WIWIWWWllfllmlfillfllWWI“illHHHI

  

_(DN®

 

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE it RETURN BOX to romovo this checkout from your record.
TO AVOID FINES mum on or baton data duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

THE DEVELOPMENT OF A METHOD TO ENCODE BODY
MOVEMENT OF AUTOMOTIVE VEHICLE OPERATORS

by
Maria M. Eppler

A Thesis

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

MASTER OF ARTS

Department of Human Environment and Design

1995

ABSTRACT

THE DEVELOPMENT OF A METHOD TO ENCODE BODY MOVEMENT OF
AN AUTOMOTIVE VEHICLE OPERATOR

by

Maria M. Eppler

Measurements of subjective comfort levels experienced by the vehicle operator
are critical when evaluating seats. The most common method of recording this
information is through the use of questionnaires. Questionnaires can be intrusive and are
time consuming for the subject when administered periodically throughout the drive.
Since comfort levels change over time, a method is needed to record these changes into
data that will allow for quantitative analysis of comfort. This paper describes a method to
encode body movement so it can be used as a tool to continuously and objectively record
the subjective comfort levels of the vehicle operator providing data for quantitative

analysis.

Copyright by
MARIA M. EPPLER

1995

Dedicated to my parents for their love and support

Acknowledgments

The author wishes to thank Dr. Tim Springer and Dr. Mac Reynolds for their guidance

and support during this study.

Special thanks to Ben Moggenberg and all the subjects for their participation.

Thanks also to Jan Loria and Cathy Chubb for their secretarial skills which were most

appreciated when dealing with the paperwork required for this degree.

TABLE OF CONTENTS

List of Tables ................................................................................................... vii
List of Figures .................................................................................................. viii
Introduction ...................................................................................................... 1
Literature Review ............................................................................................ 8
Recording Movement .......................................................................... 8
Encoding Methods ............................................................................... 9
Testing and Application ...................................................................... 14
Methods and Procedures ................................................................................. 17
Methods .............................................................................................. 17
Procedures .......................................................................................... 21
Results ............................................................................................................ 25
Discussion ...................................................................................................... 32
Future Study .................................................................................................. 37
Conclusion .................................................................................................... 40
Appendix
Body movement as recorded by each observer with respect to
body segment and time. .................................................................... 41
Bibliography ................................................................................................. 46

vi

LIST OF TABLES

Table Page

 

25
2

25
3

26
4 36
5 41
6 41
7 42
8 42
9 43
10 43
11 44
12 44
13 45

vii

Figure

LIST OF FIGURES

Page
An example of the Benesh Movement Notation for recording
choreography .............................................................................. 1 0
Corlett’s chart for recording standing work posture in industry ...... 13
Defined body segments for recording body movement ............. 20
Cumulative frequency chart for subject movement recorded in
Drive 009 .................................................................................... 26

Cumulative frequency chart for subject movement recorded in

Drive 025 ....................................................................................

Cumulative frequency chart for subject movement recorded in

Drive 031 ....................................................................................

Body movement by body segment over time, Drive 009,

Observer 1, Trial 1 ......................................................................

Body movement by body segment over time, Drive 0025,

Observer 1, Trial 1 ......................................................................

Body movement by body segment over time, Drive 031,

Observer 1, Trial 1 ......................................................................

viii

Introduction

Body movement is an accepted area of study in the social sciences.
Psychologists and communication scholars study body movement to learn more about
relationships, social influence, and social regulation. Body movement is a method of
nonverbal communication thus, it sends a message. Unlike verbal communication, body
movement is a continuous form of communication as long as visual contact is
maintained. A continuous form of communication allows observers to detect subtle
changes in the message that are often undetected in non-continuous forms of
communication. Since body movement is continuous and often subtle, the sender is
typically not aware of the message they are sending and thus it is hard for the sender to
mask the message. Body movement studies are premised “...on the view that
information about a subject’s affective state could be reliably obtained...even when he
sought to deceive the investigator” Ekman & Friesen, 1969 (cited in Weitz, 1979). Thus,
body movement gives a continuous, undiminished view of the subject’s interaction with
people and the immediate environment.

Body movement is a form of communication that relays information that may or
may not be intentional on the part of the sender. According to DeVito and Hecht, 1990,
nonverbal communication involves all messages other than words, including body
movement. They state that nonverbal communication occurs when we interpret a
message, whether or not a message was intentionally sent. A prime example of
unintended nonverbal communication was the focus of a study performed by Montepare,

Goldstein, and Clausen, (1987). The study demonstrated that “observers were able to

identify specific emotions from variations in walking styles”. By observing body
movement we are able to obtain a view as to what a person is feeling. For example if a
person touches a hot plate they immediately pull their hand away. By observation alone,
one would conclude that the surface was hot and caused pain due to the manner in which
the hand was removed from the hot plate. Since a message is interpreted this would be
considered a form of communication. By studying body movement, one would be able to
assess what a person is feeling about their surroundings due to their reaction to the
environment. Through nonverbal communication, one should be able to asses the
person’s comfort levels when the person is directly reacting to an element in the
environment such as a seat.

When defining an environment that humans desire, comfort is often listed as a
criterion. This is especially true when the user is in direct physical contact with the
environment, such as in sitting. Sitting is an activity that one engages in frequently. We
typically sit when we eat, travel, work and relax, therefore a comfortable seat is strongly
desired, (Branton and Grayson 1967; Richards 1980), but comfort in seating is hard to
define. Everyone continuously experiences levels of comfort and discomfort but when
asked to give a definition, terminology becomes vague. Unfortunately there is no
accepted theoretical definition for sitting comfort and no agreed upon method for its
measure (Lueder 1983). For the purpose of this paper comfort will be defined as a
positive sensation which is on the same continuum as discomfort. Since they are on the
same continuum, comfort will be the term used to describe both comfort and discomfort

unless discomfort is the only sensation being discussed. Comfort is a continuous

sensation (negative or positive) that is always present in a person but not always
consciously recognized. Being a continuous sensation, comfort is in constant flux.
People try to maintain a level of comfort by minimizing or eliminating levels of
discomfort.

To better understand sitting comfort it is important to understand how the seat
affects the user. Numerous studies have been conducted measuring the effects of the seat
on the subject. For example, studies focus on pressure (Shen and Galer 1993) and muscle
activity (Zimmermann et al., 1993). If the seat is in a vehicle, other factors are
considered such as vibration (Zimmermann et al., 1993). Despite objectivity in
physiological measures, they are not a direct indication of the subjective experience of
comfort, (Lueder 1983).

To study the subjective aspect of sitting comfort, questionnaires are frequently
used to assess comfort levels. Branton notes that comparisons between seats are
dependent upon the subject’s memory capacity. For short term sitting the results may be
stable but “...judgments of long-term comfort would be very unstable and unreliable.”
(Branton 1969). He also points out that comfort is “very primitive and deeply
ingrained...not readily accessible to introspection and verbalizations”. Based on
Branton’s experience, it follows that questionnaires are not the most accurate tools with
which to record changes in comfort levels over long periods of time. Another argument
against using questionnaires to track comfort levels is that sensations of comfort are not
static, but dynamic over time. Since feelings of discomfort are in continuous flux,

questionnaires which are given at one moment in time would only capture the discomfort

levels experienced at that particular moment. To more accurately assess the comfort
levels of a subject, continuous data collection is needed to allow researchers to identify
patterns of change in comfort levels over time. Both Branton (1969) and Jurgens (1980)
suggest observing subject sitting behavior, or body movement as a means of assessing
comfort over long periods of time.

When observing body movement of the seated subject over time it becomes
evident that movement is purposeful. Branton (1969) states “...behavior is not random
but purposeful, motivated or expressing needs.” One motivation for movement as stated
by Schoberth (Lueder, 1983) is “Movement serves as a pump to improve blood
circulation.” When circulation is reduced in muscle tissue this can lead to feelings of
discomfort. Discomfort caused by reduced blood circulation would typically be found in
a specific body area such as the buttocks. To alleviate or avoid this discomfort requires a
change in pressure in the buttocks which is achieved through movement. Body
movement could alleviate a general feeling of discomfort experienced by the seated
subject.

Seating comfort studies are used in the automotive industry but there is no
established method for evaluating seating comfort. Some studies use only questionnaires
to rate comfort levels (Hall, 1972). Other studies rely on many types of data such as
pressure distribution, static load and deflection as well as vibration (Kamijo etal., 1982)
to assess comfort. Lee et al., (1995) suggest using pressure distribution,

electromyography (EMG), and spinal loading to measure the physiological effects of

seating. They support in theory, the use of body movement to rate the comfort levels
achieved by vehicle operators in automotive seats.

The primary purpose of the driver’s seat is to support the occupant for vehicle
operation, however, comfort has become a selling point and can affect customer loyalty.
“As the mean comfort of the trip deteriorates, so does the proportion of passengers
willing to use the vehicle again” (Richards, 1980). Since people avoid discomfort, it is
important that the driver finds the vehicle not only operable and safe, but also
comfortable. The driver’s body is in direct contact with and supported by the seat so the
seat should not be a source of discomfort for the driver. Comfort research in automotive
seating also supports an attentiveness to the driving task without a great degree of
distraction from the inability to perform a needed activity due to discomfort (Branton
1969)

There are some major drawbacks to most automotive seat comfort studies. First,
testing is often not carried out in the driving environment thus reducing many of the
stresses of roadway driving (Kamijo, etal., 1982). Reduction of the driving task during
laboratory simulations greatly changes the requirements placed on the driver and thus
changes the driver’s interaction with the seat. F or example, automotive seat tests are
often conducted in a laboratory with the driver’s seat placed on a platform so the effects
of the packaging are eliminated along with effects caused by roadway travel such as
vibration. Even if the interior of an automobile is simulated, the conditions to which a
driver must respond when driving are reduced, or in some cases eliminated. Thus, the

driver interacts with a seat differently when pretending to drive in a laboratory than when

driving on a road. Second, the length of time people sit in an automotive driving seat is
not often recreated in a testing situation, thus eliminating the effects of time on comfort.
Continuously recording the vehicle operator’s body movement eliminates these problems
by allowing for unobtrusive data collection during on-road testing for long periods of
time. O

To continuously record the body movement of vehicle operators, video tape is
useful. Questionnaires capture an instant in time and may not be representative of the
seated occupant’s experience during long term sitting. Video tape records body
movement continuously throughout lengthy testing periods so the whole event can be
understood. A second benefit of recording data on video tape is that it does not directly
interact with the observer. When gathering subjective data with questionnaires, subjects
are often interrupted during their tasks to complete the questionnaire or are asked to
recall information after long periods of testing. These drawbacks are eliminated by
recording body movement with video tape. The third benefit of recording the seated
occupant’s body movement is that data collection can occur while operating the vehicle
on-road.

Sitting comfort is a continuous subjective phenomenon influenced in part by the
person’s contact with the seat. Thus, the purpose of this thesis is to develop a method to
encode body movement of vehicle operators. Body movement offers a means of
collecting objective, continuous, quantitative data for the purpose of investigating long
term sitting comfort as experienced by automotive operators. There are few well

documented methods for recording and encoding body movement and none for the task

of automotive operators. Therefore a methodology for encoding recorded movement of
automotive operators into a meaningful format must be developed. The developed
method will be modified from existing methods for encoding body movement. Once

refined, this methodology will be tested for both inter and intra reliability.

Literature Review

Recording Movement
Atha, (1984) outlines the process involved when analyzing movement:

1. Select the phenomenon to be examined

2. Identify a representative characteristic that can be monitored to give

insight about the phenomenon.

3. Design a suitable transducer that will record the primary information

required to analyze motion.

4. Describe and transform the data recorded into other variables so that all

the desired information may be obtained.

5. Assemble the collected data with information gathered from other

sources so it may be interpreted for the selected purposes.

Atha focuses on instruments to record movement for the sports sciences. Even
though the goals for sitting comfort are different from sports sciences, the approach to the
analysis of movement overlap and the methods of recording movement are the same.
There are two ways in which to record movement:

1. record the movements by hand at the time they occur

2. record the movement on tape and then analyze the tapes.

When movement is to be quantified, devices that allow for the play back of the movement
are optimal since data can be gathered unobtrusively, remotely and at a low cost. The
three general categories of instruments that gather movement data are:

1. Photographic and video, which are remote sensors of movement,

2. Specialized transducers, which convert one form of energy into another
for analysis; and

3. On-line movement analyzers, which are second generation multi-channel
instruments that continuously measure and automatically analyze angular or three
dimensional coordinate data.

These three methods are highly sensitive to different aspects of movement. All
three allow data to be played back as many times as needed for analysis. The mechanical
methods tend to result in less subjective evaluations of data.

Encoding Methods

Once the method for recording movement is achieved, one must encode
the data in a meaningful manner. This is especially true in dance which relies greatly on
the ability to record movement for future dancers. Two highly acknowledged methods of
dance notation are Labanotation developed by Laban, (Hutchinson, 1977) and the Benesh
Movement Notation , (cited in Kember, 1976). Developed separately, these methods
record choreography for dance. Both methods rely on symbolic representations of
movements to record dance positions and direction of movement. Movements are
recorded on a musical staff that is tied into the musical score so time is a dependent
function of the music, (Figure 1). The symbols for both methods are abstract but highly
detailed. Both methods are capable of recording small changes in position required for

dance, such as the position of fingers.

10

 

 

 

 

 

 

 

l + 7/-

 

 

 

 

-

 

 

 

 

Figure 1. An example of the Benesh Notation Method for recording choreography.

11

P. A. Kember, (1976), adapted the Benesh Movement Notation for recording
movement and posture of seated subjects. The five line stave, derived from musical
notation, is retained along with the basic signs developed by Benesh. The posture of the
body during movement is not actually recorded, but is indicated by lines that trace the
path of the body between two different postures. All the movements are recorded
independent of time. A series of signs was specifically developed to indicate the body
segments supported by the seat.

To fully master either the Benesh Movement Notation requires a minimum of
two years of intensive study. Kember estimated that “good” students could master
enough of the Benesh Movement Notation in a rigorous three month study to record
seated body movements. Kember’s article focuses on the method and training for
recording seated behavior so no mention is made of any analysis based on this recording
method.

E. N. Corlett (1979) designed a method for coding frequency, duration and
sequence of standing work postures in industrial tasks based on the standard posture
assumed by the subject. Any changes in the standard posture are considered either
vertical or directional deviations from the standard position. The vertical deviations are
indicated by concentric circles while the directional deviations are marked on the radii of
the diagram, (Figure 2). The portions of the circle may also be assigned numbers. Thus,

each of the body segments is represented by two numbers creating a total of ten pairs of

12

numbers. This method is designed for analyzing a particular movement during a task, not

several movements over an extended period of time.

l3

_ ,.—~.’—\.
E .
I
“fix

Figure 2. Corlett’s chart for recording standing work posture in industry

l4

Corlett realized that recording only the typical posture of a worker was not
enough information to assess postural work loads so he developed a 5-point discomfort
rating scale. When gathering information, subjects are first asked to indicate their
discomfort level on the 5-point scale. Then the subject is given a chart of the body
broken into segments and is asked to indicate all painful segments listing the most painful
and proceeding to the least painfirl. The purpose of this questionnaire is to gather
subjective information on bodily discomfort.

Testing and Application

An important part of recording movement through human observation is to
achieve a high level of inter-observer reliability among observers. Genaidy et al., 1993,
studied the amount of inter-observer reliability in subjects recording the estimated
amount of angular deviation of a body segment from the neutral position. Data were
recorded on video ensuring consistency in angular deviation for the encoders. Twenty
engineering students were chosen for the study and had no previous training in recording
postural angles by sight. All twenty subjects were briefed in the objective of the study
and experimental procedures were explained. A male engineering student modeled the
different shoulder angles in a standing position. Each subject was shown a video tape of
the model who held each pose for 30 seconds. There was a pause of 30 seconds between
postures. The subjects viewed a total of four video tapes, each with 18 positions. The

order of the video tapes was randomized.

15

Results for this study show that there was no significant difference in error
between the low, medium and high positions. The average perceived angle demonstrated
that subjects without prior experience tended to overrate low level positions and underrate
high level positions. The average geometric error was -1.3 degrees. In conclusion future
studies are recommended to compare the abilities of engineers to non-engineers when
estimating body segment angles.

Branton & Grayson (1967) coded subject behavior to detect and quantify the
differences in subjects’ posture while riding on a train. Two types of seats, sofi and firm,
were installed to record subjects postural response to difference in support. A numerical
system was created for coding static posture. A majority of the coding took place in real
time with researchers recording over 5000 real-time observations of 682 female subjects
and 818 male. while riding the train with the subjects. The second technique for recording
data consisted of filming 18 pre-selected subjects who were recorded continuously
throughout the five hour journey. . These subjects were instructed to move about the train
normally. The cameras were set with a time lapse unit exposing one frame every 10
seconds. After subject data were recorded on film, body posture was coded using the
same numerical coding as used on movement coded from real time subjects. High
correlations were found between the two techniques.

Branton and Grayson’s method for encoding the behavior of seated subjects on
the train recorded the subject’s use of specific seat features to support the subject’s
desired posture. A series of numbers coded the subject’s use of the headrest, armrests,

backrest, and the floor to support the body. This coding method divided the body into the

16

following regions, the head, arms, trunk and legs. All recording was done with respect to
time. The Branton and Grayson method recorded only static postures. Movement made
while moving from one posture to another was not recorded.

To quantify differences in the subjects’ postures, the frequency and duration of
the postures were calculated from recorded data. The frequency of use for particular seat
features was also calculated. An attempt was made to look at subjective comfort levels in
conjunction with subject behavior. Unfortunately there was not enough time to
administer all the questionnaires. From the information that was collected Branton and
Grayson stated that “the difference in (subject) attitudes (towards the seats) were not as
great as the differences in behavior observed.”

As shown by the collection of articles presented here, recording and encoding
body language is not a new concept, but it has not been frequently used in the scientific
investigation of seating comfort. As stated earlier, the purpose of this paper is to develop
a method to encode body movement in a meaningful format that will allow for future

investigation of sitting comfort.

Methods and Procedures

Methods
Notes from video tapes of sixteen subjects driving an automotive vehicle served as
the basis for the development of terms that define information to be encoded from body
movement. The following is a list of the boundaries, terms, and some basic insight on the
thought process for encoding body movement from video tape.
Two types of movement will be encoded:
a. Movement that occur when the subject alters the environment

1. radio

2. temperature control

3. seat position/steering wheel position

(any time the subject removed his or her hand from a
control, a second change in the environment was
recorded)
b. Movement that occurs as the subject changes position or posture in the
seat independent of a change in the environment.

This movement is further broken down into in-seat movement or out-of-seat movement.

1. Out-of-seat movement is defined as any movement that lifts the

body out of the seat.
2. In-the-seat movement is any movement that occurs in the
seat:

a. Pushing the body segment into the seat

17

18

b. A movement in which the body stays in contact with the
seat, e. g. sliding on the seat or squirming.

If the movement ended in the same posture, but there is no break in the movement then it
is counted as one movement. If the movement ended in the same posture but there is a
break in the movement, then it will be counted as two movements. Movements made in
order to operate the vehicle, such as moving the foot from the gas pedal to the break
pedal, are not recorded since they are required of the subject to safely operate the vehicle
and thus are fairly uniform for all automotive vehicle operators.

Some changes in posture are very obvious while others are small and/or subtle.
Detection of the small, subtle movements can be made by noting changes in background
shapes. This is especially true if the subject’s clothing contrasts with the interior color of
the car. For example if the subject wears dark clothing and the interior of the car is a
lighter color this creates a contrast with the dark fabric. By looking for changes in the
amount of the lighter background it is possible to tell if movement has occurred.

To study the effects of time on comfort requires that time be recorded in a systematic
method:

1. To permanently locate the body movement in the video tape, the frame number
is recorded at the first frame that a change in body position occurs.

2. The length of the movement begins as identified in #1 and continues until the
subject ceases to move. The time for this event is recorded from the VCR

display of internal tape play time.

19

To study the interaction between the subject and the seat, the first step is to
determine which body segments best describe this interaction. Body segments in direct
contact with the seat pan and seat back have been selected. These segments are further
divided symmetrically into the right and left sides of the body (Figure 3). The body
segments are:

1. Right shoulder region

2. Left shoulder region

3. Back

4. Right buttock

5. Left buttock

6. Right thigh

7. Left thigh
The following is a list of information that comes from the encoding process:

1. Segment of the body that moved

2. Side of the body that moved, (right or left)

3. Direction of the movement in a vertical plane or horizontal plane.

4. Time in relation to the drive that the movement occurred

5. Duration of the movement
From this information, one can also calculate the frequency of movement in a particular
body segment as well as look for patterns in the data which are created over time. Since
this method for recording movement is new, data are needed to test for intra-observer

error and inter-observer error.

20

Defined body segments for recording body movement.

Eigure 3.

21

To test the method a total of six drives performed by three subjects were viewed
in their entirety. The primary observer viewed all the tapes once and then the first drive
for each subject was encoded a second time by the primary observer to test for intra-
observer error. Only three drive subjects have been selected due to the time involved
with viewing and encoding the three hour video tapes of vehicle operators and due to the
quality of the tapes. These three subjects each have two complete three hour drives
recorded on video tape.

Inter-observer reliability was tested after a second observer is instructed in the
encoding method. Instruction consists of viewing two pages of notes that describe the
encoding method. For training purposes the second observer viewed two hours of video
tape not used in the analysis with the primary observer. The two observers critiqued
subject movement and then discussed the method for encoding. To test inter-observer
repeatability the second viewer encoded drive tapes of the three subjects. The first and
second drive tapes for each of the three selected subjects were viewed individually in
their entirety by both observers at separate times.

Procedures

To test the newly developed method previously described, data collected during a
fall, 1993 comfort study drive in an on-road automotive vehicle laboratory will be used.
Subject body movement was collected continuously throughout the three hour drive in a
1992 mid-sized vehicle with a NEC TI-24A CCD camera fitted with a F ujinon auto-iris
1:1.8/4.8 mm lens. The camera was mounted half way up the right B pillar, (the B pillar

supports the roof and separates the front and back door) to completely capture the

22

subject’s body movement independent of seat placement. Images collected by the camera
were recorded on VHS video tape using a Toshiba video cassette recorder model M-449.

Each subject participated in the three hour, non-stop, test drives which were
repeated one week after the first drive to determine the amount of repeatability between
the first and second drive. The subjects were paid a flat rate for their efforts after the
completion of two drives. Subjects were required to fill out a form indicating their level
of physical activity, general health, distance of typical travel in an automotive vehicle,
make of their vehicle and it seat features. Before driving, the subjects were informed of
the video camera recording their movement continuously during the drive. Subjects were
instructed to drive during testing as they would in their own vehicles. Spandex bike
shorts/running pants and cotton tank tops were required clothing for all subjects during
the test drive so body movements would not be hidden by bulky or loose fitting clothing.
It is important to note that subjects had full control over air circulation and temperature
controls in the vehicle to help reduce thermal discomfort due to the nature of the clothing
and the weather conditions. This also gave subjects a level of control over their
environment in a testing situation. None of the subjects complained of any thermal
discomfort.

During the drive other data were collected which created an environment different
from daily driving. Subjects wore 17 dime sized surface electrodes attached to four
major muscle regions on the right and left sides of their necks and backs to measure
muscle activity. Nine quarter size reflective targets marked the centers of joint rotation

on the right side of the body, along with the right anterior superior iliac spine of the

23

pelvis, the suprastemale and the mesostemum. These targets were photographed
periodically throughout the drive by four NEC TI-24A CCD cameras mounted on the
right side of the car. The cameras were visible to the subjects, but none of the subjects
said they were bothered by their presence. The camera on the right B pillar collected both
the anthropometric data and the continuous body movement data throughout the drive.
This camera was chosen to use for collecting body movement data for two reasons. First,
the image collected by camera four was a straight side view of the body that allowed the
observers to determine vertical movement. The other three cameras captured images of
the subject at angles and thus did not allow for the observers to determine horizontal
movement. Second, camera four completely captured the subject body regardless of
subject size or seat position while the other cameras only captured regions determined by
the use of video anthropometry. The collection of the video anthropometric images did
not interfere with the continuous collection of body movement on video tape. There was
no indication from the subjects or in the data that the testing procedures or equipment,
such as the cameras altered regular driving habits.

The driver’s seat in the test vehicle was a multi-adjustable seat. Once the subject
was seated in the car, ride technicians demonstrated the seat controls. The subjects were
then asked to operate each control and then adjust the seat to a comfortable position.
Subjects were informed that they could move the seat at any time during the drive.

During the drive a ride technician sat in the back seat on the left side and collected
data at set intervals with the aid of a 486 computer mounted in the trunk. This did not

interrupt the continuous recording of body movement. Subjects indicated that they were

24

not interrupted by the data collection with the exception of the verbal questionnaire. The
verbal questionnaire was administered at half hour intervals throughout the drive and
required that the subject answer questions asked by the ride technician by speaking into a
tape recorder.

Drives of three subjects, two males and one female, have been selected based on
the completeness of tapes for both the first and second drive for each subject. A total of
6, three hour tapes were viewed in their entirety and information encoded by the primary
observer. A second observer was trained to use the encoding procedures. After training
the second observer viewed and encoded the two tapes for each of the three subjects,
encoding a total of six video tapes.

As stated earlier, the purpose of this thesis is to further develop a method to
encode the body movement of automotive vehicle operators in a meaningful format
allowing for further investigation of automotive operator comfort as related to the seat.
Tapes of three hour drives will be viewed in their entirety using the developed method to
encode the body movement observed on the tapes. Statistical testing for inter and intra
observer reliability will be conducted. Based on the information gathered and the results
of the analysis, suggestions are made for the future study of body movement as related to

the comfort level of the automotive vehicle operator in relation to the seat.

Results

Correlations were performed for drives 009, 025 and 031, each of which was
performed by a different driver. Each of these drives were encoded three times and
therefore can be checked for both inter and intra-observer reliability. Data were grouped
based on the number of movements recorded for each body segment during the entire
three hour drive period. The data is ordinal so a Spearrnan correlation is used. The

results are in Tables 1, 2, and 3.

Tablel u‘lgu 3‘11“. Co ‘a‘,, 11‘ o 1'93“. uo WV. '0 Dr ‘ 0"
Observer 1, Trial 1 Observer 2, Trial l

Observer 1, Trial 1

Observer 2, Triall .400

Observer 1, Trial 2 .927 .444

TableZ Il‘e'tg‘h“l.l“‘ 0 1‘0, .n U0 ‘u‘n I" ‘0

r.
.

Observer 1, Trial 1 Observer 2, Trial 1
Observer 1, Trial 1
Observer 2, Trial 1 .438

Observer 1, Trial 2 .585 .388

25

26

Table3 or 1..:-‘A..l..,-‘. .0 i' o 1'0 no u. ‘H‘l ' Ir
Observer 1, Trial 1 Observer 2, Trial 1

Observer 1, Trial 1

Observer 2, Trial 1 .929

Observer 1, Trial 2 .955 .847

Cumulative frequency graphs show the accumulated movement for each body

segment per 15 minute interval. When plotted, (Figures 4, 5 & 6) the results demonstrate

that each observation mirrors the other two. These charts represent the total number of

movements over time.

 

 

 

 

 

Total # of Movements

 

 

 

0‘ i r M L

O 15 30 45 60 75 90 105120135150165

 

Time Interval (min.)

 

A+ (5632:7ver 1,ATFiaI1 +Observer2. Trial 1 —_,,:6bser&é?mra

 

 

 

 

Figural. Cumulative frequency chart of subject movement recorded in Drive 009

 

 

Total # of Movements

 

 

 

 

 

 

+ Observer? Trial-1 + Observer 2. Trial 1 + Observe-r1, TrialA2]

Time Interval (min.)

 

 

 

 

 

 

 

100

 

 

90

70
60 .
50
40
30
20 4

Total # of Movements

80 L -_

 

 

 

' _._AAObse7vAer 1,'ArFial CF Observer 2, Trialfl1 f; ObsAéArVér 1, Tfiéi'f ”

 

 

LL L 4

15 30 45 60 75 90 105120135150165 1

Time lnte rval (min.)

 

__ _J

 

 

 

Em. Cumulative frequency chart of subject movement recorded in Drive 031

28

Body movement charts also show the movement of each body segment in relation
to the 15 minute time period as well as the direction of the body movement in relation to
the seat, (Figures 7, 8, and 9). For example, any movement that raises the body out or
away from the seat is marked above the darkened horizontal line. Movement that pushed
the body segment into the seat or slide the segment across the surface of the seat is
recorded below the horizontal line. Movement is recorded in the order in which it
occurred. For example, movement marked closest to the horizontal line occurred first in
the 15 minute time period. Movement furthest away from the line, either above or below
the line, was the last to occur in the 15 minute period. Movements made to alter the
environment, such as adjusting the heat, were recorded and are shown coded as shades of

gray in the second, narrower column.

Legend for Body Movement Charts

1A = Right Shoulder 4A Right Thigh

18 = Left Shoulder 48 Left Thigh

2 = Back Thermal Adjustment

3A = Right Buttock lRadio Adjustment

 

38 = Lefi Buttock Seat Adjustment

29

 

 

 

 

 

 

 

 

 

 

 

 

o 15 30 45 60 75 90 105 120 135 150 165
2
2
4B 2 2
4B 2 4B 33 4B
2 2 4A 4B 2 33 3A 2
2 2 4B 4B 48 3A 2 4B
48 4A 2 4B 2 2 2 2 2 2 4B 33
2 4B 4B 2 2 2 4d 2 2 2 2 3A
4B 4B 4B 4B 4B 2 48 1A 1A 4B
2 4B 4B 4B 4B 4B 13 13
4B
2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Out of
Seat

In Seat

Eiguml. Body movement by body segment over time, Drive 009, Observer 1, Trial 1

30

 

 

lo 15 3o 45 60 75 90 105 120 135 150 165
33
3A
4B 4J4aqz 2
4B 2
2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

lln Seat

 

 

W- Body movement by body segment over time, Drive 025, Observer 1, Trial 1

31

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I0 15 30 45 60 75 90 105 120 135 150 165
2 18
13 1A
1A 2 4B
1B 4B 1B 48
1A 13 1A 4A
18 1A 38 4A
2 18 1A 3A 3B Outof
18 1A 1B 3B 48 seat
2 18 1A 3A 3B
4B 18 1A 38 3A 3A
2 -1B 1B 1B 4B . 3A 3A 2
2 1A l 1A 1A 1B 2 2 4A
1A 1A 48 1A 2 4B 48 4A
18 13 1A 18 1A 4B
18 1A 1A 4B
1B 18 4A
1A 1A 4A lnSeat
1B 1B 4B
43 1A
1B

 

 

 

 

 

 

 

 

 

 

 

 

 

 

W. Body movement by body segment over time, Drive 031, Observer 1, Trial 1

Discussion

The purpose of this thesis was to develop a method to encode body movement.
An important part of methodology development is to test the reliability of the method.
The correlations of the three drives as shown in the results' section are explained here.
Only three drives were observed for both inter and intra-observer reliability due to the
time intensive encoding process. On average it took the viewer 6 to 8 hours to encode
body movement from the video tapes.

The largest difference between the correlations for inter-observer and intra-
observer reliability was found in drive 009. The intra-observer correlation was high but
the inter-observer correlations were relatively low. Since this was the first of the three
drives to be encoded, this suggests the presence of a learning curve for the second
observer during the encoding procedure. A higher correlation for the intra-observer
reliability is attributed to the fact that the primary observer also developed the method for
encoding the data and was thus familiar with the procedure. The average number of body
movements recorded in drive 009 was 76. The range of encoded body movements
between viewers was 15. It is important to note that most of the difference in the of body
movements recorded occurred during the 75 minute to 120 minute period. This
difference may be due to observer fatigue.

Drive 025 showed the lowest correlation between both inter-observer and intra-
observer reliability. The lower correlations are attributed to the lack of body movement
exhibited by the drive subject. The average number of body movements encoded for

drive 025 per viewing was 18. The minimum number of body movements encoded was

32

33

14 while the maximum number was 21. The low number of total body movements
magnifies any difference in the encoding process thus lowering the correlations.

Drive 031 had the highest correlations for both inter-observer and intra-observer
reliability. This is due to two factors. First, drive 031 had the highest number of body
movements encoded averaging 93 per viewing. Second, the range of recorded body
movement was 5, the least of all three drives. Drive 031 was the last drive to be encoded
supporting the presence of a learning curving, especially for the second observer.

These correlations suggest inter-observer reliability and intra-observer reliability
so the method developed for encoding body movement from video tape is consistent.

The method was developed to investigate the type of information that the body movement
of a vehicle operator can give in relation to comfort. This information is depicted
graphically to show trends and patterns in the data over time. Data used for the
correlations are found in the Appendix. When studying the appendix it is important to
note the robustness of the data as well as the potential value. From this data comfort
profiles establishing a baseline of movement for each subject could be determined.

Cumulative frequency graphs show the number of body movements made by each
driver with respect to time. Body movements are grouped into 15 minute increments.
This graphical representation shows that each body movement was recorded at relatively
the same time by each observer. The graphs also show periods of more intense
movement. For example during drive 031 the driver moved the most during the 30 to 45

minute time period and the 135 to 150 minute period. Only the total number of body

34

movements per 15 minute period is presented in this graph. Specific body regions and
the direction of the movement are contained in the other graphs.

The time of each body movement in relation to the three hour drive is shown on
the cumulative frequency graphs. Time for each body movement was recorded off the
VCR. Lack of a set zero when recording the video tape made it difficult for the observers
to synchronize time. Constant rewinding of the tape also altered the time according to the
VCR by tenths of a second. This may have been due to a stretching of the actual tape.
Some differences in the number of movements per 15 minute period may be due to these
timing errors. For future encoding a continuous time stamp on each frame of the video
tape is recommended. Overall the cumulative frequency graphs show that body
movement when recorded with respect to time is consistent between viewers.

The body movement graphs show body movement in relation to the three hour
drive. The differences and patterns of movement in the level of body movement changes
over time creating a pattern. Body movement is made to maintain a level of comfort,
alleviating discomfort. Drive 031 distinctly shows increased levels of movement at the
90 minute and 150 minute interval with no movement in the following two 15 minute
time periods. This may indicate that through body movement the subject was able to
decrease the level of discomfort leading to periods of comfort which in turn are followed
by periods of great amounts of movement towards the end of the drive. These particular
charts seem to indicate that comfort levels are not constant but vary over time. This
supports the idea that sitting comfort is variable over time. Changes in comfort depicted

in the graphs might not be detected with pre- and post- evaluation questionnaires.

35

Body movement graphs show frequent movement in particular body segments
throughout the drive. This might lead to a direct application of the body movement
encoding tool. For example, in drive 031, the subject moves the right and left shoulder
much more than any other body segment, (see Appendix). High frequencies of
movement in the shoulders might indicate a level of discomfort in the shoulders. In the
design context of the seat, this would indicate that the current design of the shoulder
region in this particular seat aggravated the subject and needs to be studied and possibly
redesigned. Recording body movement in this way creates a valuable tool with which
seat designers can evaluate prototype seats.

Individual differences are reflected in both sets of graphs. Just as people have
different personalities, they also have different thresholds for discomfort which are
exhibited over time through different frequencies of movement Discomfort beyond the
subject’s threshold in situations where the subject is unable to leave the environment is
ofien manifested in movement to minimize the level of discomfort. Through movement,
subjects redistribute their weight or change their posture in the seat, attempting to
minimize the level of discomfort, possibly obtaining a level of comfort. Differences in
personal discomfort thresholds may be due, in part, to body size, level of physical and
mental well-being and tolerance of the discomfort sensations. For subject demographics
see Table 4. When personal discomfort thresholds are surpassed movement occurs in
attempts to change the body’s interaction with the seat.

It seems possible to establish a baseline discomfort threshold for an individual.

For example, an average of 18 body movements were recorded for the subject on drive

36

025 while an average of 76 body movements were recorded during drive 009. Drive 031
averaged 94 body movements during the three hour period. This does not necessarily
mean that the subject in drive 031 experienced the highest level of discomfort. Instead,
the high number of movements might indicate that this subject moves frequently, even
when comfortable. Further investigation of body movement includes establishing a
baseline of movement for each subject in a comfortable environment. Movement above
the baseline would then be attributed to levels of discomfort. This information would be

especially valuable for comparison studies of seats.

Table 4 Demographicmflthflrixuuhjccts

Sub. 1, Drive 009 Sub. 2, Drive 025 Sub. 3, Drive 031
Sex Male Male Female
Age 26 22 43
Stature 173.4 cm 177.5 cm 159.2 cm

Weight 80.8 kg 75.4 kg 51.9 kg

Future Study

Recommendations for future study include tracking and correlating changes in
seat position with body movement. Seat position was recorded in this study, but due to
the absence of a set zero time mark for both the body movement data and the seat
position data there was no way in which to accurately relate the two. The addition of seat
position data adds a third dimension to determining comfort levels through the body
movement of a subject. A change in seat position changes the position of the body as
well, thus creating body movement. By changing seat position, an immediate
environmental element that the subject is reacting to is also altered. Changes in seat
position could affect the frequency and range of movement as well as the body segment
moved. Thus one could study patterns of change in the position of the seat in relation to
body movement. Perhaps altering the environment decreases the level of discomfort
experienced by the subject.

Future studies of body movement should also include other measures of
movement and discomfort such as pressure mats and questionnaires. Pressure data
gathered from pressure mats in the seat could be collected continuously. High peak
pressures indicate that body mass is not well distributed in the seat pan causing
discomfort in the high peak areas. Low peak pressures with well distributed body mass
usually indicate a more comfortable seat. Pressure data would show changes in pressure
distribution caused by body movement. Pressure data also picks up slight changes in
pressure distribution caused by subtle movement that can not be detected by the human

eye. Recording changes in pressure should allow for a direct comparison with body

37

38

movement as encoded by humans provided a mutual start time is defined. If successful,
pressure could possibly enhance body movement data by detecting shifts in weight from
movement that is too small to be noticed on video tape.

Encoding body movement from the video is time consmning and expensive.
Future plans for the study of body movement include converting the method developed in
this study to a computer program to make the encoding process less time consuming and
less expensive. Computerizing the encoding process changes the evaluation of the
movement as encoded by humans in two ways. First, the accuracy in the detection of
movement frequency will be improved by automating the system, thus removing human
error. The length of the tapes often causes viewer fatigue which affects the accuracy of
the encodement. Second, judgments made intuitively by humans about movement
detected in the video tapes might be lost in the conversion to automation by computer.
For example light levels in the car change continuously due to weather conditions,
position of the sun and direction of travel. Humans are able to distinguish changing
shadows from human movement. Currently computer vision is not powerful enough to
distinguish the difference. Changes of scenery in the windows might also be confused as
body movement by the computer. These problems might by solved by tinting the
windows or adding more light to the interior.

Though specifically designed to encode the vehicle operator, the method
described in this paper could be adapted to other long term seated tasks. This method

tracks continuous movement over long periods of time, so with modification it could be

39

used for evaluating long term sitting required for such tasks such as data entry or word
processing.

Analyzing long periods of nonverbal behavior is time and labor intensive, time
constraints did not allow for a large enough sample size for more sophisticated statistical
analysis. Further statistical testing should be conducted to better understand the body

movement data prior to creating computer applications for encoding body movement.

Conclusion

The main objective of this study was to determine if encoded body movement of
the vehicle operator would provide information regarding comfort. The first step was the
development of a reliable method to encode body movement fi'om video tape over long
periods of time. The use of two observers demonstrated reliability of the method. Both
cumulative frequency graphs and body movement graphs show interesting patterns of
movement. These patterns suggest that comfort does change over time. When body
movement can be encoded with less effort, the result will be objective data that can be

quantitatively analyzed to further understand the changes in comfort experienced during

long term seating.

40

Appendix

Body movement as recorded by each observer with respect to body segment and time.

Table 5 WWW

Time Right Left Right Left Right Left
Shoulder Shoulder Back Buttock Buttock Thigh Thigh Totals

—L

1 2

NNNNO'I-th—lww

 

(a)
C

Table 6 WWW

Time Right Left Right Left Right Left
Shoulder Shoulder Back Buttock Buttock Thigh Thigh Totals

2 2 4

 

41

42

Table 7 WWW

Time Right Left Right Left
Shoulder Shoulder Back Buttock Buttock Thigh Thigh Totals

1 2

finnmnmANw—swwa

 

Table 8 WWW

Time Right Left Right Right
Shoulder Shoulder Back Buttock Buttock Thigh Thigh Totals

1

—l

 

43

Table 9 W5

Time Right Left Back Right I Left 1 Right Left Totals
Shoulder Shoulder Thigh Thigh

Nmo—L—soatroooa.‘

.3
A

 

Table 10 WW

Time Right Left Right Left Right Left
Shoulder Shoulder Back Buttock Buttock Thigh Thigh Totals

1

—l

2
1
0
0
5
1
3
3
1
2
3

N
N

 

44

Table 11 WW

Time Right Left Right Left Right Left

Shoulder Shoulder Back Buttock Buttock Thigh Thigh Totals

2 2

4

2

6

19

7

11

11

0

0

19

15

96

Table 12 3 I 0.

Time Right Left Right Left Right Left
Shoulder Shoulder Back Buttock Buttock Thigh Thigh Totals

5

1

 

45

Table 13 WW

Time Right Right Left Right Left
Shoulder Shoulder Back Buttock Buttock Thigh Thigh Totals

2 2

5
2
6

 

Bibliography

Atha, J. (1984) Current Techniques for Measuring Motion. AppliedErggngmigsJiA,
245-27.

Branton, P., & Grayson, G. (1967) An evaluation of train seats by observation of sitting
behavior. Erggmgmigs, _l_Q, 35-51.

Branton, P. (1969) Behavior, Body Mechanics and Discomfort. mm, 12,
316-327.

Corlett, E. N. (1979) Posture, it’s measurement and effects. University of Birmingham

Genaidy, A.M., Simmons, R.J., Guo, L. & Hidalgo J.A. (1993). Can visual
perception be used to estimate body part angles? Erggngmigs, 36(4),
323-329

Hall, M. (1972). Back pain and car-seat comfort. ApplieiErangmics, 3(2),
82-91

Hecht, J. A, & DeVito, J. A. (1990) Perspectives on Nonverbal Communication. In J. A.

DeVito and RE Denton (Ed) Ihelzlonverbalficmmunicatimlteadermp. 3- 18).
Prospect Heights, IL: Waveland Press, Inc.

Hutchinson, A. (1977). , 2 '
(3rd ed. ). New York: Theater Arts Books.

 

Jurgens, H. (1980). General reflections on the postural comfort of the driver and
passengers; consequences on seat design. In W

e. . .
21.21.! .1\ 2 1 1 I12 .12 .1 1 11 '11111 2_1'

Transportation. (pp. 249- 256). Academic Press

Kamijo, K. ,Tsujimura, H, Obara, H., & Katsumata, M. (1982). Evaluation of seating
comfort. W, Technical Paper 820761

Kember, P. A. (1976) The Benesh Movement Notation used to study sitting behavior.
Appliedlirgemmies, 1.3, 133- 136.

Lee, J., Grohs, T., & Milosic, M. (1995) Evaluation of Objective Measurement

Techniques for Automotive Seat Comfort, SggiefiLQLAmgmmierngineers
Technical Paper 950142.

Leuder, R.K. (1983). Seating comfort: a review of the construct in the office
environment. Appligifirggngmics, 25(6), 701-711.

46

47

Montepare, J. M., Goldstein, S. B. ,& Clausen, A. (1987). The Identification of
Emotions from Gait Information. In J. A. DeVito and R. E. Denton (Ed)
IheNonverbaLQommunicationReader (pp 78- 85) Prospect Heights, IL:

Waveland Press, Inc.

Rebiffe, R. (1980). General reflections on the postural comfort of the driver and
passengers; consequences on seat design. In W

21 11l211\ '2 1 1 11211211"1 11, '11111 2.11.

mm (pp. 240- 248). Academic Press

Richards L. G. (1980). On the psychology of passenger comfort. In Emceedmgmflhg

.1121 1 12.1.211l 11\"2 1 1 11211211“1’11

Ergonomiesandlransmrtation. (pp. 15- 23). Academic Press

Shen, W, & Galer, I. (1993). Development of a pressure related assessment model

of seating discomfort WWW
ErgonomicsSociebLMeeting (pp. 831 -.835)

Weitz, S. (1979). Body Movement and Gesture. In S. Weitz (Ed), NguyerhaL
ngmumgafim (2nd ed. pp. 87-111). New York: Oxford Press.

Zimmerman, C.L., Cook, T.M., & Goel, V.K. (1993). Effects of seated posture on
erector spinae EMG activity during whole body vibration. Erggngmics, 36(6),
667-675.

nrcnran STATE UNIV. LIBRQRIES
lllllllllllllllllllllllllllllllllllWill111111111111
31293014100501