L/

(\

 

LIBRARY
I‘Aichigan State

a i..-.:. narcH’V
l U'llVVIUII-J

This is to certify that the
thesis entitled

AN EVALUATION OF UHF RADIO FREQUENCY IDENTIFICATION
TECHNOLOGY FOR APPLICATIONS IN ARCHAEOLOGICAL
REPOSITORY MANAGEMENT

presented by

REED CARLSON EPPELHEIMER

has been accepted towards fulfillment
of the requirements for the

MS degree in Packaging

 

 

- Major ProIlessorle~Signature
car/a 5/;0/ 0
r /

Date

MSU is an Affirmative Action/Equal Opportunity Employer

5 - - - .. _.—.-.---—-—

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SIDS K:IProj/Aoc&Pres/ClRC/DateDue.indd

AN EVALUATION OF UHF RADIO FREQUENCY IDENTIFICATION
TECHNOLOGY FOR APPLICATIONS IN ARCHAEOLOGICAL REPOSITORY
MANAGEMENT
BY

Reed Carlson Eppelheimer

A THESIS
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Packaging

2010

ABSTRACT
AN EVALUATION OF UHF RADIO FREQUENCY IDENTIFICATION
TECHNOLOGY FOR APPLICATIONS IN ARCHAEOLOGICAL REPOSITORY
MANAGEMENT
By
Reed Carlson Eppelheimer

Radio frequency identification (RF ID) has received a great deal of attention as
an emerging automatic identification technology in supply chain management,
personal identification, asset management, and academic arenas such as
libraries and museums. Further, RFID implementation is improving informational
access to researchers and the general public alike. While RF ID has proven to be
a useful tool for public libraries and museum collections, the potential benefit of
RFID in the management of cultural resources extends beyond public facilities to
the front lines of archaeological research: the artifact repository. This research
evaluates the performance of an ultra-high frequency (UHF) RFID-enabled
repository management system (R-RMS) for archaeological artifacts.

The R-RMS evaluation utilized a custom software program to perform the
inventory and information management functions of documenting newly
discovered artifacts, recalling object-specific database entries and auditing
artifact inventories. Test scenarios were developed according to four
subcategories; individual artifacts, bags of artifact groups, storage bins filled with
artifact bags, and related reference materials. The results of on-site UHF testing

at two active repositories conclusively demonstrated the efficacy of all R-RMS

functions.

Copyright by
Reed Carlson Eppelheimer
2010

.To my parents, thank you for everything.

ACKNOWLEDGEMENTS

First and foremost I would like to thank Dr. Robb Clarke for the academic,
professional and personal guidance he has provided over the past four years. I
cannot possibly overstate the appreciation l have for the opportunities I have
been afforded through my work at the Michigan State University Auto-ID

Research & Testing Center.

I would like to express my gratitude to my thesis committee members: Dr.
Diana Twede for her teachings throughout my time at MSU, and Dr. Jon Frey for

coordinating and providing insight on this project, which I have enjoyed so much.

Dr. Tim Gregory, Director of The Ohio State University Excavations at
lsthmia, and Dr. Bill Lovis, Curator of Anthropology at the MSU Museum, also
deserve recognition for allowing access to their respective repositories for RFID

system evaluation.

A special thanks is owed to Mr. Cisco Garcia for his exceptional work in

developing the RFID evaluation software used in this research.

I want to thank my mom and dad for always pushing me to pursue higher
education and for providing the necessary resources to act on their advice.
Thanks, also, to my brothers, whose individual achievements have always

served to set the bar high around the dinner table.

And finally, to the professors, colleagues and friends who have made my

last seven years the amazing experience that is was, thank you all. Go Green!

TABLE OF CONTENTS

LIST OF TABLES ................................................ viii
LIST OF FIGURES ................................................ ix
KEY ABBREVIATIONS ............................................ xii
CHAPTER 1- INTRODUCTION ...................................... 1
The Modem Repository ....................................... 2
Relational Object Management and Cascading Errors ............... 8
Research Objective ......................................... 11
CHAPTER 2- LITERATURE REVIEW ................................. 12
Automatic Identification ....................................... 13

RFID System Components .................................... 16
Auto-ID with Radio Waves .................................... 20

An RF ID for (almost) any Situation .............................. 22
History of RFID Development .................................. 27
CHAPTER 3- METHODOLOGY ..................................... 39
, Materials .................................................. 40
Dig-Site Repository Testing ................................... 43
Museum Repository Testing .................................. 52
CHAPTER 4- RESULTS ........................................... 55
Bag Testing ................................................ 56

Bin Testing ................................................ 58

Book Testing ............................................... 65
Museum Repository Testing ................................... 69
CHAPTER 5- CONCLUSIONS AND DISCUSSION ...................... 73
Dig-House Repository: Artifact Accessioning ...................... 74
Dig-House Repository: Inventory Auditing ........................ 75
Museum Repository: Inventory Auditing ......................... 76
CHAPTER 6- THE RFID-ENABLED REPOSITORY ...................... 77
The Transition from Paper to Microchips ......................... 78
Ongoing Curation with RF ID .................................. 80
CHAPTER 7- FUTURE CONSIDERATIONS ............................ 84
Expanding the R-RMS ....................................... 85
Scalability Considerations .................................... 87
Statistical Validity Testing ..................................... 88

vi

APPENDIX I: DIG-SITE REPOSITORY ............................... 89

APPENDIX ll: MUSEUM REPOSITORY ............................... 90
APPENDIX III: DIG-SITE BAG DATABASE ............................ 91

APPENDIX IV: DIG-SITE BIN DATABASE ............................. 92
APPENDIX V: BIN INVENTORY TAG DATA ........................... 94
APPENDIX VI: BIN INVENTORY TRIAL DATA ........................ 100
APPENDIX VII: BIN TAG READ RATES BY TAG POSITION ............. 101
APPENDIX VIII: DIG-SITE BOOK DATABASE ........................ 104
REFERENCES ................................................. 105

vii

LIST OF TABLES

TABLE 1: ADDING AN ARTIFACT TO A REPOSITORY COLLECTION ...... 2
TABLE 2: ONGOING CURATION AT A REPOSITORY ................... 6
TABLE 3: COMMON RF ID READER COMMANDS ..................... 19
TABLE 4: COMMON CLOSED-LOOP APPLICATIONS FOR RFID ......... 35
TABLE 5: READ COUNT DATA FROM SIMPLE INVENTORY SCANS OF
UNITIZED BAGS ..................................... 57
TABLE 6: SUMMARIZED BIN TAG READ RATE DATA .................. 60
TABLE 7: SUMMARIZED PER-TRIAL READ RATE DATA ............... 61
TABLE 8: SUMMARIZED BIN TAG READ DATA BY BIN POSITION ........ 64
TABLE 9: AVERAGE NUMBER OF READS PER BOOK TAG .............. 67
TABLE 10: R-RMS FUNCTIONS AS RELATED TO RFID TEST SCENARIOS .721.
TABLE 11: DIG-SITE BAG DATABASE (FAUX DATA) ................... 91
TABLE 12: DIG-SITE BIN DATABASE ................................ 92
TABLE 13: ROW SCANNING AVERAGE READ RATE DATA FOR EACH TA(934.
TABLE 14: COLUMN SCANNING AVERAGE READ RATE DATA FOR EACH
TAG ............................................... 96
TABLE 15: RANDOM SCANNING AVERAGE READ RATE DATA FOR EACH
TAG ............................................... 98
TABLE 16: BIN INVENTORY AVERAGE READ RATE DATA FOR EACH TRlIgg.
TABLE 17: DlG-SITE BOOK DATABASE ............................ 104

viii

LIST OF FIGURES

FIGURE 1: PROCESS MAP FOR ONGOING CURATION TASKS IN A

REPOSITORY ......................................... 7
FIGURE 2: RFID SYSTEM HARDWARE COMPONENTS USED FOR R-RMS

TESTING ............................................ 40
FIGURE 3: LABELED ARTIFACT BAGS ............................... 44
FIGURE 4: ARTIFACT ACCESSION CARD ............................ 45
FIGURE 5: BAGGED ARTIFACTS IN OPENED TOP HDPE STORAGE BIN. .45
FIGURE 6: CLOSED TOP STORAGE BINS ON METAL SHELF ............ 45
FIGURE 7: SHELF OF HANDWRI'I'I'EN FIELD BOOKS .................. 46
FIGURE 8: RFID TAGGED ARTIFACT BAGS IN A STORAGE BIN .......... 47
FIGURE 9: TAGGED STORAGE BINS ON METAL SHELF ................ 48
FIGURE 10: COLUMN SCANNING TECHNIQUE ........................ 49
FIGURE 11: ROW SCANNING TECHNIQUE ........................... 50
FIGURE 12: TAGGED BOOKS ON SHELF ............................ 50

FIGURE 13: SUCCESS AND FAILURE TOTALS FOR FIRST ATTEMPT BAG
ACCESSION TRIALS ................................... 56

FIGURE 14: SUCCESS AND FAILURE TOTALS FOR FIRST ATTEMPT BAG
RECORDS RECALL TRIALS ............................. 56

FIGURE 15: SUCCESS AND FAILURE TOTALS FOR FIRST ATTEMPT BIN
ACCESSION TRIALS .................................. 58

FIGURE 16: SUCCESS AND FAILURE TOTALS FOR FIRST ATTEMPT BIN
RECORDS RECALL TRIALS ............................. 58

FIGURE 17: STATISTICAL ANALYSIS OF READS FOR EACH TAG BY
SCANNING METHOD .................................. 60

FIGURE 18:

FIGURE 19:
FIGURE 20:

FIGURE 21:

FIGURE 22:

FIGURE 23:

FIGURE 24:

FIGURE 25:

FIGURE 26:

FIGURE 27:

FIGURE 28:
FIGURE 29:

FIGURE 30:

FIGURE 31:

FIGURE 32:
FIGURE 33:
FIGURE 34:

STATISTICAL ANALYSIS OF READS FOR EACH TRIAL BY

SCANNING METHOD .................................. 61
OVERALL AVERAGE BIN TAG READ COUNTS .............. 63
EDGE BINS AND CENTER BINS USED FOR STATISTICAL
ANALYSIS ........................................... 63
STATISTICAL ANALYSIS OF READS FOR EACH TAG BY BIN
POSITION .......................................... 64
SUCCESS AND FAILURE TOTALS FOR FIRST ATTEMPT BOOK
ACCESSION TRIALS .................................. 65
SUCCESS AND FAILURE TOTALS FOR FIRST ATTEMPT BOOK
RECORDS RECALL TRIALS ............................ 66
RFID-TAGGED BOOKS SHOWING HIGHEST AND LOWEST
READING BOOK TAG ................................. 68
NUMBER OF ITEM-LEVEL TAGS DETECTED IN EACH
INVENTORY TRIAL ................................... 73
RFID TAG DETECTION RATES FOR ITEM-LEVEL INVENTORY
SCANS .............................................. 70
NUMBER OF SUCCESSFUL TRIALS FOR THE FIVE ITEMS THAT
WERE MISSED IN AT LEAST ONE TRIAL .................. 70
DENSELY POPULATED LOOSE ITEMS .................... 71
OVERALL INVENTORY DETECTION RATES FOR MISSED ITEM-
TAGS ............................................... 71
POINTS OF INTEGRATION FOR RFID IN ARTIFACT
PROCESSING ....................................... 78
PROCESS MAP FOR ONGOING CURATION IN USING AN R-
RMS ................................................ 80
CARVED IDENTIFIER ON BULK ARTIFACT ................. 85
TIN PLATE IDENTIFIER ON BULK ARTIFACT ............... 86

DIG-SITE REPOSITORY TESTING ENVIRONMENT DIAGRAM. . .
.................................................... 89

FIGURE 35: MUSEUM REPOSITORY TESTING ENVIRONMENT DIAGRAM. . .

................................................... 90
FIGURE 36: AVERAGE READ RATE PER BIN TAG: ROW SCANNING
TECHNIQUE ........................................ 101
FIGURE 37: AVERAGE READ RATE PER BIN TAG: COLUMN SCANNING
TECHNIQUE ........................................ 102
FIGURE 38: AVERAGE READ RATE PER BIN TAG: RANDOM SCANNING
TECHNIQUE ........................................ 103

xi

KEY TO ABREVIATIONS

Auto-ID ........................................ Automatic Identification
CPG ....................................... Consumer packaged goods
EPC .......................................... Electronic Product Code
GUI .......................................... Graphical user interface
HDPE ....................................... High-density polyethylene
HF .................................................. High frequency
lC .................................................. Integrated circuit
lFF .............................................. Identify friend or foe
LF ................................................... Low frequency
PE .................................................... Polyethylene
RF ID ..................................... Radio frequency identification
RMS ................................... Repository management system
R-RMS ...................... RFID-enabled repository management system
UHF ............................................. Ultra high frequency
UID ................................................. Unique identifier
UPC .......................................... Universal Product Code

xii

Chapter 1- Introduction

 

The term ‘repository’ refers to a range of institutions
which provide ongoing care for collections (Sullivan &
Childs, 2003). Repositories may be run by museums,
private collectors, government organizations, tribal
groups or academic research groups. Collections may
contain a wide variety of materials and objects, or may be
comprised entirely of a single type of item. The objects
themselves may have historical or cultural importance,
though this is, in no way, a requirement. In short, a
repository can be as large, diverse and historically
significant as the Smithsonian Repository, or as narrowly-
focused and irrelevant as an old man 's collection of

antique hubcaps.

Despite the differences, all repositories are unified by the
common goals of providing a proper storage environment
for a group of objects and maintaining an accurate record
of the collection, a practice generally referred to as
curation. This section provides an overview of artifact
curation basics and discusses common sources of error in

traditional repository management systems (RMS).

The Modern Repository

This section presents an overview of typical artifact curation practices in an
archaeological repository. Although the curatorial methods described in this
thesis are applicable to a wide range of repositories, for the purposes of this
research, the term repository will be used in reference to academic repositories

which generate artifact collections through archaeological field-work.

In this type of repository, the collection is generated by the same organization
that is responsible for ongoing care. As a result, a system is needed for
managing artifacts and information at every step of the archaeological process.
This process starts when an artifact is first acquired and registered in the
repository records, and continues in perpetuity with the ongoing curation of an

archaeological collection.

Creating Collections

The basic function of an RMS is to establish and maintain a link between
physical artifacts and the associated historical record (Reibel, 1997, p. 44). Table
1 summarizes the steps for establishing this link as an artifact is registered in an

RMS.

Table 1: Addino an artifact to a re osr'to collection
Step in Processing Function

   
 

 

 

 

 

 

 

Discovery Record provenience
Assign temporary identifier
Accessioning Ledger entry created
Accession card created
Cataloging All related documentation assembled
Accession card updated
Collections Objects packaged using archival materials
Storage location documented on accession card

 

2

When artifacts are found in the course of an archaeological survey, there is a
great deal of data that must be collected along with the artifact to allow for
meaningful study later. Depending upon the nature of the given excavation, the
location of the discovery is recorded based on horizontal and vertical control
points or with an established grid system (Fladmark, 1978, p. 83). Measurements
are taken according to a predetermined precision level, which can range
anywhere from a general geographic region, to GPS coordinates that specify the
exact position of discovery in three dimensional space (Sullivan 8. Childs, 2003,
p. 87). Data is recorded by archaeologists in the form of field notes, photographs
and drawings. This information is collectively referred to as provenience, and
provides a permanent record of the context in which an artifact is discovered. To
provide a link between recovered objects and provenience, artifacts are given a
temporary ID upon discovery, which is associated to the field records and
associated data. In some cases, when sharing a single set of provenience,
individual artifacts may be grouped based on predetermined divisions in space or
common characteristics, to simplify in field operations. With the exception of
large objects, such as architectural components which may be left in the field,

artifacts are then transported from to laboratories for accessioning.

Accessioning is the process of formally taking a new object into the repository
collection (Sullivan & Childs, 2003, p. 61). The first step is assigning a unique
accession number for each object. An accession number may be determined
arbitrarily, chronologically, or like an EPC number, with defined fields based on

information about an object such as the year or location of discovery. A record is

then created for the new accession number and documented, by hand or
digitally, in the repository inventory ledger. The task of accessioning is completed
by marking the object with the accession number (Reibel, 1997, p.71 ). Methods
for labeling artifacts include hanging paper labels, affixed metal tags, or
permanent ink applied directly to the objects surfaces. In the case of large
objects, accessioning may take place in the field. Marking of such objects
requires the use of weather resistant attachment methods, which have
historically included painting or carving the surface, attaching metal and plastic
tags, or simply relying on written records to identify and locate objects. The final
step of accessioning is the creation of an accession card which is used, like a
library card catalogue, to direct users to the location of desired objects and

information.

Cataloging is the process of gathering all of the relevant documentation
relating to an object, and documenting these sources in a common location
(Sullivan & Childs, 2003, p. 63). Information may include object storage location,
provenience data, and access information for any related reference materials
such as drawings, photographs, field notes, or map, all of which may be
managed using independent catalogues. Records may be electronic, although

legacy methods resembling library card catalogues are still widely used.

Though many methods for cataloguing exist, the complex relationships
between objects and information are commonly managed using a relational data
structure. For example, individual objects maybe grouped in bags based on

common features such as provenience or discovery date, and then the bags are

4

subsequently grouped in storage bins. In this situation, it is necessary to maintain
catalogues and ID systems for objects, bags, and bins, respectively. Information,
too, is managed using a tiered structure, with accession numbers linked to field
book entries, which may in turn be linked to maps, photos or any other related

media.

Collections refers to the process of readying artifacts for long term storage.
The preparation needed for long-term storage will vary based on the material,
durability, chemical reactivity, value and size of an object, as well as
environmental factors within the storage environment. Artifact packaging is not
always required, but may include cushioning materials, bags, bins, crates and
pallets. When packaging is needed, specialized archival materials are used to
prevent harmful interactions with antiquities. Archival packaging materials, such
as bags, bins and paper are manufactured to be acid-free, a property which limits
the hamiful transfer of chemicals from the packaging to artifacts, while extending
the shelf life of the package itself. When packaging material is used, a bag or bin
tag is created to allow identification of the contents without removing the primary
package. This limits the amount of human interaction with often-fragile artifacts

(Sullivan & Childs, 2003, p. 67).

Ongoing Curation

Once a link is established between collections and the associated historical
records, the next function of an RMS is to provide a system for managing the
collection in perpetuity. Table 2, at the top of the next page, summarizes these

curatorial functions of an RMS.

Table 2: On oin . curation at a re osito
Curatorial Function Purpose

  
  
 

    
   

  

 

 

 

Collections Management Provenience maintained
Inventory auditing
Ongoing Research Finding objects for research
Accessing historical records and objects

 

 

Collections management is an ongoing process which enables artifacts and
records to be readily accessed within a storage facility. This is accomplished
using some form of repository registration system, which acts like the memory of
a repository, providing a permanent record of the entire collection contents
(Reibel, 1997, p.12). The records are also used to perform inventory audits to
identify lost, stolen or misplaced objects. In modern repositories, an inventory
audit is completed through the labor intensive process of comparing a collections
ledger to the physical collection, one object at a time, a task which can take

months (Texas Instruments, 2010, Reibel, 1997, p. 100).

Enabling the use of collections for ongoing research, is another important
function of a repository, and refers to the method by which artifacts and records
are located and recalled, in order to facilitate further study. Depending on the
system, finding a certain object or record may entail the use of a paper-based
card catalogue or, in newer systems, an electronic database search. The
effectiveness of this system is directly related to the quality of the inventory and
data management system. While no specific set of regulations exist for how this
system is constructed, many organizations do require minimum performance
standards for these systems. An established method for testing the effectiveness

of a collections management system is to simply select a random accession

number from the repository catalogue and manually locate all associated records

and objects.

Figure 1 shows the processes of taking inventory and recalling records, by
tracing the pathways between objects and identifying documents as the

aforementioned curation tasks are conducted.

Figure 1: Process map for ongoing curation tasks in a repository

J Field Book Field Drawings

L

Reference Materials I
-- --(----i AccessionCard

Maps Photographs
nil Primary Package I ---(-- Object
t

 

 

 

 

 

 

 

 

 

 

-‘

 

 

 

 

 

 

(...

 

 

 

 

 

 

 

 

r g - -2 InventoryAudit ----- «b Start -0
Records Recall —-) Finish 0
Step ->

 

 

 

 

 

The final role of a repository is to manage the release of artifacts, enabling
the loan, transfer or disposal of an object. The process of deaccessioning
describes the permanent removal of an object and all associated provenience,
from a collection, and is necessary to keep an accurate inventory record when
the ownership of an artifact is transferred, or when an object must be disposed of
due to excessive deterioration, material hazards, or if it is confirmed that the
object was acquired illegally (Sullivan & Childs, 2003, p. 72). Similarly, when on
loan, inventory systems must indicate that the object is absent from the
collection. However, additional information is also needed to record where the

object has gone, who is responsible for it, and when it is expected back. A

 

subsequent, complimentary system is also needed to establish within the

inventory records when an object is returned.
Relational Object Management and Cascading Errors

Though effective if well maintained, an artifact management system can
quickly grow in size, and become unwieldy. This problem is often caused by
inadequate repository storage environments, limited funding for the indefinite
care of artifacts, and the fragile nature of paper-based inventory and data
management systems. Further compounding the problem, a paper-based system
used for relating information to an object is inevitably subject to human error. Due
to the level of interconnectedness between objects and records, if an error is
made in any part of the inventory and information management system, the
impact on provenience can be, as one archaeologist grimly described,

“cascading errors” (Frey, 2009).

To illustrate the potential for error in this system consider the following
example of an object within a paper-based system: The object is properly
accessioned, catalogued, packaged in a bag with similar objects, grouped in a
bin with similar bags, and placed on a shelf the repository warehouse. The object
is described by a two page entry in a hand-written field book and the entry
contains one additional reference to a technical drawing of the trench in which

the object was found.

Type-1 failures occur when an accession card is misplaced, lost or destroyed,
or an object itself is misplaced. While the object remains intact and identifiable,

and the records remain intact, the ability to connect the two is temporarily lost.
8

Type-1 failures can be recognized in a complete inventory audit, and
provenience may be restored if both the artifact and the accession card can be

located.

Type-2 failures occur when an objects accession number is rendered illegible.
In this case, the item can no longer be tied to permanent records and the
provenience may be lost forever. Damage to labels can be caused by physical
impacts, erosion from outdoor storage, degradation of the labeling substrates
and even interference from animals. While artifacts such as sculptures, weaponry
or other rare finds may maintain aesthetic or research value in the absence of
provenience, the bulk of collections are comprised of objects of less significance,
such as pottery shards, small animal bones or rusted metal refuse, which require
detailed documentation to be of academic benefit. Often these items are simply
discarded in the absence of provenience. Type-2 failures also result in
inconsistencies in the inventory record, which still contains a record for the lost

object, causing further inefficiencies in the process of accounting for collections.

Type-3 failures occur when an object label or catalogue record is lost or
destroyed for an object which relate to a more than one artifact. For objects like
bags and bins, if the label is rendered illegible, all objects within the bag or bin
can also be considered lost. For objects like maps, site drawings and other
general records, a labeling failure results in the inability to access certain portions
of information relating to the entire group of related objects. Type-3 errors can be
identified by an inventory audit and the provenience may be restored based on

accession numbers.

Type—4 failures are the most severe with respect to impact on provenience,
and occur when field books are lost or destroyed. When this happens, every
record for every object described within is also lost. The affect to ongoing
research is especially devastating when considering that field books may contain
thousands of individual records, and may account for a full season of excavation
and discovery. To hedge against potential threats, a secure method for managing
the library of field books is critical, though they are often managed in the same

manner as other records.

Regardless of where the inventory and data management failure occurs, the
problem can be attributed to one of two root causes; ID imperrnanence, resulting
in loss of provenience, or human error, resulting in inaccurate inventory records.
Either way, the result is the same, an inability to efficiently access the materials

needed to conduct ongoing research using archaeological repository collections.

10

Research Objective

This research evaluates the performance of an RFID system in conducting
useful inventory management functions at archaeological repositories. Based on
the curation methods described in this chapter, opportunities for beneficially
applying RFID were identified and an RFID-enabled repository management
system (R-RMS) was developed. The system included an accession function for
registering new artifacts, a data management function for retrieving digital

records, and an inventory function for performing inventory audits.

Test scenarios were designed to evaluate the basic operation of the R-RMS,
and conducted in two different artifact repositories. Testing was performed in the
environment of intended use in order to verify the operation of the RFID system,
while also accounting for all potential sources of destructive interferences in the
given repository environments. The results of the RF ID system evaluation are
used to develop a process diagram for effectively managing an artifact repository

using RFID technology.

11

Chapter 2: Literature Review

 

T he first decade of the twenty first century has seen
radio frequency identification (RFID) technology
evolve fiom relative obscurity into an entire family
of tools for automated capture of object-identifying
information. This chapter discusses RFID as a
subset of the larger family of automatic
identification (A uto-ID) technologies, and describes
the technology in terms of operational and physical
characteristics. A brief history of RFID technologv
development highlights major technological
milestones and industry trends which have resulted
in the recent growth of niche markets and closed-
loop RFID applications, such as library and

museum management systems.

12

Automatic Identification

Generally stated, the term automatic identification refers to any technology
used to facilitate the transfer of information from one source to another, without
the need for human input (Clarke, 2008). Common examples of automatic
identification technologies include magnetic stripes on credit cards used to direct
financial transactions, barcodes used to facilitate retails transactions and support
supply chain operations, and biometric fingerprint and retinal scans used for
secure access control. Each system utilizes a unique core technology to facilitate
the transfer of information, while sharing the common goals of “identifying,
tracking, recording, storing and communicating essential business, personal, or
product data... providing fast and accurate collection and entry of data”
(Association for Automatic Identificaton and Mobilty). The benefits of Auto-ID can
be illustrated by considering the role of the UPC barcode in modem retail store

check-out systems.

In a modern retail store, customer orders are processed by individually
scanning barcodes printed on, or attached to, items being purchased. When
scanned, Item-identifying information is optically captured, processed, and used
to generate a running total of a customer order in a single step, without a worker
manually keying in pricing information. The automation of order entry enables
efficient customer processing and limits problems associated with human errors
in data entry. These benefits of increased efficiency and error reduction are

universal across all forms of Auto-ID (Glover & Bhatt, 2006, p. 4).

13

Another commonality among Auto-ID technologies is the underlying method
by which data is stored, transferred and captured. Information is stored as a
machine readable binary code, which is detected by a data capture device,
converted to a digital signal and transferred to a host processor. In Universal
Product Code (UPC) barcodes, an alphanumeric identifier is stored optically, as a
series of vertical, black and white lines of varying thicknesses. The dimensional
variations correspond to a binary code, representing the identifier. When the
barcode is scanned by a laser, light is absorbed by the black lines and reflected
by the white lines. The reflected light is detected by optical sensors within the
scanner and then analyzed to extract the binary code. The resulting binary code
is then translated back into the original alphanumeric identifier and transmitted to

the host computer (Sweeny, 2005, p. 34).

Despite fundamental similarities in concept, the technologies used for storing
and transferring machine-readable information vary widely. Magnetic stripe card
technologies represent a binary code utilizing a strip of magnetically polarized
metal particles, held within a binding medium. The particles are oriented
positively or negatively, representing a binary 0 or 1, and are read by devices
which detect the magnetic orientation of the particles. Common examples of
magnetic stripe cards are access control cards, government issued ID cards, and

the vast majority of modern credit cards.

Biometric technologies, like fingerprint or retinal scanning, work by analyzing

the spatial relationships of key features in the images of scanned fingerprints or

14

retina. The result is a binary code which is decoded and used to identify an

individual.

Smart cards, despite taking the same form as magnetic stripe credit cards,
hold identifying information on a small computer chip or microprocessor with,
embedded within a plastic card. When placed into a reader, exposed electrical
contacts on the card create a direct connection between the memory of the card
. and the reader unit, enabling the data transfer. Smart cards are most commonly

seen in the forms of prepaid phone cards (Weier, 2009).

Another Auto-ID technology of note Is the two dimensional (2D) barcode. Like
UPC barcodes, 2D barcodes use optical scanning to transfer information. UPC
scanners project a single line across the code and detect the line thickness, in
one dimension. Therefore, the amount of data that can be stored by a 1D code is
determined by the length of the barcode. 2D barcode scanners, on the other
hand, project a beam of light like a flashlight, allowing information to be stored in
two dimensions. The increased complexity of the spatial relationship between
reflective and non-reflective areas allows for increased data storage density. As a
. result, the amount of data stored by a 2D code can be greatly increased, while
the size of the code increases relatively little, compared to a 1D code. 2D
barcodes codes can be printed for less than a cent and the information capacity
of is comparable to RF ID, making them an attractive alternative to RFID in some
cases. Data capacity issues and operational characteristics are further discussed

in a later section.

15

In the case of RF ID, identifying information is stored, in binary, on a tiny
computer chip, attached to the object being identified. The data is transferred
using electromagnetic energy in the form of radio waves. The following section
provides an overview of the components which make up an RFID system, and

how these components interact to perform the task of automatic identification.
RFID System Components

All RFID systems operate using some variation of four basic components: a ‘
transponder, a reader, a reader-antenna and a host computer. The functionality .
of an RFID system can be enhanced with addition of supplemental devices such I
as machine vision sensors, RFID tag printers and label applicators. In addition to
hardware, RFID middleware programs are run by the host computer in order to
efficiently direct the operation of an RF ID system and simplify information

management.

Hardware Basics
Though a great amount of diversity exists in the physical form and operation
of RFID hardware, all RFID systems include four primary components: a

transponder, a reader, a reader-antenna, and a host computer (Sweeny, 2005).

RFID transponders, or tags, are used to store information used for object
identification. A common form for RFID tags resembles a 2" x 4” mailing label.
Integrated into the label is a tiny computer chip, called an integrated circuit (IC),
attached to a metallic antenna. The IC memory holds a unique identifying
number, as well as a small amount of internal logic to facilitate higher level

functionality. The antenna is used to harness electromagnetic energy, receive
16

commands from the RFID system, and emit a response signal canying the data

stored on the IC.

The reader-antenna, used to emit and receive RF signals, is comprised of a
flat piece of metal within a protective housing, connected to a reader via coaxial
cable. When an electrical signal, generated within the reader, reaches the metal
plate within the antenna, the signal is transformed and emitted into the
surrounding environment in the form of radio waves. This signal powers the
operation of the transponder and carries operational commands from the reader.
The tag response is subsequently received by the antenna, transformed back

into an electrical signal and sent the reader.

The RFID reader is responsible for facilitating the operation of, and
communication between, RFID transponders and a computer server. The reader
contains computing hardware to store operational logic as well as radio
equipment to generate and receive RF signals. Commands are initially sent from
the computer interface to the logic portion of the reader, canying instructions for
performing functions such as writing information to tags, reading information from
tags, and providing variable setting information, such as power levels, or
specialized encoding specifications, which indicate how the basic tasks of writing

and reading are to be performed (Glover & Bhatt, 2006, p.108).

Once the reader receives a command from the computer interface, it uses the
internal radio signal generator to emit an RF signal via attached reader-antenna.

The transponder response is then returned to the reader, where the signal is

17

translated by the reader logic components and relayed to the computer interface

for further processing.

Variations on these basic components include portable RFID units, which
combine a reader, antenna and computer interface in a dingle handheld device,
and reader/antenna combination units, which eliminate the need for an external
antenna, allowing the units to be installed in areas where limited space is a
concern. Other currently available RFID-enabled devices include RFID-barcode
combination readers and a variety of RFID label printer/encoders designed to
meet the scalability and performance requirements of disparate RFID

applications.

Making Use of It All: Middleware

In order efficiently direct the operation of RF devices and extract actionable
information frem'the extensive raw data, an automated solution is needed.
Middleware is a software program on a host computer that serves as the
necessary link between an RF ID hardware network and a central information

management system.

Based on either human or event-triggered events, middleware programs
direct reader operation and perform tasks such as “RFID data dissemination,
data filtering and aggregation, reading from and writing to a tag, reader
integration in IT-service management, Privacy” (Lampe, 2009). To understand
how the RFID system is used to accomplish these inventory tasks, it necessary
to understand some of the reader commands used by the middleware program.

Frequently used commands are displayed in Table 3, at the top of the next page.

18

Table 3: Common RFID Reader Commands

 

 

 

 

Command Purpose Reader Response

Read Sends a command to acquire List of detected tags
information from all transponders
within the read field at a given point
in time.

Write Sends a command to program Notification of success or failure
specified information to the memory All data programmed to the tag
of a desired transponder.

Attenuate Changes the power level of the RF Notification of setting change
signal emitted by the reader-
antenna.

Mask A mask is a filter used to direct Notification of setting change
reader commands to specific tags,
and eliminate the generation of
excessive and undesired read data.

AutoMode Sets the reader to continually detect Notification of setting change
tags and store read data to reader when activated
memory. Complete accumulated tag list

(up to 5000 tags) when
deactivated

 

 

 

 

In order for the operation of an RFID system to be of benefit, the data

collected must be utilized to perform a useful function. In order to combine the

tasks of device management and information management, middleware

programs have two basic layers, “a general- purpose core architecture,

extensible for pervasive computing applications, and an RFID-related application
architecture that specializes in the generic agent architecture” (Thompson, 2006).

This double layered approach is necessary because, while the memory capacity

of RFID tags is always expandng, the information stored on the chip doesnot

generally include complete object information. Instead, the tag data typically acts

as a liscence plate which the middleware uses to retrieve object information from

an independent electronic database.

 

The complexity of a middleware solution is determined by the complexity of
the application, which may range from performing a single function to managing
a complete RF-enabled enterprise. In the environment of a consumer packaged
goods (CPG) supply chain, RF ID event triggers may be anything from the arrival
of a specific truck at a distribution center to a customer placing an online order.
These triggers may be tied to perform any number of tasks such as, creating a
bill of lading, verifying the contents of a customer order prior to shipment, or
updating material planning systems to reflect changes in inventory demands. In
any case, the middleware is responsible for gathering data from RF ID tags, and

then utilizing that data to direct the completion of a useful task.

The need for middleware is further demonstrated when considering that the
number of Ule reported in a normal day in a CPG supply chain could be in the
millions per hour. In a network such as this, a single middleware program must
be seamlessly networked to any number of geographically dispersed read-points,

hosted on any number of computers, mobile devices or central servers.
Auto-ID with Radio Waves

RF ID, like all forms Auto-ID technologies, offers the benefits of accelerated
data capture and reduced error. There are, however, several unique advantages
to using radio communication for identification, as opposed to magnets, optics or
other identification mediums. These benefits include the ability of RFID systems
to read tags through many common materials, acquire data quickly from multiple

transponders and hold enough data to allow for a UID.

20

Electromagnetic energy is able to propagate through many common materials
such as plastic, paper, rubber and wood. As such, RFID tags can be detected
through many materials without the requirement of a direct line-of-site. Whether
or not a tag can operate without a Iine-of—site is determined by how well radio
waves propagate through the given material, a property called radiolucency
(Clarke, 2008). This ability means that, for specialized application, tags can be
completely embedded within an object to enhance transponder durability. This
also means that individual transponders within a unitized population, such as a
stretch wrapped pallet of tagged shipping cases, can be scanned in entirety,
without removing the stretch-film. Replacing barcodes vw'th RFID tags in the
aforementioned grocery store check-out could potentially allow cashiers to scan

an entire shopping cart of goods without touching a single item.

Another unique capability of RFID is the high speed at which systems can
acquire data from transponders. Advertised read rates of tag read rates may
range from 20-1000 tags/sec, though observed read rates are typically on the
lower end of that range (Ramakrishnan & Deavours, 2007). Communication time
can become an issue, however, when there is a large population of tags in the
read field simultaneously. If the RF-antenna is not given adequate exposure time
to complete a communication cycle with each transponder, objects may be

missed.

In addition to transferring data quickly, RF ID transponders have the capability
to store up to 864 bits of data in the IC memory (Allen Technologies (II), 2008).

The extensive on-board memory capacity allows RFID tags to hold enough data

21

to provide a UID for 3.12144 different tagged objects, based on the 480-bit

memory bank of the Alien Higgs 3 chip (Alien Technologies (II), 2008).

The concept of a UID is another important difference between UPC barcodes
RFID. The data stored on UPC barcodes includes a manufacturer’s ID number
and a stock keeping unit (SKU). The code is designed to generically identify the
product by type, for example, 1Ib bag of Generic Brand Frozen Cod. RFID tags,
too, provide product and manufacturing information, but hold an additional UID
which is designed to identify each object individually among a larger population
of similar items, for example, Generic Brand Frozen Cod- 1 lb bag #5968492.
The UID can also be used to gain a more granular view of a population of objects
by allowing advanced filtering of tags and the abiltiy to selectively issue

commands.
An RFID for (almost) any Situation

When implementing an RFID solution, it is important to understand how
environmental and material interferences can impact the performance of RFID
systems. Furthermore, in selecting RF ID hardware for an application it is useful
to understand how the operational characteristics of various RFID technologies

can be used to overcome environmental hazards.

RFID, it turns out, is not a specific technology, but rather a collection of
technologies which all operate in a similar way. A leader in RFID industry news
and information, RFID Journal defines the term Radio Frequency Identification
as, “any method of identifying unique items using radio waves” (RFID Journal (II),

2010). As this definition suggests, radio waves can be utilized in a number of
22

ways to accomplish the goal of identification. Methods may differ in frequency,
tag power source and in the integration of additional technologies. Each variant
has distinct operational characteristics which are determined by the physics of
the respective method. Examples of functional differences, commonly cited to
guide application development decisions, include: read distance, communication

speed, sensitivity to water or metal, and extended functional capabilities.

Physical Limitations

In the early years of RF ID commercialization, a commonly used selling point
was that RFID does not need line-of-site to operate. This claim, however, is only
sometimes true. Certain materials can inhibit the operation of RFID systems by
absorbing, blocking or otherwise interfering with RF energy. In these instances, a
direct line-of-site may be required in order for a tag to be detected. The most
commonly cited causes of material interferences with RFID systems are due to

water and metal in the reading environment.

Water has a negative impact on RF ID communication because it absorbs
electromagnetic energy at certain frequencies (Clarke, 2008). The energy
intended for powering the tag is thereby dissipated and as a result, little or no
energy remains to power the transponder. This is a concem for RF ID system
developers because many environments contain water. Consider consumer
products such as bottled beverages, fresh produce, meats, shampoos,
windshield washing fluid. All of these products contain enough water to

negatively impact the performance of RF ID systems. Furthermore, uncontrolled

23

water in the environment such as rain or humidity can also reduce the overall

perfonnanoe of an RFID system.

Metal objects can negatively affect RF communication in two ways. First, RF
energy cannot propagate through metal. If a tag is surrounded on all sides, it will
be unreadable, unless it is specifically designed for the given application. Metal
can have a secondary impact on the performance of an RFID system by causing
electromagnetic interference capable of detuning the delicate radio components
of an RFID transponder. This detuning affect can causes a change in the
resonant frequency of a transponder (Clarke, 2008). If a tag is detuned to the
extent that it no longer responds to the specified frequency, the tag will no longer

be able to communicate with the RFID system.

Metallic objects of concern in the CPG supply chain include packaging
materials such as the common “tin” can or foil layered flexible packages,
equipment handling machinery like clamp trucks and forklifts, metal storage
racks, conveyor systems, and dock doors. RFID transponders can even be
detuned and rendered inoperable by other transponders. This is a particular
concern when considering the use of RFID for tagging small items which are

stored in a high-density population of other transponders.

Frequency

Depending on how and where an RFID system is to be applied, a number of
operational frequencies for communication exist, each offering a different set of
benefits and short falls, based on the unique properties of various RF

frequencies used for item identification.

24

When it comes to radio communication, interference can occur when to
different signals are broadcast at the same frequency. As a result, governments
allocate certain frequency bandwidths for certain communication needs. With
respect to RFID, standards for frequency usage vary according to the regulatory
body in charge of frequency allocation in a given geographic region. In the United
States, for example, radio communication at a frequency band from 902 MHz to
928 MHz (commonly stated as 915 MHz) is designated ultra high frequency
(UHF). In Europe, however, the UHF band refers to frequencies from 865- 868
MHz. For the purposes of this discussion, frequency bands from the United
States are presented to illustrate the effect of frequency variation on the
performance of an RFID system. Worldwide standards for UHF spectrum use are
provided online by GSI, a global standards organization which focuses on
standardizing new technologies for improving the efficiency and visibility of

products as the move through the supply chain (681, 2009).

Low frequency (LF) and high frequency (HF) RF ID systems operate using
magnetic coupling at frequencies of 125 KHz and 13.56 MHz, respectively. In
both cases, the effective read range is typically limited to under a meter. LF and
HF radio waves, though limited in read distance are affected very little by metal
or water, making them well suited to applications in animal tagging, asset
management and payment systems, where read distance may not be critical, or

desirable, and where metal and water may be a interfering factor.

Ultra high frequency RFID systems in the United States operate at an

average frequency of 915 MHz. Communication is accomplished using

25

propagated far-field RF signals, in a process called passive backscatter. Read
ranges for UHF systems can commonly reach 30 ft, depending upon the system,
making them ideal for large warehouse environments in the CPG supply chain.
The performance of these systems, however, is heavily influence by water and
metal. As previously suggested, this fact has proven a major stumbling block for
wide spread supply chain adoption, due to the wide spread presence of

environmental water and metal in the retail distribution environment.

Passive or Active

Another way in which RFID systems differ is based on where the transponder
gets the energy for operation. The RFID systems discussed to this point are
considered passive systems. This designation means that the energy required to
power the operation of a transponder is emitted by the reader-antenna and
harnessed by the transponder-antenna. When the tag is not within a suitable
read range, the tag will not receive enough energy to operate, and will instead
remain dormant until it moves to a position within the effective RF read field

(ASTM lntemational, 2009).

Fully active transponders are those which include a separate power source
such as an internal battery. Rather than waiting for energy from a reader-antenna
to respond, active tags are always operating. The additional energy allows the
system to overcome much of the RF interference resulting from water and metal,
and enables read ranges of over 1 kilometer (Sweeny (II), 2007). The additional
power source also allows for the incorporation of additional functionality, such as

environmental sensors or extended memory for continual data logging.

26

Depending on how active transponders are configured, battery life may be

anywhere from several months to several years.

Despite the superior performance characteristics, the size and cost of active
RF ID transponders are both significantly greater than passive system, and is, as
a result, primarily used for the management of high cost or high priority items.
Active RF ID has been used by the Department of Defense to track and locate
shipping containers in theatres of war, and in the development of industrial real-

time locating systems (RTLS) for asset tracking.
History of RFID Development

Once referred to as a solution looking for a problem, Radio Frequency
Identification (RFID) has developed dramatimlly since being introduced in
1939as a World War ll-era friendly-fire prevention tool (Kleefeld, 2005), into a
diverse family of technologies with an equally diverse set of applications. This
section focuses on the evolution of passive UHF RFID systems, from the first use
of radio waves for identification, to modern closed loop systems designed for the

purpose of academic enhancement.

A Slow Start for a New lDea

Like many modern technologies, RFID was first developed for applications
within the military. Of British design, the identify friend or foe (IFF) system was
designed to allow air base radar operators the ability to efficiently differentiate the
generic radar “blips” of incoming planes as hostile or friendly (RF ID Journal,
2005). This system used a radio signal generator at the base to broadcast a

signal at a certain frequency. Archaic transponders mounted on friendly aircrafts,
27

received this signal and then responded with an amplified signal of the same
frequency. The presence or absence of the response signal was then used to

determine the allegiance of the incoming craft.

Further developments in RF communication, throughout the 1950’s and
1960’s, resulted in the introduction of onboard internal memory for RF ID
transponders. “These types of systems often use ‘1 -bit’ tags — only the presence
or absence of a tag could be detected, but the tags could be made inexpensively”
(Landt, 2001 ). The tags were used primarily for anti-shoplifting applications.
When an item was purchased, the single bit of data on the item tag would be
changed from a 1 to a 0. If scanners at the store exit detected a tag displaying a
1, an alarm would be triggered, alerting workers that an un-purchased item was

leaving the store.

In the 1970’s and 1980’s, various iterations of RFID were developed for
specialty applications like nuclear waste management, animal identification,
keyless entry, and automated highway toll systems. Transponder functionality
was improved with increased memory capacity and the ability to reprogram tag
data. These early systems primarily utilized LF and HF frequencies, employed a
wide range of proprietary hardware, and were largely limited in scope due to the

still-infantile nature of the technology.

It was not until the early 1990’s that “IBM engineers developed and patented
an ultra-high frequency (UHF) RFID system. UHF offered longer read range (up
to 20 feet under good conditions) and faster data transfer” (RF IDJournal). The

extended read range and improved data capabilities made UHF systems a

28

primary candidate for applications in supply chain management. The idea of an
RFID-enabled supply chain for the consumer packaged goods (CPG) industry,
led to the establishment of the Auto-ID Center in 1999. Centered at the
Massachusetts Institute of Technology (MIT), The Auto-ID Center was started as
a joint effort between the Uniform Code Council, EAN lntemational, Procter &
Gamble and Gillett, to develop RFID for the CPG supply chain. By 2003, the
Center had gained the support of over 100 major corporations, including the
Department of Defense, and had led the development of critical RF ID standards,

providing the foundation upon which all future UHF systems would be built.

Standardization Nurtures RFID Growth

The development of certain operational standards was a critical step in the
expansion of RFID in the CPG sector because, in order for the vision of an RFID-
enabled supply chain to work, every organization that utilizes the tag must be
able read and understand the information contained on the chip, regardless of
who manufactured or last programmed the tag. Two standards were developed
to establish these essential guidelines: The Electronic Product Code (EPC)
standard, which established a few basic hardware definitions and a system for
representing manufacturer and product information as a UID, and the Gen2
standard which provided a functionally-optimized, universal communication
language for operation between compliant RFID transponders and readers,

regardless of manufacturer.

EPCglobal was originally founded between The European Article
Numbering/Uniform Code Council (EAN.UCC) and The Auto-ID Center at MIT. In

29

2004, EPCgIobal approved the EPC standard which is “a numbering scheme that
uniquely identifies all objects” (EPCglobal, 2006). The system is intended for
identifying RF ID tagged products within the global supply chain. Similar to the
Universal Product Code for barcodes, an EPC number identifies the
manufacturer, product class and other general identifying information. In addition
to the standard indicators found in the UPC code, the EPC code includes
information which identifies each object uniquely. So, whereas a UPC may
identify an object as being a box of cereal, an EPC identifies the box of cereal
individually among the larger population of cereal boxes from the same product
line. Without this standard numbering format, retail stores, for example, would
have to accommodate all of the specialized numbering schemes developed, for

internal use, by individual manufacturers.

Having established the format of the information contained on EPC tags, the
next standardization need was to ensure the interoperability of RFID hardware
components from disparate manufacturers. While all transponders canied the
same standardized EPC data set, the programming language used to transfer
information from the transponder to the reader remained hardware-manufacturer-
specific. This problem was a major obstacle for early RFID adoption in the CPG
supply chain because companies found it difficult to justify the high cost of
implementation without a guarantee that an RF ID hardware infrastructure would
remain relevant and compatible with future generations. The case for a
standardized communication protocol for RF ID hardware was further solidified by

the need for universal operation of RFID transponders between partnered

30

business entities, regardless of RFID manufacturer. The solution was the

development of the Gen 2 air-to-air protocol standard.

Developed by EPCglobaI and then adopted by ISO (Intemational
Organization for Standardization) in 2006, the Gen 2 standard established a
standardized protocol for facilitating communication between EPC—compliant tags
and readers, regardless of the hardware manufacturer. The standard established
minimum requirements for tag memory capacity and intemal logic capabilities, as
well as the architectural design for RFID tag memory banks and a method for
interacting with the stored data. Although Gen 2 compliant hardware components
from disparate manufacturers can interact, it should be noted that the

performance of the resulting system may not be optimized.

With the assurance of a stable operational platform, the UHF RFID industry
focus moved from proprietary hardware development to application development
and implementation. RFID mandates from the CPG giant Wal-Mart and the
Department of Defense (DoD) served as a further catalyst that brought RFID

from its state as an emerging technology, to center stage in the early 2000’s.

Mandated Jumpstart for RFID

WaI-Mart, realizing the supply chain potential of RFID and holding the power
to fuel the development of an entire industry for UHF RFID, in June 2003, issued
a mandate to Its top 100 suppliers that, beginning in 2005, all shipping containers
received at select distribution centers were required to carry an EPC tag. The
stated target of the Wal-Mart Initiative was to reduce stock-outs, by utilizing up-

to-the-minute inventory records to manage their supply chain in real time.

31

Despite several successful small scale pilot tests with major suppliers
throughout the eariy 2000’s, by October 2007, the mandate for RF ID had been
altered and reduced to “three focus areas: (1) shipments going to Sam’s Club; (2)
promotional displays and products going to Wal-Mart stores; (3) tests to see
RF lD’s impact in improving category management in select areas” (Supply Chain

Digest, 2009).

Retrospective analysis of the Wal-Mart RFID program shows that the primary
reasons for failing to meet expectations was the infantile nature of the hardware
systems due to a lack of commercialization in UHF RFID industry and the

absence of a sturdy foundation in technological standardization.

Shortly after the introduction of the Wal-Mart initiative, the DoD issued a
similar mandate to every one of its 40,000 suppliers. By 2007, all shipping cases
received at DoD facilities were required to carry an EPC tag. While the program
faced similar setbacks as the Wal-Mart initiative, a DoD policy states a
commitment to pursuing all relevant forms of Auto-ID technologies as the
preferred tools for supply chain management (Department of Defense, 2003).
This policy has allowed the DoD to continually adapt and grow with the RFID
industry, without the financial restrictions of industry. The underlying focus of the
DoD initiative was to gain greater visibility of material and supplies within
distribution and storage networks. Along with the EPC mandate, an active
tagging program was also pursued by the DoD for enhanced container tracking

and security.

32

Other less notable RFID projects followed including an initiative to use HF
RFID as part of a system for tracking of prescriptions drugs as part of the fight
against the grey-market and counterfeit pharmaceutical trade. While the first five
years of the 2000's brought sweeping standardization and an explosion in
attention for the UHF RFID industry, the five years that followed were marked by
a significant decline in interest from the consumer goods industry as pressure
from the Wal-Mart mandate diminished and businesses slowly realized the
complexity of executing a complete RF lD supply chain solution and the simple

fact that RFID was still too expensive and technologically underdeveloped.

Understanding Hype

When UHF RF ID was first brought into the spotlight, there was a flurry of
interest and disorganized activity within the industry. Slowly however, the
technological limitations were realized and the industry lost momentum towards
ubiquitous adoption. This is a common trend which is common to most emerging

technologies, and is described by the Gartner Hype Cycle (Fenn, 2009).

With respect the cycle predicts that on the road to full adoption the technology
will first experience an innovation trigger. the establishment of the Auto-ID Center
and the Wal-Mart Mandate, leading to the peak of expectations: when media
hype and additional mandates from the DoD and the healthcare industry gave
credence to the potential of UHF RFID. Following the peak however, a dramatic
decline in hype leads to the trough of disillusionment when companies realized
the costs and limitations of early UHF systems, and the mandates slowly lost

momentum. Following the trough, Gartner predicts a stage called the slope of

33

enlightenment. This stage is defined by the slow realization of technological
development efforts and a rise in small scale development projects. This is

roughly were the UHF RFID industry at the time of writing this thesis.

In summary of this concept, the author likens the development of the RFID
CPG industry to the Big Bang: a lot of shock and awe all at once, followed by a
few billion years of waiting for things to cool down before anything useful takes

shape.

Growth of Closed-Loop RFID Markets

The 2009 Gartner Hype Cycle report puts RFID at the very base of the slope
of enlightenment (Fenn, 2009). While the Gartner Cycle rating is based largely on
the CPG industry, recent proliferation of RF ID niche markets and closed-loop

applications suggest that ‘little brother“ maybe a little bit ahead, on RFID.

A report from lnforrnation Week describes the current growth the UHF
industry as “innovation-driven” (Weier, 2009), meaning that mandates are no
longer the primary driving force in the development of RFID systems. In 2009,
the largest user of RF ID tags was the smart cards sector, which includes
passports, state IDs and prepaid charge systems. The second largest user of
tags, however, was not the military, retail sales or animal tracking. Instead the
second biggest user of RF ID tags was the undefined “Other” category, indicating
a fragmented industry, engaged in developing an array of unique ways to use

RFID (Weier, 2009).

A report published by ABI Research summarizes the effect of industry

mandates on the current RFID industry, saying, “Many vendors, financial

34

analysts, and industry observers were hopeful that (compliance programs) would
provide dramatic year-over-year growth. While this did not happen, hope was
never lost. Due to those (programs), passive UHF had a chance to mature in
terms of standards, performance, form factors, and pricing” (Liard & Carlaw,
2009). These improvements paved the way for new RFID systems designed for

small closed loop applications.

Closed-loop applications for RF ID are those in which tags are only utilized by
the organization that controls the system. This differs from open-loop systems
such as CPG and military supply chains, which may attempt to utilize a single
RFID tag to track objects as they pass between multiple of trading partners. For a
system like this to function, however, information systems between partners must
be compatible and a standardized data exchange format must be determined.
Reconciling the plethora of data management systems utilized by various
organizations proved a major stumbling block to early RFID adoption. Managing
a closed-loop RF ID system is comparatively easier, since a single organization
can customize a complete RFID solution, without regard to compatibility or

compliance issues with other businesses.

Table 4 provides several existing closed-loop RFID applications and a brief

description of the basic project goals.

Table 4: Common closed-loo a lications for RFID

     

 

 

 

Baggage Tracking Efficient verification of luggage owner
Reduce lost bags
Animal Tagging Efficient identification of livestock
Provide animal-specific tracking for disease
monitoring

 

 

35

Table 4: Common closed-loop applications for RFID (cont’d)

 

Asset Management Locating items within a storage environment
Provide a record of usage and maintenance

 

Brickyard Management Inventory Control

 

 

Pallet Rentals Automate Shipping, Receiving, Repair Scheduling
Automate Billing
Production Monitoring Track work-in-process inventories
Allow for efficient customization in manufacture-on-
demand

 

 

 

 

RFlD Applied for Academic Enhancement

The spread of RFID has not been limited to projects with the specific goal of
direct financial returns. RF ID has also been applied for the purpose of improving
the management of, and access to information for both academic researchers

and the general public.

Libraries around the country have used RFID to improve operations since
2000. The American Libraries Association (ALA) offers perspectives on RFID by
highlighting the benefits of rapid charging/discharging, simplified patron self-
charging/discharging, high reliability, high-speed inventorying, and automated
materials handling (Boss, 2004). In the end, librarians spend less time organizing
and more time helping patrons, while patrons spend less time searching and
more time accessing relevant information. Library RF ID systems emerged

around 2000 and have traditionally utilized the HF band.

In 2010, The Vatican Museum Library began deploying an HF RFID system
for managing a collection of over two million books, manuscripts and other
priceless items. "RF ID improves the way librarians manage their collections,
streamlining and automating item retrieval, storage and inventory processes,"
said Bill Allen, marketing communications manager, Texas Instruments RFid

36

Systems©. When completed, the RFID system will allow a complete verification
of the museum inventory to be completed in a single day, a process which once

required shutting the museum down for an entire month out of every year.

HF RFID is also being utilized at The Malaysia National Museum, though not
for the purpose of entertaining visitors. Here, the artifact management system
has been designed for the specific purpose of caring for the artifacts, and
maintaining a permanent connection between artifacts and the associated

historical records (Berhad).

Museums, too, are finding RFID a useful in the management of artifact
collections, while also experimenting with RFID enabled exhibits to enhance user
experience and gain patronage. As early as 2004, The Museum of Natural
History, in Aarhus, Denmark, deployed an HF RFID system designed to allow
patrons to interact with exhibits using handheld FDA-style RF ID readers. The
exhibit included 50 uniquely tagged, stuffed birds that visitors could scan to gain
instant access to a variety of multimedia feature relating to the bird species,

including videos, articles and quizzes (Khan, 2004).

In 2008, Library Hi tech Journal published a research paper discussing an
application of a similar RFID library system with added features of “. .. regional
seeking and positioning for collections in mis-shelves or collections loss in library,
statistics calculation in usage frequency and length of book reading, and instant
library guiding service” (Yu, 2008). These functional additions to the ALA-
envisioned RFID system were made possible by utilizing the long read range

capabilities of UHF RFID.

37

Monetary returns, though not of primary concern, have also been realized
through the implementation of RFID library and museum systems. Through
increased efficiency in inventory management, the need for manual organization
of books and historical artifacts is significantly reduced. Libraries and museums
also benefit from the introduction of RFID as a result of the process control
requirements for operating an RF ID system. Manually driven inventory systems
are subject to unintended variation over time and inconsistent performance due
to the number of people who use the system, even without necessarily

understanding how it works.

Given the numerous examples of recent RFID implementation successes in
niche applications and closed-loop inventory management systems, as well as
the recognized capability of RFID to enhance the way people access information,
it is clear that the potential of RFID for managing archaeological collections,

reaches far beyond the museum.

38

Chapter 3: Methodology

 

T o evaluate the performance of modern UHF
technology in performing basic repository
management functions, a firlly fimctional RFID-
enabled repository management system (R-RMS)
was developed for testing at two different

archaeological repositories.

This chapter will describe the methods used to
evaluate the R-RMS at an archaeological dig-site
repository and at a university-run museum
repository. It will detail he preparation and testing
of RF ID-enabled bags, storage bins, library books
and individual artifacts. Test methods will be
provided for using the R-RMS for accessioning
objects, recalling electronic database entries and

performing inventory audits.

39

Materials

RFID System Components

Evaluation of the R-RMS was conducted using an Alien Technologies ALR-
9650, a Gen 2 compliant UHF RFID reader (Figure 2, left). The desktop reader
unit measures 9" x 9” x 2” and includes a built in circular antenna. The
combination reader/antenna is primarily designed for stationary use and was
selected for this study based on its compact form factor and ease of operation.
These features are of particular importance in a repository, where storage and
research space Is often at a premium, and system operators may not be

experienced in the use of RFID equipment.

Figure 2: RFID system hardware components used for R-RMS testing

 

For extended range inventory tasks, an additional antenna was attached to
the external antenna port on the reader. This study utilized an Alien ALR-9610-
CR circularly polarized antenna (Figure 2, right), connected via twenty foot
coaxial cable. A circularly polarized antenna was selected for orientation
insensitivity in detecting transponders, a necessary property given the

uncertainty of artifact tag orientations in storage.

The RFID transponders used in this research were also manufactured by

Allen Technologies. The ALR-9662 Squiggle-SH RFID transponder (Figure 2,
40

middle bottom) is 3 Gen 2 compliant, global use UHF tag. Global use tags are
optimized for operation across a wide range of UHF bandwidths. This capability
allows each tag to be read on any Gen 2 compliant reader anywhere in the worid,
regardless of regional UHF frequency regulations. This Squiggle-SH transponder
is manufactured with a Higgs-3 integrated circuit that was selected for its
optimized performance with Alien brand readers and the unique ability to

selectively program portions of the EPC number, independently of the whole.

Network Components

The Alien ALR-9650 was connected to a host computer over a local network,
which consisted of a Netgear 4-port wireless router (Figure 2, middle bottom). A
Toshiba Satellite laptop running Windows Vista OS was used to host the R—RMS
software and the repository information databases. The reader and the computer

were connected to the router via standard Ethernet cords.

Evaluation Software

To simplify the process of collecting test data, an R-RMS middleware
program was created specifically for this application. The Java-based evaluation
software provided a graphical user interface (GUI) for directing operation of the
RFID hardware and automatically managing an electronic database The R-RMS
middleware included software features for tagging new items, recalling item-

specific information and inventory auditing of an entire population.

The R-RMS middleware feature for tagging a new item was used to program
the EPC data of object tags and create object entries in an electronic database.

The task is accomplished by placing a new tag directly on the ALR-9650 desktop

41

unit, typing object specific information into specified data fields, and then
submitting the command using the GUI. This action triggers two distinct events in
the evaluation middleware. First, a command is sent to the reader to program the
tag EPC number. If the tag is programmed successfully, the software the updates
the appropriate spreadsheets to include the new object entry. If the tag fails to
program properly, the software will halt the operation, and an entry for the tag will
not be created. This mechanism is necessary to avoid incomplete or duplicate
object records resulting from failed or repeated attempts to program a tag for a

given object.

The feature for retrieving object information was used to access electronic
database records based on the EPC number of the RFID object tag. In operation,
a tagged object is placed directly on the Alien ALR-9650 desktop unit and the
evaluation middleware GUI is used to issue a command for the reader to perform
a single read cycle. The objects detected in this scan are displayed as an on-
screen list from which the desired item is selected. A complete digital record for

the object is then displayed.

The inventory auditing feature was used to determine the presence or
absence of every RFID tagged object in a repository collection. To perform an
audit of a population, the first step was to submit a command for the reader to
begin detecting tags in AutoMode, via the auxiliary antenna. With the reader
continuously reading, the auxiliary antenna was used to scan the entire
population of tags. Upon completion, a second command issued for the reader to

stop detecting tags and to return a complete list of the tags detected during the

42

scan. Based on an automated comparison of this list to the electronic database,
the software generated two lists. The first list showed the object tags that were
successfully detected during the inventory scan, indicating that these items were
accounted for within the storage environment. The second list displayed any
object tags that were present in the digital record, but were not detected during
the inventory scan, indicating that these objects may not be present in the

storage environment.

In summary, the evaluation middleware was developed to simplify the tasks of
tagging objects, accessing information and verifying inventory. These functions
can be applied in a repository to individual artifacts, artifact group bags, storage
bins and library books. Regardless of object type, operation of the middleware

program is unchanged.
Dig-Site Repository Testing

Description of Dig-Site Testing Environment

The Temple of Poseidon in lsthmia, Greece has been an active
archaeological site since 1952. Throughout the last half-century, researchers
from the University of Chicago, the University of Califomla Los Angeles, and The
Ohio State University have managed the extensive collection of artifacts and
information, generated through continued excavation. As the basis for artifact
management at the dig-site repository, an accession card system was used to
establish relationships between artifact bags, bins, books and other reference

materials, based on paper ID labels for each class of object.

43

Testing of the R-RMS was conducted in the long term storage annex of the
dig house which included a countertop along one wall, a bookshelf along an
adjacent wall, and two rows of floor-to-ceiling metal shelving for organizing a
collection of plastic storage bins for artifacts. A diagram of the dig-site repository
test site is provided in Appendix I. Repository management testing focused on

three classes of objects: bags, bins and books.

Bags are a class of objects that are stored in the dig-house. The most
commonly observed primary package for the long-term storage of artifacts was
archival-grade polyethylene (PE) bags. Due to the wide range of artifacts found
on-site, the sizes and contents of bags vary greatly. Less common artifacts, such
as fine pottery, toys and weapons are documented and packaged individually.
Items such as building tile fragments, which occur more frequently, are grouped
based on common provenience, documented as a single lot of items, and then

consolidated in a single bag. In many Figure 3: Labeled Artifact Bags
. a {If

    

cases, impact absorbing sheets of archival-
grade PE bubble-wrap were folded in half
and placed inside storage bags to form a

protective lining.

Bags were identified using laminated ID
cards placed Inside each bag, but outside the
layer of bubble-wrap for convenient
readability (Figure 3). Each bag tag displayed

the following data fields: Accession Number,

44

Area, Trench, Gn'd, TopE/ev, BotE/ev, ExDate, Notebook and Page. Additionally,

Figure 4: Artifact Accession Card for every artifact documented, an accession

  

card was created to aid researchers in
locating the desired objects and accessing

historical records (Figure 4,).

Bags were grouped based on common
artifact characteristics and consolidated in
high density polyethylene (HDPE) storage bins measuring either 15 x 23 x 8
inches or 15 x 23 x 6 inches (Figure 5). A standardized method for packing bags
into bins was not given due to the wide variety materials, sizes and shapes of

artifacts requiring bin storage. Figure 5: Bagged artifacts in
coened to HDPE storae bin

       

On the outside of each bin, centered along
the exposed 15 inch dimension, was attached a
hand-written tag that displayed the accession
number of each bag contained within. Metal
racking was used for bin storage, with shelves

configured to hold four bins each, two stacks of two bins (Figure 6).

Figure 5_- Closed top storage The last piece of the inventory and information
bins on metal shelf

management system was the library of hand
written field books (Figure 7, top of next page).
The books comprise every formally documented

piece of information relating to every artifact

 

Figure 7: Shelf of recovered in the history of the excavation. Entries
HandwrittenFie/d Books

    

are organized chronologically according to the

new 'Wllllllltltr _'_._A date of discovery. Books are searched using a

hand-written card catalogue system.

RFID Bag Management Evaluation

Twelve RFID test bags were created for the evaluation of bag-level RFID
inventory management operations. Assorted terracotta pottery shards from a
mass storage bin comprised of artifacts lacking provenience were placed into
archival grade PE bags, measuring seven inches by ten inches. As per repository
practice, a sheet of bubble wrap was folded in half and placed in the bag as a
padded liner. Lastly, an RF ID tag was placed inside each bag, outside the layer

of bubble wrap.

Test 1: Bag-Level Accessioning
To test the reliability of the RFID system in performing the task of
accessioning artifact bags, the evaluation middleware was used to program an

RFID tag and create a digital ledger entry for each of the twelve test bags.

Test bags were placed on the Alien ALR-9650 such that the transponder was
face down and centered over the reader's antenna component. Each bag tag
was individually tested a single time using the accessioning feature of the test
software. Data fields were populated with information used to uniquely identify

test bags. The success or failure of each trial was determined in Test 2.

46

Test 2: Recalling Bag-Specific lnforrnation
To evaluate the performance of the RF ID system in providing quick access to
bag-specific reference materials, the evaluation middleware was used to scan

bag tags and access digital ledger entries.

Test bags were placed on the Alien ALR-9650 desktop unit in the same
position as Test 1. Each bag tag was trialed one time in accordance with
operation of the information recall feature of the test software. Success or failure
of each trial was determined by verifying accuracy of the bag-identifying

information displayed on-screen.

The results of Test 2 trials were also used to determine the success or failure
of the Test 1 trials for accessioning bags. Due to the mechanism for preventing
the generation of erroneous ledger entries, only successfully programmed bag
tags will have a ledger entry. Therefore, if a bag entry was successfully recalled,

the success of the accessioning process was also demonstrated.

Test 3: Simple Bag Inventory Scan
Upon completion of Tests 1 and 2, the twelve tagged artifact bags were

Figure 8: RFID tagged placed in an HDPE storage bin, measuring 15 x
artifact bees in a stora-e bin

    

23 x 8 inches, as per standard dig-house practice
. (Figure 8). The artifact bin was then scanned
twice, using the inventory audit function of the
evaluation middleware, once with the bin in each
i of the two possible shelf positions. First, the

storage bin was tested when placed directly on the metal shelf of a storage rack,

47

in the standard orientation, with another bin stacked on top. Then the test bin
was tested when placed on top of another bin on the storage rack, again, in the
standard orientation. The number of reads per tag was recorded for each one of

the Inventory trials.

RFID Bin Management Evaluation
Test specimens for bin inventory testing consisted of an existing inventory of

plastic artifact storage bins. Using the test-software, seventy-nine RFID bin tags

Figure 9: Tagged storage bins on were created and then randomly attached,
metal shelf

one per bin. Masking tape was used to
hold tags flat against the outside of the
bin, in a horizontal orientation, centered

along the exposed 15 inch dimension of

 

the bin (Figure 9). The seventy-nine test

bins comprised one complete face of the metal storage racking.

Test 4: Tagging new bins
To test the reliability of the R-RMS in performing the task of accessioning
artifact bins, the evaluation middleware was used to program an RFID tag and

create a digital ledger entry for each of the seventy-nine test bins.

Prior to accession testing, the seventy-nine transponder specimens were
assigned a number from one to seventy-nine, and labeled using a permanent
marker. The Alien Squiggle-SH transponders were then programmed using the
same method as Test 1, accessioning bags. Data fields were left blank, with the

exception of the field for Description, which was populated with the same number

48

written on the tag being programmed. The success or failure of each trial was

determined in Test 5.

Test 5: Retrieving Bin-Specific lnfonnation
To evaluate the performance of the RFID system in providing quick access to
bin-specific reference materials, the middleware was used to scan bin tags and

access digital ledger entries.

Using the same method as Test 2, each bin tag was trialed one time in
accordance with operation of the middleware feature for recalling object
information. Success or failure of each trial was determined by verifying the
accuracy of the identifying information displayed on-screen. These results were
then used to determine the success or failure of bin accessioning trial, as

previously discussed.

Test 6: Taking bin inventory
After randomly attaching the seventy-nine newly programmed bin tags to test
bins, the evaluation middleware was used to perform bin inventory testing in
accordance with the “Take Inventory' software function. Three techniques were
evaluated for scanning the population Figure 10: Column scanning technique
of tagged bins using the auxiliary
antenna: column scanning, row

scanning and random scanning.

The first scanning technique moved

the auxiliary antenna in an alternating

 

49

up-and-down motion to scan adjacent Figure 11: Row scanning tecmhque

    

columns of bins (Figure 10). The
second scanning technique moved the
antenna in an alternating left-to-right
motion to scan adjacent rows of bins
(Figure 11). The third scanning
technique did not establish a pattem for scanning RFID bins. After each inventory
trial, the number of reads for each tag was recorded. The test was repeated thirty

times for each scanning technique to allow for statistical analysis of the test data.

RFID Book Management Evaluation

Test samples for evaluating the RFID book management system were
created using thirty-nine books from the dig-house library collection: seven
hardcover volumes and thirty-two paperback journals. An RFID tag was placed
inside the front cover of each book with the long edge of the tag wedged into the
binding, centered to the height of the book. After completing accessioning and

scanning trials, the tagged books Figure 12: Ta- oed Books on Shelf

    

were moved to a wooden bookshelf

for inventory testing (Figure 12).

Authentic handwritten notebooks
were not used for testing due to strict
regulations aimed at preventing Type-4 errors of lost field notes. However, the
thirty-two journals used for testing were taller and thinner than the authentic

books, resulting in a denser tag population than would be seen in the repository

50

library. Due to this comparatively higher-density tag population, and the effects of
tag-to-tag interference, the library auditing test scenario actually presented a
greater challenge for collecting RFID read data. Therefore, if functionality of the
R-RMS is successfully demonstrated, then the functionality of a library

management system for the authentic field notes can be inferred.

Test 7: Tagging new books
To test the reliability of the RF ID system in performing the task of
accessioning new books, the evaluation middleware was used to program an

RFID tag and create a digital ledger entry for each of the thirty nine test books.

Test books were placed on the Alien ALR-9650 such that the bound edge of
the book was centered over the reader’s antenna component, and then
programmed using the established method for object accessioning. Data fields
were populated with accurate information relating to the book being accessioned,
in order to create a complete ledger entry for each sample book. The success or

failure of each trial was determined in Test 5.

Test 8: Retrieving book-specific information
To evaluate the performance of the RFID system in providing quick access to
book-specific reference materials, the evaluation middleware was used to scan

book tags and access digital ledger entries.

Using the established method for recalling object information, each book tag
was trialed one time in accordance with operation of the R-RMS middleware.
Success or failure of each trial was determined by verifying the accuracy of the

identifying information displayed on-screen. These results were then used to

51

determine the success or failure of book accessioning trials, as previously

discussed.

Test 9: Taking Book Inventory
After accessioning, tagged books were placed back on the wooden bookshelf
for book inventory testing. The evaluation middleware was used to perform book

inventory testing in accordance with the inventory auditing software function.

Book inventory scans were conducted with the auxiliary antenna pointed
towards the books, held about two feet away from the shelf while scanning. The
antennae was moved from right to left across the shelf and then back, left to
right, for each inventory trial. Four trials were conduced and the number of reads

per trial was recorded for each tag.
Museum Repository Testing

Following the evaluation of R-RMS capabilities at the dig site repository, the
RFID system was tested further at an artifact repository for a university-run
museum. While much of the dig site testing was intended to verify several
functional capabilities of the RFID system for repository management
applications, testing at the museum repository was aimed at validating the

performance of the item-level inventory auditing feature.

Description of Museum Testing Environment
Unlike the rustic environment of the dig-house, the museum repository was
housed in a university class room. Along one wall of the room was a row of eight

metal cabinets, with inside dimensions of 27 x 30 x 72 inches each. Inside the

52

cabinets, artifacts were simply organized, often times without primary packaging,
on roll-out metal shelves, measuring 24 x 29 inches. Artifacts accession numbers
were written directly on the surface of the items with India ink, due to the lack of

packaging to facilitate attachment of paper ID cards.

The computer, network and RF ID equipment were assembled on a metal
desk which was located along a wall, adjacent to the row of metal cabinets,
approximately twelve feet from the storage cabinets to be tested. A diagram of

the museum repository testing environment is provided in Appendix II

ltem-Level RFID Management

Evaluation of the R-RMS item-level inventory auditing function was performed
at the museum repository using 157 artifacts from the MSU Artifact Teaching
Collection. Prior to on-site testing, the evaluation software was used to create an
RFID tag and an electronic ledger entry for each of the artifacts to be tested. The
information needed to create an accurate artifact ledger was provided by an
existing inventory list for the collection. The 157 artifacts were housed in and on

top of two adjacent storage cabinets, on twelve different metal shelves.

Due to the lack of primary packaging for unitizing objects and RFID
transponders, the pre-programmed tags were placed alongside the associated
artifacts, in a manner which attempted to recreate realistic variations in tag
position. Given the wide range of items, general rules were developed for placing
tags with certain types of objects. For tagging bowls and pottery, RF ID tags were
placed inside individual vessels. Arrowheads, stone hand-tools and other ancient

blades were stored flat on the metal shelves, and were relatively dispersed from

53

surrounding artifacts. As a result, tagging was accomplished by neatly laying tags
flat, on top of the artifacts. By contrast, tagging of densely populated loose items
such as Clovis artifacts or Iithic fragments was accomplished more haphazardly,

by placing tags as close to the correct artifacts as possible.

rest 10: Taking Item Level inventory

With all of the artifact tags in place, the R-RMS inventory feature was
tested in accordance with operation of the “Take Inventory” feature of the test
software. Before each trial, every pull-out shelf which held a tagged artifact was
fully extended for scanning by the auxiliary antenna. Upon starting the inventory
scan, the antenna was swept over the top of each shelf, starting at with the
highest shelf and working downward. Each shelf was scanned for three to five
seconds before being pushed back into the cabinet, revealing the next shelf
below. Inventory scans were completed when all twelve shelves had been

scanned.

The result of each trial was documented, identifying which items were
detected and which items were not. Item-level inventory trials were repeated

thirty times to ensure statistical validity.

54

Chapter 4: Results

 

This chapter presents the results of RF ID repository
management system evaluations conducted at the
dig-site repository for excavations at The Temple of
Poseidon in lsthmia, Greece, and the Michigan

State University Museum repository.

55

Bag Testing

Test 1: Tagging new bags
The results of Test 1, accessioning new bags, are displayed in Figure 13,

below. The graph shows the total number of bag tags, out of twelve, which were

successfully programmed on the first Figure 13: Success and failure totals

for first attem tba accession trials
attempt, and total numberwhich failed p g

 

15

to program on the first attempt. 12
3 10
In bag tag accession trials, 100% I:
a!
5
of the twelve bag tags were m
o
successfully programmed on the first °
Success Failure

attempt. The complete bag database

generated by the evaluation software in this test Is presented in Appendix III.

Test 2: Retrieving bag-specific information
Figure 14 depicts the results of Test 2, recalling database records from bag

tags. The graph shows the number of tags, out of twelve, which returned the

correct information from the electronic Figure 14_- Success and failure totals
for first attempt bag records recall
database after the first attempt, and

 

15

the number of tags which failed on the 12

a 10
first attempt. .2

O)

3 5

Results show that bag specific
0
information was successfully recalled 0
Success Failure

on the first attempt for 100% of the

twelve bag tags.

Test 3: Simple Bag Inventory

A single inventory scan was conducted with the bin of RFID tagged bags
placed on top of another shelved bin (off shelf), and then again with the bin
placed directly on the metal shelf beneath another bin (on shelf). Table 5 shows

the results of Test 3, inventory auditing of bag Table 5.. Read count data from
simple inventory scans of

tags. The table shows the total number of reads unitized bags
for each of the twelve bag tags during a single
1 24
inventory trial of each bin placement scenario. 2 19
3 43
The bag inventory scan results show that 4 20
5 12
100% of tags were detected an average of 5 25
7 22
22.83 times each, in the inventory scan of the 8 20
off shelf artifact storage bin. 100% of bag tags 190 :3
were also detected in the inventory scan of the 11 14
12 2o
storage bin placed directly on the metal shelf, 22.83
Std Dev 8.31
though a lower average read rate was observed Min 12

 

Max 43

of 22.33 per tag.

In the off shelf inventory scan of the bin of RF ID tagged artifact bags Tag 3
was read the most, 43 times, while Tag 5 was read the least, 12 times. When the
bin was placed directly on the metal shelf, Tag 3 was again read the most, 35
times. Tag 6, however, was only read twice despite being read 26 times in the
preceding off-shelf inventory trial. Later review of the bag tag placement revealed
that Tag 6 had shifted in the bag and been pressed flat against the bottom of the

bin by the associated artifact, the top of a broken terracotta amphora.

57

Bin Testing

Test 4: Tagging new bins
The results of Test 4, tagging new bins, are displayed in Figure 15, below.

The graph shows the total number of bin tags, of seventy—nine, which were

successfully programmed on the first Figure 15: Success and failure totals
for first attempt bin accession trials
attempt, and total number of tags 79
so
which failed to program on the first 60
a
U)
attempt. i5 40
5
. . . m 20
In bin tagging trials, 100% of the 0
o
seventy-nine bin tags were Success Failure

successfully programmed on the first attempt. The complete bin database

generated by the evaluation software in mis test is presented in Appendix IV.

Test 5: Retrieving Bin-Specific Information
Figure 16 depicts the results of Test 5, recalling database records from bin
tags. The graph shows the number of tags, of seventy-nine, which returned the

correct information from the electronic Figure 16: Success and failure totals
for first attempt bin records recall trials
database after the first attempt, and 79
so

the number of tags which failed on the 60

m
U)

first attempt. :3 40
C

u . a 20

Results show that brn-specrfic 0
0
information was successfully recalled Success Failure

on the first attempt for 100% of the bin tags.

58

Test 6: Bin Inventory

The bin inventory auditing results summarized below are based on a
statistical analysis of the average number of reads per trial for each bin tag. For
each of the three scanning techniques, the average bin tag read rate was
determined and used to compare the performance of the techniques in
performing inventory audits on the seventy-nine RFID-tagged storage bins. The
interquartile range (IQR) was also determined by the statistical analysis, and
used to determine the degree of performance variation between individual tags,

according to the scanning technique.

. Using a row scanning technique, 100% of bin tags were detected in
every trial.

. Using a column scanning technique, 100% of bin tags were detected in
every trial.

. Using a random scanning technique, 100% of bin tags were detected
in every trial.

. The row scanning technique showed the highest average read rate per
bin tag, 14.8 per trial.

. Column scanning and random scanning techniques showed similar
average read rates for bin tags, 10.7 and 10.8 reads per trial,
respectively.

. The highest read rate for a bin tag, 31.6 reads per trial, was observed
using a row scanning technique. The lowest read rate for a bin tag, 3.4
reads per trial, occurred when using a random scanning technique.

. The row scanning technique showed the most variation in bin tag read
rates, with an IQR of 8.6 reads per tag.

a The column scanning technique showed the less variation in bin tag
read rates than the row scanning technique, with an IQR of 4.6 reads

per tag.

. The random scanning technique showed the least variation in bin tag
read rates, with an IQR of 3.4 reads per tag.

59

Table 6 displays the data for the bin inventory results listed above. A
complete test summary of individual bin tag read rates is provided for each of the

three scanning techniques in Appendices V (a), (b) and (c).

Table 6: Summarized bin ta read rate data

 

 

 

 

 

 

O o o 0

Average 14.84 10.70 10.75
Median 14.07 10.37 10.87
Q1 10.23 6.83 6.85
Q3 18.85 13.77 14.10
IQR 8.62 6.93 7.25
Max 31.57 20.67 23.67
Min 5.33 4.63 3.37

 

 

 

 

 

 

Based on the data in Table 6, Figure 17 displays a box-and-whisker diagram
for each of the three inventory scanning techniques. Box-and-whisker diagrams
display distribution information for a data set. The upper and lower quartiles are
used to define a box, representing the IQR. The median value is displayed as a
line within the IQR box. It is clear from this diagram that the row scanning

technique showed the highest read rates for bin tags.

Figure 17: Statistical analysis of reads for each tag by scanning method
35

w
o

25

20

15

10

Average Read Rate per Bin Tag

U1

 

0

Row Column Random

60

While it is important to maximize the readability of tags to ensure accurate
inventory auditing, it is equally important to ensure repeatability of the results as
well. The bin inventory auditing results summarized below are based on a
statistical analysis of the average number of reads per bin tag for each trial. The
IQR is used to determine the degree of variation across thirty trials, for each of
the scanning techniques. The observed variation between trials indicates the
repeatability, and subsequently the reliability, of the three inventory auditing

methods.

. The row scanning technique showed the least variation in average per—
trial read rates, with an IQR of 1.8 reads per tag.

. The column scanning technique showed more variation across thirty
trials than the row scanning technique, with an IQR of 4.4 reads per
tag.

0 The random scanning technique showed the most variation across
thirty trials, with an IQR of 3.4 reads per tag.

Table 7 displays the data for the bin inventory results listed above. A
complete test summary of individual trial read rates is provided for each of the

three scanning techniques in Appendix VI.

T 7:

     

61

Based on the average per-trial read rates displayed in Table 7, Figure 18
displays a box-and-whisker diagram for each of the three scanning techniques.
From this diagram it is clear that the row scanning technique resulted in the least

variation, as well as the highest read rates for bin tags, across thirty trials.

Figure 18: Statistical analysis of reads for each trial by scanning method
18

16
14
12
10

Average Read Rate per Trial

 

ore-boson

Row Column Random

Regardless of bin inventory scanning technique, a trend of relatively lower
read rates was observed for bin tags located on the top row and far-right column
of the shelving unit (Edge Bins), as compared to tags more centrally located on
the shelves (Center Bins). An example of this trend can be seen in Figure 19, at
the top of the next page, which shows the average read rate of bin tags, for all
bin inventory trials combined. The diagram depicts the shelf position of each
tagged storage bin, identified by the randomly assigned bin ID number. Each bin
is patterned to indicate the average read rate of the respective tag according to
the arbitrarily determined ranges of: less than 1, 1-5, 6-10, 11-15, or 16 and
above. Similar figures for row, column and random scanning techniques are

provided in Appendices Vll (a), (b) and (c).

62

Figure 19: Overall average bin tag read counts

Avg. Read Count I Ceiling
0 rm);

 

 

 

 

 

IIeM

 

 

 

To formally evaluate this trend, a statistical analysis was conducted to
determine performance differences between edge and center bin tags, as

defined by Figure 20.

Figure 20: Edge bins and center bins for statistical analysis

 

| Ceiling

Edge Bins
Center Bins D

 

IIBM

 

 

 

 

   

63

The summarized results, below, are based on the statistical analysis which

looked at the average number of reads per trial for each bin tag.

. Tagged edge bins showed an average bin tag read rate of 8.2 reads
per trial, with an IQR of 2.2 reads per trial.

. Tagged center bins showed a significantly higher average bin tag read

rate than edge bins, with 15.2 reads per trial. Center bins also showed
greater variation than edge tags, with an IQR of 4.0 reads per trial.

Table 8 displays the full statistical analysis summary for the average bin tag

read rates of edge and center bin tags.

Table 8: Summarized bin ta . read data b bin . osition

 

 

 

 

 

 

 

 

 

Average 8.18 15.21
Standard Deviation 1.99 3.38
Median 7.82 14.71

Q1 7.02 13.09

Q3 9.25 17.13

IQR 2.23 4.04

Max 13.79 25.86

Min 4.72 6.78

 

 

 

 

 

Based on the statistical analysis data in Table 8, Figure 21 displays a box-

and-whisker diagram for edge and center bin tags, respectively.

Figure 21: Statistical analysis of reads for each tag by bin position
30

 

N
01

 

 

 

N
O

 

 

 

 

 

 

 

.3
O

 

E

Average Read Rate per Tag
01 a

 

O

 

Edge Bins Center Bins

64

Based on Figure 21, it is clear that the centrally located tagged bins perform
better, according to tag read rates, than bins located along the top and right-side

of the shelving unit.
Book Testing

Test 7: Tagging new books
The results of Test 7, tagging new books, are displayed in Figure 22. The

graph shows the total number of book tags, out of thirty-nine, which were

successfully Programmed 0” the first Figure 22: Success and failure totals for
first attempt book accession trials

attempt, and total number which failed 38

 

to program on the first attempt.

In book accession trials, 97% of

Book Tags

book tags (38 of 39) were successfully
1

 

programmed on the first attempt.
Success Failure

Though the exact cause of the single

failure is uncertain, the process of programming an RFID tag does require a

complex exchange of information between tags and readers, which can be easily

disrupted by environmental interferences or the presence of unexpected tags in

the read field.

The complete book database generated by the evaluation software in this test

is presented in Appendix Vlll.

65

Test 8: Retrieving Book-Specific Information
Figure 23 depicts the results of Test 8, recalling database records from book

tags. The graph shows the number of book tags, of thirty-nine, which returned

the correct Information from the Figure 23: Success and failure totals for
_ first attempt book records recall trials
electronic database after the first 38

4O

 

attempt, and the number of tags which 30

failed on the first attempt. 20

Book Tags

Results show that book-specific 1°
1

 

information was successfully recalled
Success Failure

on the first attempt for 97% of book

tags (38 of 39). The failure in recalling object-specific information was caused by

the initial programming failure In Test 7, as the result of a specific software

feature which prevents data from being entered into the database if a tag is not

programmed properly.

Upon discovering the tag programming failure, a second programming
attempt was made using the same RFID tag, resulting in a successfully

programmed tag and a properly updated database entry.

66

Test 9: Book Inventory
Table 9 shows the average number of reads per book tag, across a total of
four book auditing trials. Data is organized according to a random tag number,

determined when the tag is programmed.

Table 9: Avera oe number of reads oer book ta-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

67

 

 

T Average Std Dev Min
1 Volume 67: Number 2 8.00 2.83 6 10
2 Volume 63: Number 2 3.50 0.71 3 4
3 Volume 66: Number 3 5.50 2.12 4 7
4 Volume 63: Number 1 4.00 1.41 3 5
5 Volume 61: Number 4 5.50 3.54 3 8
6 Volume 61: Number 3 5.00 2.83 3 7
7 Volume 61: Number 1 3.00 2.83 1 5
8 The Metal Objects 9.50 4.95 6 13
9 Sculptures I 8.00 4.24 5 11
10 Volume 61: Number 1 5.50 3.54 3 8
11 Volume 65: Number 1 1.50 0.71 1 2
12 Volume 64: Number 2 3.50 2.12 2 5
13 Temple of Posseidon 9.00 4.24 6 12
14 Volume 65: Number 2 3.00 1.41 2 4
15 Volume 65: Number 1 2.50 2.12 1 4
16 Volume 62: Number 4 5.00 2.83 3 7
17 Volume 61: Number 1 5.50 3.54 3 8
18 Volume 61: Number 1 5.50 3.54 3 8
19 Volume 61: Number 2 5.50 3.54 3 8
20 Volume 63: Number 3 2.50 0.71 2 3
21 Volume 66: Number 4 7.50 2.12 6 9
22 Volume 63: Number 1 2.50 0.71 2 3
23 Volume 67: Number 1 8.50 2.12 7 10
24 The Hexamillion and The Fortress 9.00 4.24 6 12
25 Volume 63: Number 4 3.00 1.41 2 4
26 Sculpture ll: Marble Sculptures 9.50 3.54 7 12
27 Volume 64: Number 1 2.50 0.71 2 3
28 Topography and Architecture 7.50 3.54 5 10
29 Volume 64: Number 3 2.00 0.00 2 2
30 Volume 64: Number 4 2.50 0.71 2 3
31 Volume 64: Number 3 2.50 0.71 2 3
32 Volume 62: Number 3 3.50 0.71 3 4
33 Volume 63: Number 2 2.00 1.41 1 3
34 Volume 66: Number 1 5.00 1.41 4 6
35 Volume 66: Number 2 5.00 2.83 3 7
36 Volume 65: Number 4 4.50 2.12 3 6
37 Volume 67: Number 3 14.00 5.66 10 18
38 Volume 65: Number 3 3.50 2.12 2 5
39 The Late Bronze Age Settlement 9.50 3.54 7 12
Overall 5.26 2.84 1 18
Range
140 T 2-18

 

 

Based on the data in Table 9, 100% detection of book tags was observed in
each inventory trial, with an overall average of five reads per tag per trial. The
book tag for Hesperia Volume 67: Number 3 showed the overall highest
detection rate, with an average of 14 reads per trial, while Hesperia Volume 65:

Number 1 read the least, with an average of 1.5 reads per trial.

One reason for this discrepancy in RFID tag detection rate may be based on
the thickness and relative position of the tagged books on a shelf (Figure 24). As
previously mentioned, RFID tags can experience interference from other RFID
tags. In this tagging scenario, all of the RFID tags were placed at the same

location within the cover of the Figure 24: RFID-tagged books showing
highest and lowest reading book tag

 

books. As a result, when

shelved, the RFID book tags 4 3...... .. ..

i I u Hesperia Volume
directly align with other RFID ‘ . 5 ;.j; - 3 65: Number1
tags in adjacent books. The g figfigg Egg

T": ‘ I

poorest performing book tag,

Hesperia Volume
67: Number 3

Hesperia Volume 65: Number 1

 

was sandwiched between two

 

 

 

 

other tags. In addition, the books which contained the interfering tags were the
two thinnest books in the library collection. This means that the tag for Hesperia
Volume 65: Number 1 was closer to two adjacent tags than any of the other book
tags in the middle of the row. In contrast, the superior read rate of the book tag
for Hesperia Volume 67: Number 3 may be the result of the book being placed on

the end of the row of books. While the majority of tags have the potential for

68

interference from two adjacent tags, books on the end of rows have only one

potential source for tag-to-tag interference.
Museum Repository Testing

Test 10: Item-level Inventory

Test results displayed in this section are from item-level inventory testing
conducted at the museum repository. Figure 25 shows the number of tags, out of
157, that were detected in each of the thirty inventory trials, a total 4710 potential
tag reads overall. Results show missed tags in six of the thirty trials, with no trial

missing more than two of the 157 item tags.

Figure 25: Number of item-level tags detected in each inventory trial

ITags Read EJTags Missed
1 2 2 2 1 1

 

1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930

Percent of Tags Detected (of 157)

Trial Number

Results of item-level inventory auditing trials show a total of nine missed tags
across all thirty trials. This is equivalent to an RFID item-tag detection rate of
99.8% (4701 of 4710). Figure 26, at the top of the next page, presents this data

as a pie chart.

69

Figure 26: RFID tag detection rates for item-level inventory scans

ITags Missed DTags Read

IE!

4701

Of the 157 items, only five tags were missed in one or more times across
thirty trials. The total number of successful inventory tn'als for each of these fives

item tags is shown in Figure 27.

Figure 27: Number of successful trials for the five items that were missed in at
least one trial
30 27 28 28 29 29 __

 

Successful Trials (of 30)

 

Clovis- 73 Abrader- 106 Clovis- 68 Clovis- 90 Lithics- 126
Analysis of the five items to go undetected at least once reveals that four of
them belonged to groups of comparatively small artifacts such as Clovis items or
lithic fragments, which both were stored without primary packaging in densely
packed trays (Figure 28, at the top of the next page). Possible explanations for
this observation include tag—to—tag interference and inadequate exposure time for

detecting every tag in the tray. Despite this potential for tag-to-tag interference,

70

even the worst performing tag, assigned to Clovis- 73, was still detected in 27 of

the 30 trials, or 90% of the time.

Figure 28: Densely Populated Loose Items

 

The overall inventory detection success for each of the five missed tags is

displayed in Figure 29.

Figure 29: Overall inventory detection rates for missed item-tags
Clovis ltem- 73 Clovis ltem- 68 Clovis ltem- 90

:. 21.
Abrader- 106 Lithics-126

2 E1 Detected
Trials
I Undetected
Trials

Aside from occasional missed reads from the decidedly difficult-to-read item

 

 

 

 

tags, the only other tag to be missed in at least one trial was for the Abrader. The
Abrader was constructed entirely of stone and wood, materials which do not
interfere with RF ID readability, and the object was stored in an area of low tag

71

density Despite these favorable tagging conditions, the tag did not read in two
trials. In the course of testing it was observed that the tag had shifted from its
original placement on top of the Abrader, and was laying flat on the metal shelf
during the unsuccessful inventory scans. Once discovered, the tag was re-placed

before continuing inventory trials.

72

 

Chapter 5: Conclusions and Discussion

Based on the R-RMS evaluation, it is concluded that
UHF radio fiequency identification technology can
be successfitlly applied for the purpose of

archaeological repository management.

This chapter will present the conclusions drawn
from repository testing and discuss the suitability of
UHF technology for archaeological repository
management. The lessons learned from this
research are presented as they relate to an RFID-
enabled artifact accessioning system and a

repository management tool for inventory auditing.

73

 

Unlike previous chapters that divided discussion of the repository
management system (R-RMS) evaluation according to test numbers, this chapter
will combine test scenarios and present conclusions under three subsections:
dig-house artifact accession system, dig-house inventory management and
museum inventory management. Table 10 shows the relation of test scenarios to

R-RMS functions as described herein.

Table 10: R-RMS functions as related to RFID test scenarios
R-RMS Function Related Test Scenarios
Dig-House Repository: Test 1: Bag-Level Accessioning
Artifact Accessioning Test 2: Recalling Bag-Specific Information
Test 4: Tagging new bins
Test 5: Retrieving Bin-Specific Information
Test 7: Tagging new books
Test 8: Retrieving book-specific information

Dig-House Repository: Test 3: Simple Bag Inventory Scan
Inventory Auditing

 

Test 6: Taking bin inventory
Test 9: Taking Book Inventory

Museum Repository: Test 10: Taking Item Level Inventory
Inventory Auditing

 

 

 

 

 

Dig-House Repository: Artifact Accessioning

R-RMS testing at the dig-house repository was 992% successful (129 of 130)
in performing the functions of programming transponder EPC data, creating
object entries in an electronic database, and using RFID tag data to recall object-
specific database entries. Therefore, the R-RMS is capable of consistently

performing the basic functions of an RFID—enabled repository accession system.

74

Based on the one instance of tag programming failure, functionality was
demonstrated for the software mechanism for preventing erroneous database

entries, and the tag was successfully programmed on the second attempt.
Dig-House Repository: Inventory Auditing

An RFID system used for the task of inventory auditing must be able to collect
RFID read data from a large population of object tags and communicate the
accumulated read data back to the host computer for processing. Dig-site I
inventory testing of the R-RMS resulted in a 100% detection rate of the possible ,
7,290 tags across bag, bin and book inventory audit trials. Therefore, the RF ID
system is determined to be capable of reliably and accurately performing

inventory auditing at the archaeological dig-site repository.

In addition to validation of the RF ID inventory system as a whole, the
following conclusions were drawn from R-RMS inventory audit testing at the dig-

house repository:

. Bag tags within a closed storage bin can be successfully detected without
removing the bin’s lid.

. Metal shelving can negatively impact the readability of RF ID bag tags
when tags are positioned flat against the bottom of the storage bin.

o Readability issues associated with shelving placement can be alleviated if
attention is given to ensure that no bag tags are positioned against the
bottom of storage bins. Alternatively, a thin layer of foam padding may
provide sufficient separation from the metal surface to make the tag
readable.

. The readability of RFID tagged bins is affected by the placement of the
bins on storage shelves. Bins placed around the perimeter of a storage
shelf, near a physical structure such as a wall or ceiling, did not read as
often as bins that were located further from the barriers, towards the
center of the shelf.

75

. A complete inventory audit of seventy-nine bins can be completed in less
than twenty seconds.

. For bin inventory auditing, a row scanning technique will collect tag data
with 100% accuracy, more consistently than column or random scanning
techniques.

. Given the success of detecting tags on the test books, and that authentic
field notebooks are thicker than the books tested, the R—RMS can function
properly with authentic field notebooks.

Museum Repository: Inventory Auditing

Item-level inventory testing at the museum repository showed a 99.8%
detection rate of 157 tagged artifacts inside metal cabinets across thirty trials
(4701 of 4710). Therefore, the RF ID system is capable of effectively and

accurately performing item-level inventory audits in the museum repository.

. Item-level RFID tags functioned at 99.8% in the metal-heavy environment
of artifact storage cabinets.

a RF ID tags on artifacts that are stored without packaging in a dense
population can be successfully detected. To ensure reliability, these
objects should be scanned slower, and at a closer range than less densely
populated items.

. Using the R-RMS, a complete inventory audit of 157 tagged artifacts can
be conducted in less than one minute.

. Readability of item tags can be negatively affected by metal cabinet
shelves. To avoid this problem, an effort should be made to place tags
where they will not be in close proximity to metal shelving, metallic
artifacts, or other tags.

76

The RFID-Enabled Repository

 

This section synthesizes the findings of the RFID
repository management system evaluation in order
to develop a process map for utilizing an R-RMS to
manage archaeological collections from discovery

through ongoing curation in an RFID-enabled

repository.

77

The Transition from Paper to Microchips

Transitioning to an RFID-enabled repository management system from a
paper-based system requires updating or replacing some of the basic object
identifying documents to incorporate RFID technology. Additional updates to
supporting technologies, such as networking and data storage hardware, may
also be necessary. Figure 30 illustrates the technology updates that would be
required for implementation of an RFID accessioning and inventory system, like
the one evaluated in this research.

Figure 30: Points of integration for RFID in artifact processing

Discovery Accessioning Cataloguing Preparation Storage

‘Ar *
I Provenience I I Ledger Entry’I‘ Reference Bang m
I Material I

I Field Notes I Field Notes *
Completed Book iD
I Temporary I y,
Card

 

 

 

 

[15" = RFID integration PointI

 

 

 

 

 

 

 

 

The first change to the initial processing of artifacts is the elimination of the
hand-written ledger notebook, used during the task of accessioning. Rather than
writing a new object entry in a ledger notebook, information is entered into an R-
RMS software program and used to automatically update an electronic ledger.
The primary benefit of a digital ledger is the ability to efficiently organize entries.
For example, if two related objects are documented in a handwritten ledger at

different times, the ledger cannot be updated to display the entries next to one

78

another. By contrast, modern database systems have advanced data editing

tools for maintaining organized collections records.

The next change to an existing system would be the incorporation of an RFID
tag with bag ID labels. Using an approach similar to the “Tag New item” method
from this research, RFID tags would be programmed and then affixed to the
reverse side of the paper label. This practice, however, would require a change
in the processing of bagged artifacts. Generally, bag ID tags are created during
the ‘preparation’ stage. The R-RMS described in this thesis, however, would
require the creation of bag tags during the ‘accessioning’ stage, when the RF ID

tag is programmed.

Bin ID labels would require changes similar to RFID bag labels. Using an
approach similar to the “Tag New Item” method, RFID tags would be
programmed and then affixed to storage bins or ID labels using the pressure
sensitive adhesive tag backing. Unlike bag labels, the process of creating bin
labels would still be performed during the storage preparation stage, due to the

fact that bins are not utilized until the final stage of artifact ‘storage’.

The accession card system would be changed dramatically by an R-RMS,
and may even be eliminated. The purpose of an accession card is to document
all of the reference materials assembled during the ‘cataloguing’ stage of initial
processing. By using an electronic database, it is possible to add the accession
card information to the original ledger entry at a later time, thus eliminating the

need for me separate cards entirely. Alternatively, a paper based accession card

79

system may still be maintained in conjunction with an R-RMS as an emergency

back-up for electronic records.

RFID book tags would also be introduced for managing field notes and other
library reference material. The traditional method for labeling books, displaying a
title on the front cover and book spine, would remain unchanged. When a new
book is added to the library collection, an RFID tag would be programmed and

affixed to the book, according to a predefined tag placement method.
Ongoing Curation with RFID

The ultimate goal of an R-RMS is a perfect system for the ongoing curation of
archaeological artifacts. Using RFID to instantly recall electronic database
records and to perform periodic inventory audits is a significant step toward this
goal. As such, once an RFID system is established in a repository, complete with
RF-labels and electronic records for bags, bins and books, the next step is to
utilize the technology infrastructure for the tasks of recalling records and auditing

artifact collections, as shown in Figure 31.

Figure 31: Process map for ongoing curation in 11ng an R-RMS

 

 

 

 

IPackaging f--- R-RMS I Digital Database I

 

 

 

InventoryAudit ------ > Start -.
Records Recall —-) Finish -9
Step ->

 

 

 

 

80

To streamline to process of locating artifact-specific reference materials, the
R-RMS can be used to scan bag ID labelsand instantly access complete artifact
records from an electronic database. The intermediate step of consulting an
accession card is eliminated, reducing the potential for Type 1 failures in the

repository, and providing accelerated access to information for researchers.

Although bin and book tags would be primarily used for inventory auditing,
they could also be applied beneficially to recall database entries. Bin tags can be 5..
used to reduce unnecessary handling of antiquities by using the R-RMS to recall I!
identifying information for every item associated with a bin tag. In doing so, the I
complete contents of the storage container can be determined without ever
touching an artifact or even removing the lid. In a similar process, using the
database linkage between entries for artifacts and books, book tags could be
scanned to determine every artifact documented in a given book. Scanning book
tags using the same method evaluated in this thesis, could also power a library
check in/out feature. Using RFID personnel cards, responsibility could be
assigned for loaned book, providing accountability for all Type-4 failures in

maintaining provenience.

With pressure from government agencies, tribal heritage organizations and
self-imposed ethical standards to maintain accurate collections records, periodic
inventory auditing is a critical function of repository management. The R-RMS
eliminates the manual, line-by-line, ledger-to-object comparison method of

inventory auditing, and instead was an RF lD middleware tool for automatically

81

determining missing items from a collection based on a single RFID scan of the

storage environment.

Verifying the proper location of storage bins is an important task for
maintaining an organized repository. Bins maybe moved from their storage
location to enable further study of the contents, to gain access to bins lower in a
stack, or simply to make room for other research activities. When this happens,
there is no guarantee that a bin will be returned to the proper location. In a dig-
site repository, the bin inventory feature could be used to scan the storage
environment and identify any instances of bins missing from storage. Given the
low density distribution of bin transponders and the RF-transparence of plastic
storage bins, inventory scans of the dig-site repository could be completed in a
matter of minutes. The efficiency of the system would allow for bin inventory
verification as frequently as desired, even daily. According to statistical analysis
of bin inventory scanning techniques, a column scanning technique reliably
performs with 100% accuracy, with less variation in tag read rate. The column

method should therefore be employed for all bin inventory auditing.

Verifying the proper location of artifacts is a critical task for maintaining an
organized repository. Artifacts are constantly being accessed, studied, loaned
out, and generally moved throughout an archaeological repository environment.
Like storage bins, there is never a guarantee that an object moved will be
property retumed to storage. Considering the museum repository environment,

an item-level inventory auditing tool would be used to frequently verify the

82

 

accuracy of artifact inventory records. Unlike bins, however, item level tags may

require a more meticulous scanning technique than dispersed artifact tags.

Without a well organized library of reference materials for documenting the
provenience of every artifact in a repository collection, ongoing research efforts
can be hindered by the inability to efficiently locate relevant documentation, if at
all. Managing books with an R-RMS would allow for frequent book inventory
audits to identify missing books, a critical step towards avoiding Type-4 errors of

lost field notebooks.

In summary, the R-RMS system evaluated by this research is capable of
performing the repository management tasks of accessioning, recalling records
from an electronic database and performing inventory audits on RFID-tagged

artifacts, bags, bins and books.

83

Chapter 6: Future Considerations

 

The R-RMS system proposed and evaluated
by this research has been fitnctionally
demonstrated for applications in managing
archaeological collections. The
opportunities for RFID in archaeology,
however, are not confined to the four walls
of the repository. This section presents
potential opportunities for functionally
expanding the R-RMS to include new
hardware and artifact tracking capabilities.
Additionally, scalability issues are presented
and recommendations are made for further
developing and implementing a full-scale

RFID repository solution.

84

Expanding the R-RMS

Managing Bulk Artifacts

While smaller artifacts can be collected from a site and later stored in a
repository, bulk items are those which are too big for this research approach to
be practical. Instead, bulk items, such as architectural block fragments, are
studied in the field and then left in place. Since no bag or bin is used, all

identifying information must be attached directly to the object itself.

Throughout the history of excavations at the Corinth site, many unsuccessful
attempts have been made to develop a method for joining bulk items and
identifiers. There are two primary considerations for developing a bulk-item
identification method: aesthetic impact and lD-permanence. Any permanent
marking will negatively impact the aesthetics of an artifact, so the footprint of the
attachment mechanism must be minimized. Additionally, the objects are stored
outside, and are, therefore, subjected to weathering. If the identifier is not
designed to withstand erosion, and the identifier becomes unreadable, the
connection between the object and the documented historical record is severed.
The difficulty in balancing these to requirements for marking objects can be seen

Figure 32: Carved by looking at discontinued marking techniques from the

identifier on bulk artifact

.N"

last 50 years.

    

Carving the accession number into objects was
used to provide a very resilient identifier, but resulted
I in a major disruption to the objects appearance (Figure

32). Painting the accession number on the artifact minimized the aesthetic impact

85

 

and allowed the option for later removal if desired, but was subject to rapid
degradation from the elements. Yet another marking technique used a spike nail
Figure 33: Tin plate to attach tin plates, stamped with accession numbers,

identifier on bulk artifact
aw?“ .. to bulk—items (Figure 33). Aesthetic impact was initially

    
  

 

. limited to a single nail hole and the tin plate was
' H ; ‘ thought to be a reasonably permanent identifier. As

time progressed, however, the tin corroded, staining

 

the artifacts with rust and rendering the plates unreadable.

If an R-RMS system is to be applied for the management of bulk items, an
aesthetically acceptable method for attaching transponder to artifacts will need to
be developed and tested for performance and durability. In order to satisfy the
requirement of permanence, transponders should be attached to bulk items in
such a manner that the tag is protected from exposure to the elements, physical
impacts, and malicious destruction. Ideally, this would entail removing a small,
non-critical portion of the object and embedding the tag into the artifact itself. To
reduce the aesthetic impact of imbedding transponders, a drill might be used to
create a small diameter hole, just deep and wide enough to accommodate a tag.
The hole could then be filled with any number of epoxies or sealants, which could

be color matched to the object to further reduce visual disruption.

Specialty RFID Hardware
The functionality of the R-RMS system could also be expanded using
specialized RFID hardware such as RFID tag printers and handheld RFID reader

units. Using a compact RFID tag printer, designed for niche applications, item

86

labels could be created in a single step that would also eliminate the need to print

a separate human-readable tag.

The R-RMS system could also be enhanced by the integration of hand-held
reader units for applications in the field. In the aforementioned case of managing
bulk items, handheld readers are a necessity because the artifacts cannot be
brought to the RF ID system. In addition to enabling the management of bulk
artifacts, the handheld readers could also be used to document newly discovered
artifacts in the field, at the moment of discovery, thus reducing the risk of Type-3
failures resulting from the loss of temporary field ID tags associated to a group of

artifacts with shared provenience.
Scalability Considerations

As the number of artifacts being managed by an RMS system increases, so
does the amount of computing power needed operate the system. The evaluation
software developed for this research stored information using a set of Microsoft
(MS) Excel spreadsheets. While the Excel platform was more than capable of
managing the three small test spreadsheets, the program is primarily designed
for the purpose of displaying and analyzing arrays of data, rather than organizing
large collections of data. Efficiently managing collections larger than 100,000
objects would require a more powerful database format such as MS Access or
SQL database programs specifically designed to organize large collections of

data.

Another information management technique of note is that of cloud

networking, also called software as a service (SaaS). In a cloud network, rather
87

 

than hosting a database on a local data storage device, information is accessed
online, stored at a remote location and managed by a SaaS provider. An R-RMS
utilizing a cloud network could allow interoperability of RFID tagged artifacts from
one repository to another, and could provide the basis for improved information

sharing capabilities between researcher organizations.
Statistical Validity Testing

Statistically valid tests were performed to quantify the effect of different
scanning techniques on the performance of RFID bin inventory auditing. The
remaining evaluation scenarios for artifacts, bags and books, however, were
largely qualitative in nature. This is a product of the research being primarily
focused on the conceptual development of a specialized RFID system for
repository applications. To further understand the performance of UHF RF ID

systems in a repository, the following topics should be pursued:

o The effect of book thickness on the readability of RF ID library tags
. The effect of bin contents and position on the readability of bag tags
. Tag attachment methods for bulk artifacts

Recognizing the importance of validating a system in its entirety, the author
recommends that statistically valid tests be done, in the environment of intended
use, on all R-RMS features prior to implementing future versions of an RFID

repository management system.

88

Appendix I: Dig-Site Repository

environment

  

34:
Bookshelf Tagged Books

D

 

Counter top

-' ' " Aux.Antenna
Computer (Book Inventory)

Aux. Antenna

RFID Reader Unit (Bin Inventory)

Tagged Artifact Bins
(on metal storage racks)

Tagged Artifact Bags

(in storage bin) Additional Bins

89

 

 

 

\— 3858 382.02

 

 

 

       

£238 8&5. 88¢

 

2:25. .52.

' a2 .2500
23an0
E5 .381 Dim

Efimwfi EoE:E.S:o 95mm“ hormone; E3832 fin 05mm

  

 

83E .858

 

 

 

 

 

 

bozmomom 853.32 M: fiwcommsq.

90

Appendix III: Dig- Site Bag Database

 

Table 11.‘ Di

Item UID

Date

Trench

 

Tag lD

Description of Contents

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6.25.2009

 

 

2222 3412 DCO3 0119 2409 7637 4 6.25.2009 Shards Bin seven flat shards

2222 3412 0003 0119 2409 7580 7 6.25.2009 Shards Bin about 10 flat shards

2222 3412 DCO3 0119 2409 7620 10 6.26.2009 Dhards Bin eight flat shards and a
pottery handle

2222 3412 DCO3 0119 2409 7623 12 6.25.2009 Shards Bin three large flat shards

2222 3412 D003 0119 2409 7624 6 6.25.2009 Shards Bin Amphora top w/ broken
handles

2222 3412 DC03 0119 2409 7621 2 6.25.2009 Shards Bin 2 amphora bases

2222 3412 DCO3 0119 2409 7618 8 6.25.2009 Shards Bin Handfull of small shards.
One large curved piece

2222 3412 DCO3 0118 0826 5323 11 6.25.2009 Shards Bin seven pottery shards

2222 3412 DCO3 0119 2409 7627 3 6.25.2009 Shards Bin fifteen or so small shards

2222 3412 DCO3 0119 2409 7628 6.25.2009 four shards and an
amphora base

2222 3412 DCO3 0119 2409 7625 6.25.2009 Shards Bin three amphora bases

2222 3412 DCO3 0119 2409 7630 5 Shards Bin 8 small and 3 medium sized

shards

 

91

 

 

Appendix IV: Dig-Site Bin Database

 

Table 12: Di -Site Bin Database

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Item UID Date Tagged Area Content
1111 3411 8802 0116 1436 5599 1
11113411 B802 01161436 5600 2
1111 3411 8802 0116 1436 5601 3
111134118802 01161436 5602 4
1111 3411 8802 0116 1436 5645 5
1111 3411 8802 0116 1436 5644 6
1111 3411 8802 0116 1436 5640 7
1111 3411 8802 0116 1436 5639 8
111134118802 01161436 5637 9 ‘
111134118802 01161436 5636 10
1111 3411 B802 0116 1436 5635 11
1111 3411 8802 0116 1436 5634 12
11113411 88020116 1436 5633 13
111134118802 01161436 5629 14
1111 3411 8802 0116 1436 5628 15
111134118802 01161436 5627 16
1111 3411 8802 0116 1436 5626 17
111134118802 01161436 5625 18
111134118802 01161436 5621 19
1111 3411 8802 0116 1436 5622 20
111134118802 01161436 5624 21
111134118802 01161436 5609 22
1111 3411 8802 0116 1436 5619 23
111134118802 01161436 5608 24
111134118802 01161436 5620 25
1111 3411 8802 0116 1436 5613 26
11113411 88020116 1436 5611 27
111134118802 01161436 5612 28
111134118802 01161436 5618 29
1111 3411 8802 0116 1436 5616 30
1111 3411 8802 0116 1436 5614 31
111134118802 01161436 5605 32
111134118802 01161436 5603 33
111134118802 01161436 5646 34
111134118802 01161436 5604 35
1111 3411 8802 0116 1436 5606 36
111134118802 01161436 5607 37
111134118802 01161436 5662 38
111134118802 01161436 5661 39

 

 

 

 

 

92

Table 12: Dig-Site Bin Database (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

111134118802 01161436 5659 40
111134118802 01161436 5658 41
1111 3411 8802 01161436 5657 42
1111 3411 8802 01161436 5653 43
1111 3411 8802 0116 1436 5650 44
111134118802 01161436 5649 45
111134118802 01161436 5654 46
111134118802 01161436 5663 47
111134118802 01161436 5655 48
111134118802 01161436 5660 49
111134118802 01161436 5652 50
1111 3411 8802 01161436 5651 51
111134118802 01161436 5647 . 52
1111 3411 8802 01161436 5679 53
111134118802 01161436 5669

111134118802 01161436 5664 55
111134118802 01161436 5673

1111 3411 8802 01161436 5665 57
111134118802 01161436 5684

1111 3411 8802 01161436 5683 p 59
111134118802 01161436 5680 60
111134118802 01161436 5666 61
111134118802 01161436 5677 62
1111 3411 8802 01161436 5678 63
111134118802 01161436 5667 64
111134118802 01161436 5675 65
111134118802 01161436 5670 66
11113411 8802 01161436 5671 67
111134118802 01161436 5672 68
111134118802 01161436 5674 69
111134118802 01161436 5668 70
111134118802 01161436 5676 71
11113411 8802 01161436 5682 72
11113411 8802 01161436 5681 73
111110021003100410051011 74
1111 3411 8802 0116 1436 5731 75
1111 3411 8802 01161436 5688 76
111134118802 01161436 5687 77
11113411 8802 01161436 5686 78
11113411 8802 01161436 5685 79

 

93

Appendix V(a): Bin Inventory Tag
Data, Row Scanning Technique

 

Table 13: Row scannin- averaoe read rate data for each tao
Bin ID Average Median Q1 Q3 IQR Max Min
1 17.43 17.00 15.25 19.00 3.75 23.00 10.00

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2 17.70 17.00 16.00 19.00 3.00 24.00 12.00
3 19.77 20.00 17.25 21.00 3.75 27.00 14.00
4 10.50 11.00 9.25 12.00 2.75 13.00 5.00
5 24.00 24.00 23.00 26.00 3.00 28.00 14.00
6 8.97 9.00 8.00 10.00 2.00 13.00 5.00
7 18.40 19.00 16.25 21.00 4.75 23.00 13.00
8 9.77 9.00 8.00 11.75 3.75 16.00 4.00
9 14.97 15.00 13.00 16.75 3.75 23.00 8.00
10 11.03 11.00 9.25 12.75 3.50 16.00 7.00
11 18.57 18.50 16.25 20.00 3.75 23.00 13.00
12 23.77 23.50 22.25 26.00 3.75 31.00 18.00
13 11.40 11.00 10.00 12.00 2.00 16.00 5.00
14 9.30 9.00 8.00 10.00 2.00 14.00 7.00
15 18.10 18.00 16.00 20.00 4.00 22.00 15.00
16 8.40 8.00 7.00 9.00 2.00 12.00 6.00
17 10.17 10.00 8.00 12.00 4.00 15.00 7.00
18 31.57 29.50 25.00 36.75 11.75 45.00 21.00
19 14.00 14.00 12.00 15.00 3.00 20.00 10.00
20 26.73 27.00 24.00 28.00 4.00 34.00 18.00
21 12.87 12.00 11.00 14.75 3.75 19.00 7.00
22 5.33 5.00 5.00 6.00 1.00 9.00 3.00
23 20.23 20.00 18.25 21.75 3.50 33.00 15.00
24 15.97 16.00 15.00 17.00 2.00 22.00 12.00
25 12.70 13.00 11.00 14.00 3.00 19.00 9.00
26 10.30 10.00 9.00 11.00 2.00 16.00 6.00
27 20.13 19.50 17.25 23.00 5.75 25.00 15.00
28 11.80 11.50 11.00 13.00 2.00 15.00 9.00
29 18.03 18.00 16.25 20.00 3.75 23.00 12.00
30 8.50 8.00 7.00 9.75 2.75 14.00 5.00
31 19.13 19.00 17.25 20.00 2.75 26.00 14.00
32 8.23 8.00 7.00 9.00 2.00 13.00 5.00
33 13.13 13.00 11.00 14.75 3.75 19.00 8.00
34 17.23 17.00 16.00 19.75 3.75 22.00 12.00
35 15.67 16.00 14.00 16.75 2.75 22.00 12.00
36 17.40 18.00 15.00 19.75 4.75 22.00 12.00
37 9.27 9.00 7.25 10.00 2.75 20.00 6.00
38 14.77 15.00 13.25 16.00 2.75 20.00 10.00
39 20.30 20.00 19.00 22.00 3.00 26.00 13.00
40 13.37 13.00 12.00 14.75 2.75 22.00 9.00
41 8.03 8.00 7.00 9.00 2.00 11.00 4.00
42 14.07 14.00 13.00 15.00 2.00 20.00 9.00
43 11.47 11.50 9.00 14.00 5.00 20.00 5.00
44 20.90 21.00 19.25 22.00 2.75 28.00 15.00
45 15.23 15.00 13.25 17.00 3.75 21.00 8.00

 

94

 

 

Table 13: Row scanning average read rate data for each tag (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

46 17.80 17.50 16.00 20.00 4.00 24.00 10.00
47 9.80 9.50 9.00 10.75 1.75 13.00 7.00
48 11.27 11.00 10.00 12.00 2.00 17.00 7.00
49 7.27 7.00 7.00 8.00 1.00 10.00 4.00
50 7.00 7.00 6.00 8.00 2.00 9.00 3.00
51 9.87 10.00 8.25 11.00 2.75 14.00 5.00
52 11.63 11.50 10.00 12.75 2.75 16.00 9.00
53 13.00 12.00 12.00 14.00 2.00 20.00 8.00
54 25.23 26.00 24.25 27.00 2.75 32.00 15.00
55 15.87 15.50 14.00 17.00 3.00 22.00 11.00
56 12.33 12.00 11.00 14.00 3.00 17.00 9.00
57 6.53 7.00 5.25 7.00 1.75 10.00 4.00
58 19.80 19.50 17.00 22.00 5.00 27.00 11.00
59 19.90 20.00 19.00 21.00 2.00 26.00 12.00
60 20.00 20.00 18.25 21.75 3.50 25.00 16.00
61 17.37 18.00 16.00 19.75 3.75 24.00 11.00
62 16.80 17.00 15.25 18.75 3.50 22.00 10.00
63 6.60 6.00 5.25 8.00 2.75 9.00 4.00
64 13.63 13.00 11.00 16.75 5.75 23.00 8.00
65 20.23 20.00 18.00 22.00 4.00 26.00 14.00
66 19.27 19.00 17.00 22.00 5.00 26.00 15.00
67 19.37 20.00 17.25 21.00 3.75 26.00 13.00
68 14.07 14.00 12.00 16.00 4.00 19.00 10.00
69 23.07 23.00 21.00 26.00 5.00 27.00 15.00
70 11.63 11.00 10.00 13.00 3.00 18.00 7.00
71 16.80 17.00 15.00 18.75 3.75 22.00 12.00
72 11.23 11.00 10.00 12.00 2.00 18.00 8.00
73 20.87 21.00 18.25 22.75 4.50 26.00 16.00
74 14.13 14.00 13.00 16.00 3.00 18.00 10.00
75 27.23 27.00 26.00 29.00 3.00 37.00 16.00
76 8.40 8.50 7.25 9.00 1.75 13.00 3.00
77 9.13 9.00 8.00 10.00 2.00 15.00 7.00
78 9.73 10.00 9.00 10.00 1.00 13.00 5.00
79 7.27 7.00 6.25 8.00 1.75 12.00 5.00

 

 

95

Appendix V(b): Bin Inventory Tag
Data, Column Scannitg Technique

 

Table 14: Column scannin avera e read rate data for each tag
3 I t - . - -o . 0 O 0' .
13.87 14.00 11.00 16.00 5.00 25.00 9.00
10.90 11.00 9.25 12.00 2.75 17.00 5.00
15.33 15.50 14.25 18.00 3.75 21.00 10.00
9.07 9.00 8.00 10.00 2.00 15.00 7.00
13.73 13.00 11.25 15.00 3.75 23.00 10.00
6.30 6.00 5.00 7.00 2.00 15.00 3.00
10.37 10.00 9.00 12.00 3.00 16.00 6.00 «A
6.77 6.50 5.00 8.00 3.00 12.00 4.00
8.07 8.00 6.00 10.00 4.00 14.00 4.00
10 11.27 11.00 10.00 12.75 2.75 17.00 7.00
11 20.67 21.00 19.00 23.00 4.00 28.00 11.00 :'
12 14.47 14.00 11.00 17.75 6.75 27.00 8.00 l
13 11.43 11.00 10.00 13.00 3.00 18.00 8.00 '

 

 

 

 

 

 

 

 

OQNOUI-wa-l

 

 

 

 

 

 

 

 

14 7.40 7.00 6.00 9.00 3.00 10.00 4.00
15 16.13 16.00 13.25 19.00 5.75 25.00 9.00
16 6.57 6.50 5.25 7.75 2.50 10.00 4.00
17 5.57 5.00 4.00 6.00 2.00 18.00 3.00

 

18 25.73 26.00 22.00 30.75 8.75 39.00 14.00
19 10.00 9.00 8.25 11.00 2.75 16.00 7.00
20 19.93 19.00 17.00 22.75 5.75 30.00 14.00

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

21 5.73 5.00 5.00 7.00 2.00 9.00 4.00
22 5.47 5.00 4.00 6.00 2.00 14.00 3.00
23 9.70 9.50 8.00 12.00 4.00 16.00 4.00
24 15.87 16.50 14.00 18.00 4.00 22.00 9.00
25 6.90 6.00 5.00 8.00 3.00 14.00 4.00
26 5.63 5.00 4.00 7.00 3.00 11.00 2.00
27 16.77 16.50 14.00 20.00 6.00 30.00 9.00
28 9.47 9.00 8.00 10.75 2.75 15.00 7.00
29 11.83 12.00 9.25 14.00 4.75 20.00 6.00
30 6.23 6.00 5.00 7.00 2.00 10.00 3.00
31 12.47 12.00 10.00 15.75 5.75 20.00 5.00
32 7.53 7.00 6.00 8.00 2.00 21.00 4.00
33 9.00 9.00 7.00 11.00 4.00 14.00 5.00

 

34 11.83 11.00 10.00 13.75 3.75 19.00 8.00
35 11.33 11.00 9.25 13.00 3.75 18.00 7.00
36 11.20 10.00 8.00 14.00 6.00 19.00 6.00
37 4.67 4.00 3.00 6.00 3.00 10.00 2.00
38 11.83 12.00 10.00 14.00 4.00 20.00 7.00
39 13.80 14.00 11.00 16.00 5.00 19.00 6.00

 

 

 

 

 

 

 

 

 

40 8.83 8.00 7.00 10.00 3.00 15.00 5.00
41 5.40 5.00 4.00 6.75 2.75 9.00 2.00
42 12.57 12.00 11.00 13.75 2.75 18.00 7.00
43 5.20 5.00 3.25 6.00 2.75 10.00 3.00

 

44 18.33 17.00 15.25 22.00 6.75 30.00 8.00
45 10.50 10.50 9.00 12.00 3.00 17.00 4.00

 

 

 

 

 

 

 

 

 

 

 

96

Table 14: Column scanning average read rate data for each tag_{cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

46 8.70 8.00 7.00 10.75 3.75 18.00 5.00
47 7.37 7.00 6.00 8.75 2.75 15.00 4.00
48 10.17 10.00 9.00 11.00 2.00 15.00 7.00
49 6.23 6.00 5.25 7.00 1.75 9.00 4.00
50 6.33 6.00 5.00 7.75 2.75 14.00 4.00
51 6.60 6.00 5.00 8.00 3.00 14.00 3.00
52 10.80 10.50 10.00 12.00 2.00 14.00 7.00
53 4.77 4.00 3.25 5.75 2.50 9.00 3.00
i 54 18.57 19.50 15.25 22.75 7.50 25.00 9.00
55 8.70 8.00 7.00 10.75 3.75 16.00 5.00
56 15.97 16.00 15.00 18.00 3.00 23.00 7.00
57 4.87 5.00 4.00 6.00 2.00 8.00 3.00
58 18.57 18.50 16.00 21 .75 5.75 31.00 4.00
59 11.27 11.00 9.25 12.75 3.50 19.00 4.00
60 9.70 10.00 7.25 11.00 3.75 16.00 2.00
61 15.00 15.50 12.00 17.00 5.00 25.00 10.00
62 12.07 11.50 9.00 14.00 5.00 21 .00 5.00
63 5.53 6.00 5.00 6.00 1.00 9.00 2.00
64 9.13 10.00 7.00 11.00 4.00 13.00 4.00
65 11.93 12.00 9.00 15.00 6.00 21.00 6.00
66 14.03 13.50 11.25 16.50 5.25 26.00 6.00
67 11.37 11.00 8.00 13.75 5.75 19.00 6.00
68 16.87 17.00 16.00 18.00 2.00 24.00 4.00
69 14.83 14.00 12.00 17.75 5.75 22.00 5.00
70 6.23 6.50 5.00 8.00 3.00 9.00 3.00
71 11.47 12.00 9.00 13.00 4.00 18.00 4.00
72 7.03 6.50 6.00 8.00 2.00 11.00 3.00
73 12.00 11.00 11.00 14.75 3.75 22.00 4.00
74 16.20 16.00 15.00 18.00 3.00 24.00 9.00
75 15.63 15.50 13.00 19.00 6.00 22.00 6.00
76 7.47 7.00 6.00 9.00 3.00 13.00 4.00
77 4.63 5.00 3.25 5.00 1.75 9.00 2.00
78 7.67 8.00 6.00 8.75 2.75 11.00 3.00
79 5.20 5.00 4.00 6.00 2.00 8.00 3.00

 

 

97

1"—

Appendix V(c): Bin Inventory Tag
Data, Random Scanning Technique

 

Table 15: Random scannino avera- 6 read rate data for each tag
Bin ID Average Median Q1 QB IQR Max Min
1 13.77 14.00 11.00 16.00 5.00 21.00 5.00

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

98

2 14.00 14.00 11.25 16.00 4.75 20.00 6.00

3 16.60 17.00 14.00 20.00 6.00 22.00 10.00

4 6.97 7.00 6.00 8.00 2.00 12.00 1.00

5 14.77 14.00 13.00 17.75 4.75 25.00 6.00

6 6.03 6.00 4.25 8.00 3.75 11.00 1.00 “a;
7 13.03 13.00 11.00 15.00 4.00 22.00 8.00 ‘ i
8 5.73 5.00 4.00 7.00 3.00 12.00 2.00 E
9 10.17 11.00 7.00 12.00 5.00 20.00 6.00

10 10.43 10.50 8.25 13.00 4.75 17.00 2.00 .
11 16.30 15.50 13.00 20.00 7.00 25.00 10.00 'L 1
12 16.73 16.00 11.25 21.75 10.50 34.00 5.00

13 11.30 12.00 9.25 14.00 4.75 17.00 1.00

14 5.97 6.00 5.00 7.75 2.75 11.00 1.00

15 15.53 17.00 12.00 19.00 7.00 21.00 6.00

16 5.80 6.00 4.00 7.00 3.00 11.00 2.00

17 6.73 6.50 6.00 8.00 2.00 11.00 1.00

18 20.27 17.50 16.00 23.00 7.00 37.00 10.00

19 10.07 10.00 8.00 13.00 5.00 19.00 3.00

20 23.67 22.50 18.00 27.75 9.75 46.00 11.00

21 8.33 8.00 6.00 10.00 4.00 17.00 3.00

22 3.37 3.00 2.00 4.75 2.75 7.00 1.00

23 14.23 14.00 12.00 16.00 4.00 22.00 8.00

24 14.20 14.50 12.00 17.00 5.00 20.00 6.00

25 8.97 9.00 7.25 10.75 3.50 20.00 3.00

26 6.23 6.00 3.25 8.00 4.75 16.00 2.00

27 14.27 15.00 11.00 17.00 6.00 23.00 7.00

28 8.90 8.50 7.25 11.00 3.75 16.00 3.00

29 14.23 13.50 10.00 17.00 7.00 26.00 5.00

30 6.60 7.00 4.25 8.00 3.75 12.00 2.00

31 14.53 14.00 12.00 17.50 5.50 27.00 7.00

32 6.00 6.00 5.00 7.00 2.00 11.00 1.00

33 8.93 8.00 7.00 10.00 3.00 17.00 5.00

34 13.37 13.50 12.00 14.75 2.75 20.00 8.00

35 12.30 12.00 10.00 14.00 4.00 22.00 7.00

36 12.77 13.00 10.25 16.00 5.75 22.00 7.00

37 6.40 7.00 5.00 8.00 3.00 11.00 1.00

38 9.30 10.00 8.00 10.75 2.75 14.00 4.00

39 13.80 14.00 11.25 15.00 3.75 23.00 8.00

40 8.50 8.00 6.00 10.75 4.75 20.00 2.00

41 6.43 7.00 5.00 8.00 3.00 11.00 2.00

42 12.60 12.50 8.50 16.00 7.50 20.00 3.00

43 7.80 8.00 6.25 9.75 3.50 17.00 2.00

44 16.80 18.00 14.25 19.75 5.50 24.00 8.00

45 12.60 12.50 11.00 15.00 4.00 19.00 5.00

 

{oont’d}

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 15: Random scanning average read rate data for each tag
46 9.10 9.00 6.25 10.75 4.50 18.00 5.00
47 6.30 6.00 5.00 8.00 3.00 11.00 1.00
46 11.03 11 .00 8.25 14.00 5.75 16.00 4.00
49 5.33 5.00 4.00 6.00 2.00 10.00 1.00
so 4.60 4.00 4.00 6.00 2.00 7.00 2.00
51 7.43 8.00 5.00 9.00 4.00 15.00 2.00
52 10.87 11.00 9.00 13.00 4.00 18.00 5.00
53 6.70 6.00 4.25 8.75 4.50 16.00 3.00
54 16.50 19.00 15.25 21 .00 5.75 29.00 12.00
55 10.10 10.00 6.00 12.00 4.00 26.00 3.00
56 10.93 11.00 10.00 12.75 2.75 17.00 5.00
57 4.27 4.00 3.00 5.00 2.00 8.00 2.00
58 14.73 13.50 12.00 17.75 5.75 26.00 6.00
59 10.97 10.00 9.00 12.00 3.00 17.00 7.00
60 12.60 12.00 11.00 14.75 3.75 26.00 8.00
61 15.33 15.50 13.00 17.75 4.75 23.00 7.00
62 10.10 10.00 8.00 11.00 3.00 21.00 6.00
63 5.23 5.00 4.00 7.00 3.00 9.00 3.00
64 8.20 8.00 6.00 9.00 3.00 16.00 5.00
65 11.10 11.00 8.25 13.00 4.75 22.00 4.00
66 13.43 13.00 11.25 16.00 4.75 21.00 7.00
67 14.20 15.00 11.00 17.00 6.00 23.00 7.00
68 12.40 13.00 9.50 14.00 4.50 22.00 6.00
69 14.30 13.50 12.00 17.00 5.00 23.00 8.00
70 6.33 5.00 4.25 7.00 2.75 15.00 3.00
71 11.73 12.00 10.00 13.75 3.75 16.00 6.00
72 8.07 9.00 6.25 10.00 3.75 14.00 3.00
73 16.53 16.50 14.00 19.00 5.00 26.00 10.00
74 11.97 11 .00 9.00 14.00 5.00 24.00 5.00
75 15.43 15.00 12.00 17.00 5.00 27.00 7.00
76 6.60 6.00 5.00 8.00 3.00 12.00 3.00
77 7.07 6.50 5.25 8.75 3.50 15.00 1.00
76 7.37 7.00 6.00 8.00 2.00 13.00 4.00
79 4.63 5.00 3.00 6.00 3.00 10.00 1.00

 

 

 

 

 

 

 

 

 

 

99

Appendix VI: Bin Inventory Trial Data

 

Table 16: Bin invento avera-e read rate data for each trial
Average Read Rate per Trial

 
 
 

 

  

Thai 7: Column Scan Rev-x Scan Random Scan Overall

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 11.04 13.86 11.16 12.02

2 8.68 16.04 10.19 11.64

3 9.42 15.67 12.10 12.40

4 9.33 16.56 13.34 13.08

5 9.48 16.19 10.78 12.15

6 9.03 15.91 12.44 12.46

7 9.19 14.80 11.66 11.88

8 9.41 13.61 13.65 12.22

9 9.42 13.89 11.71 11.67

10 8.95 13.67 9.89 10.84

11 9.14 14.85 11.94 11.97

12 9.75 14.18 12.27 12.06

13 9.81 13.41 12.46 11.89

14 11.95 15.65 12.39 13.33

15 10.73 16.08 12.89 13.23

16 13.91 15.86 16.10 15.29

17 11.54 15.66 11.94 13.05

18 10.48 14.30 11.47 12.08

19 12.04 15.37 8.96 12.12

20 9.16 14.63 8.91 10.90

21 11.29 15.65 10.10 12.35

22 10.68 14.71 11.16 12.19

23 13.30 14.43 7.97 11.90

24 11.73 15.80 9.08 12.20

25 11.16 15.66 7.00 11.27

26 12.52 14.43 8.43 11.79

27 12.20 13.95 8.80 11.65

28 12.27 13.37 7.89 11.17

29 11.94 12.99 8.70 11.21

30 11.35 14.20 7.05 10.87
Average Read Count 10.70 14.84 10.75 12.10
Maximum Read Count 13.91 16.56 16.10 15.29
Minimum Read Count 8.68 12.99 7.00 10.84
Standard Deviation 1.44 1.00 2.15 0.89

 

 

100

Appendix VII(a): Bin Tag Read Rates
by Tag Position: Row Scan Technique

Figure 36: Average read rate per bin tag: row scanning technique

 

 

Avg. Read Count

0
1-5
6—10
11-15
16+

V
UHIIg
8

 

 

 

Ceiling

 

 

 

 

 

 

 

 

101

IIBM

 

 

Appendix VII(b): Bin Tag Read Rates by
Tag Position: Column Scan Technique

Figure 37: Average read rate per bin tag: column scanning technique

 

 

L Ceiling

 

Avg. Read Count
0 p - ,
1-5
6-10

11-15
16+

 

 

 

Ufllli

 

“9M

 

 

 

 

102

 

Appendix VII(c): Bin Tag Read Rates by
Tag Position: Random Scan Technique

Figure 38: Average read rate per bin tag: random scanning technique

 

 

Ceiling

u-O-

 

Avg. Read Coun
0 W
1-5 -
6-10 -
11-15 [2
:1

16+

 

 

 

 

 

 

103

Appendix VIII: Dig-Site Book Database

 

Table 17: Di .

Item UID

-Site Book Database

Book Title

Dig Site

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3333 3412 DCOB 0119 2409 7626 July 98 Volume 67: Number 3 Hesperia
3333 1002 1003 1004 1005 1006 Apr 98 Volume 67: Number 2 Hesperia
3333 3412 DCO3 0117 0611 8793 Jan 98 Volume 67: Number 1 Hesperia
3333 3412 DCO3 0117 0611 8784 Oct 97 Volume 66: Number 4 Hesperia
3333 1002 1003 1004 1005 1008 July 97 Volume 66: Number 3 Hesperia
3333 3412 DC03 0119 2409 7616 Apr 97 Volume 66: Number 2 Hesperia
3333 3412 DCO3 0119 2409 7579 Jan 97 Volume 66: Number 1 Hesperia
3333 3412 DC03 0119 2409 7617 Oct 96 Volume 65: Number 4 Hesperia
3333 3412 DC03 0119 2409 7634 July 96 Volume 65: Number 3 Hesperia
3333 3412 DC03 0117 0611 7039 April 96 Volume 65: Number 2 Hesperia
3333 3412 DC03 0117 0611 7040 Jan 96 Volume 65: Number 1 Hesperia
3333 3412 DC03 0117 0611 7036 Jan 96 Volume 65: Number 1 Hesperia
3333 3412 DC03 0117 0611 8805 July 95 Volume 64: Number 3 Hesperia
3333 3412 DC03 0117 0611 8803 Oct 95 Volume 64: Number 4 Hesperia
3333 3412 DC03 0117 0611 8801 July 95 Volume 64: Number 3 Hesperia
3333 3412 DCO3 0117 0611 7037 April 95 Volume 64: Number 2 Hesperia
3333 3412 DC03 0117 0611 8798 Jan 95 Volume 64: Number 1 Hesperia
3333 3412 0003 0117 0611 8795 Oct 94 Volume 63: Number 4 Hesperia
3333 3412 DCOS 0117 0611 8781 July 94 Volume 63: Number 3 Hesperia
3333 1002 1003 1004 1005 1007 April 94 Volume 63: Number 2 Hesperia
3333 3412 DC03 0117 0611 8792 Jan 94 Volume 63: Number 1 Hesperia
3333 1002 1003 1004 1005 1009 Jan 93 Volume 63: Number 1 Hesperia
3333 3412 DCO3 0117 0611 8807 April 93 Volume 63: Number 2 Hesperia
3333 3412 DC03 0117 0611 8806 July 93 Volume 62: Number 3 Hesperia
3333 3412 DC03 0117 0611 7043 Oct 93 Volume 62: Number 4 Hesperia
3333 3412 DC03 0117 0611 7026 Jan 92 Volume 61: Number 1 Hesperia
3333 3411 8802 0116 1436 5721 July 92 Volume 61: Number 3 Hesperia
3333 1002 1003 1004 1005 100A Oct 92 Volume 61: Number 4 Hesperia
3333 3412 DC03 0117 0611 8778 Jan 92 Volume 61: Number 1 Hesperia
3333 3412 DCO3 0117 0611 8779 Jan 92 Volume 61: Number 1 Hesperia
3333 3412 DCO3 0117 0611 8780 April 92 Volume 61: Number 2 Hesperia
3333 3412 DC03 0117 0611 7035 Jan 92 Volume 61: Number 1 Hesperia
3333 3412 DC03 0117 0611 7033 1952 The Metal Objects lsthmia

3333 3412 DC03 0117 0611 8796 1967 Sculpture ll: Marble Sculptures lsthmia

3333 3412 DC03 0117 0611 8794 ? The Hexamillion and The Fortress lsthmia

3333 3412 DC03 0117 0611 7034 1952 Sculptures l lsthmia

3333 3412 DC03 0117 0611 8799 ? Topography and Architecture lsthmia

3333 3412 DCOB 0117 0611 7038 ? Temple of Posseidon lsthmia

3333 3412 DC03 0119 2409 7638 ? The Late Bronze Age Settlement lsthmia

 

 

 

104

 

References

 

AIM Global. (2010). Association for Automatic ldentificaton and Mobilty.
Retrieved January 4, 2010, from AIM Technologies:
http:l/www.aimglobal.org/technologiesl

Alien Technologies. (2007). Product Overview: ALR-9650 Gen 2 RFID
Reader with Integrated Antenna. Morgan Hill, CA.

Alien Technologies (II). (2008, July). Higgs 3 Product Overview. Retrieved
March 2, 2010, from Alien RFID:
http://www.alientechnology.oom/docs/products/DS_H3.pdf

ASTM International. (2009). D-7435: Standard Test Method for Determining
the Performance of of Passive Radio Frequency Identification (RFID)
Transponders on Palletized or Unitized Loads. In Annual Book of ASTM
Standards Volume: 15.10 (p. 762). West Conshohocken: ASTM.

Berhad. (2009). Case Study: National Museum Malaysia. Retrieved
September 5, 2009, from Solmate RFID Asset Management:
http://www.cbs.com.my/english/solutions/solmate_ams.pdf

Boss, R. (2004, May 14). RFID Technology for Libraries. Retrieved December
2, 2009, from American Library Association:
http: l/www. ala. org/PrinterTemplate. cfm?Section=technotes&Template=/C
ontentManagementlHTMLDispIay. cfm&ContentID=68138

Clarke, R. (2008). Chapter 10: The Influence of Product and Packaging
Characteristics on Passve RFID Readability. In J. Keny, Smart Packaging
Technologies for Use in Fast Moving Consumer Goods (pp. pp. 167-
195). Wiley.

Curran, K., & Porter, M. (2007). A Primer on Radio Frequency Identification
for Libraries. Library Hi Tech , 25, 595-611.

Department of Defense. (2003). D00 Supply Chain Material Management
Regulation. Washington DC: Office of the Deputy Under Secretary of
Defense Logistics and Material Readiness.

EPCglobaI. (2006, September). RFID Implementation Cookbook. Retrieved
December 20, 2009, from RF ID and EPC Essentials:
http://www.epoglobalinc.org/what/cookbook/chapter1l002-
RFID_EPC_Essentials_v1 .pdf

Fenn, J. (2009). Inside the Hype Cycle: What’s Hot and What’s Not in 2009.
Retrieved from Gartner:

105

http://my.gartner.comfrtlcontentl1 101 800/ 1 1 01817/august12_hype_cycle_
final_jfenn .pdf

Fladmark, K. (1978). A Guide to Basic Archaeological Field Procedures.
Burnaby: Simon Frasier University.

Floerkemeier, C., & Lampe, M. (2009, June 10). RFID middleware design -
addressing application requirements and RFID. ETH Zurich, vaitzerland.

Frey, J. (2009, June 23). Personal Conversation. .
Glover, B., & Bhatt, H. (2006). RFID Essentials. Sebastopol: 0' Reilly.

G81. (2009, March 18). Regulatory status for using RFID in the UHF
spectrum. Retrieved from EPCgIobal Inc:
http://www.epoglobalinc.org/techlfreq_reg/RFID_at_UHF_ReguIations_20
090318.pdf

Hoag, J., & Thompson, C. (2006, October). Architecting RFID Middleware.
Retrieved September 2009, from IEEE lntemet Computing:
http://vw.ddns.uark.edu/contentl2006-09—PAPER—lEEE-Intemet-
Computing—Architecting-RFID-Middleware.pdf

Infotek Software & Systems. (2009). i-TEK Brochure. Retrieved December
20, 2009, from Concept Papers:
http:l/www.infoteksoftware.com/concept_papersfr-
TEK_brochure_v3_0.pdf

Khan, F. (2004, September 7). Museum Puts Tags on Stufi'ed Birds.
Retrieved from RF ID Journal:
http://www.rfidjoumal.com/article/view/t 1 10/1 I1

Kleefeld, E. (2005, October 12). Efforts to avoid “friendly fire’ spawned
ancestor of today’s RFID. Retrieved from Wisconsin Technology Network:
httpzllwistechnology.com/articles/2348/

Landt, J. (2001). Shrouds of Time: The History of RFID. Pittsburgh, PA:
Association for Automatic Identification and Mobility (AIM).

Liard, M., & Carlaw, S. (2009). RFID Annual Market Overview. New York:
ABIresearch.

National Park Service. (1983, September 29). Archaeology and Historic
Preservation. Retrieved February 22, 2010, from Secretary of the
Interior‘s Standards and Guidelines: httpzllwww.nps.gov/historyllocal-
law/arch_stnds_0.htm

106

Ramakrishnan, K., & Deavours, D. (2007). Performance Benchmarks for
Passive UHF. Lawrence, KS: lnfonnation and Telecommunications
Technology Center.

Reibel, D. (1997). Registration Methods for the Small Museum. Walnut Creek,
CA: AItaMira Press. ‘

RFID Journal. (2005). The History of RFID Technology. Retrieved August 31,
2009, from RF ID Journal:
http://www.rfidjou mal.com/articIe/articleview/ 1 338/1/1291

RFID Journal (II). (2010). Glossary of RFID Terms. Retrieved from RFID
Journal: http://www.rfidjoumal.com/articIe/glossary/3

Roduner, 0., Lampe, M., 8 Floerkemeier, C. (2007). RFID Application
Development. IEEE Systems Journal, 1-13.

Sirico, L. (2009). The Evolution of RFID Middleware to Intelligent Sensor
Network. Retrieved January 10, 2010, from lndustryWizards.com:
httpzllrfidwizards.com/Index.php?option=com_content&task=view&id=467
&Itemid=395

Sullivan, L., 8. Childs, T. (2003). Curating Archaeological Collections.
Lanham, MD: AItaMira Press.

Supply Chain Digest. (2009, February 23). RFID News: Looking Back at the
Wal-Mart RFID Time Line. Retrieved December 14, 2009, from
SupplyChainDigest.com: http://www.scdigest.com/assets/On_Target/09-
02-23-1.php

Sweeny, P. J. (2005). RFID for Dummies. Hoboken, NJ: Wiley Publishing,
Inc.

Sweeny (II), P. J. (2007). CompTlA RFID+ Study Guide. Indianapolis: Wiley.

Texas Instruments. (2010, July 7). RFID Technology Streamlines
Management of Vatican Library’s Treasured Collections. PR Newswire ,

p. 1.

Weier, M. H. (2009, November 16). Slow And Steady Progress. lnfonnation
Week, 31.

Yu, S. C. (2008). Implementation of an innovative RFID application in
libraries. Library Hi Tech, 26(3), 398-410.

107

        

     

Y LIBR

T

4855

ARIES

 

 

 

 

 

 

   

mlmlillllsllllle lililillilllil
3 1293 03063