

external low power sensors and actuators, ANDY100, Farsens, 3 2016, v09.

[133] RFMicron Magnus S3 Sensor IC, RFMicron, 2015, pB003P11.
[134] Higgs4 IC EPC Global Compliant Class-1 Generation-2, Alien, 6 2014, rev. 1.3.5.
[135] S. Naderiparizi, A. N. Parks, Z. Kapetanovic, B. Ransford, and J. R. Smith, “Wispcam: A
battery-free rfid camera,” in 2015 IEEE International Conference on RFID (RFID), April
2015, pp. 166–173.

[136] H. Huang, P.-Y. Chen, C.-H. Hung, R. Gharpurey, and D. Akinwande, “A zero power
harmonic transponder sensor for ubiquitous wireless 
l liquid-volume monitoring,” Scientific
reports, vol. 6, no. 1, pp. 1–10, 2016.

[137] S. Mondal, D. Kumar, S. Karuppuswami, and P. Chahal, “A harmonic rf phase-shifter based
wireless ph sensor,” in 2018 IEEE 68th Electronic Components and Technology Conference
(ECTC), 2018, pp. 796–801.

[138] D. Kumar, S. Mondal, Y. Deng, and P. Chahal, “Wireless battery-free harmonic
communication system for pressure sensing,” Micromachines, vol. 11, no. 12, 2020.
[Online]. Available: https://www.mdpi.com/2072-666X/11/12/1043

[139] NPI-19 Series Medium Pressure Sensor, NOVA Sensor, 12 2015.

146

[140] MIXED SIGNAL MICROCONTROLLER, Texas Instruments, 1 2012, sLAS578J.
[141] C. Ding, A. A. H. Damen, and P. P. J. van den Bosch, “Robust vibration isolation by
frequency-shaped sliding surface control with floor velocity/acceleration measurement,” in
IECON 2010 - 36th Annual Conference on IEEE Industrial Electronics Society, Nov 2010,
pp. 1627–1632.

[142] J. P. Sebastia, J. A. Lluch, J. R. L. Vizcaino, and J. S. Bellon, “Vibration detector based on
gmr sensors,” IEEE Transactions on Instrumentation and Measurement, vol. 58, no. 3, pp.
707–712, March 2009.

[143] TDC1000 Ultrasonic Sensing Analog Front End (AFE) for Level Sensing, Flow Sensing,
Concentration Sensing, and Proximity Sensing Applications, Texas Instruments, 8 2015,
sNAS648B.

[144] TDC7200 Time-to-Digital Converter for Time-of-Flight Applications in LIDAR, Magne-

tostrictive and Flow Meters, Texas Instruments, 3 2016, sNAS647D.

[145] M. M. Islam, K. Rasilainen, and V. Viikari, “Implementation of sensor rfid: Carrying sensor
information in the modulation frequency,” IEEE Transactions on Microwave Theory and
Techniques, vol. 63, no. 8, pp. 2672–2681, Aug 2015.

[146] S. M. Noghabaei, R. L. Radin, Y. Savaria, and M. Sawan, “A high-efficiency ultra-low-power
cmos rectifier for rf energy harvesting applications,” in 2018 IEEE International Symposium
on Circuits and Systems (ISCAS), May 2018, pp. 1–4.

[147] P. Chahal, F. Morris, and G. Frazier, “Zero bias resonant tunnel schottky contact diode for
wide-band direct detection,” IEEE Electron Device Letters, vol. 26, no. 12, pp. 894–896,
Dec 2005.

[148] NTC Thermistors, NCP18WM474J03RB, muRata, 8 2018, https://www.murata.com/.
[149] CDS Photoconductive Photocells, PDV-P7002, Luna Optoelectronics,

4 2016,

https://lunainc.com/wp-content/uploads/2016/06/PDV-P7002.pdf.

[150] D. Kumar, S. Mondal, S. Karuppuswami, Y. Deng, and P. Chahal, “Harmonic rfid commu-
nication using conventional uhf system,” IEEE Journal of Radio Frequency Identification,
pp. 1–1, 2019.

[151] EPC™ Radio-Frequency Identity Protocols Generation-2 UHF RFID, GS1 EPC global Inc,

4 2015, version 2.0.1.

147