.
ﬁrs; ,u 3:

.11: .145.

... am.”

a
kl...

.. a
“any“.

v. ‘? -
L .
ﬂu): I r, h

‘ . ... Q:

My:
xi...”

”NM“. .
awn 4m...

«u,
‘ an“
. . 3:5,,
,2 . $352.
.

 

 

...§§%§$ﬁ .. . ‘ n.1,? .. a?

:35 LIBRARY ‘
L004 Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

MEASUREMENT AND ANALYSIS OF HUMAN POSTERIOR
BACK AND THIGH CONTOURS FOR AUTOMOTIVE
SEAT DESIGN

presented by

Zhenyu Liu

has been accepted towards fulﬁllment
of the requirements for the

Doctoral degree in EngLneering Mechanics

 

 

 

Major Professor’s Signhture

IX riI 1005

Date

MSU is an Aﬂinnative Action/Equal Opportunity Institution

 

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

 

I ._..

 

 

11% a “gag

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 c:EIFlC/Date0ue.indd-p. I 5

 

MEASUREMENT AND ANALYSIS OF HUMAN POSTERIOR BACK AND THIGH
CONTOURS FOR AUTOMOTIVE SEAT DESIGN

By

Zhenyu Liu

A DISSERTATION

Submitted to
Michigan State University
In partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Mechanical Engineering

2005

ABSTRACT

MEASUREMENT AND ANALYSIS OF HUMAN POSTERIOR BACK AND THIGH
CONTOURS FOR AUTOMOTIVE SEAT DESIGN

By
Zhenyu Liu

To better understand and represent the seated human posterior back,
buttocks, and thigh contours, three groups of human subjects (5th percentile
small females, 50th percentile mid-sized males, and 95th percentile large males)
(Entire measured in a specially designed contour chair within typical automotive
interior environment. The postures were acquired by a human motion
measurement system. I

The data were analyzed, and then representative seated mid-sized male
posterior back, buttocks, and thigh contours were developed for both physical
hardware and computer program software for automotive seating design and
validation.

To explore the relations among contours of people with different
anthropometric dimensions and the possibility to relate the contours of different
anthropometric groups of people, a method was developed to scale the mid-
sized male contours to large male or small female subjects’ contours. Potentially
it could be used to generate the contours for people in different groups other
than small female and large male. For the ﬁrst time, the scaling methodology of
scaling contours of mid-sized male to small female or large male was veriﬁed by
experimental data. The coefﬁcient of correlation analysis produced very high

agreements between the scaled and measured data of the contours.

Both the hardware and software from this research are being utilized in
the automotive seating industries. The data collected in this research are having

signiﬁcant inﬂuences on the seating design and validation.

Copyright by
Zhenyu Liu
2005

Dedication

I would like to dedicate this dissertation to my parents and my sisters, my
parents-in-law, for their guidance and encouragement in everything and every

way, even in my pursuing beautiful dreams.

I would like to dedicate this dissertation to my family, my wife, Lan, our

wonderful children Alice Yiyang, Christine Yihuan, and Daniel Yikang, for their

inspiration, support, and sacriﬁces.

ACKNOWLEDGEMENTS

The author would like to express his sincerest gratitude to the following
people for their assistance, and encouragement in the course of preparing this
dissertation.

To my advisor and mentor, Dr. Robert Hubbard, for your guiding and
assisting me in every way possible. Your advice and encouragement will be with
me well beyond the scope of this dissertation. Thank you.

To my graduate committee: Dr. Gary Cloud, Dr. Hongyu Tsai, and Dr.
Joseph Vorro, thank you for reviewing my work.

Special thanks to Cliff Beckett, for helping me with every detail in the
contour chair building, and keeping experimental equipments in excellent
condition.

To Chris and Melissa Gedraitis, Bob Boughner, Richard Setyabudhy, and
Akram Ali, for your effort and time put into this work.

To Automotive Seat and Package Evaluation and Comparison Tools
program (ASPECT) and Human Solution, Inc., for their generous funding of my

research.

vi

TABLE OF CONTENTS

LIST OF FIGURES ............................................................................................... x
LIST OF TABLES ................................................................................................ xv
CHAPTERI INTRODUCTION ............................................................................. 1
1.1 Background and Literature Review ...................................................... 1
1.2 Statement of The Problem ................................................................. 13
1.3 Objectives .......................................................................................... 13

CHAPTER 2 ANATOMICAL BASICS AND TARGETED BONY LANDMARKS..15

2.1 Introduction ........................................................................................ 15

2.2 Posture Descriptions .......................................................................... 15

2.3 Positional Terminologies .................................................................... 19

2.4 Upper Limb ......................................................................................... 19

2.5 Spine .................................................................................................. 20

2.6 Thorax ................................................................................................ 23

2.7 Pelvis .................................................................................................. 24

2.8 The Lower Limbs ................................................................................ 25
CHAPTER 3 MEASUREMENT OF SEATED HUMAN CONTOURS AND

POSTURES ................................................................................... 27

3.1 Introduction ........................................................................................ 27

3.2 Experimental Device and Setup ......................................................... 27

3.2.1 Contour Test Chair ................................................................. 27

3.2.1.1 Seat Pan ..................................................................... 29

3.2.1.2 Seat Back .................................................................... 31

3.2.1.4 Measurement Contour Chair Accuracy ....................... 32

3.2.1.5 Seat Foam Stiffness Selection .................................... 32

3.2.2 Test Lab Space ...................................................................... 35

3.2.3 Measurement Targets and Probe ........................................... 35

vii

3.2.5 Reference Chair ...................................................................... 38

3.2.6 Contour Data Acquisition ........................................................ 40
3.2.7 Posture Data Collection .......................................................... 41

3.3 Testing Protocol ................................................................................. 43
3.4 Subject Selection Criteria ................................................................... 45
3.5 Subject Anthropometric Data Acquisition ........................................... 46
CHAPTER 4 DATA ANALYSIS APPROACHES AND RESULTS ...................... 49
4.1 Introduction ........................................................................................ 49
4.2 Initial Raw Data Treatment ................................................................. 49
4.2.1 Data Transformation Among Coordinate Systems ................. 50
4.2.1.1 Translation ................................................... . ............... 53

4.2.1.2 Rotation ....................................................................... 54

4.2.2 Generating Contour Surface from Measurement Data Points 55

4.3 Seated Posture Calculations ........................................ , ...................... 56
4.3.1 Lower Neck Joint .................................................................... 57
4.3.2 Shoulder Joints ....................................................................... 58
4.3.3 Upper Lumbar Joint ................................................................ 59
4.3.4 Lower Lumbar and Hip Joints ................................................. 60
4.3.5 Knee Joints ............................................................................. 62
4.3.6 Ankle Joints ............................................................................ 63
4.3.7 Wrist Joints ............................................................................. 63
4.3.8 Elbow Joints ........................................................................... 64

4.4 Deformed Contours Data Analysis ..................................................... 65
4.4.1 Methods .................................................................................. 65
4.4.1.1 Analyzing the Thigh and Buttocks Data Surface ......... 66

4.4.1.2 Analyzing the Back Data Surfaces .............................. 68

4.4.2 Comparing Measured Data and J826 H-point Machine. ......... 69
4.4.2.1 The Thigh and Buttocks Data ...................................... 69

4.4.2.2 The Back Contour Data ............................................... 73

4.5 Deﬁning Mid-male Contour ................................................................. 81
4.5.1 Deﬁning The ASPECT Mid-male Contours ............................. 81
4.5.1.1 Thigh and Buttocks Contours ...................................... 82

4.5.1.2 Back Contour .............................................................. 83

4.5.2 Deﬁning The RAMSIS Mid-male Contours ............................. 93

CHAPTER 5 CONTOUR SCALING FOR LARGE MALE AND SMALL FEMALE
...................................................................................................... 95

5.1 Introduction ........................................................................................ 95

5.2 Development of Scaling Factors ........................................................ 96

5.2.1 Thoracic .................................................................................. 97

5.2.2 Lumbar ................................................................................... 97

5.2.3 Pelvis ...................................................................................... 98

5.2.4 Thigh ....................................................................................... 99

5.3 Scaling Procedures and Results ........................................................ 99

5.3.1 Small Female ........................................................................ 100

5.3.2 Large Male ............................................................................ 113

5.4 Correlation Comparison Between Measured and Scaled Contours .126

5.4.1 Methods ................................................................................ 126

5.4.2 Correlation of Large Male and Small Female Contours ........ 127

5.4.2.1 Differences ................................................................ 127

5.4.2.2 Coefﬁcients of Correlation ......................................... 141

5.4.2.3 Corridors ................................................................... 145

5.5 Discussion of Scaling Results .......................................................... 154

5.5.1 Thorax .......................... 154

5.5.2 Lumbar ......................................................................... 155

5.5.3 Pelvis ........................................................................... 155

5.5.4 Thigh ............................................................................ 156

CHAPTER 6 CONCLUSIONS ........................................................................... 158

APPENDICES ................................................................................................... 162
APPENDIX A

Scaling Factors Used for Scaling Procedures ........................................ 163
APPENDIX B

Scaled and Measured Small Females Results ....................................... 166
APPENDIX C

Scaled and Measured Large Males Results ........................................... 199

LIST OF REFERENCES ................................................................................... 223

ix

LIST OF FIGURES

Images in this dissertation are presented in color.

Figure 1 J826 Two-Dimensional H-point Template and Three-Dimensional

H-point Machine ..................................................................................... 3
Figure 2 Two—Dimensional JOHN Model ............................................................... 6
Figure 3 Deﬁnition of TRA ................................................................................... 11
Figure 4 TLC ....................................................................................................... 12
Figure 5 A Typical Posture Model ....................................................................... 17
Figure 6 Subject in Contour Measurement Chair ................................................ 18
Figure 7 Targets at Lateral Epicondyle, Mid-wrist, and Acromion Probed .......... 20
Figure 8 The Spine Column, Adapted from Chafﬁn [13] ..................................... 22
Figure 9 Targeted at Manubrium and Xiphoid Process ....................................... 23
Figure 10 Pelvis: illium, ischium, pubic symphysis .............................................. 25
Figure 11 Targets on Lower Limbs ..................................................................... 26
Figure 12 Contour Chair with Targets for Local Coordinate System ................... 28
Figure 13 Light Box ............................................................................................. 30
Figure 14 Same Subject Measured on Stiff and Soft Foams .............................. 33
Figure 15 Measurement Target ........................................................................... 36
Figure 16 Measurement Probe ........................................................................... 37
Figure 17 Reference Chair .................................................................................. 39
Figure 18 Transformed Data Points .................................................................... 51
Figure 19 Lofted Data Surface ............................................................................ 52
Figure 20 From Coordinate System Ox1x2x3 Translated to Ox1’x2’x3’ .................. 53

Figure 21 Cartesian Coordinate System Ox1x2x3 Rotated to Ox1’x2'x3’ ............... 55

Figure 22 Calculation of Lower Neck Joint .......................................................... 57
Figure 23 Calculation of Shoulder Joint .............................................................. 58
Figure 24 Calculation of Upper Lumbar Joint ...................................................... 59
Figure 25 Pelvis Coordinate System ................................................................... 61
Figure 26 Calculation of Lower Limbs Joints ...................................................... 63
Figure 27 Calculation of Upper Limbs Joints ...................................................... 64

Figure 28 Sliced Thigh Curves (Relative to Left Femur Axis) and Back Curves .65

Figure 29 Contour Surface Analyzing Scheme ................................................... 67
Figure 30 Sections of J826 H-point Machine and Subjects Contours ................. 71
Figure 31 Sections of J826 H-point Machine and Subjects Contours ................. 72
Figure 32 Sections Through the ASPECT and UMTRI Contours with Postures
and Contact Edges .............................................................................. 74
Figure 33 Measured Back Contours ................................................................... 75
Figure 34 Contour Sections Up the Seat Back ................................................... 76
Figure 35 Comparison between ASEPCT and UMTRI Back Contour ................. 80
Figure 36 ASPECT Symmetrical Buttocks and Thigh Contours .......................... 82
Figure 37 Torso Length Scaled In Z-Direction .................................................... 84
Figure 38 Measured Contour Curves .................................................................. 87
Figure 39 Measured Back Contours Between Upper Lumbar Joint .................... 88
Figure 40 Average Back Contour And Posture Side View .................................. 89
Figure 41 Average Back Contour Left Rear View ............................................... 90
Figure 42 ASPECT Back, Buttocks and Thigh Contour Surfaces ....................... 91
Figure 43 ASPECT Manikin ................................................................................ 92
Figure 44 Tested and Scaled Subject 1’s Contour Curves, rear view ............... 102
Figure 45 Tested and Scaled Subject 1’s Contour Curves, oblique view .......... 103

xi

Figure 46 Tested and Scaled Contour Curves at Each Section of Small Female

Subject 1 ............................................................................................ 104
Figure 47 Measured and Scaled Subject 1’s Contour ....................................... 115
Figure 48 Measured and Scaled Subject 1’s Contour ....................................... 116

Figure 49 Tested and Scaled Contour Curves at Each Section of Large Male
Subject 1 ............................................................................................ 117

Figure 50 Typical Difference Curves between Scaled and Measured Contour. 129

Figure 51 Typical Difference Curves between Scaled and Measured Contour.133

Figure 52 Typical Bar Charts of Coefficient of Correlation ................................ 142
Figure 53 Typical Corridor Plots at Mid Thorax, Lumbar, Pelvis, and Thigh of
Large Males ....................................................................................... 146
Figure 54 Typical Corridor Plots at Mid Thorax, Lumbar, Pelvis, and Thigh of
Small Female ..................................................................................... 150
Figure 55 Pelvis Region Not Well Supported in Seat ........................................ 156
Figure 56 Female Subject 2 Rear View ............................................................ 167
Figure 57 Small Female Subject 2 lso View ..................................................... 167
Figure 58 Scaled and Measured Curves of Small Female Subject 2 ................ 168
Figure 59 Small Female Subject 3 Rear View .................................................. 171
Figure 60 Small Female Subject 3 Iso View ..................................................... 171
Figure 61 Scaled and Measured Curves of Small Female Subject 3 ................ 172
Figure 62 Small Female Subject 4 Rear View .................................................. 175
Figure 63 Small Female Subject 4 lso View ..................................................... 175
Figure 64 Scaled and Measured Curves of Small Female 4 ............................. 176
Figure 65 SmaIFemale Subject 5 Rear View .................................................... 179
Figure 66 Small Female Subject 5 lso View ..................................................... 179
Figure 67 Scaled and Measured Curves of Small Female Subject 5 ................ 180
Figure 68 Small Female Subject 6 Rear View .................................................. 183

xii

Figure 69 Small Female Subject 6 lso View ..................................................... 183

Figure 70 Scaled and Measured Curves of Small Female Subject 6 ................ 184
Figure 71 Small Female Subject 7 Rear View ...... 187
Figure 72 Small Female Subject 7 lso \ﬁew ..................................................... 187
Figure 73 Scaled and Measured Curves of Small Female Subject 7 ................ 188
Figure 74 Small Female Subject 8 Rear View .................................................. 191
Figure 75 Small Female Subject 8 lso View ..................................................... 191
Figure 76 Scaled and. Measured Curves of Small Female Subject 8 ................ 192
Figure 77 Small Female Subject 9 Rear View .................................................. 195
Figure 78 Small Female Subject 9 lso View ..................................................... 195
Figure 79 Scaled and Measured Curves of Small Female Subject 9 ................ 196
Figure 80 Large Male Subject 2 Rear View ...................................................... 200
Figure 81 Large Male Subject 2 lso View ......................................................... 200
Figure 82 Scaled and Measured Curves of Large Male Subject 2 .................... 201
Figure 83 Large Male Subject 3 Rear View ...................................................... 204
Figure 84 Large Male Subject 3 lso View ......................................................... 204
Figure 85 Scaled and Measured Curves of Large Male Subject 3 .................... 205
Figure 86 Large Male Subject 4 Rear View ...................................................... 207
Figure 87 Large Male Subject 4 lso \ﬁew ......................................................... 207
Figure 88 Scaled and Measured Curves of Large Male Subject 4 .................... 208
Figure 89 Large Male Subject 5 Rear View ...................................................... 211
Figure 90 Large Male Subject 5 lso View ......................................................... 211
Figure 91 Scaled and Measured Curves of Large Male Subject 5 .................... 212
Figure 92 Large Male Subject 6 Rear View ...................................................... 215
Figure 93 Large Male Subject 6 lso View ......................................................... 215

xiii

Figure 94 Scaled and Measured Curves of Large Male Subject 6 .................... 216

Figure 95 Large Male Subject 7 Rear View ...................................................... 219
Figure 96 Large Male Subject 7 lso View ......................................................... 219
Figure 97 Scaled and Measured Curves of Large Male Subject 7 .................... 220

xiv

LIST OF TABLES

Table 1 Speciﬁcations of the Contour Chair ...................................................... 29
Table 2 Subjects Feet Positions ....................................................................... 32
Table 3 Subjects Selection Criteria ................................................................... 46
Table 4 Calculation of Hip Joint Center ............................................................ 60
Table 5 Calculation of Lower Lumbar Joint ....................................................... 62
Table 6 Average Differences (mm) between Scaled and Measured Data ...... 137
Table 7 Standard Deviations (mm) of the Differences .................................... 138
Table 8 Average Differences (mm) between Scaled and Measured Data ...... 139
Table 9 Standard Deviations (mm) of the Differences .................................... 140
Table 10 Coefﬁcients of Correlation .................................................................. 143
Table 11 Coefﬁcients of Correlation .................................................................. 144
Table 12 Body Dimensions for Large Male Scaling .......................................... 164
Table 13 Body Dimensions for Small Female Scaling ...................................... 165

CHAPTER 1 INTRODUCTION

1.1 Background and Literature Review

Seating comfort is a growing issue in both automotive and ofﬁce seat
design. Comfort is also a vague and ubiquitous term, and comfort has no clear
and operational deﬁnition. Comfort is inﬂuenced by several factors such as
vehicle package factors, seat factors, an occupant’s sitting posture, and seat
contour. Seat factors include the geometric dimensions and mechanical
responses of seats to test manikins and to human beings, as well as the layout of
the vehicle interior components, also called vehicle package factors. The
positions of the vehicle components affect a driver’s location in the vehicle seat
and can also affect the driver’s postures. Therefore, a driver's posture results
from compromises between comfort and restrictions from interior layouts and
driving tasks. During driving, a driver may change his posture from time to time
in order to make himself comfortable. When sitting, a person’s body contours
either conform to the seat contours or the seat will not be in contact with and
support the person's body. The contours of the seat determine the possible
contours of the body and the compatibility of seat and body contours inﬂuence
contact pressures and sitting posture. Hence, seating comfort is directly related
to seat contour, including lumbar support location and amount of lumbar
prominence, associated with permissible ranges of sitting postures.

To achieve desired seat comfort, accurate and efﬁcient tools are needed

for both seat design and evaluation. To comply with the Society of Automobile

Engineer (SAE) Handbook Recommended Practice J826 [1] used since the early
1960’s, seat dimensions and shapes are determined by using an SAE J826 two-

dimensional H-point template, and an SAE J826 three-dimensional H-point

 

 

 

Figure 1 J826 Two-Dimensional H-point Template and Three-Dimensional

H-point Machine

machine, known as the J826 manikin (Figure 1). The H-point in the SAE J826
tools is a representation of the average man’s hip joint center (HJC). The SAE
handbook also speciﬁes measurements for dimensions, adjustability, and
conﬁguration of auto seating.

It is simple and practical to use physical tools in designing, conﬁrming and
evaluating the desired comfort. The survey [2] conducted by the Automotive
Seat and Package Evaluation and Comparison Tools (ASPECT) program found
a strong need for a physical tool like the J826 H-point machine. However, the
J826 2-D H-point template and 3-D H-point machine were originally developed
for locating occupants in an automobile seat within a vehicle interior. Hubbard [3]
pointed out that the misuse of the J826 2-D H-point template and 3-D H-point
machine resulted in a slumped posture and affected the comfort of seated
occupants for more than two decades.

The J826 machines simulate the mid-size male body segments by
simulating the combined pelvis and thighs, a segment of combined thorax and
lumbar, and the two feet and legs, which can be adjusted to the 10‘“, 50th and
95th percentile male adult’s leg and thigh lengths. The torso and butt-thigh
segments can articulate about a lateral axis through the H-points. The 3-D H-
point machine is usually used to determine the H-point of a seat and seat back
angle following methods speciﬁed in the SAE Handbook.

Normal human motion in a seat is very complicated, and includes, among
other factors, lumbar curvature changes as well as, thorax and pelvis articulation

around spinal joints, associated with changes in spinal curvature. The structure

of the 3-D H-point machine has a slumped posture with a straight lower back
does not represent a normal human postures change in vehicle seats. The 3-D
H-point machine does not ﬁt into seats that have highly contoured shapes with
large amount of lumbar prominence. If the J826 manikin is placed in a contoured
automobile seat with lumbar support and its lower lumbar and pelvis area is held
against-the seat back, its upper torso is in a fonrvard position without contacting
the seat back. If its upper torso is pushed against the seat back, its H-pOint
slides cut away from the seat back.

From the point view of seating design and analysis, it is very important to
describe human posture and contour as precisely and correctly as possible.
There were two very important studies in this area. One was conducted by Haas
and Hubbard at Michigan State University. They developed a two dimensional,
sagittal computer model [4], which represents the average adult male’s
geometry, including the shape and relative position of the head, thorax and
pelvis, and spinal motion. The head, thorax, and pelvis in the Haas model were
connected by a series points representing joint centers in the cervical, thoracic,
and lumbar spinal regions, as seen in Figure 2. In this model, the thorax and
pelvis articulate around upper and lower lumbar joints respectively, with oneto-
one opposite rotation. This computer model program is known as JOHN to

recognize the support from Johnson Controls, Inc.

 

 

 

Figure 2 Two-Dimensional JOHN Model

Another important anthropometric study was performed at the University
of Michigan Transportation Research Institute (UMTRI) [5] for crash dummy
development by determining how a driver is positioned within an automobile. It
was the best data resource at that time and was partially adopted in the JOHN
program. In their study, UMTRI placed subjects in production auto seats in car
interior layout environments. A combination of seat surface casting strips and
photographs of external landmarks on subjects’ skin was used to determine both
exterior contour and landmark locations. Subjects were placed into three
subgroups: mid-sized male, small female, and large male. The limitations of
UMTRI contours are as follows: 1) the production seats were used in the study
and the shape of the seat affected subject’s contours, especially in the posterior
thigh area, where soft tissue deformed substantially. The production seats were
designed and evaluated by the 2-D H-point template and 3-D H-point machine,
therefore, the shortcomings of the tools still inﬂuenced the UMTRI data; 2) the
study did not consider the lumbar curvature and foot positions, which are crucial
for seating design; 3) in Haas’s study he pointed that the UMTRI contour’s pelvis
was positioned too high. Apparently the 46 mm soft tissue under ischial
tuberosities on the bottom of the pelvis when one sits in a seat was
unreasonable.

Development of computer model programs has rapidly advanced and
cannot be overstressed in the area of seat design and computer-aided evaluation
to achieve optimal design. Also, computer modeling can reduce expensive

design and validation cycles. There are several available anthropometric

computer models for possible application to automotive seating design, which are
brieﬂy described below.

{:RAMSIS [6] is a three-dimensional, ergonomic computer program
developed by TecMath (now Human Solutions), GmbH in Kaiserslautern,
Germany, to represent people seated within automotive interior layouts for
analysis and design. RAMSIS includes a posture prediction algorithm based on
data of subjects seated in vehicle packages. The human sizes and shapes are ﬁt
to video images of individual subjects, and they are based on anthropometric
data. Thus, the RAMSIS model is positioned in a seat based on the video
images of individual subjects. Because video cameras are incapable of
capturing the deformed body segments that are contacting a seat surface, the
contour is not realistic. Therefore RAMSIS is not being used for seat comfort
design or evaluation. This present study is being performed to provide TecMath
the deformed back, buttocks, and posterior thigh contou‘rE\

JOHN has been developed under the direction of Professor Hubbard at
Michigan State University with the support of the Automotive Systems Group of
Johnson Controls, Inc. JOHN program is based upon Haas’s human model [4],
which represents the movement between the ribcage and pelvis with selected
values of lumbar curvature and the torso recline angle to describe seated
postures. JOHN represents the geometry of the skeleton, major muscles, and
back skin surfaces of a 50th percentile male.

There were several major phases in JOHN development worth mentioning

here.

For his model [4], Hass deﬁned two geometric variables to determine the
relative positions of body segments. Torso recline angle (T RA) was deﬁned as
the angle measured from a vertical line passing through the twelfth thoracic/ﬁrst
lumbar vertebrae (T 1 2/L1) and the ﬁfth lumbar/ﬁrst sacral (L5/S1) spinal joint
centers. Total Lumbar Curvature (TLC) was deﬁned as the total amount of
curvature of the lumbar region in the sagittal plane. TLC is zero when the lumbar
joints are aligned on a straight line. TLC is the angular position change of the
thorax relative to the pelvis and is positive with lordosis (concave lumbar
curvature). Both TRA and TLC are as shown in Figure 3 and Figure 4.

Boughner [7] created a three-dimensional model program for the 50th
percentile adult male using solid modeling techniques. The models of major
skeletal structures were established by using the data from several anatomy
texts and the data from the UMTRI study [5]. This model also represents the
posterior thigh muscle geometry to calculate hamstring muscle length.

Bush [8] modeled the shapes of the erector spinae muscles and the
external skin contours of the JOHN program by using data from the cadaver
studies of Koritké and Sick [9]. Bush rotated the 0° TLC contours about the
lumbar spinal joint centers to achieve other contours with respect to selected
values of TLC, -10°, 0°, 10°, 20°, 30°, and 40°.

Frost [10] developed new contours for the JOHN models by using both the
Bush and UMTRI data [5]. These contours could be articulated to any speciﬁed
amount of TRA and TLC in the sagittal plane. Then, by using scaling factors for

the individual dimensions of the thoracic, lumbar, and pelvic segments, Frost

created back skin contours for 5th percentile adult females as well as 50‘", and

95th percentile adult males.

10

 

 

TRA

 

 

Figure 3 Deﬁnition of TRA

11

 

7

Figure 4 TLC

12

TLC is positive
when lumbar
dorlosis

1.2 Statement of The Problem

The shape of the seat has a strong inﬂuence on the posture and comfort
of the seat user. Because the body articulates when a person sits in a seat, his
body changes posture so that his body surfaces against the seat conform to the
seat shape. Thus, the shape of the seat strongly inﬂuences the shape and
posture of the seat user. If the seat shape is not compatible with possible or
preferred shapes of the person, then there is geometric mismatch between the
seat and sitter. Ellis mismatch can reduce the area of contact, increase local 0(’
pressures, and be perceived as discomfortTIJSome postures may be preferred by
the seat user to function in the seated environment, such as seeing instruments
or operating controls in an automobile. Also, some postures are more healthful,
such as postures with some concave lumbar curvature (lordosis), which allows
the upper torso to rotate rearward and the head to be over the base of the neck.
This lumbar curvature reduces spinal disc pressures, and this head position
reduces neck and shoulder muscle tension to maintain head position. The shape
of the interface surface between the seat and the user is signiﬁcant for comfort

support and posture.
1.3 Objectives

The general goal is to deﬁne seated contours for people in typical driving

postures. The more speciﬁc objectives of this study are as follows:

13

r/
1.
l

{1‘- Collect data of seat contact surface contours and postures for
representative subjects of the 5"1 percentile small adult female, 50th
percentile average adult male and 95th percentile large adult male.

2. Develop contours based on analyzed measurement data that could be
used for the JOHN and RAMSIS programs.

3: Develop a methodology to represent seated contact surface contours for

I

. . /I
diverse drrver populations. 3

_,__.—$

14

CHAPTER 2 ANATOMICAL BASICS AND TARGETED BONY LANDMARKS

2.1 Introduction

In a sitting position, human body weight is mainly transferred to certain
areas of ischial tuberosities of the pelvis and their surrounding soft tissues, and
also the feet [11, 12]. The weight distribution is related to seating comfort, and
soft tissue fatigue and is usually determined by sitting posture and the supporting
seat contour. To better understand sitting posture and comfort, a brief review of

the anatomy of spine, pelvis, and thigh is presented.

2.2 Posture Descriptions

The representation of sitting posture adopted in this research was
developed in the ASPECT program [2]. ' It is based on a kinematical model of the
human body, in which the human body is modeled as an open chain of straight-
Iine segments connected at major joints [4]. The segment lengths, the
orientation of the segments, and the location of the joints have to be determined
by means of a combination of photography and calculation [2], as seen in Figure
5.

To collect posture data, the choice of appropriate bony landmarks has to

be made based upon both the data acquisition method and human anatomy [2].

0 The bony landmarks of interest have to be palpable and easily located. It is

crucial for targets at the selected landmark locations to be visible to multiple

15

video cameras so that their positions can be recorded and calculated to
determine the joint centers of interests.
A typical presentation of a targeted subject and his posture data is shown

in Figure 5 and Figure 6.

16

 

Lower Neck Joint

Wrist Joints / |\Shoulder
\ Elbow J7'nt Jornts
pper Lumbar
Knee Joints Joint
\ Front of

Pelvis A b
Lower Lum ar Joint

/ Hip Joint

Centers
Ankle Joints

Figure 5 A Typical Posture Model

17

Steering
whee

 

Figure 6 Subject in Contour Measurement Chair

2.3 Positional Terminologies

The term “superior” is used to express the direction toward the top of the
head, and “inferior” toward the feet. The direction toward the front of the body is
deﬁned as “anterior", and toward the back of the body as “posterior”.

There is a central mid-sagittal plane of symmetry passing through the
head and trunk that separates the right from left sides. On either side of the
central plane, the direction pointing to the plane is referred to as “medial”, and
the direction away from the plane is referred to as “lateral”.

A plane perpendicular to the long axis of the body and dividing its superior

portions from inferior is named the transverse or horizontal plane.

2.4 Upper Limb

In the performance of tasks and reaching controls, the positions of the
shoulders and the upper limbs inﬂuence spinal postures and comfort during
driving. Therefore, arm and shoulder positions must be taken into account for
posture [7, 13]. The three bony landmarks targeted or probed on each side were
the acromian process, lateral epicondyle, and mid-wrist (a midpoint of styloid
process of radius and ulna) to locate the shoulder joint, elbow joint, and mid-wrist
skin points, respectively. Figure 7 shows that the acromion process, lateral

epicondyle, and mid wrist were targeted or probed.

19

‘ ACromion Process probed

O
Mid—wrist

Lateral
Epicondyle

 

Figure 7 Targets at Lateral Epicondyle, Mid-wrist, and Acromion Probed

2.5 Spine

The spine is very important in biomechanical aspects of sitting, especially
the lumbar region. The spinal column consists of ﬁve segments of cervical,
thoracic, lumbar, sacral, and coccygeal vertebrae, as shown in Figure 8. There
normally are thirty-three vertebrae stacked on top of one another in a spinal
column for a normal person. Each vertebra is named with respect to its relative

position in the spine column from top to bottom, such as C1-C7 for the seven

20

vertebrae in the cervical region, T1 -T12 in the thoracic region, and L1-L5 in the
lumbar region.

Among these ﬁve segments, only cervical and lumbar vertebrae have
signiﬁcant mobility. Because the cervical vertebrae are responsible for head
movement, they are of less importance for sitting postures. In the thoracic spine,
the mobility is limited because of the connection to the rib cage [13]. The sacrum
is effectively part of the pelvis with 5 vertebrae fused together, and the coccyx (4
vertebrae) does not have meaningful function for seating posture. There are
natural curvatures in the vertebral column: cervical lordosis, thoracic kyphosis,
and lumbar lordosis resulting from the vertebrae and discs structural
characteristics when standing erect, as shown in Figure 8. In this study, markers
of the spinous processes of C7, T8, and sacrum were used to determine the
spine posture. By palpating the most posterior aspect of the spinous process,
these vertebral landmarks can be found, especially the prominent C7 vertebral

process.

21

Cervical

   

 

(Lodorsis)
Thoracic
i (Kyphosis)
Lumbar
(Lordosis)
Sacrum
Coccygeal

Figure 8 The Spine Column, Adapted from Chafﬁn [13]

22

2.6 Thorax

Thoracic vertebrae articulate with each other and with the rib cage, so
their mobility is limited [13]. This is why the thorax is considered to be a rigid
body in posture analysis [2]. The top end of the sternum is named the
manubrium. The inferior tip of the stemum is called the Xiphoid process, Figure
9. Together with the landmarks on the vertebral column, these two stemal
landmarks are used to determine the thoracic shape and position in posture

analysis.

 

Figure 9 Targeted at Manubrium and Xiphoid Process

23

2.7 Pelvis

When seated, the body is mainly supported in the area of the ischial
tuberosities of the pelvis and their surrounding soft tissues [13]. In this sense,
the pelvis is really a base for the trunk when sitting. The sacrum transfers most
of the weight from the lumbar spine through the pelvis to the ischial tuberosities
and its surrounding soft tissues when a person sits. The pelvis also articulates at
hip joint centers with the femurs of the thighs. The pelvis consists of two
hipbones, or coxal bones. Each hipbone derives from three separate irregular
bones: the illium, ischium, and pubis. The illiac crest is anteriorly prominent at
the right and left anterior superior illiac spines (ASIS), which can be palpated and
targeted easily. The most inferior prominence of each ischium is known as
ischial tuberosity, which supports the pelvis in the sitting position. The pubic
symphysis is located where the two pubic bones connect the left and right
hipbones anteriorly. During the tests, the pubic symphysis was probed to
determine its location. Also, the mid-point of two posterior superior illiac spine

(PSIS) was targeted as shown in Figure 10.

24

 

Figure 10 Pelvis: illium, ischium, pubic symphysis

2.8 The Lower Limbs
The spherical head of the femur inserts into the acetabulum of the pelvis

to form the hip joint center. The neck of the femur then extends to the greater

25

and lesser trochanters. At the distal end of the femur, there are the medial and
lateral femoral condyles of the knee.

The tibia is the shinbone, which transfers the forces between the knee and
the ankle. The ﬁbula is the slender bone lateral to the tibia. At the distal end of
the ﬁbula, the lateral malleolus is easily seen and palpated. The lateral femoral
condyle, lateral malleolus, and approximate ball of foot were selected as bony
landmarks to determine the locations of the ankle joint, knee joint, and foot

position, as shown in Figure 11.

7,. ateral
. 0006938.-

 

Figure 11 Targets on Lower Limbs

26

 

CHAPTER 3 MEASUREMENT OF SEATED HUMAN CONTOURS AND

POSTURES

3.1 Introduction

The goal of this research is to experimentally determine seat contact
surface contours for selected subgroups of driver populations in certain
permissible ranges of sitting postures within typical vehicle interior geometries.
To simplify anthropometric issues, two measurements that account for most of
the variance of body dimensions (height and weight) [14,15] were chosen to
classify the subgroups of driver populations. In this section, the. details related to

measurement will be discussed.

3.2 Experimental Device and Setup
3.2.1 Contour Test Chair

/To acquire the seat/sitter contact surface data, a special contour test chair
was designed and built, Figure 12. To represent a typical mid-sized sedan, the
speciﬁcations of the contour chair were used. See Table 1. The seat pan angle
and the seat back angle were measured by following the SAE J826 procedure

described in [1]\

. \l
/
I/V

27

f Wires & their pat

ii

1‘ .‘
2:

  

4L.

_ 3.. Light box for seat pan
1, I

 

Figure 12 Contour Chair with Targets for Local Coordinate System

28

Table 1 Speciﬁcations of the Contour Chair

 

 

 

 

Seat height 280 mm from foot support surface to H-point
Seat angle 15° with respect to horizontal
Seat back angle 23° with respect to vertical
Steering wheel angle 20° with respect to vertical

 

500 mm X 500 mm X 100 mm
Cushion foam size -

(Width by Depth by Thickness)

 

 

 

 

Seat back size 580 mm X 650 mm X 100 mm

 

3.2.1.1 Seat Pan

In the contour chair, there was a ﬂat seat pan with a foam block having
dimensions of 500 mm by 500 mm by 100 mm. A hundred stainless steel wire
cables were evenly distributed over the 500 mm x 500 mm area in a matrix of
10x10 with the wires 45 mm apart. These wires moved smoothly within plastic
tubes when one of the cable ends was displaced via soft plastic discs of 25 mm
in diameter that were on the foam surface. The other ends of these wires were
led through the plastic tubes to channels in a light box where they were aligned in
parallel channels. By means of video cameras and computer interface, the

positions of the wire ends in the light box shown in Figure 13 could be digitized

29

accurately and rapidly with a Visual Basic program. An original reference
position ﬁle (before deformation) of these ends was taken separately with respect
to the displaced position ﬁle (after deformation). Thus the deformation of the

seat pan and seat back can be recorded and further processed.

 

Figure 13 Light Box

30

3.2.1.2 Seat Back

On the seat back there is a 13 x 16 matrix of steel cables, 45 mm apart. A
maximum of four inches of working distance can be set for the wires on the seat
back. No foam pad was needed for the seat back because there was sufﬁcient
friction in the matrix of measurement wires to support the torso in the selected
posture. The seat back was ﬂat to eliminate the inﬂuence on the measured back
contour from the shape of the back support. The data acquisition principles were

the same as those for the seat pan.

3.2.1.3 Foot Support Plate and Steering Wheel

60 ﬁt the different sized groups, the foot support plate and steering wheel
were adjustable in the fore-aft direction. However, the angles were ﬁxed. The
selected angle value was 20 degrees with respect to horizontal for foot support
plate, vertical for steering wheel, respectively. Refer to Figure 6.

The horizontal distance from the heel to H-point could be changed to
adjust the relative deformation of the thighs into the seat foam, and let the
subject determine his preferred position by altering the horizontal position of the
foot support plate. Also, the design foot position was analyzed. The distances
for small size female, mid size male, and large male were listed in the Table 2.

The steering wheel was adjusted horizontally and vertically for the subject
to feel comfortable as in his normal driving condition, and the angle of the

\

steering wheel was ﬁxed. Refer to Figure 6

31

Table 2 Subject’s Feet Positions

 

Mid-male Small Female Large Male

 

 

 

Foot Position 900 mm 800 mm 1000 mm

 

 

 

 

3.2.1.4 Measurement Contour Chair Accuracy
To determine the accuracy of the contour chair, two methods were
employed to obtain the outer shape of the SAE manikin. F i_rst,_thg sheﬂgf the

SAE manikinﬁwﬂas‘scanned by a probe?" This shell could be considered as an

‘ .M‘h- - V-

. L...—

accurate one. Secondly, the SAE manikin was put into the measurement contour
chair and the shape was recorded in the system. Then, the second shell shape
was compared to the ﬁrst measurement from the scanning at the same locations.

The results showed the system error was within 3 to 5 mm.

3.2.1.5 Seat Foam Stiffness Selection

(After the primary mechanism of the contour test seat was built, a
preliminary comparison study between softer and stiffer foams was performed for
the same subjects to evaluate the effects of foam properties on measured thigh
contours. The results are shown in Figure 14. In the regions directly under the
femurs, the deformed thigh shapes were almost the same for both stiffer and
softer foams owing to the mechanical restrictions from the femur and the

presence of the hamstring muscles. For the softer foam, more of the subjects’

32

  
 

Data from stiffer foam

.0
‘1'
r .
/

//

/

Figure 14 Same Subject Measured on Stiff and Soft Foams

33

body shapes were recorded because the foam deformed more. These ﬁndings
were also veriﬁed by Setyabudhy’s ﬁnite element model [16]. On the other hand,
the foam shape itself as a seat factor was also considered. A ﬂat, soft block of
foam was selected 'as seat pan cushion to minimize the inﬂuence of seat pan

shape on the contour shape and to obtain more of the subjects’ body shapes)

3.2.2 Test Lab Space

The testing space was approximately 6m X 6m X 3m. A ﬁve-camera
MacReﬂex Qualysis high-speed video system [17] was carefully set up in the
space so that each target could be visible simultaneously at least for two
cameras, as required by the software used for calculation of the three
dimensional coordinates [17]. Through calibration of the camera system, a right-
handed SAE vehicle Cartesian system was deﬁned in the space, with positive X-
axis pointing rear ward, Y-axis from subject’s left to right, and the Z-axis upward

vertically.

3.2.3 Measurement Targets and Probe

Measurement targets were spherical wooden beads of 16 mm diameter
covered with 3M retro-reﬂective tape [18], shown in Figure 15, as required by the
camera system. The reﬂective bead was then glued to a square piece of vinyl
with approximate 20 mm by 20 mm in dimension. By means of double-sided
tape and medical tape, the target was secured to the subject’s skin over the

skeletal landmark of interest.

35

A /Tape-covered Bead »

/Vinyl Backing

 

Figure 15 Measurement Target

iii,
\/ A measurement probe used in the testing is shown in Figure 16. There

were two targets ﬁxed on a steel rod of 4 mm diameter. The centers of the
targets were on the rod centerline. Whenever it is difﬁcult to apply the targets in
some region or for two cameras to capture a target, the probe can be used to
locate the skeletal landmark.

The landmark location could be calculated from the measured centers of

bead 1 and bead 2, and the lengths of the steel rod as follow.

 

Ln = In. —x.)2 +(y. -y.)2 +(z. —z.)2 (1)

36

_ L-X2 +(L-L12)x1

X
L12 {,2

Y = L'Yz +(L—L12)YI ,‘i (2)
L12

_L-zz+(L—L,2)z1

Z
I-12

Bead 2 at
(XL-N117)

Steel Rod

 

\
\ Landmark (XXX)

Figure 16 Measurement Probe

37

3.2.5 Reference Chair

When a subject sits in the contour chair, any targets on his [FEBJVJJIILQK
be visible to the cameras. To collect the position data of C7, and T8, a non-
deforrnable thorax was assumed. This makes it possible to gather the position
data of C7 and T8 relative to the manubrium and xyphoid in a so-called reference
seat, or hard seat as presented in Figure 17.

There is a 70 mm opening down the centerline on the seat back, which
provides visibility to the targets on the subject’s spine for the cameras. The seat
angles were determined based on SAE standards J826 [1]. The ﬁrst target on
the edge is 120 mm fonrvard of the seat back, and the second target 225 mm
apart from the ﬁrst one in the rearward direction. The two targets provided a

I “r

reference coordinate system.

38

1" - r
/'
/ ,

I .

 

Figure 17 Reference Chair

39

3.2.6 Contour Data Acquisition

A reference position of the contour chair was collected to record the initial
positions of the ends of the steel cables when the seat pan foam was not
deformed. While the subject was sitting on the contour chair, the positions of the
displaced ends of the steel cables were gathered. The differences between the
two positions of the steel cables illustrated the deformed interface between the
seat foam surface and the subject.

In the collected seat contour data, only the contact region between a
subject and the seat can be taken for a subject’s body contour. To differentiate
the subject’s body contour from the non-contact part of the seat surface, a testing
probe was used to trace the border of the contact areas of a subject and the
seat. The positions of the two balls on the probe were recorded, since the
lengths of the probe were known, the tip positions of the probe could be
computed.

Ten mid-sized male subjects were measured and 9 out of 10 subjects'
data analyzed for the ASPECT manikin contour [16]. Ten of small female and
another ten large male subjects were measured and the data were analyzed.

The testing protocol is attached. The subject measurement is pictured in

Figure 6.

4O

3.2.7 Posture Data Collection

The representation of sitting posture adopted in this experiment was
developed in the ASPECT program [2]. This kinematic model of the human body
is modeled as an open chain of straight-line segments connected at major joints.
The segment length, the orientation of the segments and the location of the joints
have to be determined by means of a combination of photography and
calcirlation. Reﬂective balls of 16 mm diameter were placed at the selected skin
landmarks on subjects. The posture data were collected in both the reference
hard seat (Figure 16) and the contour test chair in Figure 6 or 11.

A ﬁve-camera MacReﬂex Qualysis high-speed video system was used to
gather the motion measurement data in the approximately 6 m x 6 m x 3 m
testing space [17]. Each camera contains an optical ﬁlter, which cuts off the
visible part of the light spectrum. Each camera is equipped with an infrared light
to illuminate the measurement area, and light is reﬂected from the targets back to
the camera. These infrared lights help to suppress the ambient light and only the
target balls can be seen and recorded. There is a video processor to calculate
the center of the ball coming with each camera. The MacReﬂex software
provides the tool to read data from the camera system and to distinguiSh one
target from another among all the targets in the image by using direct linear

'1

transformation from the object space to the image space. Each camera car: only

 

readjonly two of the’three coordinates in a three-dimensional space,“ so any
marker must be seen by two cameras to form an over-deterrnined system of four

equations for the three coordinates in the three-dimensional space. Before data

41

collection, all the cameras were calibrated to gather all the set-up constants for
the direct linear transformation. Finally the MacReﬂex converts the data to a
format proper for the subsequent processing, such as a CSV format in] our case. Y
During the measurement, the acromions were probed because of the
visibility to the cameras. Owing to the personal privacy and sensitivity concern,

the pubic symphysis was probed by the subject after instructions was given.

42

3.3 Testing Protocol

1) Instruct a subject about the measurement procedures. Have the
subject sign the consent form (lRB#96-765) approved by University Committee
on Research Involving Human Subjects (UCHRIS).

2) Target subject for the reference chair

The landmarks C7, T8, T12, Sacrum, stemal notch, xyphoid process; right
and left ASIS, right and left lateral femoral condyle, maleolous, ball of foot on left
side were targeted by the taped wooden beads on vinyl backing. The pubic
symphysis will be probed by the subject.

3) Take the reference ﬁle of the body in the reference chair in Figure 16.
Measure the distance between the ball of foot target and heel point for the
calculation purpose.

4) Target subject for contour chair

Surface landmarks stemal notch, xyphoid process; right and left lateral
humeral condyle, right and left Wrist, right and left ASIS, right and left femoral
lateral condyle, right and left maleolous and ball of foot on left [side and two seat

i I
pan targets, two seatback targets and M93?! frame targets.

Surface landmark acromion will be probed by a tester and pubic
symphysis will be probed separately by the subject during the measurement
process.

5) Decide steering wheel position

43

Position subject at the desired distance between H-point and foot support
bar, placing hands at 3 and 9 o’clock position (during all measurements), adjust
the steering wheel fore or aft for a comfortable position by the subject.

6) Decide preferred foot position

Set the distance between manikin H-point and foot support bar to
designed distance 1, let subject sit in the contour chair and select a comfortable
position 1 by moving foot rest forward, then subject leaves chair.

Set the distance between manikin H-point and foot support bar to
designed distance 2, let subject sit in the chair and select a comfortable position
2 by moving foot rest reanrvard.

Decide Preferred foot position (the mid-point 3 between positions 1 & 2).
7) Gather posterior contour of thigh and buttocks“; ' 1"

Set the distance between manikin H-point and foot support bar to
designed driving distance with the seat pan in normal position, let subject sit into
the chair with back ﬂat against seat back.

At designed driving position, wait for 20 sec., and then gather the data for
contour and posture (one ﬁle with Acromian probed, the another with Pubic
Symphysis probed).

Move the foot support to designed driving position 2, wait for 20 sec.;
gather the data of contour, posture, and perimeter.

Move the foot support to preferred position, wait for 20 sec., gather the

data of contour, posture.

Move the foot support to designed driving position 3, wait for 20 sec.,
gather the data of contour, posture.

Redo step 7) twice.

8) Gather posterior contour of back, thigh and buttocks

With 0 lumbar prominence, pull the back wires out and set the distance
between manikin H-point and foot support bar to designed driving position with
the seat pan in a fonrvard position, let subject sit on the seat pan, move the seat
pan back into the seat back support using power track, have subject sit in the
chair with back ﬂat against seat back and wait for 20 sec., gather data of contour,
posture.

Redo step 8) twice.

3.4 Subject Selection Criteria

Based upon data from U. 8. population study by Abraham, [15], body
mass and stature range from 5‘" percentile small female values are 47 kg (104
lbs.) and 151 cm (59.5 in.) to 95‘" percentile large male values of 102 kg (225
lbs.) and 187 cm (73.5 in.), respectively. A decision was made that three size
groups would be used, they are 5‘“ percentile small female, 50th percentile
average male, and 95‘" percentile large male within the U. S. population,. The

ranges of mass and stature are given in the Table 3.

45

 

 

 

 

 

Table 3 Subjects Selection Criteria
/ i
S”°j°°.t Large Male (95%) Mid-sized Male (50%) Small Female (5%)
Categorles

Range Min Max Min Max Min Max

32:: 76.5 108 67.5 78.75 45 56.25
mm,» (170) (240) (150) (175) (100) (125)
Stature
(height) 178.9 186.2 164.2 171.5 144.8 151.9

. (73) (75) (67) (70) (59) (62)

cm I (In)

 

 

 

 

 

 

 

3.5 Subject Anthropometric Data Acquisition

To detail information of participating subjects’ body dimensions, it was
necessarily to obtain anthropometric data. These data were collected by using
the deﬁnitions and methodologies speciﬁed in 1988 Anthropometric Survey Of
US. Army Personnel [14]. The list of the following anthropometric items was

gathered.

ANTHROPOMETRIC DIMENSIONS MEASURED

Deﬁnitions cited from ANSUR [18], as follow.
Stature (99, reference number in [18])

46

 

Weight (124)

Sitting Height (93)

Biacromial Breadth (10)

Mid—shoulder Height, Sitting (78)

Hip Breadth, Sitting (66)

Waist Height, sitting (121)

Bispinous Breadth (14)--------- Inter-ASIS Width
Waist Breadth (1 12) .
Chest Breadth (32)

Chest Depth (36)

lnterscye ll (70)

lnterscye l (71)

Waist Depth (115)

Waist Back Length (Natural indentation)(1 10)
Waist Back Length (Omphalion)(111)

Waist Height,sitting (Natural lndentation)(120)
Waist Height,asitting (Omphalion)(121)

Chest Circumference (33)

Chest Circumference at Scye (34)

Chest Circumference below Breast (35)

Waist Circumference (Natural indentation)(113)
Waist Circumference (Omphalion)(114)
lnter-PSIS Wldth

47

Hip Breadth (65)

Hip Breadth,sitting (66)
Buttock Depth (24)

Buttock Circumference (23)
Thigh Circumference (103)
Knee Circumference (71)
Buttock height (25)
Buttock-knee length (26)

Buttock-popliteal length (27)

48

CHAPTER 4 DATA ANALYSIS APPROACHES AND RESULTS

4.1 Introduction

The recorded posture data were tracked by using MacReﬂex software [17]
and exported in a format of CSV ﬁles. During data processing, the data were
transformed from intermediate coordinate systems to the global coordinate
system.

Because of the complexity of comparing the three-dimension data plots,
the contour data were converted into two dimension curves for analysis purposes
by creating planes relative to skeletal landmarks to intersect the contour
surfaces. In this section the methods of data analysis are discussed. Some of

the representative results are presented and discussed.

4.2 Initial Raw Data Treatment

Based on the posture descriptors deﬁned by the ASPECT program [2],
key skin landmarks on both sides of the subjects were targeted to reﬂect the
subject’s real sitting postures in the present study. The tracked data for target
location were input to the spreadsheet program [19] that has been revised for this
application to calculate the joint centers coordinates of interest. After coordinate
transformation and the contour surface data ﬁle format was changed to IGES
ﬁles, the data were imported into the IDEAS program package [20]. First, the

data points (see Figure 18) from the seat pan and seat back were splined into

49

curves, then these curves were lofted to create the three-dimensional contour

surface as seen in Figure 19.

4.2.1 Data Transformation Among Coordinate Systems

There were four coordinate systems used: 1) a global coordinate system
in the testing space set by the camera system, 2) one local coordinate system on ,
the seat pan, 3) one local coordinate system located on the seat back, and 4) a
local coordinate system on the contour chair frame. The systems 2, 3 and 4 are
as shown in the Figure 12. All the back, buttocks and posterior thigh data from
the local coordinate systems on the seat back and seat pan were transformed

into the global coordinate system [21].

50

Figure 18 Transformed Data Points

51

 

Figure 19 Lofted Data Surface

52

4.2.1.1 Translation

 

 

 

 

x1
Data point [U] = x2 in coordinate system Ox1x2x3 was translated to 6
x3
x,’
new coordinate system O’xl’xz’x3’ as [U'] = x2 . See F igue 20. The distance
x3’
a
between the two origins O and O’ are [D]: b , and the following relation will
c
hold:
x1 a
U'=U+D= x2 + b (3)
x3 c
X3
X3
X2
D 0'
X1
X2
0

X1

Figure 20 From Coordinate System Ox1x2x3 Translated to Ox1’x2’x3’

53

4.2.1.2 Rotation
Suppose that two rectangular Cartesian coordinate systems ox1x2x3 and
0x1 ’xz’xa’ originate at the same point 0, as shown in Figure 21, and the angle
between unprimed and primed axes is deﬁned by
Ail=(Xi'. Xi) i.i=1.2.3 (4)
and the direction cosines 3]] are given by
at]: cos(xi’, Xj) i,j=1,2,3 (5)

so the transformation matrix [ajj] is

an 312 813
[an]: 321 a22 a23 iri=1v2r3 (6)
a31 a32 333

x1
Write translated position vector U’ as [U'] = x2 and the rotated same point U as
x3'
x1
[0]: x2 , the following relation will hold,
X3

[U]: [aili' [U] i,j=1,2,3 (7)

54

x3'

 

 

 

Figure 21 Cartesian Coordinate System Ox1x2x3 Rotated to Ox1’x2’x3’

4.2.2 Generating Contour Surface from Measurement Data Points

( A contour surface was generated from the measurement data points of a
subject after the raw data were initially treated, including ﬁle format change,
coordinate system transformation, etc. The l-DEAS program was employed in
the surface generation. First, the primary curves were in the lateral direction, as
in Figure 19. The measurement points in the lateral direction were B-splined into
curves. These primary curves describe the major features of the contour [14].
The secondary direction in surface generation was engaged in lofting the primary

curves into a contour surface, as shown in Figure 1b

55

4.3 Seated Posture Calculations

One of the goals of this study was to represent subjects’ seated surface
contours relative to their bone structures. To accomplish this goal, the subjects’
sitting postures were estimated by using both targeting landmarks and available
scaling data [2]. Once the locations of target centers were recorded, actual bony
landmark locations could be calculated by constructing a vector from the center
of the target to the location of the actual bony landmark on the skin surface. The

length of this vector LT..>L was as follow:
1
LT->L = ‘Z’Dt + Tb + Ti (8)

Where 0:: the diameter of the target,
Tb: the thickness of the target backing,

R: the tissue thickness at the targeted location.

To determine the direction of the vector from the target center to a joint
center, another adjacent landmark target was chosen to create a line between
the two target centers. Usually the direction of the vector would be toward the
joint under the skin landmark and perpendicular to the line. For a second method
to determine a joint center, three landmark targets were needed to yield a plane
with three points. The direction of the vector was chosen to be perpendicular to

the plane, and pointing to the joint by the landmark.

56

4.3.1 Lower Neck Joint

'The location of the lower neck joint was determined using the C7 and
suprastemal surface landmarks. See Figure 22. The vector R from C7 to
suprastemal was rotated counter clockwise 8° and became R’. The lower neck

joint was the point on vector R’, away from C7 at 55 percent of the R length.

R’=0.55R

C7 /

Lower Neck Joint

    

Suprasternal

Figure 22 Calculation of Lower Neck Joint

57

C7

  
  

Acromion
Suprasternal

/
0.42R

 

Shoulder Joint

Figure 23 Calculation of Shoulder Joint

4.3.2 Shoulder Joints

The surface landmarks C7, suprastemal, and acromion were used to
calculate the shoulder joint. The Y coordinate of the shoulder is the same as that
of the acromian landmark. Through the acromion, a vector A was created to
form an angle of 67° with the vector R, as shown in Figure 23. The shoulder joint
is located on the vector A at P, which is 42 percent of the length of C7 to
suprastemal from the acromian. From the position of this point, the joint
coordinate X and Z could be decided. Because the ASPECT posture was
deﬁned as sagittal symmetry, another shoulder joint could be reﬂected in the 3-D

space.

58

4.3.3 Upper Lumbar Joint

The upper lumbar joint was calculated from the T8 and T12 positions, as
shown in Figure 24. The vector T8-T12 was rotated counter-clockwise 94° to
form vector UL, and the upper lumbar joint was located on the vector UL at 52

percent of the length of C7 to suprastemal from the T12.

T8

94°

E Upper Lumbar Joint

T12

Figure 24 Calculation of Upper Lumbar Joint

59

4.3.4 Lower Lumbar and Hip Joints

The right and left anterior-superior iliac spine (ASIS) landmarks, the pubic
symphysis (PS) landmark, and the posterior-superior iliac spine landmark (PSIS)
were used to calculate the lower lumbarjoint and hip joint centers. A pelvis
coordinate system was created to calculate the lower lumbar joint and hip joint
centers, as shown in Figure 25. The Y-axis passes through both right and left
ASIS on bone. The Z-axis was perpendicular to Y-axis and passes through the
pubic symphysis and the mid-point of right and left ASIS. The X-axis was
mutually perpendicular to the Y and Z-axes.

To locate the hip joint center (HJC) in the pelvis coordinate system, the
method exploited by Seidel, et al [22], was adopted in this study. The HJC
coordinates were scaled from subjects’ pelvic width (PW), pelvic height (PH), and
pelvic depth (PD) as tabulated in the Table 4. The computed HJC coordinates

were then transformed into the global coordinate system for further analysis.

Table 4 Calculation of Hip Joint Center

 

 

 

 

Coordinate Measurement Calculation
X Pelvic depth PD 0.34*PD
Y Pelvic width PW i 0.36*PW
Z Pelvic height PH - 0.79*PH

 

 

 

 

 

6O

Mmmumrrrv '1")V"r

. 3.“.1 can t

.1"ng-

 

Figure 25 Pelvis Coordinate System

61

The method developed by Reynolds, et al [23] was used to locate the

lower lumbarjoint center in the bone pelvis coordinate system [2] as illustrated in

 

 

 

 

 

 

 

 

Table 5.
Table 5 Calculation of Lower Lumbar Joint
Coordinate Measurement Calculation
X Pelvic depth PD 0.399*PD
Y Pelvic width PW 0
Z Pelvic height PH 0.432*PH
4.3.5 Knee Joints

To determine the knee joint locations, the landmarks at the lateral

maleolus and lateral condyle along with the calculated hip joint location were

utilized to create a plane. At the measured lateral femoral condyle landmark, a

vector was constructed perpendicular to this plane. The knee joint location was

on the vector medial to the landmark by a distance of 11.8 percent of the

measured length between lateral maleolus and the lateral femoral condyle

landmarks. See Figure 26.

62

 

4.3.6 Ankle Joints

The ankle joint could be calculated from Figure 26. At the lateral maleolus
landmark, a vector was created perpendicular to the described plane. The ankle
joint was located medial to the landmark by a distance equal to 8.5 percent of the

distance between the lateral maleolus and the lateral femoral condyle landmarks.

  
 
 
 

Lateral Femoral Condyle

  

0.118 P
Knee joint

  
  

Lateral Malleolus
Hip Joint Medial

Ankle joint 0.085 P

Figure 26 Calculation of Lower Limbs Joints

4.3.7 Wrist Joints
For this application, the wrist landmarks were adequate for accuracy. The
measured wrist point was just a mid skin point between the radial and ulnar

styloid process.

63

4.3.8 Elbow Joints

In Figure 27, the landmark on the lateral humeral condyle, mid wrist, and
calculated shoulderjoint were used to yield a plane for calculating the elbow
joint. A vector was created passing through the lateral humeral condyle
landmark, perpendicular to the plane mentioned above. The elbow joint was
located on the vector medial to the lateral humeral condyle a distance at 15.5

percent of the length between the lateral humeral condyle and wrist landmarks.

Shoulreroint

 
 
 
 

Wrist Landmark

Medial

  
  
  

Lateral
Elbow Joint

0.1550

Figure 27 Calculation of Upper Limbs Joints

64

4.4 Deformed Contours Data Analysis

4.4.1 Methods
To compare different subjects’ body shapes, the three-dimensional data

were converted to two-dimensional curves as shown in Figure 28.

 

. “ 1'
\"x 3'
.13
if.

, «x.

 

Figure 28 Sliced Thigh Curves (Relative to Left Femur Axis) and Back Curves

65

4.4.1.1 Analyzing the Thigh and Buttocks Data Surface

After examining the left and right sides of the measured data, it was found
that the differences between both sides were within the resolution range of the
contour chair, which is 5 mm. So a decision was made to section the left thigh.
A local coordinate system was created with its origin at left HJC and x-axis
through the knee joint center after the transform of the HJC together and
alignment of the left thigh axes for this purpose. Although the subjects were from
the same mid-size male subgroup, each of their body parts was not exactly
proportional to their stature. Therefore, the thigh axis was deﬁned as from HJC
to knee joint center on the same side of the body, and evenly divided into 10
sections for the comparison. Planes that were perpendicular to the left thigh axis
started from the left HJC towards knee joint center were created at ten percent of
their thigh length (HJC to knee joint) apart from each other to intersect the three-
dimensional contour surface. See Figure 29. Because of the limitation of the
length of the seat pan of the contour chair, there was seven sections on each
subjects’ data, and the seventh usually is not long enough to provide all the
information needed.

The other way to section the contour surface is by using the vertical plane
through the thigh axis to intersect the contour surface to investigate the shape
change in the longitudinal direction of the thigh.

The same procedure was applied to the data of the J826 H-Point machine.

66

Lower Neck Joint

:l
I

- WW5!!!" n _. I
“ ~‘“‘-.-- ﬂu.) E:
,~ - 988.8921: . ..
\ Upper LumbarJoint gi‘l‘.’
. . . I I
K . try-*A x9 ,
. ‘t'lrrlil’s‘sgt'rw
- ' . . ‘gIIV
Lower LumbarJornt (,K

~~€~'~\
‘ Sectioning planes

. _ _ I (I!
*4/ch 1, It,»
.317.» were perpendicular
to the linkages
f H-point \\
/
\

W‘s“

 

 

Figure 29 Contour Surface Analyzing Scheme

67

4.4.1.2 Analyzing the Back Data Surfaces

The anatomical aspects of the back and its motion are much more
complicated than the thigh. Future applications of the processed back contour
data for input into JOHN program were considered. Sectioning planes were
created to intersect the back contour surfaces that were perpendicular to the joint
linkages, such as the linkage between lower neck joint and upper lumbar joint of
the thorax, or the linkage between upper lumbar joint and lower lumbar joint of
the lumbar region. In the pelvic region, the linkage of the hip joint center (the
mid-point of right and left HJC and lower lumbar was used as the intersection
planes were created. The intersection planes were spaced at a distance of ten
percent of the linkage length as seen in Figure 29. Due to the differences in
linkage orientations caused by the different postures, using this sectioning
method sometimes left a gap between two adjacent regions as can be seen in
Figure 28. According to the thoracic and pelvis shape and their relative position
to the joint centers, the linkages were extended and a few more planes were
created to intersect the back surface to ﬁll the gaps in such a case. Meanwhile
these planes also intersected the skeletal links in the torso and lumbar to create

the points of intersection with the torso and lumbar links.

68

4.4.2 Comparing Measured Data and J826 H-point Machine
Based on the measurement data of nine mid-male subjects, a comparison

between the J826 machine and measurement data was conducted.

4.4.2.1 The Thigh and Buttocks Data

At each thigh section, there were typically nine curves (from nine mid-
sized male subjects) together with the intersection points of the thigh axis to be
analyzed. In order to compare the thigh shape relative to the skeletal linkage, the
lowest points on the curves to the left buttocks and thigh were determined by
creating straight horizontal lines tangent to the curves at the lowest 2 position in
the coordinate Y-Z plane in the thigh coordinates where X was along the thigh
axis. One of the nine lowest points was chosen as a reference point (usually one
near the center of the distribution). Then all the curves along with the
corresponding intersection points on the thigh axes were translated to coincide
with all the lowest points on the curves. From here the relative positions of the
intersection points on the thigh axes was calculated and further statistic analysis
performed. .

These sectioning and analysis methods were employed to compare the
measured data of the mid-sized subjects and J826 H-Point machine. The
following are the brief results of that part of the study as an example here.

In Figure 30, (A) are the sections under the hip joint center (HJC). The H-
point of J826 H-point machine and the average HJC of human subjects were

positioned on the same 2 level. There is 5 mm difference between the lowest

69

points on the curves of J826 H-point machine and subjects, i.e., the H-point
location of J826 H-point machine is 5 mm higher than the mean location of the
subjects’ HJC. From (B) and (C) it can be seen that the subject’s thigh contours
were more symmetrical, ﬂatter and the thigh axes closer to the bottom of the
curves compared to the J826 H-point machine. Moreover, the subjects’ contours
are more prominent under the ischial tuberosities in the medial region.

To study the contours along the length of the thigh, the thigh axes and
HJC of all subjects and the J826 H-point machine were overlaid. After the lowest
points on all curves were determined and marked, all curves were translated to
coincide with the mean lowest point and the marks left in the original positions
(Figure 31). This data veriﬁed that the H-point of J826 H-point machine was
higher than the mean HJC location of human subjects. There is a 10 mm vertical
distance between the lowest points of the J826 H-point machine and the mean
location of subject contours. The bigger difference occurred near the knees on
the curves. The main reason for this is that the subjects’ leg lengths are
different. The shorter the subjects’ lower leg, the more the thigh penetrates the
seat pan. The testing protocol controlled the horizontal distance between the H-

point and toe-bar on the contour chair.

70

Z J826 H-point

Subjects’ Average H-point

/\

 

 

 

 

More prominent under ischial tuberosities

 

(A) Sections Right Under H-point

J826 Thigh Reference Point

   

 

Subjects’ Thigh Average
Reference Point -. /

 

(B) The Fourth Sections from H-point

Figure 30 Sections of J826 H-point Machine and Subjects Contours

(Planes perpendicular to thigh axes)

71

J826 Thigh Reference Point

Z
+ Subjects’ Average Thigh
\ Y + Reference Point

More symmetrical and ﬂatter

    

(C) The Sixth Sections from H-point

Figure 30 Continued

  
 

Thigh Axes

 

Mean Lowest Point of Subject J826 Lowest Point

Figure 31 Sections of J826 H-point Machine and Subjects Contours

(Planes vertical through the thigh axes)

72

4.4.2.2 The Back Contour Data

The approach used to analyze the thigh and buttocks contours was also
applied to the back contours. The comparisons between human subject data,
the J826 H-point machine, and the UMTRI data [5] were performed.

Figure 32 shows a view from the left upper front of sections at 30 mm
intervals above the HJC of the average data from this study for the ASPECT
program (so named as ASPECT contour data) and the UMTRI back contours
with the limits of seat contact marked for each contour. The limit of seat contact
for the UMTRI contours was selected at the divergence of the UMTRI shell and
seat contours measured in this study. The region of measured seat contact for
the ASPECT contour was wider than with the UMTRI back (Figure 32).

The side views of the ASPECT and SAE back contours are very close
(Figure 33). Compared to the ASPECT back contour, the side view of the
UMTRI crash dummy seat-shell contour was convex (Figure 33).

Compared to the average posture of the ASPECT subjects, the angle from
UL to LN in UMTRI posture was 6 degrees forward. The UMTRI torso seat-shell
shape was rotated rearward about its UL joint to provide a composite back
contour such that the original UMTRI contour was rearmost below the UL and the
rotated UMTRI contour was rearmost above the UL. The side view of this
adjusted UMTRI back contour was very close (within a few millimeters) to the

ASPECT and the J826 back contours (Figure 34).

73

    
 

ASPECT seat back (red
ASPECT contact border gr -
ASPECT back (green) 16. 480mm
15, 450mm
’ 14,420mm
UMTRI dummy shell (blu = ,3. 390mm
UMTRI contact border (blue) ‘ 12. 360mm
‘ - 11, 330mm
-_ \l' 10. 300mm
ASPECT average posture (red W,
UMTRI dummy posture (blu - ‘\—z “. 9. 270mm
“.___ . ' 8. 240mm
1“ 7, 210mm
’ l 6, 180mm
’ i . 5. 150mm
/7 I 4. 120mm
/ \ 3. 90mm
i 2, 60mm
0. at HJC

Figure 32 Sections Through the ASPECT and UMTRI Contours with Postures
and Contact Edges

V
A

‘x
l . "1|
ASPECT average posture (green)

   

rJ '
x, R. .
t]. "l. \5'1'1 ix
- . 'l J826 (red), offset rearward by
. 5 'l
- ~. 5% i difference between H-pt. and HJC
I specr(
green)
UMTRI posture I rm
8. Rotated back 6° to be parallel i| UMTRI
‘0 ASPECT thorax (p'nk) i a. Parallel to ASPECT thorax
b. Original position (blue) I (pink)
I

 

b. Original position (blue)

 

 

 

 

Figure 33 Measured Back Contours

Between Upper Lumbar and Lower Neck Joint

75

   
 

 

ASPECT measured (green) UMTRI dummy (gold)
ASPECT splined (black) J826(pink)

l H '\ i +

.\ \ ‘2‘ . "a
"\ \ \
El E

{E

5 l
lg: . i

I ‘ ‘ . ‘ __ > J' . ._ .. I. _.

J

Section 2, 60mm above HJC 3. 90mm 4, 120mm 5, 150mm 6, 180mm
100mm

 

ASPECT measured (green)

UMTRI dummy (gold and blue - rotated 6° back)
ASPECT splined (black)

J826(pink \
+ + , 4'.

          

u
\
\t
(l
t
I
‘.
I
l
\.
l

mﬁ/I'c— I,”

"#0.“—
..-—-u M— .- -

‘2

 

 

f ,
Cl ‘1- ’ :-
. i i 1 ‘i :1“. E
I I I . I l.‘ ‘5‘
47' h J" I W .1 _ ‘ 45
Section 7, 210mm above HJC 8, 240mm 9, 270mm 10, 300mm

 

Figure 34 Contour Sections up the Seat Back

76

 

ASPECT measured (green) UMTRI dummy (blue - rotated 6° back)
/ ASPECT MW) \\ / J826(plnk)
x +

  

.__.__,+

 

 

I
I
I

 

i ,5 4L
. " i 17, 510mm
. +’ 4 ' g
Section 12. 360mm above HJC 13. 390mm 14, 420mm 15. 450mm 16. 480mm 100mm
Figure 34 Continued

With the side view back contours aligned as above, X-Y (transverse)

sections of the ASPECT back contour were compared to sections of the UMTRI

composite torso. A series of sections at 30 mm intervals above the H-point or

HJC through the ASPECT, UMTRI, and SAE back contours are shown in Figure

34. Section 2 is the ﬁrst section shown and it is 60 mm above HJC. For Sections

3 - 7 and 14 — 16, the sides of the UMTRI sections generally appeared to be

continuous with lateral parts of the ASPECT contours near the 50 mm X-direction

limit of the ASPECT back measurements. In the region of the shoulder on the

77

sides of Sections 8 — 13, the UMTRI contours were fonrvard of the ASPECT
contours probably because of difference in arm placement. For the UMTRI
contours the hands were on a steering wheel while the ASPECT contours were
measured with the arms closer to the torso. The middle part of the ASPECT
measured contours (within approximately Y = :90 mm) were nearly ﬂat because
of the ﬂat seat back of the measurement chair, while the UMTRI contours were
more rounded (Figure 34).

In the sections of Figures 34 and 35, the sides of the J826 back contour
become apparent in section 4, which is 120 mm above H-point and HJC. Up to
section 8 at 240 mm above HJC, just above the upper lumbar pivot, the J826 is
more rounded than the ASPECT measured contour or the UMTRI dummy shell,
yet the widths of the J826 and UMTRI contours are similar. Above the upper
lumbar pivot, the J826 contours continue to be more rounded than the ASPECT
contours and the UMTRI contours are generally between J826 and ASPECT. As
stated above, the difference between the shape of the ASPECT and UMTRI
contours above section 10 was probably due to the fonlvard UMTRI hand position
to reach the steering wheel.

As described in the next section, the sides of the UMTRI dummy shell
torso were added to the measured ASPECT back contours to provide the
ASPECT manikin torso contour. The following is a comparison of the back
widths of the J826 torso and ASPECT torso (which uses the sides of the UMTRI

dummy shell torso).

78

The narrowest part of the J826 back is 306 mm at 230 mm above H-point.
Comparable dimensions of the UMTRI dummy shell are the width lateral to the
bottom of the 10th rib which is 312 mm at 160 mm above HJC and the abdominal
breadth of 325 mm and waist breadth at the umbilicus 314 mm. These
dimensions indicate that J826 is approximately 6 mm narrower than the new
ASPECT back.

The widest part of the J826 back is 392 mm at 430 mm above H-point.

On the UMTRI dummy shell, the width between the rear upper margins of the
armpit (posterior Scye landmark) is 394 mm wide at 370 above HJC. This is
nearly the same as the J826 widest part. The dimensional differences between
J826 and UMTRI back widths are small, and the ASPECT back is ﬂatter in cross-

section (seen below in Figure 35).

79

_ xx--. Splined ASPECT contour (black)

    
    
 
 
 
  

(green) UMTRI Rotated Back _ an”: ‘9: 171 510mm, above HJC
6° (blue) "5;:16 480mm
UMTRI (maize) 15 450mm
' 14, 420mm
13. 390mm

 

12, 360mm
11, 330mm
10, 300mm
9, 270mm
8, 240mm
7, 210mm
6, 180mm
. 5, 150mm
4, 120mm
3, 90mm

1 “x 2, 60mm above
HJC

X direction, forward

right side

Figure 35 Comparison between ASEPCT and UMTRI Back Contour

80

4.5 Defining Mid-male Contour

Even though humans are not precisely symmetrical, the posture and
contours for manikins or the computer models ought to be symmetrically deﬁned
[24]. By creating a central, mid-sagittal plane of symmetry, the contour surfaces
were separated into right and left sides. The right side and its reﬂection were
used for the development of the manikin contours. After the right side of the
contour was reﬂected about the plane of symmetry, the right and left sides were
nearly identical, with a difference of only approximately :1 mm. This small
difference is less than measurement accuracy and supports the decision to
develop symmetrical contours.

There were two versions of mid-male contours that were deﬁned in this

section, and they were ASPECT mid-male and TecMath mid-male contours.

4.5.1 Deﬁning The ASPECT Mid-male Contours

In the ASPECT program, there was a need to reevaluate and revise the
J826 hardware and procedures as stated earlier in the Chapter 1. Because the
contact contours of the new manikin were very important in determining the
interactions with seat pans and resulting measurements of H-point location and
seat pan angle, a study of the contact contours between subjects and the seat
support surface was carefully conducted, which resulted in the creation of the

new ASPECT manikin contour of mid-male size.

81

4.5.1.1 Thigh and Buttocks Contours

The thigh and buttocks surface contour of the new manikin was deﬁned by
selecting one of the mid-male subject’s contours that was consistently in the
middle of the small range of variation of the nine subjects’ contours. The contour
then was separated right from left; the right side and its reﬂection were used in
the ASPECT thigh and buttocks contours. An averaged contour would have
been similar to, yet may not be as realistic as, this single subject’s contour. The

ASPECT thigh and buttocks contours are shown in Figure 36.

 

Figure 36 ASPECT Symmetrical Buttocks and Thigh Contours

82

4.5.1.2 Back Contour

Compared to the buttocks and thigh, back posture and contours are more
complex because there are a greater number of mobile segments, leading to
greater potential for postural and contour variability.

Although the subjects were chosen to be mid-size males, there were
variations in their skeletal anthropometric dimensions and postures. To reduce
the effects of these variations and to analyze the subjects’ back contours for the
generation of the single manikin contour being developed, the following analysis

steps were accomplished:

1. The measured back contour data for each subject were splined to yield a
surface.

2. The splined back contour for each subject was rotated and translated so
that the mid-HJC was aligned on the same axis and the LN was in the
same plane.

3. A coordinate system was established (Figure 37) such that:

a. The X-axis was fonrvard and perpendicular to the seat back plane of
the measurement chair. The posture data for each subject was
positioned so that the point midway between the hip joint centers
(mid-HJC) was on the X-axis. The origin of the axis system was in
the plane of the seat back of the measurement chair.

b. The Z-axis was upward in the seat back plane. The posture data for

each subject was positioned so that the perpendicular projection

83

2, up the seat back

I
\ Lower neck joints
at Z = 507mm

—- Upper lumbar joints

— Lower lumbar joints

 

   

Hip joint centers

Figure 37 Torso Length Scaled In Z-Direction

of the lower neck joint (LN) to the seat back plane was on the Z-
axis.
c. The Y-axis was perpendicular to X and Z, in the seat back plane,

and to the left.

In this coordinate system, the mid-HJC of each subject was placed on the

X-axis and the lower neck joint (LN) was placed in the X-Z plane.

4. The Z distance from the mid-HJC to the LN was averaged (Zav = 507 mm),
and the Z values of the posture and splined contour data for each subject

were scaled to the average.

By these steps, the mid-HJC of each subject was at (X., 0, 0) where Xi was
X-coordinate value of each subject’s HJC. The LN of each subject was (xi, 0,
23‘.) where X,- was the X-coordinate value of each subject’s LN and Zav is the
average Z coordinate value for all the subjects. The Y coordinate values were
not altered but were within a few millimeters of zero.

The X values of the HJC and LN were averaged and the X, Y, and Z
values of the LL and UL were averaged to yield an average torso posture for the
back contour subjects.

The lower backs of the subjects’ were all against the ﬂat seat back of the
measurement chair so that their measured contours in this region were very

similar (Figure 38). In the upper torso (Figure 39), the contours varied because

85

the upper torso postures of the subjects varied primarily due to different angles of
their thorax from vertical.

The depth of ASPECT back contour measurements was limited to 50 mm
in the X direction perpendicular from the ﬂat seat back. The measured and
splined seat back contours for each subject were sectioned with X-Y (transverse)
planes that were 30 mm apart in the Z-direction starting 60 mm above HJC,
which was near the bottom of the measured back data. At seventeen points
(sixteen segments) equally spaced along each contour section, the X and Y
values of the contours for each section were averaged to generate an average
back contour. There were very small differences within two millimeters in the Y
values between subjects, and these were averaged by this procedure. The
average contour sections are shown in Figures 40 and 41. The ASPECT back,
buttocks, and thigh contours are shown in Figure 42. ASPECT manikin is

illustrated in Figure 43.

86

Z, up the seat
back

Middle and

lower torso
contours are very
similar between
subjects.

 

Figure 38 Measured Contour Curves
from Mid Thorax to End of Lumbar

87

  
 
 

. \~i‘;\ \Ievel lower neck joint

   

Measured back contours differ
probably due to differences in
upper torso postures shown in

Figure 41. f
right side ._
\ \ \ a) -" left side
\ \ ‘>\
4]
._
X direction. fonrvard ’ \ ’ level upper
, , lumbar jomt

Figure 39 Measured Back Contours between Upper Lumbar Joint
And Lower Neck Joint

88

A Z
17, 507mm
16, 480mm
15, 450mm
14,420mm
13, 390mm
12, 360mm
1, 330mm
—— 0, 300mm
9, 270mm
8, 240mm
7, 210mm
6, 180mm
5, 150mm
4, 120mm
3, 90mm
2, 60mm
1
0

Lower neck joint

 

 

 

   
 

 

upper lumbar joint

 

 

 

 

 

lower lumbarjoint

 

 

 

 

 

hip joint center

Figure 40 Average Back Contour And Posture Side Vlew

89

 

left side right side

Figure 41 Average Back Contour Left Rear View

90

 

Figure 42 ASPECT Back, Buttocks and Thigh Contour Surfaces

91

 

Figure 43 ASPECT Manikin

92

To deﬁne the edge of contact between the seat and body, points on the
contour sections were chosen that were the ﬁrst points from the sides with an X
value that was clearly less the maximum X value for that section. The X-value of
the chosen point was typically near 40 mm. Actual contact was between this
chosen point and the next point to the side, which was at the maximum
measurable X value (near 50 mm). The contours lateral to these chosen points
would be affected by inclusion of some points that were not in contact with the
subject. However, contours splined to the edge of known contact were found to
be less than 1 mm different in X-value when compared to contours splined with
points beyond the contact edge. The contact edge for the average ASPECT
back contours is shown in Figure 32 with the edge of seat back contact for the

UMTRI dummy shells.

4.5.2 Deﬁning The RAMSIS Mid-male Contours

Boughner [7] developed a three-dimensional skeleton that represented
mid-size males and was based on the UMTRI dummy anthropometrics [5] and
articulated according to the work of Haas [4]. To his skeletal model, Boughner
added shapes to represent major muscle groups. To represent the mid-size
male, Frost [10] added skin surfaces to the Boughner model so the skin surfaces
articulated with the skeletal linkage. In developing this contour, Frost considered
both the skeletal and muscle geometries of Boughner [7] and the surface
contours of UMTRI [5]. So, the Frost contours included transverse section

curvatures that were believed to be more representative of people’s back than

93

the ﬂatter UMTRI contour that was based on the shape of an automobile seat in
contact with their human subjects. Frost also scaled the mid-male contour based
on anthropometric dimensions to represent small women and large men.
Because the ASPECT contour of this study was based on data measured
in a ﬂat-backed seat ﬁxture, the resulting contours appeared to be ﬂat. For
aesthetics in the development in this study of the contours for the RAMSIS model
[6], the ASPECT contours were adjusted by a few millimeters, less than
measurement variability. For this, the contours from Frost [10] and UMTRI [5]
and the APSECT contours from this study were Considered. The side view
(sagittal plane) shape of the Frost and UMTRI contours were nearly identical but
the Frost contour was thought to have more human-like transverse contours, so
the Frost contour was selected. To develop the RAMSIS contour, the postural
linkages for the Frost and ASPECT contours were aligned and the differences

between these contours were averaged to produce the RAMSIS contour.

94

CHAPTER 5 CONTOUR SCALING FOR LARGE MALE AND SMALL FEMALE

5.1 Introduction

Since the early 1960s, both physical and digital models of mid-sized
males, such as SAE practice J826 [1] and the JOHN model [7], have been
extensively used in seat design and evaluation. However, there are needs for
tools representing other groups of people in different height and weight ranges.
Among those most needed are models of 5th percentile small female and 95th
percentile large male for applications to conﬁrm that seats will accommodate
these extreme sizes of people. Because of the extensive and prevailing
application of the mid-sized male models, it became very meaningful to explore
the correlations or the relations between mid-male and small female, and mid-
male and large male. Frost [10] and Robbins [5] used the scaling method to
create contour models for small females and large males from the mid-male data,
but the scaling method was never validated by data from human subjects.
Therefore in this study, mid-male contours from experimental data were
employed to create contours for small females and large males by scaling.
Furthermore, the measured data of large males and small females were utilized
to validate the scaling methodology.

If a feasible scaling system could be established, then the contours of any
percentiles could be created from the mid-sized male contours by using the
scaling method. This would be a very efficient way to create a contour for any

size individual.

95

5.2 Development of Scaling Factors

The seat contact contours of three groups of human subjects (small
women, mid-size men, and large men) were measured in this study along with
anthropometric dimensions for these three size groups. As described previously,
a seat contact contour was developed to represent mid-size men and was used
in the development of the ASPECT manikin and for RAMSIS model development
[6].

In computer modeling, it is easily possible to scale the size of shapes so
that seat contact contours for different size people could be generated based on
their sizes. In this section, scaling factors are described for the purpose of
scaling the mid-size male contours to the size of the contours for small women
and large men. These scaling factors were based on values of anthropometric
dimensions measured on the surfaces of the human subjects and internal
skeletal linkage dimensions that were based on the external anthropometric
dimensions. The approach to evaluating agreement between the contours
scaled from the mid-size male to the other size people was to compare the
scaled contour for each individual subject to the measured contour for that
subject.

Scaling factor F in one speciﬁc body direction was deﬁned as the
individual subject’s body dimension D, divided by mid-sized male body
corresponding dimension Dm, i.e.,

D8
F = a” , (9)

m

96

There were three scaling factors for each body segment corresponding to
scaling in length, depth, and breadth directions, for instance, F., F... and Fb,

respectively.

5.2.1 Thoracic
In the thorax area, three scaling factors are deﬁned for body dimensions in

breadth as Fm, in depth as Ftd, and in length as F".

 

n=Dm um
Dmtb
D8

Feb—0m: (11)
0

F =—s" 12

u D"... ( )

Where Dab, Dstd, and Dst. are subject’s chest breadth (reference number 32 in
[14]), chest depth (36), and chest length, respectively. The chest length here is
the average distance between the lower neck joint and upper lumbarjoint. Dmtb,
Dmtd, and Dmu are the chest breadth, chest depth, and distance between the

lower neck and upper lumbar joint of the mid-male contour, respectively.

5.2.2 Lumbar
To scale the mid-sized male contour to the desired subject’s contour in

this area, scaling factors Flb, Fld, and F.. are used for the lumbar area scaling.

97

 

F = 13

l. D... < )
0

F = w 14

.. D... ( >
D

F-uJi 15

.. D ( )

mll

In equations 13 to 15, 05",, Dsld, and D," stand for subject’s waist breadth (112),
waist depth (115), and length between the upper and lower lumbar joints, Dmlb,
Dmtd, and Dmu for the waist breadth (112), waist depth (115), and the distance

between the upper and lower lumbarjoints of the mid-male contour in that order.

5.2.3 Pelvis

In the same fashion as for the thorax and lumbar areas, three scaling

factors are deﬁned for the pelvic area as F», de, and Fpl.

 

 

 

o
3,: ”° 06)
Dmpb
o
l=',,,,=spd 07)
0m,d
F —D” 08)
PI — Dmpl

Where 0,», Dspd, and Dsp. are used for subject’s sitting hip breadth (66), buttocks
depth (24), and the distance between the lower lumbar joint and the midpoint of
HJCs, and Dmpb, Dmpd, and Dmp. for the sitting hip breadth (66), buttocks depth
(24), and distance between the lower lumbar joint and midpoint of HJCs of the

mid-male contour, correspondingly.

98

5.2.4 Thigh
For the thighs, scaling factors Fm, Fm, and Fm are deﬁned for the thigh

breadth, depth, and length in the following way:

 

D

F . = —3°” 19
D5

Fthid = 6:: (20)
D .

F . = stl'lll 21

mthil

In the equations 17-19, Dam. is the average distance between the HJCs and knee
joints of a subject, Dmthi. is the distance between the HJC and knee joint of the
mid-male contour. For the thigh region, the pelvis’s breadth and depth are

utilized to scale the thigh shape for both the subject and the RAMSIS mid-male.

5.3 Scaling Procedures and Results

Once the scaling factors for each body segment had been deﬁned, the
values of the scaling factors were calculated and are tabulated in Table 12 for
large males and in Table 13 for small females in Appendix A.

Before the scaling procedures started, there were four local coordinate
systems to be deﬁned.

1) Thorax coordinate system was originated at upper lumbar joint, positive X-
axis pointing to lower neck joint, positive Y-axis from left to right. The Z-

axis was determined by the existing axes by the right-hand rule.

99

2) Lumbar coordinate system started with the X-axis from lower lumbarjoint
toward upper lumbar joint, Y-axis pointing to right from left, and the X and

Y directions decided the Z-axis by the right-hand rule.

3) Pelvic coordinate system origin was located midway between the two hip
joint centers with X-axis pointing to lower lumbar joint, Y-axis pointing from

left to right, and the Z-axis was then decided by these two X and Y axes.

4) Thigh coordinate system was originated at the mid point between right and
left hip joints. Positive X-axis pointing toward the midway of the two knee
joint centers, positive Y-axis pointing from left to right, and Z-axis was

decided by the X- and Y-axes.

5.3.1 Small Female

In the scaling process, the RAMSIS mid-sized male contour and posture
were scaled down to each small female subject’s contours.

First the thorax contour of the RAMSIS mid-male was translated to
coincide with the upper lumbar joint for the subject. Then the mid-male thorax
contour and the skeletal linkage were scaled down to the small subject’s contour

in the local thorax coordinate system using the previously deﬁned scaling factors.

100

Finally the scaled body segment and the posture linkage were aligned with the
subject’s linkage connecting lower neck joint and upper lumbarjoint.

The lumbar contour of the RAMSIS mid-male contour was repositioned at
the lower lumbar joint of the small female subject. In the local lumbar coordinate
system, the lumbar contour of the mid-male then was scaled down according to
the small female subject’s anthropometric data. The scaled lumbar and the
posture were then rotated to the subject’s posture linkage.

The pelvic contour and the posture linkage of the RAMSIS mid-male was
relocated at the center of the right and left H-points. The pelvic contour and
posture were scaled down to the small female subject’s dimension. The contour
and linkage were also rotated to align to the small subject’s.

All of the scaled contour curves from small female subject 1 are presented
together with the curves from measured contours for that subject in Figure 44 for
a back view and in Figure 45 for an oblique view. The curve at each individual
section is illustrated in Figure 46 in detail. By visual inspection, the differences
between the two sets of contour curves are within less than 10 mm. Considering
the accuracy of contour and skeletal landmark measurement, the two sets of
curves match with each other to an acceptable degree. The rest of the small
female subjects’ results are documented in the Appendix B. The similarity of

these curves is analyzed and discussed in section 5.4 below.

101

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Subject 1

 

Figure 44 Tested and Scaled Subject 1’s Contour Curves, rear view

102

caled

 

ETA/UPS

Figure 45 Tested and Scaled Subject 1’s Contour Curves, oblique view

103

_—

SF1 Thorax 2"d

SF1 Thorax 3rd

SF 1 Thorax 4"

Figure 46 Tested and Scaled Contour Curves at Each Section of Small Female

Subject 1

104

 

'\ "

‘ “—— ~.——-—— _———-'

SF1 Thorax 5m

SF1 Thorax 6th

Figure 46 Continued

105

 

SF1 Lumbar 18t

SF1 Lumbar 2"‘1

SF1 Lumbar 3rd

Figure 46 Continued

106

SF1 Lumbar 4th

I
\
W

—-— ——_ _‘_—

SF1 Lumbar 5m

 

 

SF1 Lumbar 6"1

Figure 46 Continued

107

SF1 Pelvic 15‘

.I ‘ _ . _.."

SF‘IPericZ"d

 
   
 

- ~ —_ —.—.—.—4’ -

SF1Pe|vic4th

Figure 46 Continued

108

 

SF1 Pelvic 5th

 

SF1 Pelvic 6th

Figure 46 Continued

109

SF1 Thigh 1st

 

SF1 Thigh 2"d

 

sr=1 Thigh 3rd

Figure 46 Continued

110

 

SF1 Thigh 4“"

 

SF1 Thigh 5th

 

SF1 Thigh 6'“

Figure 46 Continued

111

l
\.

SF1 Thigh 7th
E .
\ _g__ ° //
*Qézfﬁff
SF1 Thigh 8”1
Figure 46 Continued

112

5.3.2 Large Male

In the scaling process, the RAMSIS mid-sized male contour and posture
were scaled to each large male subject’s contours.

First the thorax contour of the RAMSIS mid-male was translated to
coincide with the upper lumbar joint of the subject. A local thorax coordinate
system was created at upper lumbar joint, X-axis pointing from upper lumbar joint
towards lower neck joint and Y axis in lateral direction. Then the mid-male thorax
contour was scaled up to the large male subject’s contour in the local thorax
coordinate system, using the previously deﬁned scaling factors. After this the
scaled contour and posture were then aligned with the subject’s thoracic skeletal
axis system.

The lumbar contour of the RAMSIS mid-male contour was repositioned at
the lower lumbarjoint of the large male subject. In the local lumbar coordinate
system, which was originated at lower lumbar joint and with X axis pointed from
lower to upper lumbar joint center and Y axis goes laterally. The lumbar contour
of the mid-male was then scaled up according to the large male subject’s
anthropometric data. The scaled lumbar and the posture were rotated to line up
with the subject’s posture linkage.

The pelvic contour of the RAMSIS mid-male was relocated at the center of
the right and left H-points. Next, the pelvic contour was scaled up to the large
male subject’s dimension in the local pelvic coordinate system. The mid-point of
the right and left H-points was the center of the local pelvic coordinate system,

and its X-axis was in the direction of the mid point to lower lumbar joint, Y in the

113

lateral direction. Then the scaled pelvic contour and the posture were brought
into the line of subject’s posture linkage.

The rest of the other large males’ results are documented in the Appendix

114

 

 

\

__// : } Tested
:F ,. W
(I

 

 

 

 

 

 

 

f¥
‘ § Scaled

V

E
L .)

 

 

 

 

 

 

 

 

 

 

 

 

ETD/UPS

Figure 47 Measured and Scaled Subject 1’s Contour

115

Scaled

.417
7' , 4414111114,] 1' '

TeSted / 'Wgw

 

2

l<:

ETA/UPS

Figure 48 Measured and Scaled Subject 1’s Contour

116

 

LM1 Thorax 1”

\ /
\ /

wh

‘_’_——~_—-—‘§__’

LM1 Thorax 2"°

 

LM1 Thorax 3”

Figure 49 Tested and Scaled Contour Curves at Each Section of Large Male

Subject 1
117

 

LM1 Thorax 4‘"

LM1 Thorax 5m

LM1 Thorax 6"1

Figure 49 Continued

118

 

LM1 Lumbar 1“

 

LM1 Lumbar 2""

I

\
/.-——-
\

 

LM1 Lumbar 3'"

Figure 49 Continued

119

LM1 Lumbar 4'"

 

LM1 Lumbar 5m

 

LM1 Lumbar 6‘"

Figure 49 Continued

120

\\ //
LM1 Pelvis 1”
l l
I l
l l
\ /
\ /
\ /
\\ //
\\ \\ ———————— ////
LM1 Pelvis 2"“

 

LM1 Pelvis 3'“

Figure 49 Continued

121

LM1 Pelvis 4m
l I
I I
\ /
\~ /
\ /
\\\\\ ///
W
LM1 Pelvis 5m

l I
I I
\ /

\ /

LM1 Pelvis 6m

Figure 49 Continued

122

\
‘~—~._—.——-ﬂ-’

LM1 Thigh 1"

LM1 Thigh 2'“I

LM1 Thigh 3'“

Figure 49 Continued

123

LM1 Thigh 4m

LM1 Thigh 5‘"

i ' ‘ . I
\ /

a”-\\' / ”

~\_/

LM1 Thigh 6tn

Figure 49 Continued

124

 

LM1 Thigh 7‘n

Figure 49 Continued

125

5.4 Correlation Comparison Between Measured and Scaled Contours

When comparing each pair of curves of measured and scaled contours,
the scaled curves typically matched to the measured contours within less than 15
mm. A further numerical correlation analysis [25] was performed to assess the
quantitative resemblance between the scaled and the tested contour curves at

selected sections of the contours.

5.4.1 Methods

Three aspects of the geometric resemblance between the scaled and
tested curves of the contours were considered. They were differences between
scaled and tested curves, shape resemblances, and corridors with one and two
standard deviations [26, 27]. The differences showed how far apart the curves
were from each other numerically and graphically. The coefﬁcient of the
correlation was employed in the analysis of the shape resemblance as deﬁned in
the following section. The corridors with standard deviations described how
closely the scaled contour curves were distributed in the space.

The scaled and tested curves have respective magnitudes X3, and XM.
So the coefﬁcient of the correlation between the scaled and tested curves was SR

deﬁned as follows [26, 27]:

]:x,dex

R = b b (22)
Jjaxgdx L xadx

 

 

or,

126

_ ZXSXMox

R ’ ,5; 73px: xgax

(23)

 

 

If SR =1, the curves with magnitudes of X3 and XM had identical shapes.
The values of SR were between 1 and 0. When SR = 0, it means that the two

compared are not related to each other.

For simplicity of comparisons and reduction of the amount of
computational work, curves from sections 3, 4, and 6 in the thoracic region, 2, 4,
and 6 in the lumbar region, and 4, 5, and 6 in the pelvis region, and 1, 4, and 7 in
the thigh region were quantitatively analyzed. The procedures (MatLab program)

[28] were used to perform the correlation analysis.
5.4.2 Correlation of Large Male and Small Female Contours

5.4.2.1 Differences

The ﬁrst comparison was to ﬁnd out the difference, d, in the X direction
between tested and scaled contoUr curves. Figure 50 and Figure 51 show typical
plots of the difference (cl) in the midsections of thorax, lumbar, pelvis, and thigh
for large males and small females. In the measured region of the body contour,
the contour data were linearly interpolated every millimeter apart to provide a
continuous curve. Then, at the same corresponding locations, the differences
between points on the measured contours and the scaled contours were found

and plotted. Almost all the differences (d) were within 10 to 12 mm. The rest of

127

the difference results are in the Table 6 for large males and Table 8 for small
females, and the standard deviations of the differences can be found in the Table

7 and Table 9 for large males and small females, respectively.

128

 

20 l l l l T l l

15*

10-

 

  

Difference (mm)
0

'5 + subj.1
+ sub]. 2
-10 _ I . + subj. 3 —

+ subj. 4 *
-¢— subj. 5
-15 i subj. 6 -

‘ > + subj. 7

 

 

 

 

 

 

l l l

50 1 00 1 50 200

 

_2 l 1 l
300 -150 -100 -50 0
Y (mm)

(A) At Mid Thorax — Large Males

Figure 50 Typical Difference Curves between Scaled and Measured Contour

at Mid sections of Large Males

129

Difference (mm)

Lumbar
20 l l l l I l l

 

   
 

 

o~ -0- subj. 1
+ subj. 2
+ subj. 3
-5 “ + subj. 4
+ subj. 5
+ subj. 6
'10 ” + subj. 7

 

 

 

 

 

 

'ggoo -150 -100 -50 o 50 100 150 200

(B) At Mid Lumbar — Large Males

Figure 50 Continued

130

Difference (mm)

Pelvis
20 I l l l l l l

 

15- 1 —

 

 

10 _ + subj. 2 g
' j ->— subj. 3

‘ —a— sub]. 4
_ - ' + subj. 5 _
15 sub]. 6
—+— sub]. 7

_2 L I l
300 -150 -100 -50 0 50 100 150 200
Y (mm)

 

 

 

 

 

 

(C) At Pelvis — Large Males

Figure 50 Continued

131

Thigh
25 i i l l T l l

 

 

 

 

 

Difference (mm)

 

 

 

 

Y (mm)

(D) At Mid Thigh — Large Males

Figure 50 Continued

132

 

 

30 l l W i i

-O- subj. 1

25’ + subj. 2 J
->- sub]. 3

20~ + subj.4 ‘
+ subj. 5

15] subj. 6 ~
- —f- '. 7
1o '- 8
'. 9

 

 

    

 

Difference (mm)
01

 

 

 

0
-5
-10
I
-15 [ 1 _
-2 I l l l I
- 50 -100 -50 50 100 150

(A) At Mid Thorax — Small Females

Figure 51 Typical Difference Curves between Scaled and Measured Contour

at Mid sections of Small Females

133

Difference (mm)

Lumbar

 

  

 

 

 

 

 

 

 

 

20 T ﬁr T l 1
15 ._ . ‘ I . . . . ._
10— . _
xi ’ 4‘ \
5- _
0- _
-5 — _
—0— _ 1
-1o — , —<— .2 -
" + sub]. 3 *
-15 — + subj. 4 _
+ sub]. 5
_ _ + subj. 6 _
20 -i- subj. 7
-+ sub]. 8
'25 h -x- subj.9 “
"3950 -100 -50 0 50 100 150
Y (mm)

(B) At Mid Lumbar - Small Females

Figure 51 Continued

Difference (mm)

Pelvis

 

 

 

, “"’"’ -O- subj. 1
-5- ‘ +subj.2 * --
->— subj. 3
+ subj.4
'10 ‘ + subj. 5

subj. 6
—i— subj. 7
-e- subj. 8
-x- subj. 9

 

 

 

 

 

 

- 50 -100 -50 0 50 100 ' 150

(C) At Mid Pelvis — Small Females

Figure 51 Continued

135

Difference (mm)

Thigh

 

   

 

  
 
 

 

 

 

 

 

 

20 l T l l i
15- -
10 1 1
5
0
+ sub]. 1
-5 r + sub]. 2 ~
->- sub]. 3
‘ + sub]. 4
-1° _ + subj. 5 ‘
subj. 6
_1 5 _ —i— subj. 7 _
i -O- subj. 8
. -*- subj. 9
_2 1 1 r r 4 r
- 50 -100 -50 0 50 100 150

Y (mm)

(D) At Mid Thigh — Small Females

Figure 51 Continued

136

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o m m m a m s o sushi
N m N m N 2 o N N 5%:
o o m N w a o o v 59.:
N m N N 2 N v N N 5::
o m. N N 2 v m a o 599:
N m m m 2 m. 2 o v 522:.
o o N N 3 m m. v N 529:
N w P v v N N a o 22$
5. N m o v m N m m 22mm.
m N w o m N N m e 223
o m m o S m o m o .853
o m N 5. NF m a v v .253
o m N m N N m o N .853
3962 N “8.35 o 603% m 60.35 v 80.35 N .835 N 8635 F 335
3.2. 093 i.

Ema 59:355. new opium comics AEEV moocmcoEo camco><

o ozﬂ

137

 

omm.o><

 

N :95

 

v 59,:

 

F :95

 

o .855...

 

v 520:...

 

N 585......

 

N 22mm

 

N NEE

 

v m_>_mn_

 

0 39:3

 

VCDNNNO‘JIDIDLDV'V'V

IDNOONFO‘DV'V’V'VIDV

V’OFV‘VNLOV'V'V'CDV’

lDwOONNNV'OSOICDCDl-O

(DIDNNNLOIDNNNNV

OOLDFFNNV‘OOIOLDCDCO

v-V'V-Fv-NCDLDMOCOOO

OOV’COOOCOCOLDCOV‘V’VV'

v 5953

 

e.

.V

m

w

m

N

N»

N

N 59:3

 

 

ommcm><

 

N 58.35

 

N 335

 

N amaze.

 

v 335

 

N smasm

 

 

N 28.35

N 8635

 

 

 

$833 was 35....

moocoLoEo or: No AEEV mcozmSoo “:3:me

n 052.

138

 

gem}.

 

N 5%:

 

v =2:

 

w 29;...

 

N EOE

 

s 52:

 

m x805.

 

N .885.

 

m £25m

 

O’NMNNOMFDD

m m_>_mn_

 

N
1-

v £25m

 

o .823

 

NO)

3223

 

MVKDNLDVVC‘OVLDV’VVV

NDNMFPQMMNMMMM

v

MVMQGCDV’COMMLOLDNV

DVVUDMNMNV'COCOVUDV

mvvrxmcocommmmcomm

MMMMNFNMNO’MC‘OWM

CONVLDC‘ONVVCDLDMMV'V’

NNO’NLDV’C‘ONLOCDCOCDCDLD

N .253

 

 

comb;

 

2835

 

N 32.0.

 

2835

 

N .835

 

N 6035

 

v 8535

 

N .835

 

N .835

 

F 80.35

 

 

 

Ema 09:35.2 new 3.3m comics AEEV $39550 53.55.

358.8 =me i

m min...

139

 

3892

 

N 59.:

 

N 5%:

 

N 5%:

 

NNNNV‘

o x805...

 

O
P

v x805.

 

m x59:

 

N x805.

 

N News

 

Nmsaa

 

v £25m

 

N .353

 

v .283

 

NNMNNNMMV‘V’MNC‘OM

(DI-OWNv-NV'CD

NVNPs—V-NMNNNFNN

NNIDC’ONOONNV’C‘OVFNO’)

v-OOMNNFOOC‘ONVNNNN

MFV’MNNNv—V‘mmmmm

NNWNMMV‘NNVMNVU)

NNNPFFNNC‘OC‘OC’DNMN

”MMNNNMMVNNFMM

NNVNFFNNNMNNMN

F .283

 

 

ouﬂm><

 

m “835

 

N N835

 

283$

 

N 28.35

 

N N833

 

3838

 

N N835

 

N N835

 

N .835

 

 

 

moo—55:5 9: No AEEV NcozmSoo 2%:me

$0533 @683 =me1

m min...

140

5.4.2.2 Coefﬁcients of Correlation
The coefﬁcient of correlation analysis results SR were falling in the range
of 0.99 ~ 0.95 for almost all section curves, as presented in Table 10 for large
male subjects, and Table 11 for small female subjects. The values of SR
indicated nearly identical shape of the pairs of scaled and measured curves.
Figure 52 shows the trends of the coefﬁcients of correlation of subjects 1

of the large male and small female subjects.

141

 

 

 

 

 

 

 

 

 

 

 

geLii-‘zwﬁg... » .1.

7 . 1:: , -»
. ' . (Tum: g“. ML;-

1
“'9 ~

 

 

I L2
I L4
u L6
I P4
I P5
I P6
I Th2
I Th3
I Th4
I The
I: Thi1
I Thi4
I Thi7
IAveragei

 

 

(B) Coefﬁcient of Correlation — Small Female Subject 1

Figure 52 Typical Bar Charts of Coefﬁcient of Correlation

142

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NNNN NNNN NNNN NNNN NNNN NNNN NNNN NNNN @9522
NNNN NNNN NNNN NNNN NNNN NNNN 83 23 N 59.:
23 «SN $3 23 83 NNNN NNNN NNNN N. 5%:
NNNN NNNN NNNN ENN NNNN NNNN NNNN NNNN N 59.:
83 NNNN NNNN «NNN NNNN 83 23 NNNN N .259:
83 NNNN NNNN 83 83 83 No.3 NNNN v 559:
£3 £3 83 83 NNNN NNNN NNNN 83 N .85.:
83 NNNN N85 23 23 $3 NNNN NNNN N 522.
NNNN NNNN NNNN 83 23 Name NNNN NNNN N N_>_Nn_
NS 3.3 83 NNNN NNNN NNNN NNNN NNNN N. 22$
23 NNNN NNNN NNNN NNNN NNNN SNNN NNNN N .253
83 «Nos NNNN N23 83 NNNN 83 82. v .253
NNNN «No.0 NNNN 23 83 N23 23 23 N .253
2593. N N836 N 3.35 N N835 N. .233 N N835 N N835 N 50.320.

 

 

 

 

 

 

 

 

 

 

9.00.33 o_mE No.3:

:o_Nm_m..oo No 358580

or wish

143

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Nde mmmd Name 83 mmmd Nde wood 83 mmmd Nmmd Noemi
N85 Sad Nde Nde N86 N86 N23 mvmd Nvmd mmwd N 59.:
N86 ammo N86 ammo Nmmd NNNN 83 mag 83 N86 N 59.:
ammo Nmmd N30 N86 83 mamd N86 vmmd N23 N85 N 59.:
Name wood ammo 83 N23 N86 N86 ammo 33 ago N 55.:
NNNN NNNN NNNN NNNN NNNN NNNN NNNN NNNN 88 83 :52:
NNNN NNNN NNNN NNNN NNNN 83 83 NNNN NNNN 83 N 55.:
Nmmd Nmmd ammo Name 83 ammo v8.0 83 N30 83 N ENE
33 NNNN NNNN 38 NS 83 NNNN 83 33 33 N 55....
Sad Nde 886 N50 NNNd Nde N :5 mag 53 Nde N N_>_Nn_
83 82 ammo NNN.o wood 23 Nmmd «8.0 v8.0 93 N. N_>_Nn_
NNNN NNNN NNNN NNNN NNNN NNNN NNNN NNNN NNNN NNNN N253
NNNN NNNN NNNN NNNN NNNN NNNNN NNNN NNNN NNNN NNNN N253
83 N86 83 33 N86 N30 ammo mamd mag mmmd N .253
253 22.35 N .833 :8.sz N523 28.35 38.35 N 2.35 N .835 22.35
$00.33 0.0EON =0Ewi
530.0560 No 256580 NN 038.

144

5.4.2.3 Corridors

There were always differences even though the contour curves were at
the same relative sections and in one subgroup of subjects. Because the
anthropometric dimensions were different from each other for the different
subjects, the variability necessitates that corridors be created to represent the
range of contour shapes at the interested sections.

The average of the tested contour curves were ﬁrst calculated at each
section of interest. Then at each point (1 mm apart), the standard deviations of
the tested data were computed, and were added to the average curves to
construct the corridors.

For the most of the scaled contour curves, they were within the corridors

of tested average with i- 0'. See Figures 53, (A), (B), (C), (D), and Figure 54, (A),

(B). (C). (D)-

145

 

 

 

 

 

 

 

 

(A) Scaled Curves and Corridors at Mid Thoraxes - Large Males

Figure 53 Typical Corridor Plots at Mid Thorax, Lumbar, Pelvis, and Thigh of

Large Males

146

 

 

 

 

 

 

 

 

 

 

(B) Scaled Curves and Corridors at Mid Lumbar -— Large Males

Figure 53 Continued

147

 

‘é’

 

 

 

 

 

 

 

 

 

 

(C) Scaled Curves and Corridors at Mid Pelvis — Large Males

Figure 53 Continued

148

X (mm)

 

 

 

 

 

 

 

 

 

(D) Scaled Curves and Corridors at Mid Thighs — Large Males

Figure 53 Continued

149

 

 

 

 

 

 

 

 

 

(A) Scaled Curves and Corridors at Mid Thoraxes - Small Females

Figure 54 Typical Corridor Plots at Mid Thorax, Lumbar, Pelvis, and Thigh of

Small Female

150

 

 

I
d
.5
O
r

_'.

N

o
T

 

 

 

.‘a.

.5

o
r

.13

0|

0
r

L

a

o
r

 

.3

N

O
r

 

 

 

 

(B) Scaled Curves and Corridors at Mid Lumbars - Small Females

Figure 54 Continued

151

X (mm)

 

 

 

 

 

 

 

 

 

-50 O 50 100 150
Y (mm)

(C) Scaled Curves and Corridors at Mid Pelvis — Small Females

Figure 54 Continued

152

 

 

 

'60” ."-e-Mean
-- subj.1
. -- subj.2
£0- 3. . , __ SUbj.3 ._ ; . _

> 2 -- subj.4 ,
g g -- subj.5 , g
_ ,. , i , -- subj.6 , . _ .. S . .
100 ' i -- subj.7 7
---- subj.8
---- subj.9

'12-’950 400 -50 o 50 100 150
Y (mm)

 

 

 

 

 

(D) Scaled Curves and Corridors at Mid Thighs — Small Females

Figure 54 Continued

153

5.5.1 Thorax

In this region, the subjects’ backs came into contact with the chair usually
after the 3rd section from the lower neck joints in this study. This is basically
decided by the position of the lower neck joint with respect to the chair back. If
the position of the lower neck joint is far away from the chair back, the subject’s
back starts contacting the chair back at a lower point vertically on the chair back.
After the 4"1 section, the contour shapes were less affected by the sitting
postures of the subjects as can be observed in the Figures 44,45, 46, 47, 48, 49,
and the rest of the ﬁgures in Appendices B and C.

From the correlation analysis of the scaled and measured contour data,
very good results could be seen from the 4th section down to the 6th section,
which was the last section in this region.

The average difference (d) of between the scaled and measured contour
curves, presented in the Table 6, and Table 8, show that the contour data agree
within 15 mm with standard deviation less than 10 mm after the 4th section.

Table 10 and Table 11 list all the coefﬁcients of the correlation of the
correlation analysis between the swled and measured data. For the thorax
region, all the coefﬁcients are more than 0.95, which means that the curves are
very highly related or resemble to each other.

In the Figure 53 (A) and Figure 54 (A), most of the scaled contour curves

are very close to the average measured curve within one standard deviation 0' .

154

5.5.2 Lumbar

Because the bodies were in full contact with the chair back in the lumbar
region, the back deformed shapes were well measured and well deﬁned in this
area, and the contour results were close to each other from the scaling and the
testing. The differences between the scaled and measured contour curves were
less than 12 mm and deviations less than 7 mm (Table 6 and 8). The correlation
analysis results show that the contour curves have excellent resemblance, with
the coefﬁcient of correlation within the range of 0.989 to 0.999 (Table 10 and 11).

The scaled contour cUrves were all within the corridors with 2 0'.

5.5.3 Pelvis

In the current seat, the body segment in this region is not well supported
by the seat. Orin this region, the body is not contacting the seat. See Figure 55.
Obviously the body shapes were not well recorded and deﬁned. Only after the
5th section in this area did the contour curves reﬂect subjects’ body shapes. That
is why the coefﬁcient of correlation started becoming close to 1 after the 4th

sections as can be seen in the Table 10 and 11 for the small female data.

155

5.5.4 Thigh

Seat back

  
     
  

Sitter and seat
surface i terface

’
”
I’
’

In this region the boo
was not well contacting Undeformed seat surface

the seat surface

Figure 55 Pelvis Region Not Well Supported in Seat

In this area, the interface between the chair and sitter were in good

contact conditions for most of the contours because of the presence of the foam.

See the Figures 44, 45, 46, 47, 48, 49, and the rest of the ﬁgures in Appendices

B and C. The differences (d) were less than 13 mm with maximum standard

deviation 9 mm for large males, and the differences were less than 10 mm with

maximum standard deviation 6 mm for the small females.

However, at the front of the chair, the thigh postures affected the

deformed thigh shape measurement simply because the thigh might not fully

156

contact the seat surface. So at the last sections of the contour curves, some
deviation could be seen. In Table 10 and 11, some of the coefﬁcients of

correlation were below 0.9.

157

CHAPTER 6 CONCLUSIONS

Both seating postures and seat contours are very important to seat design
and seating comfort. The focus of this study was to determine (relative to a
skeletal linkage) the contours between people and seats that are deformed by
contact for a useful range of human subject sizes and postures. For this study,
three representative human subject groups were chosen: 5"1 percentile small
adult female, 50"1 percentile average adult male, and 95th percentile large adult
male. Ten subjects from each anthropometric group were measured. To
determine the seated contours, the subject’s posterior contours, including back,
buttocks, and thighs, were acquired in a specially designed contour chair
simultaneously with supporting force data from several load cells and posture
data from a video system.

The contour data from average size males were analyzed by section'ng
splined surfaces into two—dimensional curves based on axes systems in major
anatomical segments, thorax, lumbar, pelvis, and thigh. Even though human
beings are not exactly symmetrical, the measured postures and contours were
interpreted to provide symmetrical contours for development of the ASPECT
manikin and RAMSIS computer models. These contours were compared to
contours of the J826 SAE manikin.

The thigh and buttocks surface contour of the ASPECT manikin was deﬁned by
selecting one subject’s contour, which was consistently nearthe middle of the

nine subjects’ contours at each of several cross-sections of the thigh and

158

buttocks region; this was done to provide a contour that was realistic because the
variation between subjects was small.

Because of the greater number of mobile segments in the torso compared
to the thigh and buttocks region, there are greater postural and contour variability
in the back area. After the data were transformed, and scaled to the average
distance between lower neck joint and mid-HJC, the back data and postures
were averaged in the X direction (generally posterior to anterior direction) in each
local segmental coordinate system to better represent the range of the variation.

For the aesthetics in the development in the back contours for the
RAMSIS model, the ASPECT back contours were adjusted a few millimeters,
less than measurement variability. To develop the RAMSIS contour, the postural
linkages of the contours by Frost and ASPECT contours were aligned and the
differences between these contours were averaged to generate the RAMSIS
contour. Both ASPECT and RAMSIS contours could be used for the applications
of the seat design and all purposes of dummy design.

Representations of mid-size males have been extensively used as the
basis and starting point for design and evaluation of seating. Thus, it is very
useful to explore the relationships between mid-size males and other commonly
used human sizes, which are small females and large males. The present study
is the ﬁrst to scale mid-male contours to the anthropometric dimensions of small
females and large males and to validate these scaled contours with the
measurement data from small females and large males. A system was

established in this study to scale the contours of mid-size males to represent

159

contours for small females and large males. In the scaling process, the body
regions of mid-male’s thorax, lumbar, pelvis, and thigh were separately scaled to
the anthropometric dimension of each individual subject in the small female or
large male subgroups, and then rotated to align with the body postural linkage of
the subject.

Veriﬁed by the analysis of correlation, the scaling results were in very
close agreement with the measurement data and most of the coefﬁcients of
correlation were within 0.95-0.99. The corridors were created to illustrate the
possible range of contour shapes at the interesting sections. First, the average
of the measurement contour curve was found at each section, and then the
standard deviations of the measured data were computed, and were used to
calculate the corridors. For most of the scaled contour curves, they were within
the corridors of the measured average with :t a. In a few extreme cases, the
maximum average difference between the scaled and measured contour curves
was within 15 mm, and maximum standard deviations were 9 mm.

Because of the closeness of the scaled contours to measured contours for
this broad range of human sizes, the contours developed in this work for average
size males can be used with conﬁdence in computer models for scaling to
represent people of different sizes such as small females and large males based
on anthropometric dimension. This is a very useful result of a very simple scaling
process to represent various human back contours. The sample sizes in the
present study are limited, and future work could add conﬁdence to representing

and scaling human back contours. However, the contours and scaling provided-

160

in this study can be easily and usefully implemented in computer models such a
RAMSIS.

The contours provided by this study represent the interface between two
deformable objects: seats and people. The back shape of the torso is highly
dependant on posture, and these can be represented by the contours and
skeletal linkage of this study.

Consideration should be given to regional deformability. For a particular
posture, the back of the torso (where the skeletal structures are close to the
contact surface) is not highly deformable relative to regions beneath the buttocks
and thighs (where pelvis and femurs are prominent only in the central regions).
Also, the deformability of seats varies from place to place. So, the contours of
this study should be used and interpreted with consideration of seat and body
deformability. Future work could add useful represent of the defonnability of the

human soft tissues and seats.

161

APPENDICES

162

APPENDIX A

Scaling Factors Used for Scaling Procedures

163

 

 

 

 

 

 

 

 

 

 

 

NNN NNN NNN NNN NNN NNN NNN NNN 5.50 52.8
NNN NNN NNN NNN NNN NNN. NNN NNN £525 528
NNN NNN NNN NNN NNN NNN NN NN 59.3 533
NNN NNN NNN NNN NNN NNN NNN NNN 520 55.5
NNN NNN NNN NNN NNN NNN vNN NNN 5.505 55.5
NNN NNN NNN NNN NNN NNN NNN NNN 59.3 55;
NNN NNN NNN VNN NNN NNN NNN NNN £20 520
NNN NNN NNN NNN NNN NNN NNN NNN 5525 520
NNN NNN NNN NNN NNN NNN NNN NNN 59.3 520
.55. .55. .55. .55. .55. .55. .55. .55. 5.22055
.13 N N... N 23 N 8:. N 3% N E: N 6.2. N 55.2-55. . Noam
5......sz 50.2w .8..sz .855 505% .855 50.35 50%....

 

 

 

 

 

 

 

 

 

 

95.08 0.5.). 093 .3 30.20.55 zoom NN 0355

164

 

 

 

 

 

 

 

 

 

 

 

8. 3. SN «Pu SN mmm mmm 58 new 8N 2.80 0.88m
$5 :5 55 :5 35 $5 «on 53, N8 85 “wwmhm
8 x. B 8 .5 om .5 8 8 B 258.. 052.8
8. 5.: 8. a: .5. 8. mm. 8. 8. SN 2.80 552.
SN ova com 58 EN 2m 8m 8m 8m 8» 2.89m 55....
mm. m: 8. 8. m: m: a. N: .5. a: 598.55....
8N EN EN 3. Bu mom ovu 2m 2m ova 2.80 .820
8m 8m .28 5. mm mum cam .8. wow mum 2585 .820
:8 8m emu 58 EN :5 mmm mmm 8m «8 2.98.. .820
.EE. .55. .55. Es. .55. .55. .55. .55. .55. .55. 585885
0.2:. 2...... .0...» Cam 8me :0: f0; .00? .053 85.2-2.2 . >80
.8530 .8530 .8530 .8530 .8530 .8530 .8530 .8530 .8530 Swim...

 

 

 

 

 

 

 

 

 

 

 

 

9.30m .2560". =me .0. 0co_mc0E_o .68 m? 035...

165

APPENDIX B

Scaled and Measured Small Females Results

166

 

 

 

 

 

    

~- —. —-—-’a-.’/’

Figure 56 Female Subject 2 Rear View

Figure 57 Small Female Subject 2 lso View

167

\

‘.\ W

W
- \ --”’

SF2 Thorax 2"(1

SF2 Thorax 3"

N

- x
l
\‘\ J”
’/
\
,ﬁ"

' ‘ t
“—- ..—— I,

SF2 Thorax 4"1

SF2 Thorax 5"

SF2 Thorax 6"1

SF2 Lumbar 1“

SF2 Lumbar 2"“

\
_ \j , .
\§\\~——‘///\i

SF2 Lumbar 3"

Figure 58 Scaled and Measured Curves of Small Female Subject 2

SF2 Lumbar 4‘" . nd
SF2 Pelvnc 2
\ /
‘ /
“k [m ‘\ ll
\Q 7/ \\ /
WI/ \\\ ’x/
W
SF2 Lumbar 5'“
SF2 Pelvic 3'“
-A \‘ I, _ l j
\ \\ [I’m \\_\ - L'r‘.
\ ‘ 2” v /
SF2 Lumbar 6th SF2 Pelvic 4m
’ l
A I | “\‘l . I
‘K 1' r... \K /./N
\ \‘. ,.
\ / W/
\\ -------- ”/ m
SF2 Pelvic 1“ SF2 Pam 5

Figure 58 Continued

169

 

SF1 Thigh 4m

\ /

SF2 Thigh 5'"

 

SF2 Thigh 6‘"
\\ ,
Wk
SF2 Thigh 7'h
l \ ' " ' I
/ \ —————— //
x W
/
SF2 Thioh 8‘"
Figure 58 Continued

170

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

J... k
i“?
M)

Figure 59 Small Female Subject 3 Rear View

Figure 60 Small Female Subject 3 lso View

171

\
|
I
II
\
\X x/

I
W’,

~—

SF3 Thorax 4m

\\ I
\\\

SF3 Thorax 5‘"

SF3 Thorax 6"‘

I
l
_ ' 'N—
k /
\ /
’/

~-.—-—--—’

SF3 Pelvic 1"

l i
i
—\i\ - if.—
I."
\t I

\ l
\W

SF3 Lumbar 2'”

SF 3 Lumbar 3'“

Figure 61 Scaled and Measured Curves of Small Female Subject 3

 

 

\“ /
w”!
SF3 Lumbar 4"
SF3 Pelvic 3'"
‘ I
ﬂ ‘\ ' [If-‘—
WW,
SF3 Lumbar 5"
SF3 Pelvic 4'"
\
“\‘X‘ . I; |\ I!
\ \
‘ —— ‘\§‘\‘\
W
SF3 Lumbar 6m SF3 Pelvic 5“
l . ,
H ‘ ‘ I, \ I
‘x.\ 12‘.— I
\§\\ // \\\\ \—-//
SF3 Pelvic 1“ SF3 Pelvic 6"
l‘ l
_\ \ I
c \\ //"~ \\ /)
\\“__._ ____ ’// \\\\‘\‘
SF3 Pelvic 2"“ SF3 Thigh 1“

Figure 61 Continued

173

SF3 Thigh 2"1

SF3 Thigh 3"

_—---
‘

SF3 Thigh 4"

SF3 Thigh 5"

SF3 Thigh 6‘”

 

SF3 Thigh 7‘h

'—~.

      

\-_—’ \ I
\c——.—’

SF3 Thigh 8"

Figure 61 Continue

 

 

 

 

 

 

 

 

 

 

Figure 62 Small Female Subject 4 Rear View

/

i/-,
- . /
/ /

‘1.“ /W

  
 

Figure 63 Small Female Subject 4 lso View

175

    

‘ﬁ-p". \’ s ’

SF4 Thorax 1"

SF4 Thorax 2"“

/"—‘\
,
\\\____,

sr=4 Thorax 3'"

'-
~—’

sr=4 Thorax 4'"

SF4 Thorax 5‘“

SF4 Lumbar 1"

SF4 Lumbar 2‘”

SF4 Lumbar 3'‘1

l

l .
\J

SF4 Lumbar 4"

-———ﬁ

1
l
\\

\\

sr=4 Lumbar 5"

Figure 64 Scaled and Measured Curves of Small Female 4

 

"\X //‘“
.w
SF4 Lumbar 6"1 SF4 Pelvic 6th
: I
\‘i . if. K I)
\X / \\\\\\ ’I’l

14W ~ _-——

SF4 Fame 1“ SF4 Thigh 1::

l
i i K I.
at l LIE ‘\
i I \
\x X
SF4 Pelvic 2"" SF4 Thigh 2nd
I. \
\ // k g/
“2%— ‘ ‘\~ - -”
SF4 POMC 5m SF 4 Thigh 3rd
\
LE ‘\\\\ //l
\\ ’1, ‘ \Vty/
SF4 Pelvic 3rd
Figure 64 Continued

177

 

SF4 Thigh 5m
\\ ,
SF4 Thigh 6""

\
\ ——————— /
I v

SF4 Thigh 7‘"
\ '__‘ //
SF4 Thigh 8'"
Figure 64 Continued

178

 

 

 

 

 

 

 

 

 

 

 

 

Figure 65 SmalFemale Subject 5 Rear View

 

4 I
l /”
I
z/I’ 1/1/
x /’ " ' ' '
I
// ,/’/’ I / I [I
I 1’ , / II
a“ ’ I, I

Figure 66 Small Female Subject 5 lso View

179

/ ~ \ /
\\ ,7/l \ /
Th “ SF5 Lumbar 1"
SF5 orax 1
“I: l: M _\\{ I,
\ 4:'/ \\ “’//
SF5 Lumbar 2"“
SF5 Thorax 2"6

SF5 Thorax 3'“ SF5 Lumbar 3'“
l l
._\\ _ I, —\ j j
\\\ 1:1, / \~ -..-.._..-/
SF5 Thorax 5‘"
SF5 Lumbar 4'h

ﬁ' . /————— —_\\ '
\\\\ I //’ M

‘W'
”'1
SF5 Thorax 4 SF5 Lumbar 5‘"

Figure 67 Scaled and Measured Curves of Small Female Subject 5

180

 

—\\\\ I’l/ ._ __
SF5 Lumbar 6'” SF5 peMc 5th
a: I: ,/"~_— \ //
\\ //’/ ‘; I’II’,
SF5 Pelvic 1st SF5 Pelvic 6'“

SF5 Pelvic 2"0 SF5 Thigh 1st

I
\
ia—

\

- th
SF5 Pelvrc 4 SF5 Thigh 2nd

Figure 67 Continued

181

   

SF5 Thigh 3'“ SF5 Thigh 6‘“

 

W“‘—T‘J SF5 Thigh 7'“

SF5 Thigh 4th

 

SF5 Thigh 8‘"

SF5 Thigh 5m

Figure 67 Continued

182

 

 

 

 

 

 

 

 

 

 

 

 

Figure 68 Small Female Subject 6 Rear View

at

\_

2

7

75

Figure 69 Small Female Subject 6 lso View
183

\ 11,/P—
\\ --
SF6 Thorax 1"
i: :
\‘ I‘, f“—
4,
\-_ /
SP6 Thorax 2"d
\
ir—

SF6 Thorax 3'‘1

~—

/'

SF6 Thorax 4m

SF6 Thorax 5"1

l
l
I

l
. [rm——
’1

SF6 Lumbar 11“

SP6 Lumbar 2"“

SF6 Lumbar 3'6

SF6 Lumbar 4'h

SF6 Lumbar 5th

Figure 70 Scaled and Measured Curves of Small Female Subject 6

184

SF6 Lumbar 6th

SF6 Pelvic 1st

SF6 Pdvic 2"6

x ”

‘—_—_—-"

SF6 Pelvic 3“1

SF5 Pelvic 4th

 

SF6 Pelvic 5'h

 

SP6 Pelvic 6'”

SF6 Thigh 2"“

SF6 Thigh 3"

Figure 70 Continued

185

 

SF6 Thigh 4m SF6 Thigh 7'"

/
\‘ ’ W
x I
’ ‘

SP6 Thigh 6'"

SF6 Thigh 5‘"

.0...“ ’/

 

SF6 Thigh 6th SF6 Thigh 9'"

Figure 70 Continued

186

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 71 Small Female Subject 7 Rear View

 

Figure 72 Small Female Subject 7 lso View

187

SF7 Thorax 1“

SF7 Thorax 3rd

I
l

\

\\\ /
“A "

SF7 Thorax 4'"

‘R

SF7 Thorax 5m

SF7 Lumbar 1"

SP7 Lumbar 2"“

SF7 Lumbar 3'"

SF7 Lumbar 4‘"

SF7 Lumbar 5m

Figure 73 Scaled and Measured Curves of Small Female Subject 7

SF7 Lumbar 6m

\\ r“—
\\ /

\
W-.. ——-- ’

SF7 Pelvic 18t

SF7 Pelvic 2"d

 

SF 7 Pelvic 4‘"

 

SF7 Pelvic 5'"

 

SF7 Pelvic 6‘"
L.—
SF7 Thigh 1"t
' l
\\ /’
SF7 Thigh 2"“
Figure 73 Continued

189

 

l ' ' - l
I

\\
\~\/\/

. rd
SF7 Thigh 3 SF7 Thigh 7a..

 

 

 

 

. 4m
1‘ . . ' I
W \\ /"-.‘\ //’
SF7 Thigh 5‘" . m
SF7 Thigh 9
\ \\‘~ ””””” \ /
SF7 Thigh 6'h
Figure 73 Continued

190

 

 

 

 

 

 

 

 

 

 

Figure 74 Small Female Subject 8 Rear View

. Figure 75 Small Female Subject 8 lso View

191

SF7 Thorax 1“

___\‘ _
\\ /’/
RN ,- _

SF7 Thorax 2"d

\LQJ
\

SF7 Thorax 3'u

0
———~‘ \
I;
1
I
~V’ /,

SF7 Thorax 4'"

SF7 Thorax 5‘"

ﬁ‘ . ’
\\ ‘

SF7 Lumbar 18t

SF7 Lumbar 2"“

__~__'—\
\ .
‘. '\‘ /

SF7 Lumbar 3'“

M b , ”M

SF7 Lumbar 4'’1

SF7 Lumbar 5'“

Figure 76 Scaled and Measured Curves of Small Female Subject 8

 

_ \.‘ . N— \\ . ,’
.‘ t ,' \ l
x \\ I /
’ \‘\ ’ //

\——--

SF7 Lumbar 6"1
SF7 Pelvic 4th

_.\ . ﬂ \ .
\ ’

 

SF7 Pelvic 1’" SF8 Pelvic 5m
ﬂ \i 8 'I’ . V
- nd
SF7 Pe'V'C 7- SF8 Pelvic 6'"
\\|\ "I '\\ [I
x _________ ;’//"— \‘ -—__ ,2 I
- h
SF7 P6|VIC 3‘ SF8 Thigh 15:
Figure 76 Continued

193

r— _ or

SF8 Thigh 2"(1 SP8 Thigh 6th

   

SFB Thigh 3rd SFB Thigh 71h
- m
SP8 Thigh 4 SF8 Thigh 8‘“
\\ — ”// \\\\‘WM
SP8 Th' h 51”
'9 SP8 Thigh 9'“
Figure 76 Continued

194

 

 

 

 

 

 

Figure 77 Small Female Subject 9 Rear View

\
/,

<j/””

Figure 78 Small Female Subject 9 lso View

195

 

SF9 Thorax 1"

SF9 Thorax 2"‘1

SF9 Thorax 3'‘1

SF9 Thorax 4'“

 

SP9 Thigh 7'h

 

\. -

\'~ ’1
\ .

\ - V4,

\ ‘\ /’/

‘
‘—--~’—--—’

SF9 Lumbar 1"

SF9 Lumbar 2"“

SF9 Lumbar 3'‘1

SF9 Lumbar 4m

SP9 Lumbar 5m

Figure 79 Scaled and Measured Curves of Small Female Subject 9

 

 

 

\\ IZ/Ir—
SF9 Lumbar 6'" m
SF9 Pelvic 6
AK Ij/ '— Jg j.—
\\ x” \- //
SF9 Pelvic 1" SF9 Thigh 1“
a 1 - :r—ﬁ I ‘ ' ‘ |
‘i l“ e‘ I
\\\_____ ”1,,” \\\\‘)/_N/
SF9 Pelvic 2"d . m
SF9 Thigh 2
xx //
SF9 Pelvic 4‘h SF9 Thigh 3'“

   

SF9 Pelvic 5‘"
SP9 Thigh 4‘“

Figure 79 Continued

197

x I

\W

SP9 Thigh 5m

 

SF9 Thigh 6th

Figure 79 Continued

198

SF9 Thigh 7‘“

SP9 Thigh 8'"

APPENDIX C

Scaled and Measured Large Males Results

 

199

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

72.3;
: \x
E -2;
l- A $5
jg:
{5.91
‘4: 1.

Figure 80 Large Male Subject2 Rear View

 

Figure 81 Large Male Subject 2 lso View

200

 

+

‘ '--

LM2 Thorax 1"

 

\ //
\ /
\ ___________ ,1
LM2 Thorax 2'“I
l l
l l
\ I
\ /
*g /"
\ _____________ /
rd
LM2 Thorax 3

 

LM2 Thorax 4"

 

\‘_ ____-"’

LM2 Lumbar 1“

\ w I
Vw—w—nﬂ’

LM2 Lumbar 2"“

LM2 Lumbar 5"

Figure 82 Scaled and Measured Curves of Large Male Subject 2

 

 

ll

\
r'
31“—
. / ”7 \
L“ 5 /
‘>.}‘ - __—_-g-ﬂ-IP

r’
.i.
LM2 Pelvis 3'"

 

LM2 Lumbii‘r 4“

I i
‘ l
\ I
4 .
/ . x} \
LM2 Pelvis 4"

LM2 Lamb?!r 5..

 

 

\‘ —l—— ”I
\ .
A, i f,"— ‘77
\ / q . 7‘wzu’l
\\x\\.\~ a. //
LM2 Pelvis 5...
LM2 Lumbl‘ir 6“
I ‘ \
i I \
\ I, x “I
\ ‘7‘» ..-r;.”’
J‘V + // ,——- \w:-':-'::"’
\ //
________ LM2 POWis 6“
LM2 PelViS 1“

\\
LM2 T1139h 1“

LM2 pelvis 2"“

Figure 82 Continued

202

C-‘ \ x «ﬂ‘o- 4
~‘_" 2. .—2’ x ...... ' ~~~~~ 0
\‘ \ ;’ \\~—u— ”-\ \v”’
‘_._--—__t “_—” \‘-_”
LM2 Thigh 2"“ LM2 Thigh 5"
J I
/ i l
l/ \ . '—‘~ I
,‘ ‘§ \‘. /
\“\\__-" ‘~- -11" \ N“ I"\ -- .— z

- rd
LM2 Tr“9h 3 LM2 Thigh 6'“

i
I
I
I
i
\
\
\
/
I
I
I
|
\
\
\
. Ju'ul-j.

.“J;

- 3|
LM2 Thigh 4" LM2 “"9" 7

Figure 82 Continued

203

 

 

 

 

 

 

 

 

 

Figure 83 Large Male Subject 3 Rear View

 

Figure 84 Large Male Subject 3 lso View

204

 

 

LM3 Thorax 2"“

LM3 Thorax 6m

LM3 Lumbar 1"
LM3 Thorax 3"

 

LM3 Lumbar 2"“
LM3 Thorax 4'”

 

m LM3 Lumbar 3”
LM3 Thorax 5

Figure 85 Scaled and Measured Curves of Large Male Subject 3

205

\x / \ _ l
/ I " v
'\“‘...—_._——u—’" ‘~- \‘ .’- 4* ’..i‘

s ’

“ %—-_—_——‘M"

"I

LM3 Lumbar 6'" LM3 Thigh 5"

l |
I .
‘\
v \ ‘ l l

‘. ._\ ,. \ ' ' . I

u} .I \ ,_.—: I
\ 'I"~. / \\ ,-' ’1 ’/

‘—' -A‘- ’9 ‘ur \ 1“ I ’

\‘-—--———-" ‘*~~_')~".~’ N""/'

' It

’ -~~‘~. \ I \
K ‘s ./ '~. \ I", I \p ’_
""--"-" “-‘_>-t—" ‘3': ‘7'

LM3 Thigh 2"“
LM3 Thigh 7m

LM3 Thigh 3'“

Figure 85 Continued

206

 

 

Figure 86 Large Male Subject 4 Rear Wew

 

Figure 87 Large Male Subject 4 lso View

207

 

LM4 Thorax 3"l

 

LM4 Thorax 4m

, \ / /
W’
LM4 Thorax 6‘“
l l
l l
A \\ I
‘l\“\\\\_____._.——\_////
LM4 Lumbar 1st

W70
LM4 Lumbar 2'”

LM4 Lumbar 3

Figure 88 Scaled and Measured Curves of Large Male Subject 4

208

LM4 Lumbar 4"1
l l
l l l
l - I
l l
\ /
\\ //
\\ ————————— ////
LM4 Pelvic 2"d

\‘\\ /;
\‘M/
LM4 Lumbar 5‘“

LM Pelvic 1“

 

LM4 Pelvic 3'“

LM4 Pelvic 4‘"

l I
\ /
\ /

A\§\ //
/
\\\\\_’/

LM4 Pelvic 5'"

‘~-————"

LM4 Pelvic 6'“

Figure 88 Continue

209

LM4 Thigh 2"“ LM4 Thigh 5'"

- ill
LM4 Thigh 3'd LM4 ““9“ 6

 

LM4 Thigh 4th

Figure 88 Continued

210

 

 

 

 

 

Figure 89 Large Male Subject 5 Rear View

”3

ﬂ "
, c/ATW/I/IW
’- \ 44.22/60
:« . I

- ‘
. , - a

Figure 90 Large Male Subject 5 lso View

211

 

 

 

\\ /
"—‘\ //-———-/
W
LM5 Thorax 2"“
l
l i
l I
.. /
\\\\ ///// '
LM5 Thorax 3’d
/ \
l l
l\ I
‘\ \\ //,//
\\“~- J
LM5 Thorax 4'”
I -— \
l - l
\ /
‘7‘» \ / /—-
“~\ / .
‘¥ 4,/
LM5 Thorax 5"

I
'\
.ﬁ \ .
\ ,1/

\\
\W

LM5 Thorax 6'"

LM5 Lmbar 1“

LM5 Lumbar 2"“

LM5 Lumbar 3'“

Figure 91 Scaled and Measured Curves of Large Male Subject 5

i
I l i
I \ - I
\ ’/"/—‘ \ /’
Kw _//{l \\ ’1/
‘ \\\\~_-”"”/
LM4 Lumbar 4‘" LM5 Pelvic 2'“1
i
\

 

LM5 Pelvic 3”1

 

LM5 Lumbar 6m LM5 Pelvic 4m
' i
ask [II /~_ \\\\ l/II/
LM5 Pelvic 1“t I-Nl5 Pelvic 5'"

Figure 91 Continue

213

l I
r . . l l
\ . I \ I
. \ / \ /
/

\ —— /
‘ \ /
\ J \\ I I’
‘ ~ _,
"
..

LM5 Thigh 5"

 

LM5 Thigh 1"

’_———‘

LM5 Thigh 2“‘1 LM5 Thigh 6'h

LM5 Thigh 3'" LM5 Thigh 7‘"

 

LM5 Thigh 4‘"

Figure 91 Continued

214

 

 

 

 

Figure 92 Large Male Subject 6 Rear View

 

'--/ 4714'”- 1.. _

Figure 93 Large Male Subject 6 lso View

215

 

‘ \
\‘\ é/ //

\_.—_-””

LM6 Thorax 3n1 LM6 Lumbar 1“

LM6 Lumbar 2"d
LM6 Thorax 4'“

Figure 94 Scaled and Measured Curves of Large Male Subject 6

216

/

 

LM6 Pelvic 2nd
' I
‘ I
W// \ I,
\\ I
LM6 LUmbar 4m \:\\ //
QQ‘”?=_//
LM6 Pelvic 3rd
\\.\ . III
l\ ll
LM6 Lumbar 51h \ \\ /
t\\:‘\\ /H/
W
LM6 '38ch 41!:
i
i
. l I'—
+ I”
\\ ///
LM6 Lumbar 6m
LM6 Pelvic 5‘"
Figure 94 Continued

217

 

 

 

\ \\\\“‘~ -- ””/y
\_‘\\ /,/ \\\ Ill
\ ___ ”/,/’ \ v”
LM6 Thigh 1x LM6 Thigh 5'"
i
ll . |
//,. \\ _ //
\ "’ “ ‘ ‘9’;
LM6 Thigh 2"“ LM6 Thigh 6"
\\ . I,
LM6 Thigh 3m LM6 Thigh 7‘"
Figure 94 Continued

218

 

 

 

Figure 95 Large Male Subject 7 Rear View

 

Figure 96 Large Male Subject 7 lso View
219

 

 

‘-,.——"" ~’_~‘_ ﬂ"

 

 

LM7 Thorax 4"

’\

LM7 Thorax 6'"

//_'

\\ / /

LM7 Lumbar 1“t

LM7 Lumbar 2"“

LM7 Lumbar 3rd

Figure 97 Scaled and Measured Curves of Large Male Subject 7

220

LM7 Lumbar 4'"

LM7 Lumbar 5‘"

I

\ I

i . II /

\‘\\ 7/
w

LM7 Lumbar 6'"

LM7 Pelvic 1“t

LM7 Pelvic 3'“

LM7 Pelvic 4"

LM7 Pelvic 5'"

Figure 97 Continued

221

LM7 Pelvic 6‘"
I . t
i I
\\ l ' x/
\\ 4/
LM7 Thigh 1"
\\ //
LM7 Thigh 3"
l I
i I
\\ I
\ \ //
W’”
LM7 Thigh 2"d

\\‘\\ "— ’/”//,
LM7 Thigh 4'h

\ - , /

\\ //

\\\ ’,/’ ~\ \\\\ 7
LM7 Thigh 5‘"

 

LM7 Thigh 6m

 

LM7 Thigh 7‘"

Figure 97 Continued

222

LIST OF REFERENCES

1 S.A.E. Handbook 1980 - Part 2, Society of Automotive Engineers,
Warrendale, PA.

2 Hubbard, R. P., M. P. Reed, etc., “ASPECT Task 1: Foundational Activities,
Interim Report”, February 1997.

3 Hubbard, R. P., W. A. Haas, R. L. Boughner, R. A. Canole and N. J. Bush,
“New Biomechanical Models for Automobile Seat Design”, paper no.
9301 10, lntemational Congress, Society of Automotive Engineers, 1992.

4 Haas, W. A., “Geometric Model and Spinal Motions of the Average Male in
Seated Postures,” Master Thesis, Michigan State University, 1989.

g Robbins, D. H., “Anthropometric Speciﬁcations for the Mid-Sized Male
V Dummy, Volume 2”, UMTRl—83—53-3, US. Department of Transportation,
National Highway Trafﬁc Safety Administration, Washington, 0.0., 1983.

6 Seidl, A., “RAMSIS - A New CAD-Tool for Ergonomic Analysis of Vehicles
Developed by Order of the German Automotive Industry,” TecMath GmbH,
January 1994.

7 Boughner, R. L., “A Model of Average Adult Male Human Skeletal and Leg
Muscle Geometry and Hamstring Length for Automotive Seat Designers,”
Master Thesis, Michigan State University, 1991.

8 Bush, N. J., “two-dimensional Drafting Template and Three-Dimensional
Computer Model Representing the Average Adult Male in Automotive
Seated Pastures", Master Thesis, Michigan State University, 1992.

9 Koritké, J.C., and H. Sick, “Atlas of Sectional Human Anatomy", Urban &
Schwarzenberg, Baltimore, Md. 1983.

10 Frost, B. L. III, “A Three-Dimensional Computer Model of Human Back
Contours for Automotive Seat Design”, Master Thesis, Michigan State
University, 1999.

11 Schakle, B., K. D. Chidsey, and P. Shipley, “The Assessment of Chair
Comfort," in E.Grandjean, Ed., Sitting Posture, Taylor, London, 1969.

12 Bush, TR, “Posture and Force Measures of Mid-Sized Men in Seated
Positions”, Ph. D. Dissertation, Michigan State University, 2000.

13 Chaffin D., Andersson G. “Occupational Biomechanics: Second Edition.
John Wiley and Sons, New York. 1991.

223

yt/ Claire C. Gordon, etc.,”1988 Anthropometric Survey of U. S. Army

15

16

17

18

19

20

21

f/

23

Personnel: Summary Statistics Interim Report”, March 1989.

Abraham, 8., Johnson, C. L., and Najjar, F., “Weight and Height of adults 18-
74 Years of Age, United States, 1971-74 Vital and Health Statistics, Series
11". DHEW Publications Number 79-1656.

Setyabudhy, R., A. Ali, R. Hubbard, C. Beckett, and R. Averil, “Measurement
and Modeling of Human Soft Tissue and Seat Interaction,” paper 970593,
SAE, 1997.

Qualisys Mac Reﬂex Software and Instruction Manual, Qualisys, 148
Eastern Blvd. Ste. 110, Glastonbury CT 06033.

3M Product Literature on 7610 High Gain Tape for Motion Systems. Screen
Products, Industrial Optics Project/3M St. Paul, MN 55144. Ph. 612-733-
4403 and industrial Tape Division, 3M P. O. Box 358 Southﬁeld, MI
48037, Ph. 1-800-241-0184.

Melissa Sloan Gedraitis, “Experimental Determination of a Spinal Linkage
Model for the Seated Human Torso”, Master Thesis, Michigan State
University, 1998.

SDRC l-DEAS VI, “Solid Modeling Package“, Structural Dynamics Research
Corporation, Milford, Ohio, 1991.

Mase GE, Mase GT, “Continuum Mechanics for Engineers”, CRC Press,
Inc., 1992.

Seidel, GK, Marchinda, DM, Dijkers, M, and Soutas-Little, RW, 1995, “Hip
Joint Center Location from Palpable Bony Landmarks - A Cadaver Study".
J. Biomechanics, 28(8): 995-998.

Reynolds, HM, Snow, CC, and Young, JW, (1981 ), “Spatial Geometry of the
Human Pelvis (Memorandum Report AAC-119-81-5)”. Oklahoma City, OK:
federal Aviation Administration, Civil Aeromedical Institute.

24 RH. Setyabudhy, Z. Liu, R.P. Hubbard, “Measurement and Analysis of

Human Thigh and Buttocks Contours for ASPECT Manikin Development”,
SAE paper no. 1999-01-0964, 1999.

25 Ayyub, BM, McCuen, RH, “Probability, Statistics, & Reliability for Engineers”,

CRC Press Inc., 1997.

26 Xu, L, Agaram, V, etc., “Repeatability Evaluation of the Pro-Prototype

NHTSA Advanced Dummy Compared to the Hybrid lll”, SAE paper no.
200-01-0165. 2000.

224

27 Nusholtz, GS, Xu, L, etc., “Comparison of the Pre-Prototype NHTSA
Advanced Dummy to the Hybrid lll", SAE paper no. 971141, 1997.

28 MatLab Manual: “The Language of Technical Computing”. Version 5; The
MathWorks, Inc. 2004

225

\)l

“-21

 

 

 

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

iiiili‘lii

27 36

         

3 1293