Are you sure?
This action might not be possible to undo. Are you sure you want to continue?
COMPUTATIONAL MODEL TO DETERMINE TIBIOFEMORAL FORCES AND
MOMENTS DURING KNEELING
by
Jonisha P. Pollard
B.S., University of Pittsburgh, 2005
Submitted to the Graduate Faculty of
the Swanson School of Engineering in partial fulfillment
of the requirements for the degree of
Master of Science
University of Pittsburgh
2008
ii
UNIVERSITY OF PITTSBURGH
SWANSON SCHOOL OF ENGINEERING
This thesis was presented
by
Jonisha P. Pollard
It was defended on
November 11, 2008
and approved by
Jean L. McCrory, PhD, Research Assistant Professor, Health and Physical Activity
Richard E. Debski, PhD, Associate Professor, Department of Bioengineering
Rakié Cham, PhD, Associate Professor, Department of Bioengineering
Thesis Advisor: Mark S. Redfern, PhD, Professor, Department of Bioengineering, Associate
Dean for Research
iii
Copyright © by Jonisha P. Pollard
2008
iv
Osteoarthritis affects more than 27 million Americans and cost nearly $5700 per person
annually. [1, 2] It commonly affects the knee joint and has been linked to work involving
prolonged knee bending. [3, 4] In restricted vertical working heights such as lowseam coal
mines and aircraft baggage compartments, workers are forced to assume stooped, kneeling, or
squatting postures to perform work. In order to protect the knees in these postures, we must have
an understanding of what the internal knee structures experience under these conditions. A finite
element model is being developed to quantify the stresses and strains in the tissues in static
kneeling postures. The accuracy of any finite element model is heavily dependent on the input
parameters (i.e. forces and moments). Therefore, the objective of this work was to develop a 3D
computational model which may be used to determine the net forces and moments applied to the
knee joint during static kneeling. The developed model uses inverse dynamics to determine the
net forces, net moments, and joint angles for subjects while kneeling near full flexion, kneeling
on one knee, kneeling near 90° flexion, and squatting. Motion data, ground reaction forces, and
pressures between the thigh and calf and heel and gluteal muscles were inputs into this model.
Additionally the thighcalf contact force, which was shown to be significant [5], and the heel
gluteus contact force, which had not been previously investigated, were inputs in this model.
Data from two subjects were analyzed with and without the subject wearing kneepads. Kneeling
near full flexion and squatting created sagittal joint moments 3 to 5 times larger than standing in
COMPUTATIONAL MODEL TO DETERMINE TIBIOFEMORAL FORCES AND
MOMENTS DURING KNEELING
Jonisha P. Pollard, M.S.
University of Pittsburgh, 2008
v
one subject. Moments of this magnitude may be significant to cause cartilage damage. It was
also found that the moments caused by the thighcalf and heelgluteus contacts act to extend the
knee, thereby reducing knee moments in fully flexed postures.
vi
TABLE OF CONTENTS
NOMENCLATURE ................................................................................................................. XIV
1.0 INTRODUCTION ........................................................................................................ 1
1.1 BACKGROUND .................................................................................................. 2
1.2 KNEE STRUCTURES AND INJURIES ........................................................... 2
1.3 OCCUPATIONAL KNEE INJURIES .............................................................. 5
1.3.1 Bursitis .............................................................................................................. 5
1.3.2 Meniscal Injuries ............................................................................................. 5
1.3.3 Osteoarthritis ................................................................................................... 7
2.0 MOTIVATION ............................................................................................................. 9
2.1 PREVIOUS MODELLING OF KNEE FORCES AND MOMENTS ............ 9
2.2 SPECIFIC AIMS AND SIGNIFICANCE ....................................................... 13
2.2.1 Specific Aim #1............................................................................................... 13
2.2.2 Specific Aim #2............................................................................................... 14
2.2.3 Specific Aim #3............................................................................................... 14
2.2.4 Future Significance ........................................................................................ 15
3.0 RESEARCH DESIGN AND METHODS ................................................................ 16
3.1 EQUIPMENT ..................................................................................................... 16
3.1.1 Laboratory Equipment ................................................................................. 16
vii
3.1.2 Computational Model.................................................................................... 17
3.1.2.1 Model Assumptions ............................................................................. 18
3.2 SUBJECT TESTING......................................................................................... 19
3.2.1 ThighCalf and HeelGluteus Contact Measurements ............................... 19
3.2.2 Subject Instrumentation ............................................................................... 19
3.2.3 Experimental Data Collection ...................................................................... 21
3.3 DATA ANALYSIS ............................................................................................. 23
3.3.1 Construction of Coordinate Systems ........................................................... 23
3.3.2 Joint Angle Estimation .................................................................................. 28
3.3.3 Joint Forces and Moments ............................................................................ 29
4.0 RESULTS ................................................................................................................... 34
4.1 THIGHCALF AND HEELGLUTEUS CONTACT FORCES ................... 34
4.1.1 Near Full Flexion ........................................................................................... 34
4.1.2 Squat ............................................................................................................... 36
4.2 KNEE ANGLES................................................................................................. 37
4.3 KNEE FORCES ................................................................................................. 39
4.4 KNEE MOMENTS ............................................................................................ 41
4.5 SUMMARY SUBJECT 1 .................................................................................. 44
4.6 SUMMARY SUBJECT 2 .................................................................................. 45
4.7 STATISTICS ...................................................................................................... 47
5.0 SENSITIVITY ANALYSIS ....................................................................................... 51
5.1 VARYING MODEL PARAMETERS ............................................................. 51
5.2 VARYING KNEE JOINT CENTER LOCATION ........................................ 60
viii
6.0 MODEL VALIDATION ............................................................................................ 70
7.0 DISCUSSION ............................................................................................................. 72
7.1 KNEEPADS........................................................................................................ 72
7.2 MODEL SENSITIVITY ................................................................................... 73
7.2.1 Model Parameters .......................................................................................... 73
7.2.2 Knee Joint Center Location .......................................................................... 74
7.3 SUBJECT VARIABILITY ............................................................................... 75
7.4 LIMITATIONS .................................................................................................. 76
7.5 IMPLICATIONS OF FINDINGS .................................................................... 79
7.5.1 Muscles Activity ............................................................................................. 80
7.5.2 Ligament Recruitment .................................................................................. 80
7.5.3 Meniscal Loading ........................................................................................... 82
7.5.4 Osteoarthritis Progression ............................................................................ 83
7.5.5 Postures Associated with Osteoarthritis ...................................................... 83
7.6 ADVANCEMENTS ........................................................................................... 84
8.0 FUTURE WORK ....................................................................................................... 86
9.0 CONCLUSION ........................................................................................................... 87
APPENDIX .................................................................................................................................. 89
BIBLIOGRAPHY ..................................................................................................................... 107
ix
LIST OF TABLES
Table 1: Average knee angle (degrees), Subject 1 ........................................................................ 37
Table 2: Average knee angles (degrees), Subject 2 ...................................................................... 38
Table 3: Average net external forces normalized by body weight, Subject 1 .............................. 39
Table 4: Average net external forces normalized by body weight, Subject 2 .............................. 40
Table 5: Average net external knee moments normalized by Body weight*Height, Subject 1 ... 42
Table 6: Average net external knee moments normalized by Body weight*Height, Subject 2 ... 42
Table 7: Resulting pvalues for ANOVA, Subject 1 .................................................................... 48
Table 8: Resulting pvalues for ANOVA, Subject 2 .................................................................... 49
Table 9: Resulting pvalues for ANOVA of Kneepad*Posture Interaction, Subjects 1&2 .......... 49
Table 10: Resulting pvalues for ANOVA, Subject comparison .................................................. 50
Table 11: Sensitivity of sagittal moments to changes in model parameters for all postures ........ 58
Table 12: Percent change in moments due to varying KJC for kneeling near 90 degrees flexion 62
Table 13: Percent change in moments due to varying KJC for kneeling on one knee ................. 64
Table 14: Percent change in moments due to varying KJC for kneeling near full flexion ........... 66
Table 15: Percent change in moments due to varying KJC for squatting ..................................... 68
Table 16: Sensitivity of moments to KJC location for all postures .............................................. 69
x
LIST OF FIGURES
Figure 1: Front view of knee anatomy [11] .................................................................................... 3
Figure 2: Anatomical marker set ................................................................................................... 20
Figure 3: Measured marker set ..................................................................................................... 20
Figure 4: Postures assumed by subjects during testing ................................................................. 22
Figure 5: Articulating kneepads commonly worn by lowseam coal miners ............................... 22
Figure 6: Pelvis coordinate system highlighting the location of the right hip joint center ........... 24
Figure 7: Orientation of the ATCS and ASCS .............................................................................. 26
Figure 8: Diagram of external shank forces and reaction forces and moments for kneeling near
full flexion with respect to the GCS ................................................................................. 30
Figure 9: Diagram of external shank forces and reaction forces and moments for kneeling near
full flexion with respect to the ASCS ............................................................................... 30
Figure 10: External force diagrams with respect to the anatomical shank coordinate system ..... 31
Figure 11: Thighcalf contact pressure distributions for kneeling near full flexion ..................... 35
Figure 12: Heelgluteus pressure distribution for kneeling near full flexion, Subject 1 ............... 35
Figure 13: Thighcalf contact pressure distributions for squat ..................................................... 36
Figure 14: Joint angles for Subject 1 kneeling near full flexion without kneepads ...................... 38
Figure 15: Net external forces normalized by body weight for Subject 1 kneeling near full
flexion without kneepads .................................................................................................. 40
xi
Figure 16: Moment contributions normalized by Bodyweight*Height for Subject 1 kneeling near
full flexion without kneepads ............................................................................................ 43
Figure 17: Varied forces and moment arms and corresponding sagittal knee moments for
kneeling near 90° flexion. The forces and moment arms were varied for the ground
reaction force at the toes (F1), the ground reaction force at the right knee (F2), and the
weight of the foot+shank (low leg). Note that all moment arm values are shown in cm,
all forces are shown in N, and all moments are not normalized and are shown in Nm.
Varying the zcomponent of the F1 moment arm by 3 cm changed the sagittal knee
moment by .2 Nm. When F1 was varied by 6 N, the moment changed by 3.5 Nm.
Varying the zcomponent of the F2 moment arm by 3 cm was sufficient to more than
triple the magnitude of the sagittal moment and change its interpretation. Varying the y
component of the F2 force by 6 N, changed the sagittal moment by .1 Nm. Varying the z
component of the COM of the lower leg by 3 cm changed the moment by 1.1 Nm.
Varying the ycomponent of the low leg weight changed the moment by 1.2 Nm. ......... 53
Figure 18: Varied forces and moment arms and corresponding sagittal knee moments for
kneeling on the right knee. The forces and moment arms were varied for the ground
reaction force at the toes (F1), the ground reaction force at the right knee (F2), and the
weight of the foot+shank (low leg). Varying the zcomponent of the F1 moment arm by
3 cm, changed the sagittal moment by .5 Nm. Varying the ycomponent of F1 by 6 N,
changed the sagittal moment by 3.3 Nm. Decreasing the zcomponent of the F2 moment
arm by 3 cm more than doubled the sagittal moment. When this moment arm was
increased by 3 cm, an extension moment was created. Varying the ycomponent of F2 by
6 N changed the moment by .1 Nm. Varying the zcomponent of the low leg COM by 3
cm and the ycomponent of the low leg weight by 6 N changed the moment by 1.2 Nm. 54
Figure 19: Varied forces and moment arms and corresponding sagittal knee moments for
kneeling near full flexion. The forces and moment arms were varied for the ground
reaction force at the toes (F1), the ground reaction force at the right knee (F2), and the
weight of the foot+shank (low leg), the thighcalf contact force, and the heelgluteus
contact force. Varying the zcomponent of the F1 moment arm by 3 cm changed the
moment by 4.4 Nm. Varying the ycomponent of F1 by 6 N changed the moment by 3.7
Nm. Varying the zcomponent of the F2 moment arm by 3 cm changed the moment by
5.7 Nm. Varying the ycomponent of F2 by 6 N changed the moment by .3 N. Varying
the zcomponent of the low leg COM by 3 cm changed the moment by 1.2 Nm. Varying
the ycomponent of the low leg weight by 6 N changed the moment by 1.2 Nm ............ 55
xii
Figure 20: Varied forces and moment arms and corresponding sagittal knee moments for
kneeling near full flexion, continued. The forces and moment arms were varied for the
ground reaction force at the toes (F1), the ground reaction force at the right knee (F2),
and the weight of the foot+shank (low leg), the thighcalf contact force, and the heel
gluteus contact force. Varying the zcomponent of the thighcalf moment arm by 3 cm
changed the moment by 3.7 Nm. Varying the thighcalf contact force by 6 N changed the
moment by .8 Nm. Varying the heelgluteus moment arm by 3 cm changed the moment
by 1.5 Nm. Varying the heelgluteus contact force by 6 N changed the moment by 2.5
Nm. .................................................................................................................................... 56
Figure 21: Varied forces and moment arms and corresponding sagittal knee moments for
squatting. The forces and moment arms were varied for the ground reaction force at the
foot (F1), the weight of the foot+shank (low leg), and the thighcalf contact force.
Varying the zcomponent of the F2 moment arm by 3 cm, changed the moment by 9 Nm.
Varying the ycomponent of F2 by 6 N, changed the moment by 2.7 Nm. Varying the z
component of the low leg COM by 3 cm, changed the moment by .9 Nm. Varying the y
component of the low leg weight by 6 N, changed the moment by 1.2 Nm. Varying the
zcomponent of the thighcalf moment arm by 3 cm, changed the moment by 5.1 Nm.
Varying the ycomponent of the thighcalf contact force by 6 N, changed the moment by
.9 Nm. ................................................................................................................................ 57
Figure 22: Varied KJC locations and resulting knee moments for kneeling near 90° flexion.
Varying the x, y, and zcomponents of the KJC location had no effect on the sagittal,
frontal, and transverse moments, respectively. Varying the ycomponent by 3 cm
changed the sagittal moment by 1.3 Nm. Varying the zcomponent by 3 cm more than
doubled the sagittal moment. Varying the xcomponent by 3 cm changed the frontal
moment by 1.3 Nm. Varying the zcomponent by 3 cm changed the adduction moment
by 1.7 Nm. Varying the xcomponent by 3 cm and varying the ycomponent by 3cm
more than doubled the transverse moment and changed the interpretation of this moment.
........................................................................................................................................... 61
Figure 23: Varied KJC locations and resulting knee moments for kneeling on right knee.
Varying the x, y, and zcomponents of the KJC location had no effect on the sagittal,
frontal, and transverse moments, respectively. Varying the ycomponent by 3 cm
changed the flexion moment by 2.7 Nm. Varying the zcomponent by 3 cm changed the
sagittal moment by 13.6 Nm and changed it interpretation. Varying the xcomponent by
3 cm changed the adduction moment by 2.7 Nm. Varying the zcomponent changed the
frontal moment by 1.5 Nm. Varying the xcomponent by 3 cm changed the transverse
moment by 9.1 Nm. Varying the ycomponent by 3 cm changed the transverse moment
by 1.5 Nm. ......................................................................................................................... 63
xiii
Figure 24: Varied KJC locations and resulting knee moments for kneeling near full flexion.
Varying the x, y, and zcomponents of the KJC location had no effect on the sagittal,
frontal, and transverse moments, respectively. Varying the ycomponent by 3 cm
changed the flexion moment by 1.6 Nm. Varying the zcomponent by 3 cm changed the
flexion moment by 3.6 Nm. Varying the xcomponent by 3 cm changed the adduction
moment by 1.6 Nm. Varying the zcomponent by 3 cm changed the adduction moment
by 1.1 Nm. Varying the xcomponent by 3 cm changed the transverse moment by 3.6
Nm and changed its interpretation. Varying the ycomponent by 3 cm changed the
transverse moment by 1.4 Nm and changed its interpretation. ......................................... 65
Figure 25: Varied KJC locations and resulting knee moments for squatting. Varying the x, y,
and zcomponents of the KJC location had no effect on the sagittal, frontal, and
transverse moments, respectively. Varying the ycomponent by 3 cm changed the flexion
moment 5.6 Nm. Varying the zcomponent by 3 cm changed the flexion moment by 3
Nm. Varying the xcomponent by 3 cm changed the adduction moment by 5.6 Nm.
Varying the zcomponent by 3 cm changed the adduction moment by 0.6 Nm. Varying
the xcomponent by 3 cm changed the transverse moment by 3 Nm. Varying the y
component by 3 cm changed the external rotational moment by .6 Nm. ......................... 67
xiv
NOMENCLATURE
COM – Center of mass
COP – Center of pressure
KJC – Knee joint center
AJC – Ankle joint center
HJC – Hip joint center
ATCS – Anatomical thigh coordinate system
ASCS – Anatomical shank coordinate system
GCS – Global coordinate system
MCS – Measured coordinate system
MSCS – Measured shank coordinate system
MTCS – Measured thigh coordinate system
F
t/c
– Thighcalf contact force
F
h/g
– Heelgluteus contact force
T
TGA
– Transformation matrix from ATCS to GCS
T
SGA
– Transformation matrix from ASCS to GCS
T
TMA
– Transformation matrix from ATCS to MTCS
T
SMA
– Transformation matrix from ASCS to MSCS
T
ST
– Transformation matrix from ATCS to ASCS
xv
F
knee
– Net external force applied to the knee
M
knee
– Net external moment applied to the knee
1
1.0 INTRODUCTION
Osteoarthritis affects more than 27 million Americans and cost nearly $5700 per person
annually. [1, 2] Certain occupational activities as well as heredity, age, trauma, or repetitive
stress have been associated with osteoarthritis in the knee. Studies have suggested that the risk
of developing knee osteoarthritis is increased by work involving prolonged knee bending. [3, 4]
A large project is underway at the NIOSH Pittsburgh Research Laboratory to gain a
better understanding of the biomechanics of the knee joint while kneeling, squatting, and
crawling in order to develop guidelines for the manufacture of kneepads, which may be used to
alleviate stress while kneeling. Subjects will assume kneeling postures typically seen in the
mining industry. Kinematics and kinetics data (motion capture, force plate, electromyography,
and pressure sensor) will be collected. The ultimate goal of the project is to develop a finite
element model of the knee in kneeling postures, which will provide a further understanding of
the biomechanics of knee structures while kneeling and crawling.
The goal of this Master’s thesis project was to develop a computational model which will
be used to determine the dynamic net forces and moments imposed on the knee joint during
static kneeling postures with and without knee pads. Motion data, force plate data, and pressure
sensor data were inputs into this model. The results of this thesis work will be incorporated in
the finite element model.
2
1.1 BACKGROUND
Musculoskeletal disorders vary by occupation. Different jobs expose workers to differing
ergonomic risk factors. People who work in restricted spaces, such as underground coal mines
or aircraft baggage compartments, are forced to assume awkward postures (such as kneeling,
stooping, and crawling) due to the restricted vertical height of their working environment. Other
occupations such as plumbers, carpet layers, roofers, housemaids, and agricultural workers may
not have this restriction in vertical height, but may also assume awkward postures to perform
work. This background section focuses on musculoskeletal injuries due to the awkward postures
of kneeling and squatting. First, the musculoskeletal structures that are affected by kneeling and
squatting are discussed, after which a review of occupational factors related to knee injuries is
presented.
1.2 KNEE STRUCTURES AND INJURIES
The knee is a complex weight bearing joint that connects the femur, patella, and tibia forming the
patellofemoral and tibiofemoral joints. It is also comprised of various ligaments, tendons, and
cartilage that function to add stability, motion, and weight bearing (Figure 1). (For more
information on the knee structures and their function see [6]) Any structure of the knee is
susceptible to injury; however, many occupational related disorders affect the bursae, meniscus,
and cartilage. This is particularly true for kneestraining postures such as kneeling and squatting.
[7  10]
3
Figure 1: Front view of knee anatomy [11]
The bursa is a fluid filled sac that acts as a cushion between a tendon and a bone or a
tendon and a muscle. Within the knee there are numerous bursae, both deep and subcutaneous.
There are three large subcutaneous bursae that may become injured from repetitive motions or
trauma. The suprapatellar bursa is between the femur and the tendon of the quadriceps femoris
muscle. This bursa facilitates full flexion and extension of the knee and may become injured
during acute trauma. The prepatellar bursa is between the patellar tendon and the skin. It acts as
a cushion, reducing the friction between the patella and the skin while kneeling. The
subcutaneous infrapatellar bursa is located between the skin and the tibial tuberosity. It allows
the skin to move easily over the tibial tuberosity and withstands pressure when kneeling with the
trunk upright. Any of these bursae may become irritated and inflamed, resulting in bursitis.
Bursitis generally results in a swollen bursa and restricted joint motion. In some cases the bursa
may become infected, resulting in a serious condition, septic bursitis.
The meniscus is a cartilaginous structure that provides a number of basic functions: load
bearing, stability, lubrication, and shock absorption. The meniscus transmits between 5070 %
4
of the compressive loads in the knee. [12] During the stance phase of gait the mean contact
stress on the meniscus is 20 klbs per cm
2
. [13] The meniscus provides stability by interacting
with the Anterior Cruciate Ligament (ACL), assisting with rotational stability. The meniscus
also prevents friction between the femur and tibia and allows diffusion of joint fluid into the
articular cartilage. The meniscus increases the load carrying area of the joint by 200 %, thereby
lowering the stress applied to the articular cartilage and acting as a shock absorber. [12]
The meniscus can be damaged by both traumatic and degenerative mechanisms. The
most common mechanism of traumatic meniscal tear occurs predominantly in athletes when a
twisting moment is applied to the weight bearing knee in a semiflexed position. This form of
meniscal tear is commonly concurrent with an ACL injury. [14] Meniscus tears without ACL
injury are commonly the result of degeneration of the meniscus. [15]
The articular cartilage is another structure that is commonly injured. Osteoarthritis is a
chronic condition characterized by the breakdown of the joint’s articular cartilage. This
breakdown in cartilage affects the lubrication process within the knee, and results in direct
contact between the bones causing stiffness, pain, and loss of motion. It also reduces the shock
absorbing capacity of the knee, leading to microfractures of the subchondral bone. [16]
Osteoarthritis is the most common joint disorder associated with major disability, affecting more
than 27 million Americans. [1, 17, 18] The risk of developing this disease increases with age and
it is estimated that by 2030, 20% of Americans (about 72 million people), will be at a high risk
for this disease. [18]
5
1.3 OCCUPATIONAL KNEE INJURIES
Working in certain awkward postures has been to shown to affect musculoskeletal disorders of
the lower extremities. [19] Of all lower extremity injuries, knee injuries most commonly result
from worker position or motion. [20] Kneeling and crawling can cause knee injuries ranging
from minor skin irritations to bursitis or a torn meniscus. [21]
1.3.1 Bursitis
Frequent kneeling produces large stresses on the patella which may thicken the walls of the
bursa. [9] This thickening can lead to irritation and inflammation. Prepatellar bursitis is the most
common form of bursitis and known as “housemaid’s knee” or “miner’s knee”. While this form
of bursitis does not typically lead to other knee injuries, it does account for days lost from work
and may lead to infection if left untreated. Bursitis usually does not lead to other knee problems
when properly treated.
1.3.2 Meniscal Injuries
Frequent kneeling also has been linked to an increased laxity in the ACL and development of
meniscal injuries. Sharrard and Liddell (1962) investigated hospital records of 957
meniscectomies, performed between January 1958 and June 1960, from five hospitals in one of
the largest British coalfields. After reviewing records for men between 15 and 64, the authors
determined that miners are more likely than others to suffer cartilage damage of any type. Upon
interviewing 200 of the men, the authors determined that cartilage tears were more common in
6
miners who knelt, but the tears normally occurred when walking. To gain a better understanding
of why the tears occurred in nonkneeling posture, the authors examined 80 patients with
meniscal damage. From these examinations, they determined that ACL laxity resulted from
kneeling at work. The authors proposed that this laxity decreases the stability of the knee in non
restricted postures such as standing and walking, leading to meniscal tears.
Along with kneeling, aging plays a role in the development of meniscal injuries. [22],
[23] Drosos and Pozo (2004) conducted a study on 392 patients between the ages of 18 and 60
with meniscal injuries from the general population. 32.4% of patients had sportsrelated
meniscal injuries, 38.8% had nonsportsrelated injuries and 28.8% had no identifiable injury or
no identifiable cause of injury. The most frequently reported mechanism of injury for the non
sporting group was rising from a squatted position. The average age of the patients with non
sporting related meniscal tears was 41. The authors felt that this age may reflect the
degenerative changes in the meniscus, thereby requiring less force to create an injury. Smillie
(1978) proposed that meniscus degeneration starts in the third decade of life. This degeneration
reduces the elasticity of the meniscus, thereby increasing the susceptibility to injury. Drosos and
Pozo deduced that degeneration may be the result of the repetitive microtrauma and mechanical
stresses of everyday life. These factors, in conjunction with the agerelated disappearance of
elastic fibers, may precede the development of many meniscal tears.
Previous research shows that there is a significant risk for meniscus damage associated
with an increase in age and frequent and prolonged kneeling. [7, 10, 22, 23] Some meniscus tears
can be repaired, while others will have to be removed in a meniscectomy. Although this
procedure is necessary to ensure proper joint motion, it leaves the articular cartilage of the femur
and tibia without a proper intermediate tissue. The weight bearing area of an intact meniscus
7
varies from 11 cm
2
at 90° flexion to 20 cm
2
at full extension. After meniscectomy the area
ranges from 6 cm
2
to 12 cm
2
, respectively. [13] This decreased area increases the peak local
contact stresses transmitted to the articular cartilage by 65235 %, causing degeneration. [24]
Thus, meniscal injuries have been linked to the longterm development of knee osteoarthritis. [25
 28]
1.3.3 Osteoarthritis
Osteoarthritis has been associated with a number of personal and activityrelated risk factors.
Personal risk factors include heredity, age, diabetes, smoking, and obesity. [3] Activityrelated
risk factors are trauma, sports activities, and occupation. [8, 17, 29, 30, 31] Within the
occupational factors, a knee bending variable has been examined, including levels of knee
bending, such as stooping, kneeling, crouching, or crawling. Of the 5,193 people surveyed, 315
were found to have radiographic osteoarthritis of the knee. Multiple linear regressions found
knee bending demand of the job was associated with osteoarthritis of the knee (OR=0.32)
amongst both men and women in the 5564 years age group. The increased risk for osteoarthritis
in those with physically demanding occupations was seen mostly in people 55 years and over
(OR=2.45 in men, OR=3.49 in women). The authors concluded that the strong association with
occupations in which knee bending is prominent suggests work activity may play a major
causative role in osteoarthritis of the knee. [3]
8
Subsequent studies have suggested that the risk of developing knee osteoarthritis is
increased by work involving prolonged knee bending and squatting. [3, 4, 17, 32] This is evident
in the high prevalence of knee osteoarthritis among Asian populations who spend a lot of time
performing floor activities. [17] The postures examined were squatting, sideknee bending,
kneeling, and lotus position. Of these postures, squatting, lotus position, and sideknee bending
were found to increase the risk of developing knee osteoarthritis.
9
2.0 MOTIVATION
Much attention has been given to the biomechanics of the knee joint during gait but the
same attention has not been given to the knee joint while in kneeling, stooped, or squatted
postures. Although there has been research on restricted postures, much of this research is aimed
at the low back forces and causes of low back pain. There are few studies available on squatting
and high flexion activities; however, no studies of kneeling on one knee or near 90° flexion were
found.
2.1 PREVIOUS MODELLING OF KNEE FORCES AND MOMENTS
Dahlkvist et al. (1982) developed subjectspecific 2D models of the lower leg of six male
subjects to determine the forces at the knee during squatting and rising from a deep squat.
Anthropometric data and xrays were used to create the model. External force reactions and
points of application, lower leg configurations, and electromyographic data were measured. For
three subjects, forces were calculated throughout going into a deep squat and then rising to a
standing position. For the other three subjects rising from a deep squat were compared for a
slow rise trial and a fast rise trial. Joint forces varied across subjects and during activities.
Patellofemoral joint forces were shown to be the largest, with average maximums between 4.7
times bodyweight for slow ascent and 7.6 times bodyweight for fast descent.
10
Perry et al. (1975) performed a cadaver study to show significantly increased forces on
the joint surfaces with flexion of greater than 30°. Forces in the quadriceps, patella, and tibia
during simulated weightbearing flexed knee stance were determined in a cadaver model. The
specimen was instrumented to measure the quadriceps force, patellofemoral compression, and
tibiofemoral loading. Flanged transducers were inserted into the tibia to measure the
tibiofemoral joint stress. Quadriceps muscle action was simulated by a cable linked between the
vastus intermedialusrectus femoris to the intertrochanteric region of the femur. A ring
tensiometer, aligned parallel to the femur, measured the forces in the cable as loads were applied
to the femoral head. Compressive forces of the patella were also measured. The knee was set to
six positions for measurement: 0°, 5°, 15°, 30°, 45°, and 60°. The quadriceps force necessary to
stabilize the knee was found to be directly proportional to the femoral load and angle of flexion.
At the 30° position, the required force was 210% body weight and at 60° the required force was
410% body weight. The tibiofemoral forces also increased in direct proportion to the load on the
head of the femur when in full extension. When flexed, the tibiofemoral forces were a
summation of the load on the femoral head and the quadriceps force. The patellar forces were
found to be a function of quadriceps tension and dependent on the angle of flexion and the
applied femoral load. [34]
11
Nagura et al. (2002) determined the forces during squatting activities using an inverse
dynamics approach and found similar high contact forces. Subjects were separated into two
groups, the ambulation group and the deep flexion group. Net moments and posterior forces
were shown to be significantly higher in deep flexion than in walking or stair climbing. The net
posterior force increased by 50% over walking and stair climbing. Net moments and inferior
forces were significantly smaller in deep flexion than in walking or stair climbing. The authors
concluded that the increase in extensor force during deep flexion increases the stress in the
patellar tendon and joint contact forces. [35]
Sharrard (1965) examined the forces and pressures at the knee joint of a miner simulating
a shoveling task while kneeling. He placed scales under the knees and toes of the miner. When
the 189 lbs subject rested sitting on his heels, the weight on the left knee, right knee, and toes
were 46 lbs, 56 lbs, and 84 lbs respectively. During the shoveling task, the weight on the knees
and toes rose and declined rapidly. Results showed that the pressures can rise as much as 200
pounds per square inch while shoveling. At that time, the average miner shoveled once every 2.5
seconds. The pressures at the tibial tubercle, the interval between the tibial tubercle and the
patella, the distal pole of the patella, and the body of the patella were also measured. These
pressures were shown to vary throughout the shoveling task as well. Considerable variation was
shown in the pressures and weightbearing surfaces of the knee while kneeling and performing
work. [9]
Previous biomechanical models of the knee used to evaluate deep flexion and squatting
postures did not account for the contact between the thigh and the calf. [33, 34] Nagura et al.
stated that neglecting this contact was a limitation of their study which may cause an
overestimation of net forces and moment. Caruntu et al. (2003) created a model of the knee for
12
deep flexion which included the contact between the thigh and the calf. They reported a 700 N
overestimation in quadriceps force when this contact was neglected. They also reported a 50%
increase in medial collateral ligament forces when this contact was considered. [36] Zelle et al.
(2007) measured the pressure distribution of thighcalf contact for subjects squatting and
kneeling with mean knee angles of 151.8 and 156.4 respectively. Results showed the resultant
contact force to be >30% body weight and located within 17 cm of the epicondylar axis. [5]
Thighcalf contact has been shown to be significant, and neglecting these parameters may result
in overestimation of joint forces and moments. For these reasons the contact force between the
thigh and the calf was not neglected from this computational knee model.
In summary previous research on knee biomechanics in restricted postures have failed to
examine kneeling near 90° flexion and kneeling on one knee, which may be commonly utilized
when performing work. Research on higher flexion postures such as kneeling near full flexion
and squatting did not include thighcalf or heelgluteus contact and may result in the
overestimation of joint forces and moments. In this research a computational model was develop
to determine the net dynamic forces and moment imposed on the knee during static kneeling
postures. This model may be applied to other postures in the future and accounts for the contact
between the thigh and the calf as well as between the heel and the gluteal muscles which may
occur when kneeling in high flexion.
13
2.2 SPECIFIC AIMS AND SIGNIFICANCE
2.2.1 Specific Aim #1
Develop algorithms and a 3D computational model to calculate the net dynamic forces and
moments at the knee joint from motion capture and force plate data
To date, there has not been an analysis of the forces and moments at the knee during
kneeling. Previous research has determined the net forces and moments at the knee joint for
squatting and end range flexion. [33 36] The role of this study was to develop a computational
model that may be used to determine the net forces and moments of the knee during not only
squatting and deep flexion, but also for kneeling near 90° flexion and kneeling on one knee. The
use of one computational model to analyze different postures will allow for better comparisons
and interpretation of these forces and moments. Also, unlike traditional models based on inverse
dynamics, this model accounts for the contact force between the thigh and calf, which has been
determined to be >30% body weight. [5] At this level of force, this contact may be sufficient to
have a considerable effect on the forces, moments, and stresses in knee structures. This model
also accounts for the contact between the heel and the gluteal muscles which occurs in some
people when kneeling and sitting on their heels.
14
2.2.2 Specific Aim #2
Apply the developed model to data collected from a minimum of one subject.
The effect of protective equipment (i.e. knee pads) on knee forces and moments has not
been investigated. In this study the subject was evaluated with and without knee pads. The
forces were not expected to change significantly due to the knee pads however, the knee pads
were expected to change the location of the center of pressure for the force at the kneeground
interface, thereby affecting the moments at the knee. Force plate, motion capture, and pressure
data were analyzed via the developed model for two subjects. This model will be used in the
future on a larger dataset currently being collected. The resulting forces and moments will add
to the knowledge on kneeling biomechanics. In addition, these forces and moments will be
utilized in the finite element model of the knee currently being developed.
2.2.3 Specific Aim #3
A sensitivity analysis for the input variables (anthropometric variables, joint center location
estimation, ground reaction forces, thighcalf contact force, and heelgluteal contact force) will
be implemented.
Numerous factors contribute to the forces and moments at the knee joint while kneeling.
A sensitivity analysis was performed to determine which variables contribute the most to the
forces and the moments at the knee. This analysis was also useful in determining which
parameters introduce the largest sources of error in the model.
15
2.2.4 Future Significance
The forces and moments determined using this computational model will be inputs to a finite
element model (FEM) of the knee being developed. This FEM will be used to evaluate the
stresses and strains in tissues while kneeling, crawling, and stooping. Results of the FEM will
provide a better understanding of the internal biomechanics of the knee structures while in
restricted postures. These results will be used to provide interventions, such as a novel knee pad
design, in hopes of reducing the prevalence and severity of occupational knee injuries in low
seam coal miners.
16
3.0 RESEARCH DESIGN AND METHODS
Motion analysis and force data were collected to track body positions and ground
reaction forces for simulated postures. This data was used to determine the segment positions,
segments lengths, and ground reaction forces. Pressure data were also collected to determine
the magnitude of the force transmitted from the thigh to the calf and from the heel to the gluteal
muscles when kneeling. These parameters were inputs into the computation model, allowing the
estimation of knee forces, moments, and joint angles.
3.1 EQUIPMENT
3.1.1 Laboratory Equipment
Motion data is captured using eleven Infrared cameras (Eagle, Motion Analysis Corporation,
Santa Rose, California, USA) which track the motion of ½” reflective markers. The cameras are
arranged around the room such that all markers are visible during testing. Motion data was
collected at 60 Hz.
17
Two force platforms (Model OR65, Advanced Mechanical Technology, Inc., Newton,
Massachusetts, USA) were used to measure the reaction forces at the groundknee, groundfoot,
or groundkneepad interface. The force platforms were aligned in parallel and were level with
the floor. Force data was collected at 1020 Hz.
A clinical seating pressure assessment system (ClinSeat
®
, Tekscan Inc., South Boston,
Massachusetts, USA) was used to measure the contact between the thigh and calf and the heel
and gluteal muscles. This 19.2 in x 16.8 in pressure sensor uses resistive technology to measure
the pressure between surfaces. The spatial resolution is 1 sensel/cm
2
with a pressure range of 0
30 PSI. The supplied software (Advanced ClinSeat, Tekscan Inc., South Boston, Massachusetts,
USA) generated pressure maps, total pressure, total force, and center of pressure. Total force
and center of pressure locations from the supplied software were used in the computational
model.
3.1.2 Computational Model
The computational model was developed in MATLAB
®
(The Mathworks Inc., Natick, MA) on a
personal computer. It is based on an inverse dynamics method [37] which uses measured ground
reaction forces and anthropometric measurements to determine the net external forces and
moments at the right knee joint. In this linear model the upper leg (femur) and the lower leg
(tibia, fibula, and foot) are modeled as rigid bodies attached via a pin joint with three rotational
degrees of freedom. The global reference frame was oriented such that when the subject was
standing in standard anatomical position, the xaxis was in the medial/lateral direction, the yaxis
was in the anterior/posterior direction, and the zaxis was in the superior/inferior direction. The
origin of this system was on the laboratory floor at the upper left corner of the first force plate.
18
3.1.2.1 Model Assumptions
The developed computational model is based on several assumptions.
• The knee is assumed to be a frictionless pinjoint. This allows all forces to pass
directly through the joint center.
• Segments are assumed to be rigid with mass concentrated at the center of mass
locations. This allows one center of mass to represent the weight of the segment.
• Linear relationship between external forces and moments applied to the knee.
This allows a planar model to be used to determine the external forces and
moments applied to the knee.
• The relative movement of pelvic bones is negligible. This allows approximation
of the hip joint center from palpable pelvic landmarks.
• Thickness of subcutaneous tissue between bone and skin is minimal. This allows
the assumptions that markers placed on palpable landmarks are directly on that
landmark.
• The measured thighcalf and heelgluteus contact forces are concentrated at the
measured center of pressure location. This allows the contact forces to be
represented as a single resultant force, opposed to a pressure distribution.
• Affect of patellar tendon and tibial tubercle on forces externally applied to the
tibia is negligible. This assumption allows ground contact forces measured at the
groundknee or groundkneepad interface to be assumed to act at some distance
away from the knee joint center and not be affected by the patellar tendon or tibial
tubercle.
19
3.2 SUBJECT TESTING
3.2.1 ThighCalf and HeelGluteus Contact Measurements
Thighcalf and heelgluteus contact forces (F
t/c
and F
h/g
) were measured prior to motion and force
data collection. The subject was instructed to kneel near 90° flexion and the pressure sensor was
placed on their lower leg. The subject was then instructed to kneel into full flexion while
pressure data was collected for a period of 5 seconds. The distance from the top of the sensor to
the lateral epicondyle was measured and recorded while data was collected. This was repeated
with the subject squatting.
3.2.2 Subject Instrumentation
Each subject was fitted with 41 reflective markers using a modified version of the Cleveland
Clinic Foundation’s marker set, Figure 2. The marker included one segment marker and a three
marker cluster of the thighs and shanks. This marker set was chosen to be compatible with
SIMM for use in a future study. The subject was instructed to perform a standing Tpose with
their right foot making contact with force plate 2 and their left foot off of the force plates. Data
was collected for 5 seconds. This data was used to create the anatomical coordinate system for
the thigh and shank.
20
Figure 2: Anatomical marker set
Figure 3: Measured marker set
21
Following recording of the standing Tpose, eight markers were removed due to their
high risk for falling off or becoming covered (i.e. they are virtual markers which were
reconstructed during data processing). This yielded the measured marker set shown in Figure 3.
The subject was instructed to perform another standing Tpose as well as a range of motion.
During the range of motion, the subject went through a series of motions with a laboratory
assistant. They began in a standing Tpose then performed lateral bending, twisting at waist,
raising knees up to chest, squatting, and knee flexion/extension, crossing arms, rotating head
then dropped down into a kneeling posture. In this kneeling posture the subject performed lateral
bending, twisting at waist, kneeling down into full flexion, kneeling on left knee, kneeling on
right knee, ankle rotations, kneeling on all fours, and ended in a standing Tpose. The range of
motion data was then postprocessed and used to extend the motion capture template.
3.2.3 Experimental Data Collection
Once the template was extended, experimental data collection began. A mesh roof in the
laboratory was lowered to 48”, simulating a restricted working height. The subject was
instructed to perform a static trial of each posture for 10 seconds. The postures studied were
kneeling near 90° flexion, kneeling on the right knee, kneeling near full flexion, and squatting.
Subjects were not given specific instructions on these postures. A poster in the laboratory
showed schematics of the postures studied, Figure 4. Subjects were instructed to simulate the
postures shown in the schematics. They were not given specific instructions on kneeling
postures as to not affect their normal kneeling postures. They were, however given instructions
on where to place their knees and feet when kneeling. When testing the kneeling near 90°
flexion, kneeling on right knee, and kneeling near full flexion, subjects were instructed to kneel
22
with their right knee on force plate 2 and their right foot on force plate 1. Their left side was to
remain off of the force plates. When testing the squatting posture, subjects were instructed to
kneel with their right foot on force plate 2 and their left foot off of the force plates. At the start
of testing, each subject was provided a new pair of orange articulating knee pads commonly
worn by underground coal miners, Figure 5. The subject performed each posture with and
without knee pads for a total of eight trials. Motion data was lowpass filtered using a 4
th
order
Butterworth filter at 15 Hz to smooth instrumental errors.
Figure 4: Postures assumed by subjects during testing
Figure 5: Articulating kneepads commonly worn by lowseam coal miners
23
3.3 DATA ANALYSIS
Data was analyzed for a 5 five second period of the trial after the subject appeared to have
reached a balanced posture and showed minimal instability. This portion of the static trial was
used to determine the forces, moments, and joint angles.
3.3.1 Construction of Coordinate Systems
For each segment, an anatomical and measured coordinate system was created from the motion
capture data. The anatomical system was created from the anatomical standing Tpose and
allowed the location of anatomical landmarks to be linked to the global reference frame. It was
also used to determine the location of the ankle joint center (AJC), knee joint center (KJC), and
hip joint center (HJC) as well as the location of the lower leg center of mass. A measured
coordinate system was created from the anatomical standing Tpose as well as from each static
trial and was used to link the testing markers to the locations of the markers that were removed.
The anatomical coordinate system of the thigh (ATCS) was created using the left and
right anterior superior iliac spine (L.ASIS & R.ASIS), knee, and thigh markers.
axis z r r r
axis y r r r
KJC HJC
KJC HJC
r
axis x
medial knee lateral knee
l knee media al knee later
r
4
3
2
1
− × =
− × =
−
−
=
−
−
−
=
3 1
1 2
24
The KJC was assumed to be midway between the medial and lateral epicondyles of the
femur, measured by the medial and lateral knee markers. The location of the HJC was
approximated using regression equations proposed by Bell et al. (1990) and adapted to fit the
global reference frame of the laboratory. [38] (Figure 6)
Figure 6: Pelvis coordinate system highlighting the location of the right hip joint center
−
−
+
=
− =
+
=
PW Origin
PW Origin
PW Origin
HJC
LASIS RASIS PW
RASIS LASIS
Origin
x x
* 3 .
* 19 .
* 36 .
2
25
The transformation matrix from the global reference frame to the anatomical thigh
coordinate system (T
TGA
) was created from the unit direction vectors of the ATCS.
=
4 3 1
0 0 0 1
r r r HJC
T
TGA
The anatomical coordinate system of the shank (ASCS) was determined using the knee,
ankle, and shank markers. The AJC was assumed to be midway between the medial and lateral
malleoli, measured by the medial and lateral ankle markers. The transformation matrix from the
GCS to the ASCS (T
SGA
) was created from the unit direction vectors of the ASCS.
axis x r r r
axis y r r r
axis z
AJC KJC
KJC  AJC
r
medial knee lateral knee
l knee media al knee later
r
1 3 4
2 1 3
2
1
− × =
− × =
−
−
=
−
−
=
=
1 3 4
0 0 0 1
r r r KJC
T
SGA
The ATCS and ASCS were oriented such that when standing the systems aligned with the
GCS and the positive xaxis is in the lateral direction of the right leg, the positive zaxis is in the
proximal direction, and the positive yaxis is in the anterior direction. (Figure 7)
26
Figure 7: Orientation of the ATCS and ASCS
A measured coordinate system (MCS) was created for the thigh and the shank using the
marker clusters on the segments. The measured coordinate system of the thigh (MTCS) was
created from the thigh, thigh front, and thigh rear markers. The transformation matrix from the
GCS to the MTCS (T
TGM
) was also created with the right thigh front marker as its origin.
axis y r r r
axis x r r r
rear thigh front thigh
rear thigh front thigh
r
axis z
front thigh thigh
front thigh thigh
r
4
3
2
1
− × =
− × =
−
−
=
−
−
−
=
3 1
1 2
=
1 4 3
0 0 0 1
r r r front thigh
T
TGM
27
The measured coordinate system of the shank (MSCS) was created from the shank, shank
front, and shank rear markers. The transformation matrix from the MSCS to the GCS (T
SGM
) was
created with the right shank front marker as its origin.
axis y r r r
axis x r r r
rear shank front shank
rear shank front shank
r
axis z
front shank shank
front shank shank
r
4
3
2
1
− × =
− × =
−
−
=
−
−
−
=
3 1
1 2
=
1 4 3
0 0 0 1
r r r front shank
T
SGM
To determine measured marker locations in the ATCS and ASCS, T
TMA
and
T
SMA
were
created respectively.
   
   
SGA SGM SMA
TGA TGM TMA
T T T
T T T
1
1
−
−
=
=
28
3.3.2 Joint Angle Estimation
Joint angles were determined for each trial using Euler Angle Decomposition. The largest joint
rotations occurred about the medial/lateral xaxis (α – extension/flexion) followed by the
posterior/anterior yaxis (β  varus/valgus), and the distal/proximal zaxis (γ  internal/external
rotation), yielding an Euler order of Xy’z’’. The transformation matrix from the anatomical
thigh to the anatomical shank coordinates T
ST
was created to determine the rotation matrix, R
ST
.
This matrix was used to determine the Euler angles. Therefore, motion of the thigh was in
respect to the shank.
   
   
   
TGA
1
SGA ST
TMA TGM TGA
SMA SGM SGA
T T T
T T T
T T T
*
*
*
−
=
=
=
−
−
− =
1 0 0
0 ) cos( ) sin(
0 ) sin( ) cos(
*
) cos( 0 ) sin(
0 1 0
) sin( 0 ) cos(
*
) cos( ) sin( 0
) sin( ) cos( 0
0 0 1
γ γ
γ γ
β β
β β
α α
α α
ST
R
+ −
− − +
−
=
) cos( ) cos( ) cos( ) sin( ) sin( ) sin( ) cos( ) cos( ) cos( ) sin( ) sin( ) sin(
) sin( ) cos( ) sin( ) sin( ) sin( ) cos( ) cos( ) sin( ) cos( ) sin( ) cos( ) sin(
) sin( ) sin( ) cos( ) cos( ) cos(
α β γ α β α γ α γ β α γ
α β γ α β α γ α γ β α γ
β γ β γ β
ST
R
29
( )
( )


.

\

=


.

\

=


.

\
 −
=


.

\
 − −
=


.

\
 −
=


.

\
 − −
=
− −
− −
− −
1 , 1
3 , 1
3 , 3
3 , 2
1 , 1
2 , 1
) cos(
tan
) cos( ) cos(
) sin( ) cos(
tan
tan
) cos( ) cos(
) sin( ) cos(
tan
tan
) cos( ) cos(
) sin( ) cos(
tan
1 1
1 1
1 1
ST
ST
ST
ST
ST
ST
R
R
R
R
R
R
γ
γ β
β γ
β
α β
α β
α
γ β
γ β
γ
3.3.3 Joint Forces and Moments
Ground reaction forces, segment weight, F
t/c
, and F
h/g
were inputs into the computational model.
External force diagrams for kneeling near full flexion with respect to the GCS and the ASCS are
shown in Figure 8 and Figure 9. The center of mass location and weight of the shank+foot were
determined using equations from Clauser et al., 1969 which were adjusted to use the knee joint
center and ankle joint center in this model. [40] The reaction forces and moments were assumed
to act in the positive directions.
Squatting and kneeling creates a contact force between the thigh and the calf. Kneeling
near full flexion also creates this contact, and in some subjects there is additional contact
between the heel and the gluteal muscles. F
t/c
and F
h/g
were modeled as resultant forces whose
line of action was in the anterior direction of the shank. The locations of these forces were
determined from the center of pressure locations on the pressure sensor. Forces at the foot (F
1
),
forces at the knee (F
2
) and the weight of the lower leg were measured with respect to the GCS.
(Figure 8) Thighcalf and heelgluteus contact forces were measured with respect to the ASCS.
30
Figure 9 shows the orientation of the forces and moments as presented in this research, with
respect to the ASCS. External force diagrams for all postures are shown in Figure 10.
Figure 8: Diagram of external shank forces and reaction forces and moments for kneeling near full
flexion with respect to the GCS
Figure 9: Diagram of external shank forces and reaction forces and moments for kneeling near full
flexion with respect to the ASCS
31
Figure 10: External force diagrams with respect to the anatomical shank coordinate system
32
Joint equilibrium was assumed, therefore the sum of all forces and moments at the knee
were calculated and summed to equal zero. The sum of the external forces in the x, y, and z
directions were all summed to zero to determine the reaction force, R, necessary to stabilize the
knee due to the application of the net external forces (F
knee
).
x x x
x x x
x
F F R
R F F
F
2 1
2 1
0
0
− − =
= + +
=
∑
y y y
y y y
y
F F R
R F F
F
2 1
2 1
0
0
− − =
= + +
=
∑
g h c t x x LL z
g h c t LL z z z
z
F F F F W R
F F W R F F
F
/ / 2 1
/ / 2 1
0
0
− − − − =
= − − − + +
=
∑
) (
z y x knee
R R R F + + − =
33
The sum of the external moments at the knee joint in the x, y, and, z directions were also
summed to equal zero. The net external knee moment applied to the knee joint (M
knee
) was also
determined.
   
   
   
   
   
LL / /
LL / /
/ /
/ / / / /
/ / / / /
2 2 2 2 2 2 2
1 1 1 1 1 1 1 1
)
0
0 0
0 0
0 0
M M M M M M M
M M M M (M M
M M M M M M
0 M
F COP KJC COP KJC COP KJC M
F COP KJC COP KJC COP KJC M
W COM KJC COM KJC COM KJC M
F F F COP KJC COP KJC COP KJC M
F F F COP KJC COP KJC COP KJC M
F1 F2 g h c t KJC knee
F1 F2 g h c t KJC
LL F1 F2 g h c t KJC
KJC
g h g h F z g Fh y g h F x g h
c t c t F z c t F y
x c t F x c t
LL LL
z y x z z y y x x F
z y x z z y y x x F
z y x
z y
z LL z y LL y x LL x
+ + + + = − =
+ + + + − =
= + + + + +
=
− × − − − =
− × − − − =
− × − − − =
× − − − =
× − − − =
∑
34
4.0 RESULTS
Data from two subjects were analyzed using the developed computation model. Subject
1 was a 1.8 m, 675 N, 19 year old male. Subject 2 was a 1.56 m, 720 N, 59 year old female.
Neither subject had a history of knee injuries or pathologies. All forces and moments presented
are in reference to the anatomical shank coordinate system.
4.1 THIGHCALF AND HEELGLUTEUS CONTACT FORCES
4.1.1 Near Full Flexion
The mean F
t/c
in the kneeling near full flexion posture for subjects 1 and 2 were 127.5 N and 164
N, respectively. The centers of pressure were 0.14 m, and 0.17 m along the long axis of the tibia
from the medial epicondyle of the femur. Pressure distributions, showing the centers of pressure
for this posture are shown in Figure 11. Subject 2’s posture did not cause any contact between
the heel and gluteal muscles. Subject 1 had a mean F
h/g
of 56.1 N with a center of pressure 0.43
m along the long axis of the tibia from the medial epicondyles of the femur. This pressure
distribution is shown in Figure 12.
35
Figure 11: Thighcalf contact pressure distributions for kneeling near full flexion
a) Subject 1 b) Subject 2
Figure 12: Heelgluteus pressure distribution for kneeling near full flexion, Subject 1
a) b)
Proximal Shank
Heel
36
4.1.2 Squat
The mean F
t/c
in the squatting posture for subjects 1 and 2 were 173 N and 195 N, respectively.
The centers of pressure were located 0.15 m, and 0.14 m along the long axis of the tibia from the
medial epicondyle of the femur. Pressure distributions, showing the centers of pressure for this
posture are shown in Figure 13.
Figure 13: Thighcalf contact pressure distributions for squat
a) Subject 1 b) Subject 2
a) b) Proximal Shank
37
4.2 KNEE ANGLES
Mean varus/valgus, internal/external rotation, and included knee angles for Subjects 1 and 2 are
shown in Table 1 and Table 2, respectively. Standard deviations across 5 second portions of the
trials are shown in parentheses. The angles varied over time with standard deviations ranging
from .221.44° for the included angles, .368.69° for varus/valgus, and .251.98° for int/ext
rotation. Both subjects had very small standard deviations in the included joint angles over
time. Varus/valgus deviations were the largest for both subjects. Subject 1 ranged from 15°
varus to 15° valgus when wearing kneepads and kneeling near 90° flexion. This 30° range in
angles caused a standard deviation of 8.55°. Subject 2 ranged from 3° varus to 3° valgus when
wearing kneepads and kneeling on one knee, creating a standard deviation of 2.1°. A typical
time series plot of the joint angles is shown in Figure 14.
Table 1: Average knee angle (degrees), Subject 1
Subject 1
With Kneepads Without Kneepads
Included
+ Valgus
 Varus
+ Int rot
 Ext rot
Included
+ Valgus
 Varus
+ Int rot
 Ext rot
Near 90°
Flexion
83.1
(1.35)
0.35
(8.55)
4.7
(0.73)
82.7
(.25)
7.3
(3.0)
7.3
(0.28)
One Knee
80.2
(0.39)
1.0
(3.41)
14.0
(0.42)
75.9
(0.30)
8.7
(3.45)
10
(0.44)
Near Full
Flexion
23.0
(0.21)
1.2
(0.29)
6.5
(0.47)
23.0
(0.16)
2.1
(0.60
5.6
(0.32)
Squat
24.0
(0.78)
0.9
(2.47)
5.6
(1.55)
22.2
0.43)
1.3
(1.87)
5.7
(1.07)
Standing
173.6
(.57)
2.6
(.22)
.78
(.07)
38
Table 2: Average knee angles (degrees), Subject 2
Subject 2
With Kneepads Without Kneepads
Included
+ Valgus
 Varus
+ Int rot
 Ext rot
Included
+ Valgus
 Varus
+ Int rot
 Ext rot
Near 90°
Flexion
84.9
(.40)
.81
(.51)
5.28
(.31)
38.9
(.5)
9.1
(1)
5.9
(.27)
One Knee
40.6
(.48)
.40
(2.1)
4.7
(1.4)
35.8
(.32)
4
(1.7)
.28
(.69)
Near Full
Flexion
33.7
(.25)
.58
(.69)
13.9
(.25)
31.9
(.1)
1.9
(1.6)
1.9
(.31)
Squat
36.3
(.46)
.09
(.26)
12.7
(.55)
37.9
(.47)
.63
(1.35)
7.5
(.77)
Standing
179.1
(.34)
2.35
(.14)
.05
(.02)
Figure 14: Joint angles for Subject 1 kneeling near full flexion without kneepads
39
4.3 KNEE FORCES
All forces are with respect to the anatomical shank coordinate system and normalized to body
weight, by dividing the force in N, by the body weight in N, and multiplying by 100% yielding
values in percentage of body weight (%BW). Average net forces for Subjects 1 and 2 are shown
in Table 3 and Table 4, respectively. The largest lateral forces were found in Subject 1 when
kneeling on one knee and the largest medial forces were shown in Subject 2 when squatting
without kneepads. The largest posterior forces were shown when kneeling on one knee in both
subjects. The largest proximal forces occurred in Subject 1 when squatting and in Subject 2
when kneeling on one knee. Time series of the net forces for Subject 1 kneeling near full flexion
without kneepads are shown in Figure 15.
Table 3: Average net external forces normalized by body weight, Subject 1
Subject 1
With Kneepads, % BW Without Kneepads, % BW
+ Lateral
 Medial
+ Anterior
 Posterior
+ Proximal
 Distal
+ Lateral
 Medial
+ Anterior
 Posterior
+ Proximal
 Distal
Near 90°
Flexion
6.00
(.74)
43.08
(.78)
6.50
(.74)
8.61
(.34)
41.29
(.41)
6.21
(.20)
One Knee
16.70
(.87)
74.62
(1.07)
13.58
(.57)
7.33
(.41)
67.80
(1.61)
13.63
(.33)
Near Full
Flexion
5.95
(.30)
16.57
(.78)
8.23
(.26)
5.61
(.56)
17.67
(.69)
7.65
(.68)
Squat
5.70
(.38)
13.23
(1.27)
27.53
(1.10)
2.94
(.78)
14.86
(1.54)
27.86
(.78)
Standing
.57
(.14)
11.58
(.62)
39.89
(1.51)
40
Table 4: Average net external forces normalized by body weight, Subject 2
Subject 2
With Kneepads Without Kneepads
+ Lateral
 Medial
+ Anterior
 Posterior
+ Proximal
 Distal
+ Lateral
 Medial
+ Anterior
 Posterior
+ Proximal
 Distal
Near 90°
Flexion
1.81
(.39)
45.80
(.29)
8.30
(.28)
8.18
(.21)
43.79
(.31)
23.06
(.18)
One Knee
3.07
(.92)
53.40
(.51)
12.75
(.51)
8.31
(.72)
55.88
(.60)
28.99
(.37)
Near Full
Flexion
4.49
(.16)
19.93
(.28)
16.12
(.19)
6.24
(.29)
22.22
(.19)
18.85
(.26)
Squat
6.74
(.61)
16.56
(.79)
15.14
(.73)
10.52
(.46)
11.42
(.80)
16.86
(.41)
Standing
.66
(.08)
5.12
(.29)
42.20
(.69)
Figure 15: Net external forces normalized by body weight for Subject 1 kneeling near full flexion without
kneepads
41
4.4 KNEE MOMENTS
All moments are in respect to the anatomical shank coordinate system and normalized to a
percentage of the body weight*height to reduce the differences in moments due to gender. [41]
The body weight in Newtons was multiplied by the height in meters, yielding a value in Nm. All
moments were divided by this body weight*ht value, yielding moments with units in %BW*Ht.
Average net moments for each posture with and without kneepads are shown in Table 5 and
Table 6. All sagittal moments imposed on the knee due to static kneeling were flexion moments.
Adduction moments occurred in Subject 1 for all postures. Adduction moments were created in
Subject 2 when kneeling near 90° flexion and kneeling on one knee with kneepads. Abduction
moments were shown in all postures without kneepads and when kneeling near full flexion and
squatting with kneepads. Subject 1 showed external rotation moments when kneeling near full
flexion without kneepads and when squatting. Kneeling near 90° flexion, kneeling on one knee,
and kneeling near full flexion resulted in internal rotation moments. Subject 2 showed internal
rotation moments for all postures. Time series showing the sagittal moment contributions for
Subject 1 kneeling near full flexion without kneepads are shown in Figure 16.
42
Table 5: Average net external knee moments normalized by Body weight*Height, Subject 1
Subject 1
With Kneepads Without Kneepads
Flexion Adduction
+ Internal
 External
Rotation
Flexion Adduction
+ Internal
 External
Rotation
Near 90°
Flexion
.14
(.11)
.07
(.09)
.81
(.03)
.46
(.09)
.21
(.05)
.26
(.01)
One Knee
.60
(.23)
.71
(.38)
.92
(.04)
1.00
(.19)
.86
(.05)
.32
(.04)
Near Full
Flexion
4.42
(.15)
1.16
(.04)
.39
(.03)
4.27
(.23)
.91
(.06)
.05
(.03)
Squat
5.52
(.31)
1.30
(.08)
.25
(.03)
5.85
(.28)
.79
(.16)
.32
(.04)
Standing
1.76
(.08)
.12
(.03)
.07
(.01)
Table 6: Average net external knee moments normalized by Body weight*Height, Subject 2
Subject 2
With Kneepads Without Kneepads
Flexion
Moment
+ Abduction
 Adduction
Internal
Rotation
Flexion Abduction
Internal
Rotation
Near 90°
Flexion
.63
(.10)
.24
(.01)
.53
(.02)
3.71
(.06)
.33
(.03)
.54
(.01)
One Knee
6.45
(.11)
.69
(.21)
.34
(.04)
6.05
(.07)
.06
(.1)
.08
(.04)
Near Full
Flexion
3.08
(.05)
.42
(.02)
.25
(.03)
2.93
(3.01)
.57
(.58)
.32
(.33)
Squat
5.51
(.19)
1.21
(.12)
.17
(.03)
4.49
(.12)
1.90
(.08)
.19
(.03)
Standing
.17
(.06)
.28
(.02)
.04
(.01)
43
Figure 16: Moment contributions normalized by Bodyweight*Height for Subject 1 kneeling near full flexion
without kneepads
44
4.5 SUMMARY SUBJECT 1
Kneeling near 90° flexion showed the smallest flexion moments, followed by kneeling on one
knee, kneeling near full flexion, and squatting. When kneeling near 90° flexion about 47 %BW
was placed on the right knee and another 0.27 %BW was placed on the toes. The moments
generated by these forces were almost balanced by the weight of the lower leg, creating a net
flexion moment of 0.3% BW*Ht. This posture had the smallest abduction moments: 3 Nm
without kneepads and 1 Nm with kneepads. External rotation moments were higher with
kneepads (10 Nm) than without (3.15 Nm). The higher rotation moment with kneepads was
almost entirely due to the moment created by the force at the knee (9.87 Nm). This increase in
moment was due to the kneepad changing the COP of the force at the knee. The force in the z
direction moved from 1.4 cm to 3.5 cm medial to the KJC, increasing the abduction moment by
nearly 7 Nm.
When kneeling on one knee a predominant amount of weight, 75 %BW, was placed on
the knee and a small amount of weight, .54 %BW, was placed on the toes. These forces created
a flexion moment which was reduced by the extension moment generated by the weight of the
lower leg, creating a mean net flexion moment of 0.8 % BW*Ht. Adduction moments in this
posture were lower with kneepads (9 Nm) than without (10 Nm). Internal rotation moments
were larger with kneepads (11 Nm) than without (3 Nm). This increase in internal rotation
moment was also due to the kneepad changing the location of the COP of the force at the knee
from 1.4 cm to 2.9 cm medial to the KJC.
When kneeling near full flexion the weight at the knee was 28 %BW. The weight at the
toes was higher than any other kneeling posture, 21 %BW, due to the subject sitting on their
heels, applying a force of 8 %BW. The contact between the thigh and the calf was 18 %BW.
45
The thighcalf and heelgluteus contact forces along with the weight of the lower leg created
extension moments which acted to stabilize the knee joint in this high flexion posture. The
flexion moment generated by the weight at the toes and the knees were still sufficient to create a
net flexion moment at the knee of 4 %BW*Ht. Adduction moments were larger with kneepads
(14 Nm) than without (11 Nm). When wearing kneepads a 5 Nm internal rotation moment was
created and without kneepads a 1 Nm external rotation moment was created. This difference in
transverse moments may also be attributed to the COP of the force at the knee moving from 1.3
cm to 4.2 cm medial to the KJC.
Squatting applied a force of 44 %BW to the foot, and created a thighcalf contact force of
25 %BW. The thighcalf contact force along with the weight of the lower leg created an
extension moment in opposition to the flexion moment created by the ground reaction force at
the foot, with a net flexion moment of 6 %BW*Ht. Adduction moments were higher with
kneepads (16 Nm) than without (10 Nm). The source of this difference was the xcomponent of
the force at the foot which increased from 20 N without kneepads to 40 N with kneepads. This
difference is most likely due to the subjects kneeling posture which may or may not have been
directly related to the kneepad. External rotation moments were similar for squatting with (5.5
Nm) and without (5.9 Nm) kneepads.
4.6 SUMMARY SUBJECT 2
Subject 2 showed high variability in knee forces, moments, and angles between kneepad states.
When not wearing kneepads the smallest flexion moment was 34 Nm when kneeling near full
flexion, followed by 42 Nm when kneeling near 90° flexion, 51 Nm when squatting and the
46
highest flexion moment was 68 Nm when kneeling on one knee. The largest abduction moments
were created when squatting, followed by keeling near full flexion, kneeling near 90° flexion,
and kneeling on one knee. Internal rotation moments were largest when kneeling near 90°
flexion, followed by kneeling near full flexing, squatting, and kneeling on one knee. When
wearing kneepads kneeling near 90° flexion had the smallest flexion moment of 7 Nm, followed
by 35 Nm when kneeling near full flexion, 62 Nm when squatting, and 73 Nm when kneeling on
one knee. Squatting and kneeling near full flexion created adduction moments and kneeling one
knee and kneeling near 90° flexion created adduction moments. Internal rotation moments were
largest when kneeling near 90° flexion, followed by kneeling on one knee, kneeling near full
flexion, and squatting.
Squatting and kneeling on one knee had the largest flexion moments for both knee pad
states. These postures were found to have similar flexion angles: 37° for squatting and 38° for
kneeling on one knee. Kneeling on the right knee showed the largest flexion moments because
most of their body weight was distributed to their right side. The left leg may have acted to
provide balance and stability. Some 60 %BW was distributed to the right side with 40 %BW at
the right knee and another 20 %BW at the right toes. This created thighcalf contact which was
not accounted for in the analysis of their data and may have caused an overestimation of the
flexion moments. To determine the effect of thighcalf contact on this moment, the thighcalf
contact moment created in the squatting posture was applied to this posture. When including this
contact the flexion moment decreased from 73 Nm to 46 Nm. This new moment is less than
squatting (62 Nm), but still higher than that during full flexion (35 Nm).
47
4.7 STATISTICS
In this study two subjects were asked to simulate 4 postures with and without kneepads while
measurements were recorded. To determine the effect of the kneepads on the forces and
moments at the knee, multivariate ANOVA was performed for each subject. The mean values of
the forces, moments, and angles were compared between kneepad states. A pvalue of 0.05 was
used to demonstrate statistical significance. Table 7 and Table 8 show the resulting pvalues
from ANOVA where significant values are in bold font.
No significant differences were found between kneepad states for Subject 1. This was
expected as Subject 1’s posture did not change much (04.3° change in joint angles) between
kneepad states. Subject 1 showed a significant difference in sagittal moment (p <.001), frontal
moment (p = .037), posterior/anterior force (p <.001), distal/proximal force (p <.001), included
knee angle (p <.001), and internal/external rotation angle (p = .001) between postures. This was
also expected due to the differences in joint angles. Kneeling near 90° flexion created flexion
angles of 83°, compared to the 23° of kneeling near full flexion. Internal/external rotation angles
also varied between postures and these varied joint angles may have largely contributed to the
differences in joint forces and moments between postures.
Subject 2 showed a significant difference in medial/lateral force magnitude between
kneepad states (p = .021). This difference is not thought to be a result of the kneepad itself, but
of Subject 2 not repeating the postures correctly. When originally instructed to kneel near 90°
flexion, their included angle was 85°. When later instructed to kneel their included angle
decreased to 39°. This 46° increase in knee flexion, along with the 8° changes in valgus angles
may have largely contributed to the significant difference in medial/lateral forces. Subject 1
showed significant differences in frontal moment (p = .040), posterior/anterior force (p = .000),
48
and distal/proximal force (p = .027) between postures. These differences were also expected as
kneeling near 90° flexion and kneeling on one knee created increased posterior forces, and
squatting created proximal forces and abduction moments compared to the distal forces and
adduction moments created in some other postures.
Table 7: Resulting pvalues for ANOVA, Subject 1
Subject 1
Kneepad Posture
M
o
m
e
n
t
s
Sagittal .908 <.001
Frontal .727 .037
Transverse .221 .148
F
o
r
c
e
s
Medial/Lateral .443 .288
Posterior/Anterior .934 <.001
Distal/Proximal .982 <.001
A
n
g
l
e
s
Included .948 <.001
Varus/Valgus .801 .485
Int/Ext Rotation .984 .001
49
Table 8: Resulting pvalues for ANOVA, Subject 2
Subject 2
Kneepad Posture
M
o
m
e
n
t
s
Sagittal .795 .074
Frontal .387 .040
Transverse .776 .065
F
o
r
c
e
s
Medial/Lateral .021 .692
Posterior/Anterior .967 <.001
Distal/Proximal .546 .027
A
n
g
l
e
s
Included .339 .392
Varus/Valgus .276 .564
Int/Ext Rotation .125 .521
Data from Subject 1 and Subject 2 were combined to see if there was a significant
difference in forces, moments, and joint angles due to the interaction of kneepad and posture.
No significant differences were found, Table 9.
Table 9: Resulting pvalues for ANOVA of Kneepad*Posture Interaction, Subjects 1&2
Kneepad*Posture
Interaction
M
o
m
e
n
t
s
Sagittal .903
Frontal .993
Transverse .758
F
o
r
c
e
s
Medial/Lateral .970
Posterior/Anterior .964
Distal/Proximal .805
A
n
g
l
e
s
Included .802
Varus/Valgus .153
Int/Ext Rotation .929
50
ANOVA was also performed to determine if significant differences existed between
Subject 1 and Subject 2. Significant differences in all moments, posterior forces, and included
angles were shown to exist, Table 10.
Table 10: Resulting pvalues for ANOVA, Subject comparison
pvalue
M
o
m
e
n
t
s
Sagittal <.001
Frontal <.001
Transverse <.001
F
o
r
c
e
s
Medial/Lateral .65
Posterior/Anterior <.001
Distal/Proximal .16
A
n
g
l
e
s
Included <.001
Varus/Valgus .06
Int/Ext Rotation .52
51
5.0 SENSITIVITY ANALYSIS
To determine the major sources of error in the moment calculations, sensitivity analyses
were performed. An analysis of the model parameters was performed for each posture to
determine the effect of varying the moment arms and forces on the sagittal moments. Another
analysis was performed, varying the location of the knee joint center and measuring its effect on
the knee moments. All mean values used in this analysis were from Subject 1 without kneepads.
5.1 VARYING MODEL PARAMETERS
The mean values of the zcomponents of the moment arms and the ycomponents of the forces
were varied to determine their effect on the sagittal knee moments. Although static postures
were studied there was some expected motion of the markers due to motion artifact and the
subject maintaining balance. The motion of the markers was shown to be less than 1.5 cm over
the course of a squatting trial. To include this source of error the moment arms were varied by ±
13 cm in 1 cm increments. One force plate in the laboratory had a threshold of approximately 3
N. To account for this error, the forces were varied by ± 26 N in 2 N increments. External
force diagrams for all postures are shown in Figure 16. Plots of the varied parameters and
resulting moments for kneeling near 90° flexion, kneeling on one knee, kneeling near full
flexion, and squatting are shown in Figure 17, Figure 18, Figures 19 and 20, and Figure 21,
52
respectively. In all figures the blue diamond represents the mean value of the model parameter
and its corresponding sagittal moment. The pink squares represent the sagittal moments
resulting from varying the parameter. The green line connecting the green triangles represents
the standard deviation of the model parameter. This illustrates the amount of variability of the
sagittal moment that is expected due to the model parameter. Table 10 shows the sensitivity of
the sagittal moments to the model parameters. Varied moment arms resulted in sensitivities with
units of Nm/cm. Varied moment arm results in change in moment per centimeter and varied
forces results in change in moment per Newton. Also shown are the percentages of change in
sagittal moment per centimeter or per Newton. Parameters creating the largest changes in
moments are shown in bold. Parameters with the least change in moment are shown in italics.
53
20
15
10
5
0
5
6 4 2 0 2
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of F2 moment arm, cm
20
15
10
5
0
5
30 35 40 45
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent low leg weight, N
20
15
10
5
0
5
64 62 60 58 56 54
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of F1 moment arm, cm
20
15
10
5
0
5
15 12 9 6 3 0
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent of F1 force, N
20
15
10
5
0
5
318 315 312 309 306 303 300
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent of F2, N
20
15
10
5
0
5
25 23 21 19 17 15
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of low leg moment arm, cm
Mean Value Varied SD
Figure 17: Varied forces and moment arms and corresponding sagittal knee moments for kneeling near 90°
flexion. The forces and moment arms were varied for the ground reaction force at the toes (F1), the ground
reaction force at the right knee (F2), and the weight of the foot+shank (low leg). Note that all moment arm
values are shown in cm, all forces are shown in N, and all moments are not normalized and are shown in Nm.
Varying the zcomponent of the F1 moment arm by 3 cm changed the sagittal knee moment by .2 Nm. When
F1 was varied by 6 N, the moment changed by 3.5 Nm. Varying the zcomponent of the F2 moment arm by 3
cm was sufficient to more than triple the magnitude of the sagittal moment and change its interpretation.
Varying the ycomponent of the F2 force by 6 N, changed the sagittal moment by .1 Nm. Varying the z
component of the COM of the lower leg by 3 cm changed the moment by 1.1 Nm. Varying the ycomponent
of the low leg weight changed the moment by 1.2 Nm.
54
30
20
10
0
10
23 20 17 14 11 8
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent of F1, N
30
20
10
0
10
30 33 36 39 42 45
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent low leg weight, N
30
20
10
0
10
60 58 56 54 52 50
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of F1 moment arm, cm
30
20
10
0
10
6 4 2 0 2
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of F2 moment arm, cm
30
20
10
0
10
490 485 480 475 470 465
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent of F2, N
30
20
10
0
10
25 23 21 19 17 15
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of low leg moment arm, cm
Mean Value Varied SD
Figure 18: Varied forces and moment arms and corresponding sagittal knee moments for kneeling on the
right knee. The forces and moment arms were varied for the ground reaction force at the toes (F1), the
ground reaction force at the right knee (F2), and the weight of the foot+shank (low leg). Varying the z
component of the F1 moment arm by 3 cm, changed the sagittal moment by .5 Nm. Varying the ycomponent
of F1 by 6 N, changed the sagittal moment by 3.3 Nm. Decreasing the zcomponent of the F2 moment arm by
3 cm more than doubled the sagittal moment. When this moment arm was increased by 3 cm, an extension
moment was created. Varying the ycomponent of F2 by 6 N changed the moment by .1 Nm. Varying the z
component of the low leg COM by 3 cm and the ycomponent of the low leg weight by 6 N changed the
moment by 1.2 Nm.
55
60
55
50
45
40
65 64 63 62 61 60 59 58
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of F1 moment arm, cm
60
55
50
45
40
155 150 145 140 135
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent of F1, N
60
55
50
45
40
30 35 40 45
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent of low leg weight, N
60
55
50
45
40
25 23 21 19 17 15
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of low leg moment arm, cm
60
55
50
45
40
200 195 190 185 180
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent of F2 , N
60
55
50
45
40
9 7 5 3 1
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of F2 moment arm, cm
Mean Value Varied SD
Figure 19: Varied forces and moment arms and corresponding sagittal knee moments for kneeling near full
flexion. The forces and moment arms were varied for the ground reaction force at the toes (F1), the ground
reaction force at the right knee (F2), and the weight of the foot+shank (low leg), the thighcalf contact force,
and the heelgluteus contact force. Varying the zcomponent of the F1 moment arm by 3 cm changed the
moment by 4.4 Nm. Varying the ycomponent of F1 by 6 N changed the moment by 3.7 Nm. Varying the z
component of the F2 moment arm by 3 cm changed the moment by 5.7 Nm. Varying the ycomponent of F2
by 6 N changed the moment by .3 N. Varying the zcomponent of the low leg COM by 3 cm changed the
moment by 1.2 Nm. Varying the ycomponent of the low leg weight by 6 N changed the moment by 1.2 Nm.
56
60
55
50
45
40
45 50 55 60
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent of Heelgluteus Force, N
60
55
50
45
40
48 46 44 42 40 38
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
z component of Heelgluteus moment arm, m
60
55
50
45
40
110 115 120 125 130 135
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
y component Thighcalf Force, N
60
55
50
45
40
20 18 16 14 12 10 8
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of Thighcalf moment arm, cm
Mean Value Varied SD
Figure 20: Varied forces and moment arms and corresponding sagittal knee moments for kneeling near full
flexion, continued. The forces and moment arms were varied for the ground reaction force at the toes (F1),
the ground reaction force at the right knee (F2), and the weight of the foot+shank (low leg), the thighcalf
contact force, and the heelgluteus contact force. Varying the zcomponent of the thighcalf moment arm by 3
cm changed the moment by 3.7 Nm. Varying the thighcalf contact force by 6 N changed the moment by .8
Nm. Varying the heelgluteus moment arm by 3 cm changed the moment by 1.5 Nm. Varying the heel
gluteus contact force by 6 N changed the moment by 2.5 Nm.
57
90
80
70
60
160 165 170 175
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
ycomponent of Thighcalf force , N
90
80
70
60
48 47 46 45 44 43 42 41
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
z component of F2 moment arm, m
90
80
70
60
315 310 305 300 295 290 285
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
y component of F2, N
90
80
70
60
25 23 21 19 17 15
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
z component of low leg moment arm, m
90
80
70
60
24 26 28 30 32 34 36 38
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
y component low leg weight, N
90
80
70
60
18 16 14 12 10
S
a
g
i
t
t
a
l
M
o
m
e
n
t
,
N
m
zcomponent of Thighcalf moment arm, cm
Mean Value Varied SD
Figure 21: Varied forces and moment arms and corresponding sagittal knee moments for squatting. The
forces and moment arms were varied for the ground reaction force at the foot (F1), the weight of the
foot+shank (low leg), and the thighcalf contact force. Varying the zcomponent of the F2 moment arm by 3
cm, changed the moment by 9 Nm. Varying the ycomponent of F2 by 6 N, changed the moment by 2.7 Nm.
Varying the zcomponent of the low leg COM by 3 cm, changed the moment by .9 Nm. Varying the y
component of the low leg weight by 6 N, changed the moment by 1.2 Nm. Varying the zcomponent of the
thighcalf moment arm by 3 cm, changed the moment by 5.1 Nm. Varying the ycomponent of the thighcalf
contact force by 6 N, changed the moment by .9 Nm.
58
Table 11: Sensitivity of sagittal moments to changes in model parameters for all postures
Varied Parameter Model Sensitivity
Near 90°
Flexion
F1 moment arm
.06 Nm/cm
1.13 %/cm
F1
.55 Nm/N
10.5 %/N
F2 moment arm
3.1 Nm/cm
55 %/cm
F2
.03 Nm/N
.44 %/N
low leg moment arm
.35 Nm/cm
6.2 %/cm
low leg weight
.2 Nm/N
3.6 %/N
One Knee
F1 moment arm
.16 Nm/cm
1.2 %/cm
F1
.55 Nm/N
4.8 %/N
F2 moment arm
4.8 Nm/cm
37 %/cm
F2
.02 Nm/N
.13 %/N
low leg moment arm
.39 Nm/cm
3 %/cm
low leg weight
.2 Nm/N
1.6 %/N
Squat
F2 moment arm
3 Nm/cm
4.2 %/cm
F2
.4 Nm/N
.55 %/N
low leg moment arm
.31 Nm/cm
.44 %/cm
low leg weight
.2 Nm/N
.27 %/N
Thighcalf moment arm
1.7 Nm/cm
2.4 %/cm
Thighcalf Force
.15 Nm/N
.2 %/N
59
Table 11 (continued).
Near Full
Flexion
F1 moment arm
1.5 Nm/cm
2.8 %/cm
F1
.61 Nm/N
1.7 %/N
F2 moment arm
1.9 Nm/cm
3.7 %/cm
F2
.05 Nm/N
.1 %/N
low leg moment arm
.39 Nm/cm
.75 %/cm
low leg weight
.2 Nm/N
.38 %/N
Thighcalf moment arm
1.2 Nm/cm
2.4 %/cm
Thighcalf force
.14 Nm/N
.27 %/N
Heelgluteus moment arm
.44 Nm/cm
.84 %/cm
Heelgluteus force
.48 Nm/N
.9 %/N
In the kneeling near 90° flexion model, the location of the force at the knee (F2)
increasing the sagittal knee moment by 3.1 Nm per every centimeter this force was moved
distally. The sagittal moment was the least sensitive to the force at the knee; increasing by.03
Nm for every 1 N increase in force. Similarly to the kneeling near 90 flexion model, the
kneeling on one knee model was the most sensitive to the location of the force at the knee;
increasing by 4.8 Nm with every 1 cm this force was moved distally. The sagittal moment was
the least sensitive to the force at the knee, increasing by .02 Nm with a 1 N increase in force. In
the squatting model, the sagittal moment was the most sensitive to the location of the force at the
foot, increasing by 3 Nm for every 1 cm this force was moved distally. The sagittal moment was
the least sensitive to the thighcalf contact force; decreasing by .15 Nm with every 1 N this force
60
was increased. In the kneeling near full flexion model, the sagittal moment was the most
sensitive to the location of the force at the knee increasing by 1.9 Nm for every 1 cm this force
was moved distally. The sagittal moment was least sensitive to the force at the knee; increasing
by .05 Nm for every 1 N increase in force. Moving the location of the thighcalf contact force 1
cm distally decreased sagittal moments by more than 1 Nm in the kneeling near full flexion and
squatting models.
5.2 VARYING KNEE JOINT CENTER LOCATION
Knee joint center (KJC) locations are dependent on the placement of surface markers by
laboratory researchers. Although one researcher was responsible for palpating the medial and
lateral epicondyles of the femur, some variation in the placement of these markers was expected.
This variation in marker placement will affect the location of the KJC, thereby affecting all
moment calculations. A sensitivity analysis was performed on the KJC location to determine the
effect of its variation on the knee moments. The KJC was varied by ±3 cm along the anatomical
shank axes. Changes in the KJC locations resulted in linear changes in the joint moments. Plots
showing varied KJC locations and resulting moments for kneeling near 90° flexion, kneeling on
one knee, kneeling near full flexion and squatting are shown in Figure 22, Figure 23, Figure 24,
and Figure 25, respectively. Percent change in moments resulting from variation in KJC
locations for these postures are shown in Table 8, Table 9, Table 10, and Table 11, respectively.
Although no moments changed by an order of magnitude, some variations were sufficient to
change the interpretation of the moments.
61
Figure 22: Varied KJC locations and resulting knee moments for kneeling near 90° flexion. Varying the x, y,
and zcomponents of the KJC location had no effect on the sagittal, frontal, and transverse moments,
respectively. Varying the ycomponent by 3 cm changed the sagittal moment by 1.3 Nm. Varying the z
component by 3 cm more than doubled the sagittal moment. Varying the xcomponent by 3 cm changed the
frontal moment by 1.3 Nm. Varying the zcomponent by 3 cm changed the adduction moment by 1.7 Nm.
Varying the xcomponent by 3 cm and varying the ycomponent by 3cm more than doubled the transverse
moment and changed the interpretation of this moment.
62
Sagittal moments were most sensitive to the zlocation of the KJC when kneeling near
90° flexion. This moment increased from a 5.7 Nm to a 14 Nm flexion moment when the KJC
was moved 3 cm anterior. When the KJC was moved 3 cm posterior the moment decreased to a
2.6 Nm extension. Adduction moments were most sensitive to the zlocation of the KJC,
increasing from 2.5 Nm to 4.3 Nm when the KJC was moved 3 cm posterior, and decreasing to
.8 Nm when the KJC was moved 3 cm anterior. Transverse moments were most sensitive to the
xposition of the KJC increasing from 3.3 Nm to 11.6 Nm when the KJC was moved 3 cm
lateral. This moment became an external rotation moment of 5.1 Nm when the KJC was moved
3 cm medial.
Table 12: Percent change in moments due to varying KJC for kneeling near 90 degrees flexion
Varied by, m
Change in Moment, %
Sagittal Frontal Transverse
[.01, 0, 0] 0.00 16.46 %/cm 84.87 %/cm
[.02, 0, 0] 0.00 32.91 169.73
[.03, 0, 0] 0.00 49.37 254.60
[.01, 0, 0] 0.00 16.46 84.87
[.02, 0, 0] 0.00 32.91 169.74
[.03, 0, 0] 0.00 49.37 254.60
[0, .01, 0] 7.29 0.00 17.70
[0, .02, 0] 14.59 0.00 35.39
[0, .03, 0] 21.88 0.00 53.09
[0, .01, 0] 7.29 0.00 17.70
[0, .02, 0] 14.59 0.00 35.39
[0, .03, 0] 21.88 0.00 53.09
[0, 0, .01] 48.52 22.83 0.00
[0, 0, .02] 97.04 45.66 0.00
[0, 0, .03] 145.56 68.48 0.00
[0, 0, .01] 48.52 22.83 0.00
[0, 0, .02] 97.04 45.65 0.00
[0, 0, .03] 145.57 68.48 0.00
63
Figure 23: Varied KJC locations and resulting knee moments for kneeling on right knee. Varying the x, y,
and zcomponents of the KJC location had no effect on the sagittal, frontal, and transverse moments,
respectively. Varying the ycomponent by 3 cm changed the flexion moment by 2.7 Nm. Varying the z
component by 3 cm changed the sagittal moment by 13.6 Nm and changed it interpretation. Varying the x
component by 3 cm changed the adduction moment by 2.7 Nm. Varying the zcomponent changed the frontal
moment by 1.5 Nm. Varying the xcomponent by 3 cm changed the transverse moment by 9.1 Nm. Varying
the ycomponent by 3 cm changed the transverse moment by 1.5 Nm.
64
Flexion moments were most sensitive to the zlocation of the KJC when kneeling on one
knee. This moment increased from 12.8 Nm to 26 Nm when the KJC was moved 3 cm anterior.
When the KJC was moved 3 cm posterior the moment became a .78 Nm extension moment.
Adduction moments were most sensitive to the xlocation of the KJC. Increasing from 10.6 Nm
to 13.3 Nm when the KJC was moved 3 cm anterior, and decreasing to 7.8 Nm when the KJC
was moved 3 cm posterior. Transverse moments were most sensitive to the xposition of the
KJC. The internal rotation moment changed from 3.5 Nm to 17.1 Nm when the KJC was moved
3 cm medial. This moment became an external rotation moment of 10.1 Nm when the KJC was
moved 3 cm lateral.
Table 13: Percent change in moments due to varying KJC for kneeling on one knee
Varied by, m
Change in Moment, %
Sagittal Frontal Transverse
[.01, 0, 0] 0.00 8.64 129.27
[.02, 0, 0] 0.00 17.28 258.54
[.03, 0, 0] 0.00 25.91 387.81
[.01, 0, 0] 0.00 8.64 129.27
[.02, 0, 0] 0.00 17.28 258.54
[.03, 0, 0] 0.00 25.91 387.81
[0, .01, 0] 7.13 0.00 14.28
[0, .02, 0] 14.27 0.00 28.56
[0, .03, 0] 21.40 0.00 42.84
[0, .01, 0] 7.13 0.00 14.28
[0, .02, 0] 14.27 0.00 28.56
[0, .03, 0] 21.40 0.00 42.84
[0, 0, .01] 35.36 4.73 0.00
[0, 0, .02] 70.71 9.46 0.00
[0, 0, .03] 106.07 14.19 0.00
[0, 0, .01] 35.36 4.73 0.00
[0, 0, .02] 70.71 9.46 0.00
[0, 0, .03] 106.07 14.19 0.00
65
Figure 24: Varied KJC locations and resulting knee moments for kneeling near full flexion. Varying the x, y,
and zcomponents of the KJC location had no effect on the sagittal, frontal, and transverse moments,
respectively. Varying the ycomponent by 3 cm changed the flexion moment by 1.6 Nm. Varying the z
component by 3 cm changed the flexion moment by 3.6 Nm. Varying the xcomponent by 3 cm changed the
adduction moment by 1.6 Nm. Varying the zcomponent by 3 cm changed the adduction moment by 1.1 Nm.
Varying the xcomponent by 3 cm changed the transverse moment by 3.6 Nm and changed its interpretation.
Varying the ycomponent by 3 cm changed the transverse moment by 1.4 Nm and changed its interpretation.
66
Flexion moments were most sensitive to the zlocation of the KJC when kneeling near
full flexion. This moment increased from 51.75 Nm to 55 Nm when the KJC was moved 3 cm
anterior and decreased to 48.2 Nm when the KJC was moved 3 cm posterior. Adduction
moments were most sensitive to the xlocation of the KJC, increasing from 11.1 Nm to 12.6 Nm
when the KJC was moved 3 cm lateral, and decreasing to 9.5Nm when the KJC was moved 3 cm
medial. Transverse moments were most sensitive to the xposition of the KJC. The external
rotation moment increased from .65 Nm to 4.2 Nm when the KJC was moved 3 cm medial. This
moment became an internal rotation moment of 2.93 Nm when the KJC was moved 3 cm lateral.
Table 14: Percent change in moments due to varying KJC for kneeling near full flexion
Varied by, m
Change in Moment, %
Sagittal Frontal Transverse
[.01, 0, 0] 0.00 4.66 184.53
[.02, 0, 0] 0.00 9.35 369.06
[.03, 0, 0] 0.00 14.02 553.59
[.01, 0, 0] 0.00 4.67 184.55
[.02, 0, 0] 0.00 9.35 369.07
[.03, 0, 0] 0.00 14.02 553.60
[0, .01, 0] 0.99 0.00 58.59
[0, .02, 0] 1.99 0.00 117.17
[0, .03, 0] 2.99 0.00 175.76
[0, .01, 0] 1.00 0.00 58.60
[0, .02, 0] 2.00 0.00 117.19
[0, .03, 0] 3.00 0.00 175.77
[0, 0, .01] 2.31 3.43 0.00
[0, 0, .02] 4.61 6.85 0.00
[0, 0, .03] 6.92 10.28 0.00
[0, 0, .01] 2.30 3.43 0.00
[0, 0, .02] 4.61 6.85 0.00
[0, 0, .03] 6.91 10.28 0.00
67
Figure 25: Varied KJC locations and resulting knee moments for squatting. Varying the x, y, and z
components of the KJC location had no effect on the sagittal, frontal, and transverse moments, respectively.
Varying the ycomponent by 3 cm changed the flexion moment 5.6 Nm. Varying the zcomponent by 3 cm
changed the flexion moment by 3 Nm. Varying the xcomponent by 3 cm changed the adduction moment by
5.6 Nm. Varying the zcomponent by 3 cm changed the adduction moment by 0.6 Nm. Varying the x
component by 3 cm changed the transverse moment by 3 Nm. Varying the ycomponent by 3 cm changed the
external rotational moment by .6 Nm.
68
Flexion moments were most sensitive to the ylocation of the KJC when squatting. This
moment increased from 71.2 Nm to 76.9 Nm when the KJC was moved 3 cm superior. When
the KJC was moved 3 cm inferior the moment decreased to 65.6 Nm. Adduction moments were
most sensitive to the xlocation of the KJC. Increasing from 9.59 Nm to 15.2 Nm when the KJC
was moved 3 cm medial, and decreasing to 3.9 Nm when the KJC was moved 3 cm lateral.
Transverse moments were most sensitive to the xposition of the KJC. The external rotation
moment increased from 3.8 Nm to 6.8 Nm when the KJC was moved 3 cm medial and decreased
to .8 Nm when the KJC was moved 3 cm lateral.
Table 15: Percent change in moments due to varying KJC for squatting
Varied by, m
Change in Moment, %
Sagittal Frontal Transverse
[.01, 0, 0] 0.00 19.61 26.12
[.02, 0, 0] 0.00 39.22 52.23
[.03, 0, 0] 0.00 58.82 78.35
[.01, 0, 0] 0.00 19.61 26.12
[.02, 0, 0] 0.00 39.21 52.23
[.03, 0, 0] 0.00 58.82 78.35
[0, .01, 0] 2.64 0.00 5.16
[0, .02, 0] 5.28 0.00 10.32
[0, .03, 0] 7.92 0.00 15.48
[0, .01, 0] 2.64 0.00 5.16
[0, .02, 0] 5.28 0.00 10.32
[0, .03, 0] 7.92 0.00 15.48
[0, 0, .01] 1.41 2.06 0.00
[0, 0, .02] 2.82 4.13 0.00
[0, 0, .03] 4.22 6.20 0.00
[0, 0, .01] 1.41 2.07 0.00
[0, 0, .02] 2.82 4.13 0.00
[0, 0, .03] 4.22 6.20 0.00
69
Summaries of model sensitivities to KJC locations for all postures are shown in Table 16.
Values are shown due to varying KJC locations by 1 cm in the lateral, anterior, and proximal
directions. Sensitivity values are shown in Nm changes per cm and percentage of moment
change per cm. Greatest changes in moments are shown in bold. The largest sensitivities where
shown in the transverse moments due to moving the KJC 1 cm laterally in all postures.
Table 16: Sensitivity of moments to KJC location for all postures
Posture Varied direction
Sensitivity (Nm/cm & %/cm) of Moments
Sagittal Frontal Transverse
Kneeling Near
90° Flexion
lateral no change
.42 Nm/cm
16.5 %/cm
2.8 Nm/cm
84.9 %/cm
anterior
.42 Nm/cm
7.3 %/cm
no change
.58 Nm/cm
17.7 %/cm
proximal
2.8 Nm/cm
48.5 %/cm
.58 Nm/cm
22.8 %/cm
no change
Kneeling on One
Knee
lateral no change
.9 Nm/cm
8.6 %/cm
4.5 Nm/cm
129.3 %/cm
anterior
.92 Nm/cm
7.1 %/cm
no change
.5 Nm/cm
14.3 %/cm
proximal
4.5 Nm/cm
35.4 %/cm
.5 Nm/cm
4.7 %/cm
no change
Squatting
lateral no change
1.89 Nm/cm
19.6 %/cm
1 Nm/cm
26.1 %/cm
anterior
1.9 Nm/cm
2.6 %/cm
no change
.20 Nm/cm
5.2 %/cm
proximal
1 Nm/cm
1.4 %/cm
.2 Nm/cm
2.1 %/cm
no change
Kneeling Near
Full Flexion
lateral no change
.52 Nm/cm
4.7 %/cm
1.2 Nm/cm
184.5 %/cm
anterior
1 Nm/cm
1 %/cm
no change
.38 Nm/cm
58.6 %/cm
proximal
1.2 Nm/cm
2.3 %/cm
.38 Nm/cm
3.4 %/cm
no change
70
6.0 MODEL VALIDATION
The developed computational model is based on techniques which have been previously
used to determine the 3D joint moments in gait analysis. However, due to the complexity of
using segment coordinate systems, the model was validated by several means. To ensure
accuracy of the anatomical coordinate system calculations, the origin of the anatomical
coordinate system was verified to equal to the KJC. The KJC location in the GCS was
multiplied by [T
GAshank
]
1
. The result was point [0 0 0], thereby validating the KJC as the origin
of the ASCS.
Due to the lack of research on kneeling knee forces and moments, no direct comparisons
could be made to force and moment values reported in literature. Instead the force and moment
calculations were validated by determining the forces and moments during quiet standing and
comparing these values to published research. Quiet standing produced a mean flexion moment
of .27 Nm/kg, a mean adduction moment of .15 Nm/kg, and a mean internal rotational moment
of .04 Nm/kg applied to the knee joint. These values were also comparable to published
research, thereby validating the model. [42, 43]
Joint angles were compared to values reported in published research. Hemmerich
reported flexion, external rotation, and adduction angles between of 153157°, 911°, and 68°,
respectively for squatting and 144155°, 1112°, and 710°, respectively for kneeling near full
flexion. [44] Although flexion angles agree with those in their research, the adduction and
71
external rotation angles are somewhat larger than those in this study. This is mostly due to the
differences in kneeling postures. In this study, Subject 1 performed kneeling near full flexion
with their buttocks resting on their heels. Subject 2 did not rest their buttocks on their heels, but
did kneel with their buttocks over their heels. In Hemmerich’s study, subjects performed
kneeling in full flexion without placing any weight on their heels and in a more upright posture.
This may have resulted in increased abduction moments. Also the orientation of the feet may
have created external rotation in their subjects, and internal rotation in Subjects 1 and 2 of this
study. Knee rotations also agree with the 11.1 ± 6.7° of internal rotation found in passive knee
flexion to 150°. [45] The knee joint angles when standing were also compared to published
research. The right knee had an included angle of 174179° with 2.42.6° valgus and .1.8° of
internal rotation, which is comparable to published research. [46]
72
7.0 DISCUSSION
This discussion section includes the sensitivity of the computational model to the model
parameters and knee joint center location which are important factors when applying this model
to future research. The effect of kneepads on the force and moments, subject variability, the
implications of findings, and the advancements are explained. Also the limitations of this study
which may restrict the extent to which these findings may be generalized are discussed.
7.1 KNEEPADS
Kneepads were not expected to significantly affect the knee joint angles and no statistically
significant differences in joint angles were found within subjects. Kneepads were also not
expected to significantly affect the forces at the knee. However, they were expected to affect the
moments at the knee joint. The knee pad was expected to change the location of the center of
pressure at the knee, which would change the moment arm of that force, thereby affecting the
moment. Although no statistically significant differences were found between kneepad states,
there were differences in the transverse moments which may have been caused by the kneepads.
In Subject 1 when kneeling near 90° flexion without kneepads, the moment arm of the ground
reaction force at the knee was [.014, .031, .025], contributing 2.14 Nm to the internal rotational
moment of the knee. When wearing kneepads this moment arm changed to [.036, .04, .021],
73
contributing 9.86 Nm. This implies that the COP of the groundreaction force at the knee was
shifted 2.2 mm medially when wearing kneepads. Forces were also different between kneepad
states [66.87, 308.91, 40.73] N and [51.25, 327.52, 47.90] N without and with kneepads,
respectively. However, the differences in lateral forces and distal distances accounted for merely
.07 Nm of the 7.72 Nm increase in internal rotational moment. The other 7.65 Nm were due to
the proximal force and its medial distance from the KJC. Similar changes in moments were seen
in the kneeling on one knee and kneeling near full flexion postures for this subject. The design
of the kneepad (articulating with a hard, contoured outer shell) is thought to have contributed to
this increase in transverse moment. This magnitude of torque may introduce significant changes
in the stresses and strains experienced by the internal stabilizing structures of the knee.
7.2 MODEL SENSITIVITY
7.2.1 Model Parameters
The developed computational model showed variations in sagittal moments as a result of
changes in moment arms and forces. No variation in moment arm or force changed the moment
by an order of magnitude, however some variations were sufficient to change the interpretation
of the moment from flexion to extension. When analyzing the kneeling on one knee and kneeling
near 90° flexion postures, the model was most sensitive to the location of the ground reaction
force at the knee, doubling or tripling the moment when varying the location by 3 cm. In
kneeling near full flexion and squatting, the model was most sensitive to the moment arm of the
74
thighcalf contact force, in some cases doubling the knee moment due to varying the COP by 3
cm. Sagittal moments changed from flexion to extension when the zcomponent of the COP at
the knee was increased by 3 cm for kneeling near 90° flexion and kneeling on one knee. All
efforts should be made to ensure proper positioning of all moment arms, especially the moment
arms of the thighcalf contact force and the COP of the force at the knee to reduce errors
associated with these model parameters.
7.2.2 Knee Joint Center Location
Variations in knee joint center location had large effects on the moment calculations and varied
between postures. Internal rotational moments were increased when the KJC was moved
laterally or posteriorly. Flexion moments were decreased when the KJC was moved posteriorly
and increased when the KJC was moved proximally. Adduction moments were increased when
the KJC was moved laterally or proximally. Due to the high sensitivity of moments to KJC
location care must be taken to ensure proper placement of surface markers on the medial and
lateral epicondyles. In this study the location of the medial and lateral epicondyle was marked
while the subject sat with their knee at 90°. One subject was responsible for palpating and
marking these anatomic landmarks. It is possible that changes in marker placement may create
error in moment calculations and joint angles. However the error in marker placement is
expected to be less than the 3 cm variations studied.
75
7.3 SUBJECT VARIABILITY
While the model was only tested on two subjects, there did appear to be differences between
them. Subject 1 and Subject 2 showed significantly different joint angles, forces, and moments
during kneeling. This variability may be due to their difference in stature as well as kneeling
postures. Inclusion of more subjects will be needed to determine any anthropometric effects on
the results with and without kneepads.
A comparison of the two subjects yields some insights into potential intersubject
variability. When kneeling near 90° flexion both subjects had similar forces and moments with
kneepads. Without kneepads, Subject 2 was in a more flexed posture creating increased flexion
moments. When kneeling on one knee Subject 1 was in a more upright posture, kneeling close to
90° with the right leg and the supporting left leg. Subject 2 was in a much more crouched
posture with joint angles closer to that in full flexion and squatting. This caused increased
ground reaction forces at the foot and decreased ground reaction forces at the knee leading to
increased flexion moments in Subject 2. When kneeling near full flexion, Subject 1 sat on their
heels with their feet rotated laterally. Subject 2 did not sit on their heel and kept their feet in a
vertical position with minimal rotation. This accounted for the decreased varus angles and frontal
and transverse moments of Subject 2. When squatting, Subject 1 had a wider stance which may
have accounted for the adduction moments created compared to the abduction moments in
Subject 2.
76
7.4 LIMITATIONS
The limitations of the developed model are associated with the complexity of kneeling near the
end range of flexion, as well as with the use of inverse dynamics. The assumptions necessary to
use the computational model may not be valid in all circumstances. The state of the joint
cartilage may null the assumption of a frictionless joint, as degenerative cartilage increases the
friction at the articulating surfaces. Also, the distribution of mass in any body segment is not
uniform nor is it concentrated at one location. The external applied forces are not applied
directly to the tibia, thereby causing forces which may not be linearly related to the resulting
moments due to tissue deformation. The patellar tendon and tibial tubercle may change the
center of pressure location and magnitudes of forces applied to the tibia during kneeling thereby
affecting the force and moment calculations.
The use of reflective markers to track the motion of palpable landmarks may have
introduced some sources of error. Motion of the skin to which reflective markers are attached,
also known as soft tissue artifact, strongly affects the estimation of joint angles characterized by
small range of motion. [47] Markers were placed on the medial and lateral epicondyles of the
femur while the subject was in a standing Tpose for 5 seconds. These markers were then
removed and recreated in data analysis. Motion of these markers was minimal with standard
deviations between .9 and 1.8 mm for Subject 1. Therefore the relative motion of these markers
with respect to the anatomical landmarks is not expected to introduce error in the knee joint
center estimation.
A Tekscan ClinSeat
®
system was used to measure the thighcalf and heelgluteus contact
forces. The use of this type of system, a system which uses resistive technology, has been found
to introduce creep in the measurement of static forces varying the force measurements by 10 %
77
to +20 %. [48] Also calibration of this system may reduce the accuracy and repeatability of its
measurements. [49] The recommended calibration is performed by applying a constant pressure
to the sensor by sitting on the sensor, thereby calibrating the sensor to bodyweight. It is possible
that the values reported from this system are inaccurate due to the method of calibration. Every
attempt was made to ensure repeatability of the calibrated values. One researcher calibrated the
system by sitting on the sensor atop a table with their left leg on the portion of the sensor that is
placed under the subject’s lower leg.
Along with the issues associated with measuring the contact forces, there are limitations
associated with the representation of these forces in the model. The thighcalf and heelgluteus
contact forces were represented as single resultant forces, however these contact forces are more
complex. Thighcalf contact creates a pressure distribution which has higher contact pressures
closer to the popliteal region. Heelgluteus contact creates a pressure distribution with pressures
highest in the middle of the contact area. Both contact pressures create tissue deformation which
may act to distribute the axial load in multiple directions. However, the system used to measure
these contact force was only capable of measuring axial pressure. Therefore shear forces
resulting from this contact were neglected from the model. Also due to the design of the
pressure sensor, the full contact surface could not be measured. The active sensing units on the
ClinSeat
®
system are located approximately 1.5 cm away from the top of the sensor. This
distance was accounted for when determining the center of pressures of these contact forces.
However the lack of pressure measurements in these areas decreased the measured total force.
This decrease in force may have been significant to increase the calculated flexion moments,
thereby creating an overestimation in sagittal moments. Future studies on thighcalf contact
should use measurement tools capable of quantifying the entire contact area and measuring
78
forces in multiple axes. Also, when modeling this contact tissue deformation should be
accounted for to reduce error in frontal and transverse moments.
Knee joint center location estimates have been shown to affect the interpretation of joint
moments in gait studies with small moment magnitudes. [50] Although the sagittal joint
moments in this model are of a higher magnitude, there are limitations associated with the KJC
location used in this model. Femoral “rollback” causes the KJC to move during knee flexion.
In the lateral compartment the femur moves 20 mm posterior from 0 to 120° of flexion and an
additional 10 mm when kneeling into a deep squat. In the medial compartment there is little to
no movement until 120° of flexion when there is a posterior displacement of 9 mm. [51] This
movement accounts for the increase in varus/valgus and internal/external rotations at higher knee
flexion. In this study, imaging techniques were not employed which would have been necessary
to determine the location of the knee joint center for all joint postures studied. Future studies on
knee biomechanics in which imaging techniques can be engaged should account for this change
in joint center location.
Thighcalf and heelgluteus contact force measurements were made for kneeling near full
flexion and squatting, prior to collection of motion capture and force data. Although this was
necessary to determine the moments for these postures, it neglects the possibility that similar
contact may be present in the other postures. Since subjects were not given specific instructions
on kneeling postures, it is possible that they may assume the kneeling near 90° flexion and
kneeling on one knee posture with higher flexion angles than anticipated. When kneeling on one
knee and kneeling near 90° flexion without kneepads, Subject 2 had included angles similar to
those during their squatting postures. These angles introduced thighcalf contact which was not
measured nor accounted for in the computational model. Neglecting this contact resulted in
79
overestimations of the flexion moments for these postures. In future studies it may be necessary
to instruct subjects to assume very specific postures such that all contact forces may be
accounted for and comparisons may be made between kneepads states.
Due to the small sample size, N=2, concrete conclusions cannot be drawn on the forces
and moments created in restricted postures. The small sample size creates a very small power
and statistics may not have been appropriate to show significant differences between postures or
subjects. Although data from two subjects were presented, the goal of this master’s thesis
project was the development of the model to determine the external forces and moments, and not
to characterize the forces and moments seen in kneeling and squatting. In the future this model
will be applied to data from a much larger dataset and resulting forces and moments will be
published.
7.5 IMPLICATIONS OF FINDINGS
Kneeling and squatting create tibial loading conditions which differ substantially from standing.
Increased flexion moments open the anterior aspects of the joint, increasing the loads to the
posterior tibia. Adduction moments open the medial joint space, stressing the medial soft tissues
and applying increased loads to the lateral compartment of the tibia. Abduction moments open
the lateral joint space, increasing stresses in lateral tissues and loads transmitted to the medial
compartment. Internal and external rotational moments will increase the loading to the medial
and lateral compartments of the knee, respectively. These complex loading conditions were
created in the kneeling postures in this study. The implications of these increased joint torques on
the muscles, ligaments, meniscus, and articular cartilage are presented.
80
7.5.1 Muscles Activity
To resist the externally applied forces and moments and reduce tibial translations, active
and passive internal knee stabilizers are used. Active knee stabilizers include the muscles and
tendons and passive stabilizers include the ligaments and soft tissues. In this study muscle
activity was not used as a model input and it is expected that the results would change
significantly if muscle activity were included in the force and moment calculations. When
considering the forces applied to the knee, it is assumed that the quadriceps muscle group is
responsible for resisting the externally applied flexion moments. However, although not
included in the results, EMG data of the quadriceps and hamstrings muscles were collected.
Results showed minimal activity, < 5% MVC, in these static postures. This implies that the knee
stability is mostly achieved by the passive stabilizers. Thus, the implications of this study are that
the high moments and forces calculated in the model may be transferring to the passive
stabilizers of the knee. The transference of the results of this model to the FEM model being
developed as part of the larger effort within the project will give an understanding of the
magnitude of these passive tissue loads.
7.5.2 Ligament Recruitment
The passive stability due to the ligaments and soft tissues will depend on knee orientation
and loading. Studies of passive knee motion reveal interesting findings of the recruitment
patterns of the passive knee stabilizers when kneeling near 90° flexion. Internal rotation recruits
the anterior bundles of the anterior cruciate ligament (ACL) along with the anterior and posterior
bundles of the posterior cruciate ligament (PCL) with primary stability from the posterior
81
bundles of the PCL. The application of a posterior force recruits, the anterior bundles of the PCL
and MCL. [52] Although adduction moments are generally stabilized primarily by the ACL and
the MCL, when kneeling near 90° flexion the MCL is the primary stabilizer with secondary
restraint from the PCL. [53, 54] Kneeling near 90° flexion and kneeling on one knee created
posterior forces along with adduction and internal rotational moments in both subjects. To
restrict tibial translations and rotations when in these postures, the PCL is expected to be the
primary stabilizer with assistance from the ACL and MCL. A maximum 500 N posterior force,
10 Nm adduction moment, and 11 Nm external rotational moment were applied to the tibia when
kneeling in these postures. Fukuda and colleagues found an insitu ACL force of 19.5 N when
subjected to a 10 Nm adduction moment at 90° flexion. [55] Carlin and colleagues found an in
situ PCL force of 95 N when subjected to a 100 N posterior force at 90° flexion. [56] If findings
from these insitu studies are correlated to the forces and moments from this study, it is not
expected that kneeling postures requiring approximately 90° of knee flexion (kneeling near 90°
flexion and kneeling on one knee) will cause detriment to the ligaments.
The stability of the knee in the fully flexed postures of kneeling near full flexion and
squatting is more complicated. In vivo studies of weight bearing knees in high flexion show that
up to 120° the PCL plays a major role in providing knee stability from posterior translation. [52,
57] However, in high flexion (above 120°) the PCL does not contribute substantially to knee
stability. It is believed that the posterior soft tissues of the knee (posterior horns of the meniscus,
posterior capsule, hamstrings muscles, fat, and skin) not the PCL provide stability in high
flexion. [58]
82
7.5.3 Meniscal Loading
During walking 70% of the total knee load is distributed to the medial compartment of
the knee joint. [59] This increases the load on the medial meniscsus. When an abduction moment
is added to the tibia, the load increases. In the kneeling near 90° flexion and kneeling on one
knee postures, the contact area between the tibia and the meniscus is thought to be sufficient to
withstand the forces created under these conditions without damage. However, with increasing
joint flexion, as in the kneeling near full flexion and squatting postures, there is posterior
translation of the tibia. This translation is also accompanied by the posterior translation of the
meniscus with the lateral horn translating more than the medial horn. This posterior translation
functions to increase the contact area between the tibia and the meniscus and may play a crucial
role in distributing compressive loads in full flexion. [60] Compared to extension, in high
flexion, the contact between the tibia and the femur occur at the posterior aspects of the knee,
decreasing the contact area resulting in increased stresses in the posterior meniscus.
Along with knee flexion playing a major role in the loads and contact areas of the
meniscus, adduction and abduction moments contribute to the loading of the meniscus.
Adduction and abduction moments increase loading to the lateral and medial compartments of
the knee. [61] In Subject 1, adduction moments were created when kneeling near full flexion and
squatting. The adduction moments in postures with high knee flexion may cause detrimental
effects on the lateral meniscus. The adduction moments which were as much as 16 Nm in
conjunction with the posterior forces of up to 120 N, increase the shear loads transmitted to the
meniscus. In Subject 2, abduction moments increased the loading to the medial meniscus. The
moments which were up to 21 Nm in combination with 160 N posterior forces, increased the
shear load to the meniscus. These loads are then transferred to the articular cartilage with
83
maximal shear stress occurring at the cartilagebone interface away from the center of contact.
[62] This creates areas of high stress on the articular cartilage and when sustained for prolonged
periods could lead to the deterioration of the meniscus and articular cartilage damage.
7.5.4 Osteoarthritis Progression
Articular cartilage damage may occur as a result of biological and mechanical factors.
Two forms of articular cartilage damage occur from loading. Type 1 articular cartilage damage
is characterized by damage without disruption of the underlying bone or calcified cartilage layer.
Type 2 damage is characterized by bone fracture with or without damage to the overlying
cartilage. [63] Excessive shear stress, tensile stress, and principal strain have been suggested as
mechanisms of articular cartilage damage. [63  65] Thambyah and colleagues found articular
cartilage stresses of 14 MPa during gait and these stresses increased by more than 80% when in
deep flexion. This increased stress reached the damage limit of the cartilage and was thought to
increase the risk of articular cartilage damage. [66] Thus, the implications are that the high
moments and forces associated with kneeling near full flexion and squatting may cause excessive
stress on the articular cartilage of the tibiofemoral joint thereby increasing the risk for knee
osteoarthritis.
7.5.5 Postures Associated with Osteoarthritis
Epidemiologic studies have determined postures associated with increased prevalence of
knee osteoarthritis including squatting, kneeling, stair climbing, sideknee bending, and sitting
crossedlegged. [17, 32] Stair climbing produces forces and moments greater than that during
84
walking. [67, 68] Stair descent creates flexion, varus/valgus, and external rotational moments
greater than level walking or stair ascent, making it a more stressful activity. [68] In this study all
postures created frontal moments and posterior forces greater than standing. Although kneeling
with joint flexion close to 90° is not expected to cause damage to the articular cartilage, it is
expected that postures with higher joint flexion will. The combination of increased frontal
moments, increased posterior force, and high knee flexion, create high loads on the medial and
lateral posterior compartments of the knee and may create the biomechanical stresses necessary
to initiate the developments and progression of medial and lateral compartment knee
osteoarthritis.
7.6 ADVANCEMENTS
The developed computational model accounts for the contact between the heel and the gluteal
muscles which until now, had been neglected from other models. Thighcalf contact has been
shown to significantly affect the forces in the quadriceps and it is expected that the heelgluteus
contact will have similar effects on muscle force estimations. In this study the heelgluteus
contact force was 54 N which was almost half of the 124 N thighcalf contact force. If a force of
124 N is significant to effect muscle force estimations than an additional force of 54 N should
also be significant to affect these forces.
Results of the sensitivity analysis revealed that the location of the thighcalf contact has a
large effect on the knee moments when kneeling near full flexion and squatting. Although when
using this model this may be a limitation, when considering the design of interventions this may
be an important factor. Early interventions to reduce the stresses and strains at the knee joint
85
have focused on the outside of the knee, i.e. knee pads. However, slightly changing postures or
the distribution of body weight may be useful. Newer devices should consider maximizing the
thighcalf and heelgluteus contact forces in hopes of distributing more body weight along the
lower leg, thereby creating an extensor moment to stabilize the knee joint in full flexion. It was
also found that the kneepad acted to change the COP of the force at the knee by as much as 2.9
cm in the medial/lateral direction which created an 8 Nm increase in transverse moments. This
increase in rotational moment may be detrimental to the internal knee structures. It therefore
becomes important to consider the effect of new devices on the transverse and frontal moments
of the knee which are often neglected due to their smaller magnitudes when compared to sagittal
moments.
86
8.0 FUTURE WORK
In the near future the forces and moments determined using this model will be used as
inputs into a finite element model of the knee. In addition to motion capture, force data, and
thighcalf and heelgluteus contact measurements, electromyography and knee surface pressure
data were collected as part of this project. Electromyography was used to determine which
muscles were active during different stages of kneeling, crawling, and stooping. Subjects were
also instrumented with a custom designed pressure sensor on their right knee under their
kneepad. This pressure sensor was used to measure the pressure distributions at the patella,
patellar tendon, and tibial tubercle with and without kneepads. Also a second commonly used
kneepad with a different design (nonarticulating, soft outer shell) was used. Gaining a better
understanding of the pressures, forces, and moments at the knee joint and in the internal
stabilizing structures, will allow development of better solutions to the complex knee problems
facing many lowseam coal miners. The endproduct of this project will be the design of a new
intervention which will be effective in reducing the stress and strains in the internal knee
structures caused by frequent and prolonged kneeling, stooping, and crawling.
87
9.0 CONCLUSION
In this work a 3D computational model based on inverse dynamics and segment
coordinate systems was developed in MATLAB to determine the joint angles and net force and
moments applied to the knee in static kneeling postures. This model was custom fitted to each
subject and based off their individual anthropometry, ground reaction forces, thighcalf contact
force, and heelgluteus contact force measurements. Sensitivity analysis revealed that varying
the location of the knee joint center affected the sagittal, frontal, and transverse moments in all
postures. In some cases the interpretation of the joint moment changed due to varying the KJC
by 3 cm. This model was developed for use on a much larger dataset currently being collected.
However, data from two subjects were presented in this paper. These subjects revealed that
kneeling creates high external knee joint moments which may be as much as 10 to 40 times that
during quiet standing. There were also large variations in the moments between postures and
kneepads states. Postures requiring larger flexion angles generally created larger flexion
moments at the knee. The more upright postures of kneeling near 90° flexion and kneeling on
one knee created increased posterior forces, up to 500N. Also ankle positioning appeared to play
a large role in the increased frontal moments associated with kneeling. One subject internally
rotated their ankles when kneeling which caused large knee adduction moments, up to 16 Nm.
This increased adduction moment in combination with increased posterior forces, increases the
forces applied to the medial compartment of the knee and may be significant to create cartilage
88
damage. The other subject had large abduction moments, 21 Nm, when squatting which
increased loading to the lateral aspects of the meniscus. It was also found that the kneepad used
in this study changed the center of pressure of the force at the knee, in some cases increasing the
transverse moment by as much as 8 Nm. A larger dataset will be necessary to determine the
effect of these kneeling postures on the internal structures of the knee. However, this study
suggests that kneeling increases the flexion moments applied to the knee as well as the adduction
moments and posterior forces and may be significant to initiate meniscal and articular cartilage
damage.
89
APPENDIX
MATLAB SOFTWARE
The MATLAB
®
software developed to determine the net forces, net moments, and joint angles at
the knee consists of eight MATLAB
®
files. Three files are used to calculate the forces, moments,
and angles. The other files are necessary for importing data and exporting results. The three
files: transform_m2a_shank_thigh.m, knee_angle.m, and forces_moments.m, perform all
necessary calculations and are included in this appendix.
A.1 TRANSFORM_M2A_SHANK_THIGH.M
Transform_m2a_shank_thigh.m is used to calculate the transformation matrix from the global
coordinate system to the anatomical coordinate system, the transformation matrix from the
global coordinate system to the measured coordinate system, and the transformation matrix from
the measured coordinate system to the anatomical coordinate system.
90
function [filename]=transform_m2a(results_directory)
%%%%%%%%%%%%%%%%%%%%%%%%%%% transform_m2a.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%
%%%% This function is designed to produce the transformation matrix from
%%%% testing coordinates to anatomical coordinates. This function uses
%%%% only the anatomical (calibration) data
%%%% 1. Assign data to specific markers
%%%% Part 1  Shank
%%%% 2. Determine the anatomical coordinate system for shank and
%%%% calculate the transformation matrix from global to anatomical
%%%% coordinates
%%%% 3. Determine the measured/testing coordinate system and calculate
%%%% the transformation matrix from global to measured
%%%% coordinates
%%%% 4. Determine the transformation matrix from measured to anatomical
%%%% and get the mean value of this matrix
%%%% Part 2  Thigh
%%%% 24 for thigh
%%%% 5. Save transformation matrix TMAshank to mat file
%%%%
%%%% Jonisha Pollard, JPollard@cdc.gov, 8/19/08
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% Load Anatomical Data
cd(results_directory);
load anatomical.mat
%% Determine length of data
[n o]=size(anatomical_data);
%% 1. Assign data to specific markers
right_knee=anatomical_data(:,45:47);
right_knee_medial=anatomical_data(:,78:80);
right_ankle=anatomical_data(:,51:53);
right_ankle_medial=anatomical_data(:,81:83);
right_shank=anatomical_data(:,48:50);
right_shank_front=anatomical_data(:,111:113);
right_shank_rear=anatomical_data(:,114:116);
right_thigh=anatomical_data(:,42:44);
right_thigh_front=anatomical_data(:,93:95);
right_thigh_rear=anatomical_data(:,96:98);
left_knee=anatomical_data(:,63:65);
left_knee_medial=anatomical_data(:,84:86);
left_ankle=anatomical_data(:,69:71);
left_ankle_medial=anatomical_data(:,87:89);
left_shank=anatomical_data(:,66:68);
left_shank_front=anatomical_data(:,120:122);
left_shank_rear=anatomical_data(:,123:125);
left_thigh=anatomical_data(:,60:62);
left_thigh_front=anatomical_data(:,102:104);
left_thigh_rear=anatomical_data(:,105:107);
rasis=anatomical_data(:,33:35);
lasis=anatomical_data(:,36:38);
91
%% Determine location of joint centers in global coordinates
AJC=(right_ankle + right_ankle_medial)/2; % Global Right AJC
KJC=(right_knee + right_knee_medial)/2; % Global Right KJC
AJC_L=(left_ankle + left_ankle_medial)/2; % Global Left AJC
KJC_L=(left_knee + left_knee_medial)/2; % Global Left KJC
% Global HJC Calculation. Bells method
% Leardini et al. 1999, Bell et al. 1990
HJC_origin=(lasis+rasis)/2;
PW=abs(lasis(:,1)rasis(:,1)); % inter Asis distance
HJC(:,1)=HJC_origin(:,1)+.36*PW;
HJC(:,2)=HJC_origin(:,2).19*PW;
HJC(:,3)=HJC_origin(:,3).3*PW;
HJC_L(:,1)=HJC_origin(:,1).36*PW;
HJC_L(:,2)=HJC_origin(:,2).19*PW;
HJC_L(:,3)=HJC_origin(:,3).3*PW;
%%%%%%%%%%%%%%%%%%%%%% Part 1  Shank %%%%%%%%%%%%%%%%%%%%%%%%%%
%% 2. Determine anatomical coordinate system for shank and calculate
%% transformation matrix from global to anatomical coordinates
% All coordinate systems were chosen such that hey align with the
% Global coordinate system at Standard anatomical position
% Global & Anatomical Coordinate Systems
% x is medial/lateral
% y is anterior/posterior
% z is superior/inferior
% Right
for i=1:n
r1=KJCAJC; % local zaxis
r2=right_knee  right_knee_medial; % in xdirection
r3=cross(r1,r2); % local yaxis
r4=cross(r3,r1); % local xaxis
x(i,:)=r4(i,:)/norm(r4(i,:));
y(i,:)=r3(i,:)/norm(r3(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGAshank=zeros(4,4,n);
for i=1:n
TGAshank(1,:,i)=[1 0 0 0];
TGAshank(2:4,1,i)=KJC(i,:); %right KJC is origin
TGAshank(2:4,2,i)=x(i,:);
TGAshank(2:4,3,i)=y(i,:);
TGAshank(2:4,4,i)=z(i,:);
end
92
clear r1 r2 r3 r4 x y z % clear variables
% calculate AJC and KJC in anatomical shank coordinates
for i=1:n
G_ajc(i,2:4)=AJC(i,:);
G_ajc(:,1)=1;
G_kjc(i,2:4)=KJC(i,:);
G_kjc(:,1)=1;
asAJC_all(i,:)=inv(TGAshank(:,:,i))*G_ajc(i,:)'; %Shank AJC
asKJC_all(i,:)=inv(TGAshank(:,:,i))*G_kjc(i,:)'; %Shank KJC
end
asAJC=mean(asAJC_all(:,2:4));
asKJC=mean(asKJC_all(:,2:4)); % verified to equal [0 0 0]
% Left
for i=1:n
r1=KJC_LAJC_L; % local zaxis
r2=left_knee_medial  left_knee; % in xdirection
r3=cross(r1,r2); % local yaxis
r4=cross(r3,r1); % local xaxis
x(i,:)=r4(i,:)/norm(r4(i,:));
y(i,:)=r3(i,:)/norm(r3(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGAshank_L=zeros(4,4,n);
for i=1:n
TGAshank_L(1,:,i)=[1 0 0 0];
TGAshank_L(2:4,1,i)=KJC_L(i,:); % left KJC is origin
TGAshank_L(2:4,2,i)=x(i,:);
TGAshank_L(2:4,3,i)=y(i,:);
TGAshank_L(2:4,4,i)=z(i,:);
end
clear r1 r2 r3 r4 x y z % clear variables
% calculate AJC and KJC in anatomical shank coordinates
for i=1:n
G_ajc_L(i,2:4)=AJC_L(i,:);
G_ajc_L(:,1)=1;
G_kjc_L(i,2:4)=KJC_L(i,:);
G_kjc_L(:,1)=1;
asAJC_all_L(i,:)=inv(TGAshank_L(:,:,i))*G_ajc_L(i,:)'; %Shank AJC
asKJC_all_L(i,:)=inv(TGAshank_L(:,:,i))*G_kjc_L(i,:)'; %Shank KJC
end
asAJC_L=mean(asAJC_all_L(:,2:4));
asKJC_L=mean(asKJC_all_L(:,2:4));
%% 3. Determine the measured coordinate system for shank and calculate the
%% transformation matrix from global to measured
%% Global & Anatomical Coordinate Systems
93
%% x is medial/lateral
%% y is anterior/posterior
%% z is superior/inferior
% Right shank
for i=1:n
r1=right_shankright_shank_front; % local zaxis
r2=right_shank_front  right_shank_rear; % in y direction
r3=cross(r2,r1); % local xaxis
r4=cross(r1,r3); % local yaxis
x(i,:)=r3(i,:)/norm(r3(i,:));
y(i,:)=r4(i,:)/norm(r4(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGMshank=zeros(4,4,n);
for i=1:n
TGMshank(1,:,i)=[1 0 0 0];
TGMshank(2:4,1,i)=right_shank_front(i,:); % origin
TGMshank(2:4,2,i)=x(i,:);
TGMshank(2:4,3,i)=y(i,:);
TGMshank(2:4,4,i)=z(i,:);
end
clear r1 r2 r3 r4 x y z % clear variables
% Left shank
for i=1:n
r1=left_shankleft_shank_front; % local zaxis
r2=left_shank_front  left_shank_rear; % in y direction
r3=cross(r2,r1); % local xaxis
r4=cross(r1,r3); % local yaxis
x(i,:)=r3(i,:)/norm(r3(i,:));
y(i,:)=r4(i,:)/norm(r4(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGMshank_L=zeros(4,4,n);
for i=1:n
TGMshank_L(1,:,i)=[1 0 0 0];
TGMshank_L(2:4,1,i)=left_shank_front(i,:); % origin
TGMshank_L(2:4,2,i)=x(i,:);
TGMshank_L(2:4,3,i)=y(i,:);
TGMshank_L(2:4,4,i)=z(i,:);
end
clear r1 r2 r3 r4 x y z % clear variables
%% 4. Calculate transformation matrix from measured to anatomical coords
% Right
TMAs_alldata=zeros(4,4,n); % TMAshank for all data points
for i=1:n
TMAs_alldata(:,:,i)=inv(TGMshank(:,:,i))*TGAshank(:,:,i);
94
end
% calculate the mean of TMAshank_alldata
TMAshank=zeros(4,4);
for i=2:4
for j=1:4
TMAshank(i,j)=mean(TMAs_alldata(i,j,:));
TMAshank(1,1)=1;
TMAshank(1,2:4)=0;
end
end
% Left
TMAs_alldata_L=zeros(4,4,n); % TMAshank_L for all data points
for i=1:n
TMAs_alldata_L(:,:,i)=inv(TGMshank_L(:,:,i))*TGAshank_L(:,:,i);
end
% calculate the mean of TMAshank_L_alldata
TMAshank_L=zeros(4,4);
for i=2:4
for j=1:4
TMAshank_L(i,j)=mean(TMAs_alldata_L(i,j,:));
TMAshank_L(1,1)=1;
TMAshank_L(1,2:4)=0;
end
end
%%%%%%%%%%%%%%%%%%%%%% Part 2  Thigh %%%%%%%%%%%%%%%%%%%%%%%%%%
%% 2. Determine anatomical coordinate system for thigh and calculate
%% transformation matrix from global to anatomical coords
%% Global & Anatomical Coordinate Systems
%% x is medial/lateral
%% y is anterior/posterior
%% z is superior/inferior
% Right Thigh
for i=1:n
r1=HJCKJC; % local zaxis
r2=right_knee  right_knee_medial; % in xdirection
r3=cross(r1,r2); % local yaxis
r4=cross(r3,r1); % local xaxis
x(i,:)=r4(i,:)/norm(r4(i,:));
y(i,:)=r3(i,:)/norm(r3(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGAthigh=zeros(4,4,n);
for i=1:n
TGAthigh(1,:,i)=[1 0 0 0];
TGAthigh(2:4,1,i)=HJC(i,:); %right HJC is origin
TGAthigh(2:4,2,i)=x(i,:);
TGAthigh(2:4,3,i)=y(i,:);
TGAthigh(2:4,4,i)=z(i,:);
95
end
clear r1 r2 r3 r4 x y z % clear variables
% calculate HJC and KJC in anatomical thigh coordinates
for i=1:n
G_hjc(i,2:4)=HJC(i,:);
G_hjc(:,1)=1;
G_kjc(i,2:4)=KJC(i,:);
G_kjc(:,1)=1;
atHJC_all(i,:)=inv(TGAthigh(:,:,i))*G_hjc(i,:)'; %Thigh HJC
atKJC_all(i,:)=inv(TGAthigh(:,:,i))*G_kjc(i,:)'; %Thigh KJC
end
atHJC=mean(atHJC_all(:,2:4)); % verified to equal [0 0 0]
atKJC=mean(atKJC_all(:,2:4));
% Left Thigh
for i=1:n
r1=HJC_LKJC_L; % local zaxis
r2=left_knee_medial  left_knee; % in xdirection
r3=cross(r1,r2); % local yaxis
r4=cross(r3,r1); % local xaxis
x(i,:)=r4(i,:)/norm(r4(i,:));
y(i,:)=r3(i,:)/norm(r3(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGAthigh_L=zeros(4,4,n);
for i=1:n
TGAthigh_L(1,:,i)=[1 0 0 0];
TGAthigh_L(2:4,1,i)=HJC_L(i,:); %left HJC is origin
TGAthigh_L(2:4,2,i)=x(i,:);
TGAthigh_L(2:4,3,i)=y(i,:);
TGAthigh_L(2:4,4,i)=z(i,:);
end
clear r1 r2 r3 r4 x y z % clear variables
% calculate HJC and KJC in anatomical thigh coordinates
for i=1:n
G_hjc_L(i,2:4)=HJC_L(i,:);
G_hjc_L(:,1)=1;
G_kjc_L(i,2:4)=KJC_L(i,:);
G_kjc_L(:,1)=1;
atHJC_all_L(i,:)=inv(TGAthigh_L(:,:,i))*G_hjc_L(i,:)'; %Thigh HJC
atKJC_all_L(i,:)=inv(TGAthigh_L(:,:,i))*G_kjc_L(i,:)'; %Thigh KJC
end
atHJC_L=mean(atHJC_all_L(:,2:4)); % verified to equal [0 0 0]
atKJC_L=mean(atKJC_all_L(:,2:4));
96
%% 3. Determine the measured coordinate system for thigh and calculate
%% transformation from global to measured
%% Global & Anatomical Coordinate Systems
%% x is medial/lateral
%% y is anterior/posterior
%% z is superior/inferior
% Right
for i=1:n
r1=right_thighright_thigh_front; % local zaxis
r2=right_thigh_frontright_thigh_rear; % in y direction
r3=cross(r2,r1); % local xaxis
r4=cross(r1,r3); % local yaxis
x(i,:)=r3(i,:)/norm(r3(i,:));
y(i,:)=r4(i,:)/norm(r4(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGMthigh=zeros(4,4,n);
for i=1:n
TGMthigh(1,:,i)=[1 0 0 0];
TGMthigh(2:4,1,i)=right_thigh_front(i,:); %origin
TGMthigh(2:4,2,i)=x(i,:);
TGMthigh(2:4,3,i)=y(i,:);
TGMthigh(2:4,4,i)=z(i,:);
end
clear r1 r2 r3 r4 x y z % clear variables
% Left
for i=1:n
r1=left_thighleft_thigh_front; % local zaxis
r2=left_thigh_frontleft_thigh_rear; % in y direction
r3=cross(r2,r1); % local xaxis
r4=cross(r1,r3); % local yaxis
x(i,:)=r3(i,:)/norm(r3(i,:));
y(i,:)=r4(i,:)/norm(r4(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGMthigh_L=zeros(4,4,n);
for i=1:n
TGMthigh_L(1,:,i)=[1 0 0 0];
TGMthigh_L(2:4,1,i)=left_thigh_front(i,:); %origin
TGMthigh_L(2:4,2,i)=x(i,:);
TGMthigh_L(2:4,3,i)=y(i,:);
TGMthigh_L(2:4,4,i)=z(i,:);
end
clear r1 r2 r3 r4 x y z % clear variables
%% 4. Calculate transformation matrix from measured to anatomical coords
97
% Right
TMAt_alldata=zeros(4,4,n); % TMAthigh for all data point
for i=1:n
TMAt_alldata(:,:,i)=inv(TGMthigh(:,:,i))*TGAthigh(:,:,i);
end
% calculate the mean of TMAthigh_alldata
TMAthigh=zeros(4,4);
for i=2:4
for j=1:4
TMAthigh(i,j)=mean(TMAt_alldata(i,j,:));
TMAthigh(1,1)=1;
TMAthigh(1,2:4)=0;
end
end
% Left
TMAt_alldata_L=zeros(4,4,n); % TMAthigh for all data point
for i=1:n
TMAt_alldata_L(:,:,i)=inv(TGMthigh_L(:,:,i))*TGAthigh_L(:,:,i);
end
% calculate the mean of TMAthigh_alldata
TMAthigh_L=zeros(4,4);
for i=2:4
for j=1:4
TMAthigh_L(i,j)=mean(TMAt_alldata_L(i,j,:));
TMAthigh_L(1,1)=1;
TMAthigh_L(1,2:4)=0;
end
end
%% 5. Save transformation matrixes KJC, AJC and HJC mat file
save anatomical.mat TMAshank TMAthigh asAJC TMAshank_L TMAthigh_L ...
asAJC_L append
end % end of function transform_m2a.m
98
A.2 KNEE ANGLE.M
Knee_angle.m is used to calculate the orientation of the knee joint. Using Euler angle
decomposition, with a Euler order of X’y’z”, the included joint angle, the angle of varus/valgus,
and the angle of internal/external rotation are calculated.
function knee_angle(posture, results_directory)
%%%%%%%%%%%%%%%%%%%%%%%%%%% knee_angle.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%% This function is designed to determine the knee angle by calculating
%%%% the angle between the thigh coordinate system and the shank coordinate
%%%% system.
%%%% 1. Load testing marker data and assign data to specific markers
%%%% 2. Determine measured coordinate system for shank and thigh
%%%% 3. Calculate anatomical coordinate system from measured system and
%%%% determine transformation from thigh to shank Tts
%%%% 4. Decompose transformation matrix for Euler angles
%%%% 5. Save joint angles to *posture*.mat
%%%%
%%%% Jonisha Pollard, JPollard@cdc.gov, 7/23/08
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
cd(results_directory)
load(posture) %load mat file with data for specific postures
load anatomical %load mat file with anatomical markers and TMAshank
%% 1. Load testing marker data and assign data to specific markers
[n a]=size(testing_data);
right_shank=testing_data(:,39:41);
right_shank_rear=testing_data(:,48:50);
right_shank_front=testing_data(:,45:47);
right_thigh=testing_data(:,27:29);
right_thigh_front=testing_data(:,33:35);
right_thigh_rear=testing_data(:,36:38);
left_shank=testing_data(:,81:83);
left_shank_rear=testing_data(:,90:92);
left_shank_front=testing_data(:,87:89);
left_thigh=testing_data(:,69:71);
left_thigh_front=testing_data(:,75:77);
left_thigh_rear=testing_data(:,78:80);
99
%% 2. Determine measured coordinate system for shank and thigh
%% Global Coordinate System
%% x is medial/lateral
%% y is anterior/posterior
%% z is superior/inferior
% Right shank
for i=1:n
r1=right_shankright_shank_front; % local zaxis
r2=right_shank_front  right_shank_rear; % in y direction
r3=cross(r2,r1); % local xaxis
r4=cross(r1,r3); % local yaxis
x(i,:)=r3(i,:)/norm(r3(i,:));
y(i,:)=r4(i,:)/norm(r4(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGMshank=zeros(4,4,n);
for i=1:n
TGMshank(1,:,i)=[1 0 0 0];
TGMshank(2:4,1,i)=right_shank_front(i,:); % origin
TGMshank(2:4,2,i)=x(i,:);
TGMshank(2:4,3,i)=y(i,:);
TGMshank(2:4,4,i)=z(i,:);
end
clear r1 r2 r3 r4 x y z
% Right thigh
for i=1:n
r1=right_thighright_thigh_front; % local zaxis
r2=right_thigh_frontright_thigh_rear; % in y direction
r3=cross(r2,r1); % local xaxis
r4=cross(r1,r3); % local yaxis
x(i,:)=r3(i,:)/norm(r3(i,:));
y(i,:)=r4(i,:)/norm(r4(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGMthigh=zeros(4,4,n);
for i=1:n
TGMthigh(1,:,i)=[1 0 0 0];
TGMthigh(2:4,1,i)=right_thigh_front(i,:); % origin
TGMthigh(2:4,2,i)=x(i,:);
TGMthigh(2:4,3,i)=y(i,:);
TGMthigh(2:4,4,i)=z(i,:);
end
clear r1 r2 r3 r4 x y z
% Left shank
for i=1:n
r1=left_shankleft_shank_front; % local zaxis
r2=left_shank_front  left_shank_rear; % in y direction
100
r3=cross(r2,r1); % local xaxis
r4=cross(r1,r3); % local yaxis
x(i,:)=r3(i,:)/norm(r3(i,:));
y(i,:)=r4(i,:)/norm(r4(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGMshank_L=zeros(4,4,n);
for i=1:n
TGMshank_L(1,:,i)=[1 0 0 0];
TGMshank_L(2:4,1,i)=left_shank_front(i,:); % origin
TGMshank_L(2:4,2,i)=x(i,:);
TGMshank_L(2:4,3,i)=y(i,:);
TGMshank_L(2:4,4,i)=z(i,:);
end
clear r1 r2 r3 r4 x y z
% Left thigh
for i=1:n
r1=left_thighleft_thigh_front; % local zaxis
r2=left_thigh_frontleft_thigh_rear; % in y direction
r3=cross(r2,r1); % local xaxis
r4=cross(r1,r3); % local yaxis
x(i,:)=r3(i,:)/norm(r3(i,:));
y(i,:)=r4(i,:)/norm(r4(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGMthigh_L=zeros(4,4,n);
for i=1:n
TGMthigh_L(1,:,i)=[1 0 0 0];
TGMthigh_L(2:4,1,i)=left_thigh_front(i,:); % origin
TGMthigh_L(2:4,2,i)=x(i,:);
TGMthigh_L(2:4,3,i)=y(i,:);
TGMthigh_L(2:4,4,i)=z(i,:);
end
clear r1 r2 r3 r4 x y z
%% 3. Calculate anatomical coordinate system from measured system and
%% determine transformation from thigh to shank Tts
for i=1:n
% Right
TGAshank(:,:,i)=TGMshank(:,:,i)*TMAshank;
TGAthigh(:,:,i)=TGMthigh(:,:,i)*TMAthigh;
Tts(:,:,i)=inv(TGAthigh(:,:,i))*TGAshank(:,:,i);
Rts(:,:,i)=Tts(2:4,2:4,i); % rotation matrix
% Left
TGAshank_L(:,:,i)=TGMshank_L(:,:,i)*TMAshank_L;
TGAthigh_L(:,:,i)=TGMthigh_L(:,:,i)*TMAthigh_L;
101
Tts_L(:,:,i)=inv(TGAthigh_L(:,:,i))*TGAshank_L(:,:,i);
Rts_L(:,:,i)=Tts_L(2:4,2:4,i); % rotation matrix
end
%% 4. Decompose transformation matrix for Euler angle
% Global Coordinate System
% 1st axis  flexion/ext  X
% 2nd axis  varus/valgus  Y
% 3rd axis  inversion/eversion  Z
% Euler order Xy'z''
for i=1:n
z_angle(i,:)=((180/pi) * atan2(Rts(1,2,i),Rts(1,1,i)));
x_angle(i,:)=180 + ((180/pi) * atan2(Rts(2,3,i),Rts(3,3,i)));
y_angle(i,:)=((180/pi) * atan2(cos(z_angle(i,:))*Rts(1,3,i),...
Rts(1,1,i)));
z_angle_L(i,:)=((180/pi) * atan2(Rts_L(1,2,i),Rts_L(1,1,i)));
x_angle_L(i,:)=180 + ((180/pi) * atan2(Rts_L(2,3,i), ...
Rts_L(3,3,i)));
y_angle_L(i,:)=((180/pi) * atan2(cos(z_angle_L(i,:)) ...
*Rts_L(1,3,i),Rts_L(1,1,i)));
end
%% 5. Save angles
save(posture,'x_angle','y_angle','z_angle','Tts','TGAthigh',...
'TGAshank','x_angle_L','y_angle_L','z_angle_L','Tts_L', ...
'TGAthigh_L','TGAshank_L','append')
end % end of function knee_angle.m
102
A.3 FORCES_MOMENTS.M
Forces_moments.m is used to calculate the net external forces, moments, and moment
contributions at the knee joint in the global coordinate system. The forces and moments are then
transformed into the anatomical coordinate system.
function forces_moments(weight, height, posture, results_directory)
%%%%%%%%%%%%%%%%%%%%%%%%%%% forces_moments.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%% This function is designed to calculate the net forces and moments at
%%%% the knee joint for squatting and full flexion
%%%% 1. Load testing marker data and assign data to specific markers
%%%% 2. Load force plate data
%%%% 3. Determine measured coordinate system for shank
%%%% 4. Get anatomical shank coordinates from testing coordinates
%%%% 5. Determine COP and weight of segments and joint rotation centers
%%%% 6. Calculate the joint reactive forces at the right knee in Global
%%%% coordinates
%%%% 7. Calculate the moment at the right knee in Global coordinates
%%%% 8. Calculate forces and moments in Anatomical coordinates
%%%% 9. Normalize forces and moments
%%%% 10. Saves forces and moments to *posture*.mat
%%%%
%%%% Jonisha Pollard, JPollard@cdc.gov, 7/23/08
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
cd(results_directory)
load(posture) %load mat file with data for specific postures
load anatomical %load mat file with anatomical markers and TMAshank
%% 1. Load testing marker data and assign data to specific markers
[n a]=size(testing_data);
right_shank=testing_data(:,39:41);
right_shank_rear=testing_data(:,48:50);
right_shank_front=testing_data(:,45:47);
%% 2. Load force plate data
F1=forces_data(:,2:4);
COP1=forces_data(:,5:7);
F2=forces_data(:,9:11);
COP2=forces_data(:,12:14);
103
%% 3. Determine measured coordinate system for shank
%% Global Coordinate System
%% x is medial/lateral
%% y is anterior/posterior
%% z is superior/inferior
for i=1:n
r1=right_shankright_shank_front; % local zaxis
r2=right_shank_front  right_shank_rear; % in y direction
r3=cross(r2,r1); % local xaxis
r4=cross(r1,r3); % local yaxis
x(i,:)=r3(i,:)/norm(r3(i,:));
y(i,:)=r4(i,:)/norm(r4(i,:)); % rotation matrix
z(i,:)=r1(i,:)/norm(r1(i,:));
end
TGMshank=zeros(4,4,n);
for i=1:n
TGMshank(1,:,i)=[1 0 0 0];
TGMshank(2:4,1,i)=right_shank_front(i,:); % origin
TGMshank(2:4,2,i)=x(i,:);
TGMshank(2:4,3,i)=y(i,:);
TGMshank(2:4,4,i)=z(i,:);
end
%% 4. Get anatomical shank coordinates from measured coordinates
for i=1:n
TGAshank(:,:,i)=TGMshank(:,:,i)*TMAshank; % transformation matrix
RGAshank(:,:,i)=TGAshank(2:4,2:4,i); % rotation matrix
end
%% 5. Determine COM, COP, joint rotation centers, and contact forces in
%% global coordinates
for i=1:n
% calculate location of AJC and KJC in global coordinates
g_AJC(i,:)=TGAshank(:,:,i)*[1 asAJC]';
g_KJC(i,:)=TGAshank(:,:,i)*[1 0 0 0]';% KJC is origin of ACS [0 0 0]
% calculate thighcalf and heelgluteus COP in global coordinates
% these contact forces are located along the z axis of the shank coord
COP_tc(i,:)=TGAshank(:,:,i)*[1 0 0 tc_dist]';
COP_hg(i,:)=TGAshank(:,:,i)*[1 0 0 hg_dist]';
% Thighcalf and Heelgluteus contact forces in global coordinate system
gtc_cont(i,:)=RGAshank(:,:,i)*[0 tc_cont 0]';
ghg_cont(i,:)=RGAshank(:,:,i)*[0 hg_cont 0]';
end
104
gAJC=g_AJC(:,2:4); %eliminate '1' in 1st col
gKJC=g_KJC(:,2:4);
COPtc=COP_tc(:,2:4);
COPhg=COP_hg(:,2:4);
% Lower leg COM and weight are modified from Clauser et al 1969 and
% Hinrichs 1990
% 47.5% of distance from proximal to distal
COM_lowleg=gAJC+.525*(gKJC  gAJC);
W_lowleg=zeros(n,3);
W_lowleg(:,3)=.058*weight;
%% 6. Calculate the joint reactive forces at the right knee in Global
%% Coordinates
% Sum of all forces at knee should equal zero
% Sum forces = Force_r_knee + F1 + F2 + W_lowleg + tc + hg = 0
% Knee reaction force = (F1 + F2 + W_lowleg + tc + hg)
% Force imposed on knee =F1 + F2 + W_lowleg + tc + hg
% Force imposed on Knee (reaction force will be )
Fknee= F1 + F2 + W_lowleg + gtc_cont + ghg_cont;
%% 7. Calculate moment at the right knee joint in Global Coordinates
% Sum of all moments at knee should equal zero
% Moments will be calculated using the vector cross product
% Sum of Moments at knee= Knee_reaction_Moment+ MF1 + MF2 + Mlowleg + Mtc
% + Mhg= 0
% Knee_reaction_Moment = (MF1 + MF2 + Mlowleg + Mtc + Mhg)
% Moment imposed on knee= MF1 + MF2 + Mlowleg + Mtc + Mhg
% moment arm vectors in m
r_lowleg=(COM_lowleggKJC)/1000;
r_FP1=(COP1gKJC)/1000;
r_FP2=(COP2gKJC)/1000;
r_tc=(COPtcgKJC)/1000;
r_hg=(COPhggKJC)/1000;
% calculate moments in Nm
for i=1:n
M_lowleg(i,:)=cross(r_lowleg(i,:),W_lowleg(i,:));
MF1(i,:)=cross(r_FP1(i,:),F1(i,:));
MF2(i,:)=cross(r_FP2(i,:),F2(i,:));
Mhg(i,:)=cross(r_hg(i,:),ghg_cont(i,:));
Mtc(i,:)=cross(r_tc(i,:),gtc_cont(i,:));
end
% Moment imposed on knee. Muscle moment will be ()
Mknee= M_lowleg + MF1 + MF2 + Mtc + Mhg;
105
%% 8. Calculate forces, moments, and moment arms in Anatomical shank
coordinates
% Forces
% FA = [RGA]^1 * FG
for i=1:n
aF1(i,:)=inv(RGAshank(:,:,i))*F1(i,:)';
aF2(i,:)=inv(RGAshank(:,:,i))*F2(i,:)';
aW_lowleg(i,:)=inv(RGAshank(:,:,i))*W_lowleg(i,:)';
% COP locations
aCOPlowleg(i,:)=(inv(TGAshank(:,:,i))*[1 COM_lowleg(i,:)]');
aCOP1(i,:)=(inv(TGAshank(:,:,i))*[1 COP1(i,:)]');
aCOP2(i,:)=(inv(TGAshank(:,:,i))*[1 COP2(i,:)]');
end
atc_cont=zeros(n,3);
atc_cont(:,2)=tc_cont;
ahg_cont=zeros(n,3);
ahg_cont(:,2)=hg_cont;
% Total Force imposed on Knee
aFknee= aF1+ aF2 + aW_lowleg + atc_cont + ahg_cont;
% Moment arms
ar_lowleg=aCOPlowleg(:,2:4)/1000;
ar_FP1=aCOP1(:,2:4)/1000;
ar_FP2=aCOP2(:,2:4)/1000;
ar_tc=[0 0 tc_dist]/1000;
ar_hg=[0 0 hg_dist]/1000;
% Moments
for i=1:n
aM_lowleg(i,:)=cross(ar_lowleg(i,:),aW_lowleg(i,:));
aMF1(i,:)=cross(ar_FP1(i,:),aF1(i,:));
aMF2(i,:)=cross(ar_FP2(i,:),aF2(i,:));
aMhg(i,:)=cross(ar_hg,[0 hg_cont 0]);
aMtc(i,:)=cross(ar_tc,[ 0 tc_cont 0]);
end
% Moment imposed on knee
aMknee=aM_lowleg + aMF1 + aMF2 + aMtc + aMhg;
106
%% 9. Normalize forces and moments
% forces will be normalized to body weight
% moments will be normalized to bodyweight*height
% Global forces
nF1=(100/weight)*F1;
nF2=(100/weight)*F2;
nW_lowleg=(100/weight)*W_lowleg;
ngtc_cont=(100/weight)*gtc_cont;
nghg_cont=(100/weight)*ghg_cont;
nFknee=(100/weight)*Fknee;
% Anatomical Forces
naF1=(100/weight)*aF1;
naF2=(100/weight)*aF2;
naW_lowleg=(100/weight)*aW_lowleg;
natc_cont=(100/weight)*atc_cont;
nahg_cont=(100/weight)*ahg_cont;
naFknee=(100/weight)*aFknee;
nMF1=(100/(weight*height))*MF1;
nMF2=(100/(weight*height))*MF2;
nM_lowleg=(100/(weight*height))*M_lowleg;
nMtc=(100/(weight*height))*Mtc;
nMhg=(100/(weight*height))*Mhg;
nMknee=(100/(weight*height))*Mknee;
naMF1=(100/(weight*height))*aMF1;
naMF2=(100/(weight*height))*aMF2;
naMtc=(100/(weight*height))*aMtc;
naMhg=(100/(weight*height))*aMhg;
naMknee=(100/(weight*height))*aMknee;
naM_lowleg=(100/(weight*height))*aM_lowleg;
%% 10. Save forces and moments to *posture*.mat
cd(results_directory)
save(posture,'F1','F2', 'W_lowleg', 'gtc_cont', 'ghg_cont', 'Fknee',...
'nF1', 'nF2','nW_lowleg', 'ngtc_cont', 'nghg_cont','nFknee', ...
'aF1','aF2', 'aW_lowleg', 'atc_cont', 'ahg_cont', 'aFknee',...
'naF1', 'naF2','naW_lowleg', 'natc_cont', 'nahg_cont', 'naFknee',...
'r_FP1', 'r_FP2','r_lowleg', 'r_tc', 'r_hg', ...
'ar_FP1', 'ar_FP2', 'ar_lowleg', 'ar_tc', 'ar_hg', ...
'MF1', 'MF2', 'M_lowleg', 'Mtc', 'Mhg', 'Mknee', ...
'nMF1', 'nMF2', 'nM_lowleg', 'nMtc', 'nMhg', 'nMknee', ...
'aMF1', 'aMF2', 'aM_lowleg', 'aMtc', 'aMhg', 'aMknee', ...
'naMF1', 'naMF2', 'naM_lowleg', 'naMtc', 'naMhg', 'naMknee', ...
'append')
end % end of function forces_moments.m
107
BIBLIOGRAPHY
[1] Lawrence RC, Felson DT, Helmich CG, et al. Estimates of the Prevalence of Arthritis
and Other Rheumatic Conditions in the United States. Part 2. Arthritis & Rheumatism.
2008;58(1):2635
[2] Maetzel A, Li LC, Pencharz J, Tomlinson F, Bombardier C. The economic burden
associated with osteoarthritis, rheumatoid arthritis, and hypertension: A comparative
study. Ann Rheum Dis, 2004;63(4):395 401
[3] Anderson JJ & Felson DT. Factors Associated With Osteoarthritis of the Knee in the First
National Health and Nutrition Examination Survey (HANES I). American Journal of
Epidemiology. 1988;128(1):179189
[4] Kivimäki J, Riihimäki H, Hänninen K. Knee disorders in carpet and floor layers and
painters. Scand J Work Environ Health. 1992;18(5):310316
[5] Zelle J, Barink M, Loeffen R, De Waal Malefijt M, Verdonschot N. Thighcalf contact
force measurements in deep knee flexion. Clinical Biomechanics. 2007;22:821826
[6] Welsh RP. Knee Joint Structure and Function. Clinical Orthopaedics and Related
Research. 1980;147;714
[7] Baker P, Reading I, Cooper C, and Coggon D. Knee disorders in the general population
and their relation to occupation. Occup Environ med. 2003;60;794797
[8] Jensen, Lilli K. Kneestraining work activities, selfreported knee disorders and
radiographically determined knee osteoarthritis. Scand J Work Environ Health. 2005;31
suppl 2:6874
[9] Sharrard, WJW. Pressure Effects on the Knee in Kneeling Miners. Ann R Coll Surg Engl.
1965;36;30924
[10] Sharrard WJW, Liddell FDK. Injuries to the Semilunar cartilage of the knee in miners.
Brit J. Industr. Med, 1962;19;195202
[11] http://www.tricountyortho.com/mmg/pated/knee_problems/anatomy/quads.jpg
Retrieved April 17, 2008
108
[12] Seedhom BB, Dowson D, Wright, V. Proceedings: Functions of the menisci: a
preliminary work. Ann Rheum Dis 1974:33(1):111
[13] Macquet Paul G, Van de Berg André J, Simonet Jacques C. Femorotibial WeightBearing
Areas. The Journal of Bone and Joint Surgey. 1975;57A(6):766771
[14] Smith JP, Barrett GR. Medial and Lateral Meniscal Tear Patterns in Anterior Cruciate
LigamentDeficient Knees. American Journal of Sports Medicine. 2001;29(4):415419
[15] Meister K, Indelicato PA, Spanier S, Franklin J, Batts J. A Comparison of Histologic
Differences in Meniscal Tissue Between Tears in Anterior Cruciate LigamentIntact and
Anterior Cruciate LigamentDeficient Knees. The American Journal of Sports Medicine.
2005;32(6):14791483
[16] Hoshino A, Wallace WA. ImpactAbsorbing Properties of the Human Knee. The Journal
of Bone and Joint Surgery. 1987;69B;807811
[17] Tangtrakulwanich B, Chongsuvivatwong V, Geater AF. Habitual Floor Activities
Increase Risk of Knee Osteoarthritis. Clinical Orthopaedics and Related Research.
2006;454:147154
[18] National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS).
www.niams.nih.gov
[19] Lavender SA and Andersson GBJ. Ergonomic principles applied to prevention of injuries
to the lower extremity. In W Karwowski & WS Marras (Eds.), The Occupational
Ergonomics Handbook. (pp 883893). Boca Raton, FL. CRC Press
[20] United States Bureau of Labor Statistics. Incidence rates for nonfatal occupational
injuries and illnesses involving days away from work per 10,000 fulltime workers by
part of body and selected sources of injury or illness, 2006.
[21] McMillan G, Nichols L. Osteoarthritis and meniscus disorders of the knee as
occupational diseases of miners. Occup Environ Med. 2005;62:567575
[22] Drosos GI, Pozo JL. The causes and mechanisms of meniscal injuries in the sporting and
nonsporting environment in an unselected population. The Knee. 2004;11:143149
[23] Smillie IS. Injuries to the knee joint. Edinburgh, London and New York: Churchill
Livingstone, 1978. pp. 71149
[24] Baratz ME, Fu FH, Mengato R. Meniscal tears: the effect of meniscectomy and of repair
on intrarticular contact areas and stress in the human knee: a preliminary report. Am J
Sports Med 1986:14(4):270
[25] Hawkins A, McNicholas MJ. Meniscectomy and arthritis. CME Orthopaedics
2001;2(2):5156
109
[26] Biswal S, Hastie T, Andriacchi TP, Bergman GA, et al. Risk Factors for Progressive
Cartilage Loss in the Knee. Arthritis & Rheumatism. 2002;46(11):28842892
[27] Hunter DJ, Zhang YQ, Niu JB, et al. The Association of Meniscal Pathologic Changes
with Cartilage Loss in Symptomatic Knee Osteoarthritis. Arthritis & Rheumatism.
2006;54(3):795801
[28] Felson DT & Zhang Y. An Update on the Epidemiology of Knee and Hip Osteoarthritis
with a View to Prevention. Arthritis & Rheumatism. 1998;41(8):13431355
[29] Anderson JA. Athrosis and its relation to work. Scand J Work Environ Health.
1984;10(6):429433
[30] Kellgren JH & Lawrence JS. Rheumatism In Miners Part II: XRay Study. Brit. J.
Industr. Med. 1952;9:197207
[31] Cooper C, McAlindon T, Coggon D, Egger P, Dieppe P. Occupational activity and
osteoarthritis of the knee. Ann Rheum Dis. 1994;53:9093
[32] Zhang Y, Hunter DJ, Nevitt MC, Xu L et al. Association of squatting with increased
prevalence of radiographic Tibiofemoral knee osteoarthritis. The Beijing osteoarthritis
study. Arthritis & Rheumatism. 2004;50(4):11871192
[33] Dahlvist NJ, Mayo P, Seedhom BB. Forces during squatting and rising from a deep squat.
Eng Med 1982;11:6976
[34] Perry J, Antonelli D, Ford W. Analysis of kneejoint forces during flexedknee stance. J
Bone Joint Surg Am. 1975;57:961967
[35] Nagura T, Dyrby CO, Alexander EJ, Andriacchi TP. Mechanical loads at the knee joint
during deep flexion. Journal of Orthopaedic Research. 2002;20:881886
[36] Caruntu DI, Hefzy MS, Goel VK, Goitz HT, Dennis MJ, Agrawal V. Modeling the Knee
Joint in Deep Flexion: ”Thigh and Calf” Contact. Summer Bioengineering Conference.
June 2529, 2003. Key Biscayne, Florida
[37] Winter, D.A. (1990). Biomechanics and motor control of human movement. Wiley, New
York
[38] Bell AL, Brand RA, Pedersen DR. A comparison of the accuracy of several hip center
location prediction methods. Journal of Biomechanics. 1990;23;617621
[39] Clauser CE, JT McConville, JM Young (1969) Weight,volume and center of mass of
segments of the human body. AMRLTR6970, WrightPatterson Air Force Base, Ohio.
[40] Hinrichs, RN. Adjustments to the segment center of mass proportions of Clauser et al.
(1969). J Biomechanics. 1990;23(9):949951
110
[41] Moisio KC, Summer DR, Shott S, Hurwitz DE. Normalization of joint moments during
gait: a comparison of two techniques. Journal of Biomechanics. 2003;36:599603
[42] Allard P, Lachance R, Aissaoui R, Duhaime M. Simultaneous bilateral 3D ablebodied
gait. Human Movement Science. 1996;15:327346
[43] Eng JJ & Winter DA. Kinetic analysis of the lower limbs during walking: what
information can be gained from a threedimensional model? Journal of Biomechanics.
1995;28(6):753758
[44] Hemmerich A, Brown H, Smith S, Marthandam SSK, and Wyss UP. Hip, knee, and
ankle kinematics of high range of motion activities of daily living. Journal of
Orthopaedic Research. 2006;24(4):770781
[45] Li G, Zayontz S, DeFrate LE, et al. Kinematics of the knee at high flexion angles: an in
vitro investigation. Journal of Orthopaedic Research. 2004;22(1):9095
[46] Kadaba MP, Ramakrishnan HK, Wootten ME. Measurement of lower extremity
kinematics during level walking. Journal of Orthopaedic Research. 1990; 8:383392
[47] Stagni R, Fantozzi S, Capello A. Propagation of anatomical landmark misplacement to
knee kinematics: Performance of single and double calibration. Gait & Posture.
2006;24:137141
[48] Martinelli L, Hurschler C, Rosenbaum D. Comparison of capacitive versus resistive joint
contact stress sensors. Clinical and Orthopaedics and related research. 2006;447:214220
[49] Morin EI, Bryant JT, Reid SA, Whiteside RA. Calibration Issues of Tekscan Systems for
Human Pressure Assessment. Presented at RTO HFM Specialists Meeting on “Soldier
Mobility: Innovations in Load Carriage Systems Design and Evaluation”. Kingston,
Canada, June 2729, 2000
[50] Holden JP, Stanhope SJ. The effect of variation in knee center location estimates on net
knee joint moments. Gait & Posture.1998;7;16
[51] Williams A, Logan M. Understanding TibioFemoral motion. The Knee. 2004;11;8188
[52] Blankevoort L, Huiskes R, de Lange A. Recruitment of Knee Joint Ligaments. Journal of
Biomechanical Engineering. 1991;113:94103
[53] Matsumoto H, Suda Y, Otani T, Niki Y et al. Role of the anterior cruciate ligament and
the medial collateral ligament in preventing valgus instability. Journal of Orthopaedic
Sciences. 2001;6(1):2832
[54] Toshiyuki M, Hideo M, Yasumfumi S, Yoshihiro S et al. Function of the knee cruciate
ligaments as restraints on varusvalgus torque. Simultaneous measurement of tension
along the ACL and PCL. Proceedings of Annual Meeting of Japanese Society for
Orthopaedic Biomechanics. 1999;20:151156
111
[55] Fukuda Y, Woo S, Loh J, Tsuda E et al. A quantitative analysis of valgus torque on the
ACL: a human cadaveric study. Journal of Orthopaedic Research. 2003;21(6):11071112
[56] Carlin GJ, Livesay GA, Harner CD et al. Insitu Forces in the Human Posterior Cruciate
Ligament in Response to Posterior Tibial Loading. Annals of Biomedical Engineering.
1996;24(2):193197
[57] Papannagari R, DeFrate LE, Nha KW et al., Function of posterior cruciate ligament
bundles during in vivo knee flexion. American Journal of Sports Medicine.
2007;38:15071512
[58] Li G, Most E, DeFrate LE, et al. Effect of posterior cruciate ligament on posterior
stability of the knee in high flexion. Journal of Biomechanics. 2004;37(5):779783
[59] Hurwitz DE, Sumner DR, Andriachhi TP, and Sugar DA. Dynamic knee loads during
gait predict proximal tibial bone distribution. Journal of Biomechanics. 1998;31:423430
[60] Yao J, Lancianese SL, Hovinga KR, Lee J, Lerner AL. Magnetic resonance image
analysis of meniscal and tibiomeniscofemoral contact in deep knee flexion. Journal of
Orthopaedic Research. 2008;26(5):673684
[61] Schipplein OD and Adriacchi TP. Interaction between active and passive knee stabilizers
during level walking. Journal of Orthopaedic Research. 1991;9:113119
[62] Wilson W, Rietbergen B, Donkelaar CC, and Huiskes R. Pathways of loadinduced
cartilage damage causing cartilage degeneration in the knee after meniscectomy. Journal
of Biomechanics. 2003;36:845851
[63] Atkinson TS, Haut RC, and Altiero NJ. Impactinduced fissuring of articular cartilage:
An investigation of failure criteria. Journal of Biomechanical Engineering. 1998:120:181
187
[64] Eberhardt A, Lewis J, and Keer L. Normal contact of elastic spheres with two elastic
layers as a model of joint articulation. Journal of Biomechanical Engineering.
1991:113:410417
[65] Kelly PA and O’Connor JJ. Transmission of Rapidly Applied Loads Through Articular
Cartilage Part 1: Uncracked Cartilage. Jounral of Engineering Medicine. 1996;210:2737
[66] Thambyah A, Goh JCH, Das De S. Contact stresses in the knee joint in deep flexion.
Medical Engineering and Physics. 2005;27:329335
[67] Costigan PA, Deluzio KJ, Wyss UP. Knee and hip kinetics during normal stair climbing.
Gait and Posture. 2002;16:3137
[68] Kaufman KR, Hughes C, Morrey BF, Morrey M, An K. Gait characteristics of patients
with knee osteoarthritis. Journal of Biomechanics. 2001;34:907915
UNIVERSITY OF PITTSBURGH SWANSON SCHOOL OF ENGINEERING
This thesis was presented
by
Jonisha P. Pollard
It was defended on November 11, 2008 and approved by Jean L. McCrory, PhD, Research Assistant Professor, Health and Physical Activity Richard E. Debski, PhD, Associate Professor, Department of Bioengineering Rakié Cham, PhD, Associate Professor, Department of Bioengineering Thesis Advisor: Mark S. Redfern, PhD, Professor, Department of Bioengineering, Associate Dean for Research
ii
Copyright © by Jonisha P. Pollard 2008
iii
4] In restricted vertical working heights such as lowseam coal mines and aircraft baggage compartments. [1. and pressures between the thigh and calf and heel and gluteal muscles were inputs into this model. ground reaction forces. Kneeling near full flexion and squatting created sagittal joint moments 3 to 5 times larger than standing in iv . kneeling near 90° flexion. we must have an understanding of what the internal knee structures experience under these conditions. or squatting postures to perform work.S. and squatting. net moments. [3. kneeling. Therefore. Pollard. and the heelgluteus contact force. workers are forced to assume stooped. Data from two subjects were analyzed with and without the subject wearing kneepads. The accuracy of any finite element model is heavily dependent on the input parameters (i.e.COMPUTATIONAL MODEL TO DETERMINE TIBIOFEMORAL FORCES AND MOMENTS DURING KNEELING Jonisha P. 2] It commonly affects the knee joint and has been linked to work involving prolonged knee bending. Motion data. and joint angles for subjects while kneeling near full flexion. forces and moments). which was shown to be significant [5]. Additionally the thighcalf contact force. which had not been previously investigated. A finite element model is being developed to quantify the stresses and strains in the tissues in static kneeling postures. University of Pittsburgh. In order to protect the knees in these postures. kneeling on one knee. The developed model uses inverse dynamics to determine the net forces. M. were inputs in this model. 2008 Osteoarthritis affects more than 27 million Americans and cost nearly $5700 per person annually. the objective of this work was to develop a 3D computational model which may be used to determine the net forces and moments applied to the knee joint during static kneeling.
Moments of this magnitude may be significant to cause cartilage damage. It was also found that the moments caused by the thighcalf and heelgluteus contacts act to extend the knee. thereby reducing knee moments in fully flexed postures. v .one subject.
..........................................................................0 1...2............................................................................................................................ 5 Bursitis ..... 15 RESEARCH DESIGN AND METHODS .............................2 2.......................0 3.......... 16 vi ......1.............................................................................................................................. 13 Specific Aim #2...........2.1 INTRODUCTION .........................................3 2.....0 2..... XIV 1............................................1 1..................... 9 PREVIOUS MODELLING OF KNEE FORCES AND MOMENTS .................... 7 MOTIVATION................................. 16 Laboratory Equipment ..............................3 2.....................2...3...........................................1 2......................................... 2 OCCUPATIONAL KNEE INJURIES ......4 3.....................................................................................................3 1........1 1.2..... 5 Meniscal Injuries . 2 KNEE STRUCTURES AND INJURIES ....... 9 SPECIFIC AIMS AND SIGNIFICANCE ..............3................................................. 16 EQUIPMENT ..................... 14 Specific Aim #3..................................... 5 Osteoarthritis ............................1 2................................................................................ 14 Future Significance .................................................................................................................................3....................2 2.....................................2 1................................................ 1 BACKGROUND .....................................................................................2 1...............TABLE OF CONTENTS NOMENCLATURE ....................................................................... 13 Specific Aim #1.................1 3............................................................
........................................................................................................................................7 5...................................................1 Model Assumptions ..................... 23 Construction of Coordinate Systems ....................................2 Computational Model......................2 3.............. 37 KNEE FORCES ............ 47 SENSITIVITY ANALYSIS .......................3 3........2 4...................................................... 19 Subject Instrumentation ......3 4.....2.3 4.................................................. 60 vii ........................................................................................................................................2 4.............. 51 VARYING MODEL PARAMETERS .....................3..0 4.......6 4.............2 3..3...............................1 4...........................................................................................................................1 3......................2 SUBJECT TESTING................1 5.. 45 STATISTICS ....................1.....2................................................. 23 Joint Angle Estimation ................................................................................... 17 3.....................1 3.........2. 34 Squat .......................................0 5............5 4............... 34 THIGHCALF AND HEELGLUTEUS CONTACT FORCES ...........................1.................................1 4................3................................................... 36 KNEE ANGLES................................................................................3 3.......... 21 DATA ANALYSIS .......1.................................................. 19 Experimental Data Collection .........................................1.............. 18 3....... 39 KNEE MOMENTS ..............................................2 3................................................. 44 SUMMARY SUBJECT 2 ........................... 34 Near Full Flexion ..........................4 4........... 51 VARYING KNEE JOINT CENTER LOCATION ............................................................2.......................... 19 ThighCalf and HeelGluteus Contact Measurements ......... 28 Joint Forces and Moments ..................... 29 RESULTS ............................................................................................ 41 SUMMARY SUBJECT 1 ...................................................3...............
.................6 Model Parameters...........................................................................................5 7............................................................ 74 SUBJECT VARIABILITY ..1 7...................................................5 7...............2............ 89 BIBLIOGRAPHY .....5.................0 APPENDIX .................................................................................................................................... 84 FUTURE WORK ..........................2 7. 83 ADVANCEMENTS ................................................ 107 viii ................................................................................................................................................................................................5....................................................... 70 DISCUSSION ...... 87 8................0 7......0 9......................................................5..................4 7...............................................................................1 7.............. 76 IMPLICATIONS OF FINDINGS ................4 7............ 75 LIMITATIONS .......................... 80 Ligament Recruitment ................................................................................................................................................................3 7...............................1 7..3 7............................................ 73 Knee Joint Center Location .......0 7............................................................ 80 Meniscal Loading..5..........................6....................................................................................................... 79 Muscles Activity ..2 7............................................................................................. 86 CONCLUSION.......................... 73 7.....................................2...........................2 MODEL VALIDATION....... 83 Postures Associated with Osteoarthritis .................................................. 72 KNEEPADS............................................................... 72 MODEL SENSITIVITY ....5.. 82 Osteoarthritis Progression .................................................
.................. Subject 1 ........ 48 Table 8: Resulting pvalues for ANOVA.......... 50 Table 11: Sensitivity of sagittal moments to changes in model parameters for all postures ............... Subject 1 ..................... Subjects 1&2 ................... Subject 2 .... 58 Table 12: Percent change in moments due to varying KJC for kneeling near 90 degrees flexion 62 Table 13: Percent change in moments due to varying KJC for kneeling on one knee ................... Subject 2 ..LIST OF TABLES Table 1: Average knee angle (degrees)...... Subject 1 ................. Subject 1 ................................................................... 39 Table 4: Average net external forces normalized by body weight............... 49 Table 9: Resulting pvalues for ANOVA of Kneepad*Posture Interaction.................................................... 68 Table 16: Sensitivity of moments to KJC location for all postures ... 66 Table 15: Percent change in moments due to varying KJC for squatting ........ Subject comparison ........................................... 42 Table 6: Average net external knee moments normalized by Body weight*Height...................................................... 38 Table 3: Average net external forces normalized by body weight.. 40 Table 5: Average net external knee moments normalized by Body weight*Height.... Subject 2 ................ 64 Table 14: Percent change in moments due to varying KJC for kneeling near full flexion ............................................................. 69 ix ................... Subject 2 ............... 37 Table 2: Average knee angles (degrees). 42 Table 7: Resulting pvalues for ANOVA........... 49 Table 10: Resulting pvalues for ANOVA........................
.......................... 40 x ................................................. 35 Figure 13: Thighcalf contact pressure distributions for squat ................ 35 Figure 12: Heelgluteus pressure distribution for kneeling near full flexion................. 31 Figure 11: Thighcalf contact pressure distributions for kneeling near full flexion ........... 3 Figure 2: Anatomical marker set........ 36 Figure 14: Joint angles for Subject 1 kneeling near full flexion without kneepads .................... 30 Figure 9: Diagram of external shank forces and reaction forces and moments for kneeling near full flexion with respect to the ASCS .................................................................................................................................................................................................................... 26 Figure 8: Diagram of external shank forces and reaction forces and moments for kneeling near full flexion with respect to the GCS ........................................................ 20 Figure 3: Measured marker set ..................................................... Subject 1........................ 20 Figure 4: Postures assumed by subjects during testing......LIST OF FIGURES Figure 1: Front view of knee anatomy [11] ............................................................................... 24 Figure 7: Orientation of the ATCS and ASCS........................................................................................ 38 Figure 15: Net external forces normalized by body weight for Subject 1 kneeling near full flexion without kneepads ............ 22 Figure 5: Articulating kneepads commonly worn by lowseam coal miners ............................................................................... 22 Figure 6: Pelvis coordinate system highlighting the location of the right hip joint center .................................................. 30 Figure 10: External force diagrams with respect to the anatomical shank coordinate system .......................................
2 Nm......... and the weight of the foot+shank (low leg). the ground reaction force at the right knee (F2)..... Note that all moment arm values are shown in cm...3 N............ 53 Figure 18: Varied forces and moment arms and corresponding sagittal knee moments for kneeling on the right knee..5 Nm. all forces are shown in N. Varying the zcomponent of the COM of the lower leg by 3 cm changed the moment by 1...54 Figure 19: Varied forces and moment arms and corresponding sagittal knee moments for kneeling near full flexion...5 Nm..1 Nm. 43 Figure 17: Varied forces and moment arms and corresponding sagittal knee moments for kneeling near 90° flexion...... Varying the zcomponent of the F1 moment arm by 3 cm changed the sagittal knee moment by .. Varying the ycomponent of the F2 force by 6 N. changed the sagittal moment by .7 Nm.......... and the weight of the foot+shank (low leg).. Varying the ycomponent of F1 by 6 N changed the moment by 3.. 55 xi . the thighcalf contact force. The forces and moment arms were varied for the ground reaction force at the toes (F1)... the ground reaction force at the right knee (F2). When F1 was varied by 6 N. When this moment arm was increased by 3 cm. Varying the zcomponent of the low leg COM by 3 cm and the ycomponent of the low leg weight by 6 N changed the moment by 1.1 Nm.2 Nm....2 Nm. The forces and moment arms were varied for the ground reaction force at the toes (F1)...... Varying the zcomponent of the F1 moment arm by 3 cm changed the moment by 4. the moment changed by 3. an extension moment was created..... changed the sagittal moment by . Varying the zcomponent of the F2 moment arm by 3 cm was sufficient to more than triple the magnitude of the sagittal moment and change its interpretation.. Varying the zcomponent of the low leg COM by 3 cm changed the moment by 1..... and all moments are not normalized and are shown in Nm.. Varying the ycomponent of the low leg weight changed the moment by 1.......2 Nm . Varying the ycomponent of F1 by 6 N.... and the heelgluteus contact force. Varying the ycomponent of F2 by 6 N changed the moment by .2 Nm..... changed the sagittal moment by 3.. and the weight of the foot+shank (low leg)..7 Nm..Figure 16: Moment contributions normalized by Bodyweight*Height for Subject 1 kneeling near full flexion without kneepads..4 Nm........ Decreasing the zcomponent of the F2 moment arm by 3 cm more than doubled the sagittal moment. Varying the ycomponent of F2 by 6 N changed the moment by .......1 Nm.... Varying the zcomponent of the F1 moment arm by 3 cm.. .... Varying the zcomponent of the F2 moment arm by 3 cm changed the moment by 5. the ground reaction force at the right knee (F2). Varying the ycomponent of the low leg weight by 6 N changed the moment by 1... The forces and moment arms were varied for the ground reaction force at the toes (F1)...3 Nm..
... and the weight of the foot+shank (low leg)......... Varying the zcomponent of the F2 moment arm by 3 cm................. changed the moment by . Varying the heelgluteus contact force by 6 N changed the moment by 2............. changed the moment by 5........ Varying the xcomponent by 3 cm changed the adduction moment by 2.......... changed the moment by ..... respectively....7 Nm...........1 Nm.. changed the moment by 9 Nm... and zcomponents of the KJC location had no effect on the sagittal.....................3 Nm. 56 Figure 21: Varied forces and moment arms and corresponding sagittal knee moments for squatting. Varying the zcomponent of the low leg COM by 3 cm..................... the thighcalf contact force........... and zcomponents of the KJC location had no effect on the sagittal.......7 Nm....5 Nm. ...................9 Nm... Varying the xcomponent by 3 cm and varying the ycomponent by 3cm more than doubled the transverse moment and changed the interpretation of this moment....2 Nm....... y............ Varying the ycomponent by 3 cm changed the flexion moment by 2. Varying the zcomponent by 3 cm changed the adduction moment by 1......................... the weight of the foot+shank (low leg).. 57 Figure 22: Varied KJC locations and resulting knee moments for kneeling near 90° flexion............. respectively...Figure 20: Varied forces and moment arms and corresponding sagittal knee moments for kneeling near full flexion.... and the heelgluteus contact force...... Varying the x.... Varying the thighcalf contact force by 6 N changed the moment by ........... Varying the heelgluteus moment arm by 3 cm changed the moment by 1......... Varying the ycomponent of the low leg weight by 6 N................ Varying the zcomponent of the thighcalf moment arm by 3 cm changed the moment by 3.7 Nm............. changed the moment by 2... Varying the zcomponent by 3 cm more than doubled the sagittal moment...................9 Nm... and transverse moments....... Varying the xcomponent by 3 cm changed the frontal moment by 1....... continued... and transverse moments...................8 Nm............. Varying the ycomponent by 3 cm changed the sagittal moment by 1.. The forces and moment arms were varied for the ground reaction force at the toes (F1)......... y.................... changed the moment by 1.... Varying the ycomponent of F2 by 6 N. frontal..1 Nm..........6 Nm and changed it interpretation..........5 Nm.. Varying the zcomponent changed the frontal moment by 1..... Varying the x................5 Nm.......7 Nm.. the ground reaction force at the right knee (F2)......... frontal.......5 Nm... Varying the ycomponent by 3 cm changed the transverse moment by 1................... The forces and moment arms were varied for the ground reaction force at the foot (F1)........ Varying the ycomponent of the thighcalf contact force by 6 N....... Varying the zcomponent of the thighcalf moment arm by 3 cm..7 Nm.3 Nm..... Varying the zcomponent by 3 cm changed the sagittal moment by 13.................. and the thighcalf contact force....... Varying the xcomponent by 3 cm changed the transverse moment by 9.... 61 Figure 23: Varied KJC locations and resulting knee moments for kneeling on right knee................. 63 xii ..............
frontal..6 Nm... Varying the xcomponent by 3 cm changed the adduction moment by 5. Varying the ycomponent by 3 cm changed the external rotational moment by ... and zcomponents of the KJC location had no effect on the sagittal........... Varying the zcomponent by 3 cm changed the flexion moment by 3... 65 Figure 25: Varied KJC locations and resulting knee moments for squatting. and transverse moments. 67 xiii .6 Nm....6 Nm....... and transverse moments. Varying the zcomponent by 3 cm changed the flexion moment by 3 Nm... Varying the ycomponent by 3 cm changed the flexion moment by 1.....6 Nm. y....6 Nm.. frontal. Varying the xcomponent by 3 cm changed the transverse moment by 3 Nm.6 Nm and changed its interpretation... y.1 Nm. . Varying the x. Varying the x.4 Nm and changed its interpretation. and zcomponents of the KJC location had no effect on the sagittal.. Varying the zcomponent by 3 cm changed the adduction moment by 0.....Figure 24: Varied KJC locations and resulting knee moments for kneeling near full flexion.... Varying the ycomponent by 3 cm changed the transverse moment by 1.. Varying the xcomponent by 3 cm changed the adduction moment by 1..... respectively..6 Nm....... .. Varying the zcomponent by 3 cm changed the adduction moment by 1. Varying the xcomponent by 3 cm changed the transverse moment by 3. Varying the ycomponent by 3 cm changed the flexion moment 5...6 Nm.... respectively.
NOMENCLATURE COM – Center of mass COP – Center of pressure KJC – Knee joint center AJC – Ankle joint center HJC – Hip joint center ATCS – Anatomical thigh coordinate system ASCS – Anatomical shank coordinate system GCS – Global coordinate system MCS – Measured coordinate system MSCS – Measured shank coordinate system MTCS – Measured thigh coordinate system Ft/c – Thighcalf contact force Fh/g – Heelgluteus contact force TTGA – Transformation matrix from ATCS to GCS TSGA – Transformation matrix from ASCS to GCS TTMA – Transformation matrix from ATCS to MTCS TSMA – Transformation matrix from ASCS to MSCS TST – Transformation matrix from ATCS to ASCS xiv .
Fknee – Net external force applied to the knee Mknee – Net external moment applied to the knee xv .
1 . and pressure sensor) will be collected. [1. The goal of this Master’s thesis project was to develop a computational model which will be used to determine the dynamic net forces and moments imposed on the knee joint during static kneeling postures with and without knee pads. Studies have suggested that the risk of developing knee osteoarthritis is increased by work involving prolonged knee bending. Subjects will assume kneeling postures typically seen in the mining industry. which will provide a further understanding of the biomechanics of knee structures while kneeling and crawling. 4] A large project is underway at the NIOSH Pittsburgh Research Laboratory to gain a better understanding of the biomechanics of the knee joint while kneeling. Kinematics and kinetics data (motion capture. which may be used to alleviate stress while kneeling. squatting. trauma.0 INTRODUCTION Osteoarthritis affects more than 27 million Americans and cost nearly $5700 per person annually. electromyography. and pressure sensor data were inputs into this model. or repetitive stress have been associated with osteoarthritis in the knee. force plate data. The results of this thesis work will be incorporated in the finite element model. The ultimate goal of the project is to develop a finite element model of the knee in kneeling postures.1. [3. Motion data. force plate. and crawling in order to develop guidelines for the manufacture of kneepads. age. 2] Certain occupational activities as well as heredity.
stooping. tendons. This is particularly true for kneestraining postures such as kneeling and squatting. Different jobs expose workers to differing ergonomic risk factors. housemaids.1. First. however.2 KNEE STRUCTURES AND INJURIES The knee is a complex weight bearing joint that connects the femur. and cartilage that function to add stability. and crawling) due to the restricted vertical height of their working environment. patella. (For more information on the knee structures and their function see [6]) Any structure of the knee is susceptible to injury. but may also assume awkward postures to perform work. many occupational related disorders affect the bursae. and tibia forming the patellofemoral and tibiofemoral joints. after which a review of occupational factors related to knee injuries is presented. motion. and cartilage. Other occupations such as plumbers. are forced to assume awkward postures (such as kneeling. the musculoskeletal structures that are affected by kneeling and squatting are discussed.1 BACKGROUND Musculoskeletal disorders vary by occupation.10] 2 . meniscus. such as underground coal mines or aircraft baggage compartments. [7 . This background section focuses on musculoskeletal injuries due to the awkward postures of kneeling and squatting. It is also comprised of various ligaments. roofers. and agricultural workers may not have this restriction in vertical height. People who work in restricted spaces. 1. carpet layers. and weight bearing (Figure 1).
Any of these bursae may become irritated and inflamed. septic bursitis.Figure 1: Front view of knee anatomy [11] The bursa is a fluid filled sac that acts as a cushion between a tendon and a bone or a tendon and a muscle. resulting in bursitis. lubrication. Bursitis generally results in a swollen bursa and restricted joint motion. and shock absorption. This bursa facilitates full flexion and extension of the knee and may become injured during acute trauma. It allows the skin to move easily over the tibial tuberosity and withstands pressure when kneeling with the trunk upright. both deep and subcutaneous. The subcutaneous infrapatellar bursa is located between the skin and the tibial tuberosity. The prepatellar bursa is between the patellar tendon and the skin. It acts as a cushion. reducing the friction between the patella and the skin while kneeling. Within the knee there are numerous bursae. resulting in a serious condition. The meniscus is a cartilaginous structure that provides a number of basic functions: load bearing. There are three large subcutaneous bursae that may become injured from repetitive motions or trauma. stability. In some cases the bursa may become infected. The meniscus transmits between 5070 % 3 . The suprapatellar bursa is between the femur and the tendon of the quadriceps femoris muscle.
[12] During the stance phase of gait the mean contact stress on the meniscus is 20 klbs per cm2. This breakdown in cartilage affects the lubrication process within the knee. 18] The risk of developing this disease increases with age and it is estimated that by 2030. [14] Meniscus tears without ACL injury are commonly the result of degeneration of the meniscus. [1. thereby lowering the stress applied to the articular cartilage and acting as a shock absorber. pain. [13] The meniscus provides stability by interacting with the Anterior Cruciate Ligament (ACL). and results in direct contact between the bones causing stiffness. This form of meniscal tear is commonly concurrent with an ACL injury. leading to microfractures of the subchondral bone. 17. [18] 4 . [15] The articular cartilage is another structure that is commonly injured. affecting more than 27 million Americans. 20% of Americans (about 72 million people). Osteoarthritis is a chronic condition characterized by the breakdown of the joint’s articular cartilage. [16] Osteoarthritis is the most common joint disorder associated with major disability. The most common mechanism of traumatic meniscal tear occurs predominantly in athletes when a twisting moment is applied to the weight bearing knee in a semiflexed position. The meniscus increases the load carrying area of the joint by 200 %. will be at a high risk for this disease. The meniscus also prevents friction between the femur and tibia and allows diffusion of joint fluid into the articular cartilage. It also reduces the shock absorbing capacity of the knee. assisting with rotational stability.of the compressive loads in the knee. [12] The meniscus can be damaged by both traumatic and degenerative mechanisms. and loss of motion.
Prepatellar bursitis is the most common form of bursitis and known as “housemaid’s knee” or “miner’s knee”.2 Meniscal Injuries Frequent kneeling also has been linked to an increased laxity in the ACL and development of meniscal injuries. After reviewing records for men between 15 and 64.1. knee injuries most commonly result from worker position or motion. the authors determined that cartilage tears were more common in 5 . [20] Kneeling and crawling can cause knee injuries ranging from minor skin irritations to bursitis or a torn meniscus. [9] This thickening can lead to irritation and inflammation. from five hospitals in one of the largest British coalfields.3. Sharrard and Liddell (1962) investigated hospital records of 957 meniscectomies. Bursitis usually does not lead to other knee problems when properly treated. the authors determined that miners are more likely than others to suffer cartilage damage of any type. [19] Of all lower extremity injuries. 1. [21] 1.1 Bursitis Frequent kneeling produces large stresses on the patella which may thicken the walls of the bursa. it does account for days lost from work and may lead to infection if left untreated. performed between January 1958 and June 1960.3. Upon interviewing 200 of the men.3 OCCUPATIONAL KNEE INJURIES Working in certain awkward postures has been to shown to affect musculoskeletal disorders of the lower extremities. While this form of bursitis does not typically lead to other knee injuries.
22. Previous research shows that there is a significant risk for meniscus damage associated with an increase in age and frequent and prolonged kneeling. aging plays a role in the development of meniscal injuries. in conjunction with the agerelated disappearance of elastic fibers. 10. but the tears normally occurred when walking. The most frequently reported mechanism of injury for the nonsporting group was rising from a squatted position. thereby increasing the susceptibility to injury. may precede the development of many meniscal tears.8% had nonsportsrelated injuries and 28. This degeneration reduces the elasticity of the meniscus. they determined that ACL laxity resulted from kneeling at work. [22]. Drosos and Pozo deduced that degeneration may be the result of the repetitive microtrauma and mechanical stresses of everyday life. From these examinations. 38. it leaves the articular cartilage of the femur and tibia without a proper intermediate tissue. Smillie (1978) proposed that meniscus degeneration starts in the third decade of life.4% of patients had sportsrelated meniscal injuries. The weight bearing area of an intact meniscus 6 . the authors examined 80 patients with meniscal damage. Although this procedure is necessary to ensure proper joint motion.miners who knelt. [7.8% had no identifiable injury or no identifiable cause of injury. thereby requiring less force to create an injury. Along with kneeling. 32. To gain a better understanding of why the tears occurred in nonkneeling posture. These factors. The authors proposed that this laxity decreases the stability of the knee in nonrestricted postures such as standing and walking. [23] Drosos and Pozo (2004) conducted a study on 392 patients between the ages of 18 and 60 with meniscal injuries from the general population. leading to meniscal tears. while others will have to be removed in a meniscectomy. 23] Some meniscus tears can be repaired. The authors felt that this age may reflect the degenerative changes in the meniscus. The average age of the patients with nonsporting related meniscal tears was 41.
28] 1. [25 . [13] This decreased area increases the peak local contact stresses transmitted to the articular cartilage by 65235 %. causing degeneration. Of the 5. or crawling. and occupation.45 in men. smoking. kneeling. crouching. [8.3 Osteoarthritis Osteoarthritis has been associated with a number of personal and activityrelated risk factors. The increased risk for osteoarthritis in those with physically demanding occupations was seen mostly in people 55 years and over (OR=2. [3] 7 . a knee bending variable has been examined.32) amongst both men and women in the 5564 years age group. OR=3. sports activities. 17.193 people surveyed. such as stooping. 30. The authors concluded that the strong association with occupations in which knee bending is prominent suggests work activity may play a major causative role in osteoarthritis of the knee. including levels of knee bending. age.3. respectively.49 in women). After meniscectomy the area ranges from 6 cm2 to 12 cm2. 31] Within the occupational factors. and obesity. 29. meniscal injuries have been linked to the longterm development of knee osteoarthritis. [24] Thus. Multiple linear regressions found knee bending demand of the job was associated with osteoarthritis of the knee (OR=0. [3] Activityrelated risk factors are trauma. Personal risk factors include heredity. 315 were found to have radiographic osteoarthritis of the knee. diabetes.varies from 11 cm2 at 90° flexion to 20 cm2 at full extension.
lotus position. 4. and sideknee bending were found to increase the risk of developing knee osteoarthritis. [3. kneeling. and lotus position. 32] This is evident in the high prevalence of knee osteoarthritis among Asian populations who spend a lot of time performing floor activities. 17. Of these postures. sideknee bending.Subsequent studies have suggested that the risk of developing knee osteoarthritis is increased by work involving prolonged knee bending and squatting. [17] The postures examined were squatting. 8 . squatting.
9 . For three subjects. External force reactions and points of application. Anthropometric data and xrays were used to create the model.2. lower leg configurations. 2.6 times bodyweight for fast descent. There are few studies available on squatting and high flexion activities. much of this research is aimed at the low back forces and causes of low back pain. or squatted postures. forces were calculated throughout going into a deep squat and then rising to a standing position.7 times bodyweight for slow ascent and 7. stooped.1 PREVIOUS MODELLING OF KNEE FORCES AND MOMENTS Dahlkvist et al. Joint forces varied across subjects and during activities. no studies of kneeling on one knee or near 90° flexion were found. (1982) developed subjectspecific 2D models of the lower leg of six male subjects to determine the forces at the knee during squatting and rising from a deep squat. however. Patellofemoral joint forces were shown to be the largest. with average maximums between 4.0 MOTIVATION Much attention has been given to the biomechanics of the knee joint during gait but the same attention has not been given to the knee joint while in kneeling. and electromyographic data were measured. For the other three subjects rising from a deep squat were compared for a slow rise trial and a fast rise trial. Although there has been research on restricted postures.
The quadriceps force necessary to stabilize the knee was found to be directly proportional to the femoral load and angle of flexion. Flanged transducers were inserted into the tibia to measure the tibiofemoral joint stress. 5°. 45°. The knee was set to six positions for measurement: 0°. Compressive forces of the patella were also measured. [34] 10 . 30°. aligned parallel to the femur. A ring tensiometer.Perry et al. Forces in the quadriceps. (1975) performed a cadaver study to show significantly increased forces on the joint surfaces with flexion of greater than 30°. and 60°. the required force was 210% body weight and at 60° the required force was 410% body weight. 15°. Quadriceps muscle action was simulated by a cable linked between the vastus intermedialusrectus femoris to the intertrochanteric region of the femur. patellofemoral compression. At the 30° position. The patellar forces were found to be a function of quadriceps tension and dependent on the angle of flexion and the applied femoral load. patella. and tibiofemoral loading. the tibiofemoral forces were a summation of the load on the femoral head and the quadriceps force. measured the forces in the cable as loads were applied to the femoral head. and tibia during simulated weightbearing flexed knee stance were determined in a cadaver model. The specimen was instrumented to measure the quadriceps force. The tibiofemoral forces also increased in direct proportion to the load on the head of the femur when in full extension. When flexed.
the distal pole of the patella. the average miner shoveled once every 2. The net posterior force increased by 50% over walking and stair climbing. and 84 lbs respectively. stated that neglecting this contact was a limitation of their study which may cause an overestimation of net forces and moment. Net moments and posterior forces were shown to be significantly higher in deep flexion than in walking or stair climbing. Subjects were separated into two groups. At that time. and the body of the patella were also measured. [33. Caruntu et al. (2003) created a model of the knee for 11 . The authors concluded that the increase in extensor force during deep flexion increases the stress in the patellar tendon and joint contact forces. right knee. the weight on the left knee. These pressures were shown to vary throughout the shoveling task as well. During the shoveling task. The pressures at the tibial tubercle. [35] Sharrard (1965) examined the forces and pressures at the knee joint of a miner simulating a shoveling task while kneeling. (2002) determined the forces during squatting activities using an inverse dynamics approach and found similar high contact forces. the interval between the tibial tubercle and the patella. the ambulation group and the deep flexion group. He placed scales under the knees and toes of the miner. and toes were 46 lbs. Net moments and inferior forces were significantly smaller in deep flexion than in walking or stair climbing.Nagura et al. the weight on the knees and toes rose and declined rapidly. Results showed that the pressures can rise as much as 200 pounds per square inch while shoveling. [9] Previous biomechanical models of the knee used to evaluate deep flexion and squatting postures did not account for the contact between the thigh and the calf. 56 lbs.5 seconds. When the 189 lbs subject rested sitting on his heels. Considerable variation was shown in the pressures and weightbearing surfaces of the knee while kneeling and performing work. 34] Nagura et al.
12 . Research on higher flexion postures such as kneeling near full flexion and squatting did not include thighcalf or heelgluteus contact and may result in the overestimation of joint forces and moments. which may be commonly utilized when performing work. Results showed the resultant contact force to be >30% body weight and located within 17 cm of the epicondylar axis. This model may be applied to other postures in the future and accounts for the contact between the thigh and the calf as well as between the heel and the gluteal muscles which may occur when kneeling in high flexion.8 and 156.deep flexion which included the contact between the thigh and the calf. [5] Thighcalf contact has been shown to be significant. They also reported a 50% increase in medial collateral ligament forces when this contact was considered. For these reasons the contact force between the thigh and the calf was not neglected from this computational knee model. (2007) measured the pressure distribution of thighcalf contact for subjects squatting and kneeling with mean knee angles of 151. and neglecting these parameters may result in overestimation of joint forces and moments.4 respectively. They reported a 700 N overestimation in quadriceps force when this contact was neglected. [36] Zelle et al. In summary previous research on knee biomechanics in restricted postures have failed to examine kneeling near 90° flexion and kneeling on one knee. In this research a computational model was develop to determine the net dynamic forces and moment imposed on the knee during static kneeling postures.
this model accounts for the contact force between the thigh and calf. Also. this contact may be sufficient to have a considerable effect on the forces. The use of one computational model to analyze different postures will allow for better comparisons and interpretation of these forces and moments. unlike traditional models based on inverse dynamics. 13 . [5] At this level of force.2 SPECIFIC AIMS AND SIGNIFICANCE 2. but also for kneeling near 90° flexion and kneeling on one knee. there has not been an analysis of the forces and moments at the knee during kneeling. [33 36] The role of this study was to develop a computational model that may be used to determine the net forces and moments of the knee during not only squatting and deep flexion. which has been determined to be >30% body weight.2. and stresses in knee structures. Previous research has determined the net forces and moments at the knee joint for squatting and end range flexion.2.1 Specific Aim #1 Develop algorithms and a 3D computational model to calculate the net dynamic forces and moments at the knee joint from motion capture and force plate data To date. moments. This model also accounts for the contact between the heel and the gluteal muscles which occurs in some people when kneeling and sitting on their heels.
knee pads) on knee forces and moments has not been investigated. thereby affecting the moments at the knee. and heelgluteal contact force) will be implemented. This analysis was also useful in determining which parameters introduce the largest sources of error in the model. In this study the subject was evaluated with and without knee pads. motion capture. The effect of protective equipment (i. Force plate.2 Specific Aim #2 Apply the developed model to data collected from a minimum of one subject. joint center location estimation. The resulting forces and moments will add to the knowledge on kneeling biomechanics. these forces and moments will be utilized in the finite element model of the knee currently being developed. Numerous factors contribute to the forces and moments at the knee joint while kneeling.2. the knee pads were expected to change the location of the center of pressure for the force at the kneeground interface. 14 . This model will be used in the future on a larger dataset currently being collected. 2. The forces were not expected to change significantly due to the knee pads however.2. thighcalf contact force. A sensitivity analysis was performed to determine which variables contribute the most to the forces and the moments at the knee. and pressure data were analyzed via the developed model for two subjects.e.3 Specific Aim #3 A sensitivity analysis for the input variables (anthropometric variables. In addition.2. ground reaction forces.
These results will be used to provide interventions. 15 . crawling.4 Future Significance The forces and moments determined using this computational model will be inputs to a finite element model (FEM) of the knee being developed. Results of the FEM will provide a better understanding of the internal biomechanics of the knee structures while in restricted postures.2.2. This FEM will be used to evaluate the stresses and strains in tissues while kneeling. and stooping. in hopes of reducing the prevalence and severity of occupational knee injuries in lowseam coal miners. such as a novel knee pad design.
Motion data was collected at 60 Hz. Pressure data were also collected to determine the magnitude of the force transmitted from the thigh to the calf and from the heel to the gluteal muscles when kneeling. The cameras are arranged around the room such that all markers are visible during testing. allowing the estimation of knee forces. California. USA) which track the motion of ½” reflective markers. Santa Rose.1. This data was used to determine the segment positions. segments lengths. Motion Analysis Corporation.1 EQUIPMENT 3. 16 . moments.1 Laboratory Equipment Motion data is captured using eleven Infrared cameras (Eagle. 3.0 RESEARCH DESIGN AND METHODS Motion analysis and force data were collected to track body positions and ground reaction forces for simulated postures. These parameters were inputs into the computation model. and ground reaction forces. and joint angles.3.
fibula. Massachusetts. 3. Advanced Mechanical Technology. total force. It is based on an inverse dynamics method [37] which uses measured ground reaction forces and anthropometric measurements to determine the net external forces and moments at the right knee joint. Total force and center of pressure locations from the supplied software were used in the computational model. Massachusetts. Inc. USA) were used to measure the reaction forces at the groundknee. Tekscan Inc. This 19. Tekscan Inc. In this linear model the upper leg (femur) and the lower leg (tibia. and the zaxis was in the superior/inferior direction. MA) on a personal computer.8 in pressure sensor uses resistive technology to measure the pressure between surfaces. Massachusetts.2 Computational Model The computational model was developed in MATLAB® (The Mathworks Inc.. South Boston.. USA) was used to measure the contact between the thigh and calf and the heel and gluteal muscles. 17 . or groundkneepad interface.Two force platforms (Model OR65. USA) generated pressure maps. The supplied software (Advanced ClinSeat. The force platforms were aligned in parallel and were level with the floor. Newton. groundfoot... the xaxis was in the medial/lateral direction. Force data was collected at 1020 Hz.2 in x 16. The global reference frame was oriented such that when the subject was standing in standard anatomical position. A clinical seating pressure assessment system (ClinSeat®. and center of pressure. and foot) are modeled as rigid bodies attached via a pin joint with three rotational degrees of freedom. Natick. South Boston. The spatial resolution is 1 sensel/cm2 with a pressure range of 030 PSI.1. The origin of this system was on the laboratory floor at the upper left corner of the first force plate. total pressure. the yaxis was in the anterior/posterior direction.
• Affect of patellar tendon and tibial tubercle on forces externally applied to the tibia is negligible. This allows one center of mass to represent the weight of the segment. • The knee is assumed to be a frictionless pinjoint. opposed to a pressure distribution. • The measured thighcalf and heelgluteus contact forces are concentrated at the measured center of pressure location.3. This allows the contact forces to be represented as a single resultant force. This assumption allows ground contact forces measured at the groundknee or groundkneepad interface to be assumed to act at some distance away from the knee joint center and not be affected by the patellar tendon or tibial tubercle. This allows the assumptions that markers placed on palpable landmarks are directly on that landmark. • The relative movement of pelvic bones is negligible.2. This allows approximation of the hip joint center from palpable pelvic landmarks. • Thickness of subcutaneous tissue between bone and skin is minimal. This allows all forces to pass directly through the joint center. • Linear relationship between external forces and moments applied to the knee. This allows a planar model to be used to determine the external forces and moments applied to the knee.1 Model Assumptions The developed computational model is based on several assumptions.1. • Segments are assumed to be rigid with mass concentrated at the center of mass locations. 18 .
The marker included one segment marker and a three marker cluster of the thighs and shanks.3. This data was used to create the anatomical coordinate system for the thigh and shank. The distance from the top of the sensor to the lateral epicondyle was measured and recorded while data was collected. 19 . This was repeated with the subject squatting.2 Subject Instrumentation Each subject was fitted with 41 reflective markers using a modified version of the Cleveland Clinic Foundation’s marker set.2. The subject was then instructed to kneel into full flexion while pressure data was collected for a period of 5 seconds. This marker set was chosen to be compatible with SIMM for use in a future study. Figure 2.1 ThighCalf and HeelGluteus Contact Measurements Thighcalf and heelgluteus contact forces (Ft/c and Fh/g) were measured prior to motion and force data collection. The subject was instructed to perform a standing Tpose with their right foot making contact with force plate 2 and their left foot off of the force plates. The subject was instructed to kneel near 90° flexion and the pressure sensor was placed on their lower leg.2. 3. Data was collected for 5 seconds.2 SUBJECT TESTING 3.
Figure 2: Anatomical marker set Figure 3: Measured marker set 20 .
and squatting. kneeling on right knee. twisting at waist. however given instructions on where to place their knees and feet when kneeling. kneeling on all fours. They began in a standing Tpose then performed lateral bending. During the range of motion. When testing the kneeling near 90° flexion.e.3 Experimental Data Collection Once the template was extended. simulating a restricted working height. ankle rotations. and knee flexion/extension. 3. they are virtual markers which were reconstructed during data processing). A mesh roof in the laboratory was lowered to 48”. kneeling down into full flexion. crossing arms. The subject was instructed to perform a static trial of each posture for 10 seconds. They were not given specific instructions on kneeling postures as to not affect their normal kneeling postures. subjects were instructed to kneel 21 . raising knees up to chest. squatting. This yielded the measured marker set shown in Figure 3. kneeling on the right knee. Figure 4. kneeling on right knee. rotating head then dropped down into a kneeling posture. and ended in a standing Tpose.Following recording of the standing Tpose. The subject was instructed to perform another standing Tpose as well as a range of motion. eight markers were removed due to their high risk for falling off or becoming covered (i. kneeling on left knee. They were.2. Subjects were instructed to simulate the postures shown in the schematics. twisting at waist. experimental data collection began. A poster in the laboratory showed schematics of the postures studied. kneeling near full flexion. Subjects were not given specific instructions on these postures. In this kneeling posture the subject performed lateral bending. The postures studied were kneeling near 90° flexion. and kneeling near full flexion. the subject went through a series of motions with a laboratory assistant. The range of motion data was then postprocessed and used to extend the motion capture template.
Figure 4: Postures assumed by subjects during testing Figure 5: Articulating kneepads commonly worn by lowseam coal miners 22 . At the start of testing. Motion data was lowpass filtered using a 4th order Butterworth filter at 15 Hz to smooth instrumental errors. subjects were instructed to kneel with their right foot on force plate 2 and their left foot off of the force plates. When testing the squatting posture. Figure 5. The subject performed each posture with and without knee pads for a total of eight trials. each subject was provided a new pair of orange articulating knee pads commonly worn by underground coal miners.with their right knee on force plate 2 and their right foot on force plate 1. Their left side was to remain off of the force plates.
ASIS).3 DATA ANALYSIS Data was analyzed for a 5 five second period of the trial after the subject appeared to have reached a balanced posture and showed minimal instability.ASIS & R. The anatomical coordinate system of the thigh (ATCS) was created using the left and right anterior superior iliac spine (L. knee. The anatomical system was created from the anatomical standing Tpose and allowed the location of anatomical landmarks to be linked to the global reference frame. knee joint center (KJC).3. and joint angles. r1 = r2 = knee lateral − knee medial knee lateral − knee medial HJC − KJC HJC − KJC x − axis r3 = r2 × r1 r4 = r1 × r3 y − axis z − axis 23 .3. moments. an anatomical and measured coordinate system was created from the motion capture data. This portion of the static trial was used to determine the forces.1 Construction of Coordinate Systems For each segment. It was also used to determine the location of the ankle joint center (AJC). and thigh markers. 3. and hip joint center (HJC) as well as the location of the lower leg center of mass. A measured coordinate system was created from the anatomical standing Tpose as well as from each static trial and was used to link the testing markers to the locations of the markers that were removed.
The location of the HJC was approximated using regression equations proposed by Bell et al. [38] (Figure 6) Figure 6: Pelvis coordinate system highlighting the location of the right hip joint center Origin = LASIS + RASIS 2 PW = RASIS x − LASIS x Origin + . (1990) and adapted to fit the global reference frame of the laboratory.36 * PW HJC = Origin − . measured by the medial and lateral knee markers.The KJC was assumed to be midway between the medial and lateral epicondyles of the femur.19 * PW Origin − .3 * PW 24 .
1 T TGA= HJC r 1 0 r 3 0 0 r 4 The anatomical coordinate system of the shank (ASCS) was determined using the knee. The transformation matrix from the GCS to the ASCS (TSGA) was created from the unit direction vectors of the ASCS.AJC KJC − AJC z − axis y − axis x − axis 0 r 4 0 r 3 0 r 1 r3 = r1 × r2 r4 = r3 × r1 1 T SGA= KJC The ATCS and ASCS were oriented such that when standing the systems aligned with the GCS and the positive xaxis is in the lateral direction of the right leg. the positive zaxis is in the proximal direction. (Figure 7) 25 . ankle. The AJC was assumed to be midway between the medial and lateral malleoli. r1 = r2 = knee lateral − knee medial knee lateral − knee medial KJC .The transformation matrix from the global reference frame to the anatomical thigh coordinate system (TTGA) was created from the unit direction vectors of the ATCS. and shank markers. measured by the medial and lateral ankle markers. and the positive yaxis is in the anterior direction.
Figure 7: Orientation of the ATCS and ASCS A measured coordinate system (MCS) was created for the thigh and the shank using the marker clusters on the segments. r1 = thigh − thigh front thigh − thigh front thigh front − thigh rear thigh front − thigh rear z − axis r2 = r3 = r2 × r1 r4 = r1 × r3 1 T TGM = thigh front 0 r 3 0 r 4 0 r 1 x − axis y − axis 26 . and thigh rear markers. thigh front. The transformation matrix from the GCS to the MTCS (TTGM) was also created with the right thigh front marker as its origin. The measured coordinate system of the thigh (MTCS) was created from the thigh.
r1 = shank − shank front shank − shank front shank front − shank rear shank front − shank rear z − axis r2 = r3 = r2 × r1 r4 = r1 × r3 1 T SGM = shank front r 3 0 r 4 0 0 r 1 x − axis y − axis To determine measured marker locations in the ATCS and ASCS. shank front. and shank rear markers. The transformation matrix from the MSCS to the GCS (TSGM) was created with the right shank front marker as its origin.The measured coordinate system of the shank (MSCS) was created from the shank. TTMA = [TTGM ] [TTGA ] −1 TSMA = [TSGM ] [TSGA ] −1 27 . TTMA and TSMA were created respectively.
yielding an Euler order of Xy’z’’. and the distal/proximal zaxis (γ . This matrix was used to determine the Euler angles. motion of the thigh was in respect to the shank. Therefore.3.varus/valgus).3. The largest joint rotations occurred about the medial/lateral xaxis (α – extension/flexion) followed by the posterior/anterior yaxis (β .internal/external rotation). TSGA = [TSGM ] * [TSMA ] TTGA = [TTGM ] * [TTMA ] TST = [TSGA ] * [TTGA ] −1 RST 0 0 cos( β ) 0 sin( β ) cos(γ ) − sin(γ ) 0 1 0 cos(α ) − sin(α ) * 0 = 1 0 * sin(γ ) cos(γ ) 0 0 sin(α ) cos(α ) − sin( β ) 0 cos( β ) 0 0 1 RST cos( β ) cos(γ ) sin( β ) − cos( β ) sin(γ ) sin(γ ) cos(α ) + sin( β ) cos(γ ) sin(α ) cos(γ ) cos(α ) − sin( β ) sin(α ) sin(γ ) − cos( β ) sin(α ) = sin(γ ) sin(α ) − sin( β ) cos(γ ) cos(α ) cos(γ ) sin(α ) + sin( β ) sin(α ) cos(γ ) cos( β ) cos(α ) 28 . The transformation matrix from the anatomical thigh to the anatomical shank coordinates TST was created to determine the rotation matrix. RST.2 Joint Angle Estimation Joint angles were determined for each trial using Euler Angle Decomposition.
Squatting and kneeling creates a contact force between the thigh and the calf.1 cos( β ) cos(γ ) α = tan −1 β = tan −1 3. and in some subjects there is additional contact between the heel and the gluteal muscles. 29 . 1969 which were adjusted to use the knee joint center and ankle joint center in this model. The locations of these forces were determined from the center of pressure locations on the pressure sensor. Kneeling near full flexion also creates this contact. Ft/c and Fh/g were modeled as resultant forces whose line of action was in the anterior direction of the shank.3.3 Joint Forces and Moments Ground reaction forces. Forces at the foot (F1). forces at the knee (F2) and the weight of the lower leg were measured with respect to the GCS. 2 − (− cos( β ) sin(γ ) ) = tan −1 R ST cos( β ) cos(γ ) 1 . 3 − (− cos( β ) sin(α ) ) = tan −1 R ST cos( β ) cos(α ) 3. Ft/c.. (Figure 8) Thighcalf and heelgluteus contact forces were measured with respect to the ASCS. segment weight. External force diagrams for kneeling near full flexion with respect to the GCS and the ASCS are shown in Figure 8 and Figure 9. 3 cos(γ ) sin( β ) = tan −1 R ST1.γ = tan −1 − R ST 1. and Fh/g were inputs into the computational model. 3 cos(γ ) R ST1. [40] The reaction forces and moments were assumed to act in the positive directions. The center of mass location and weight of the shank+foot were determined using equations from Clauser et al.1 − R ST2 .
Figure 8: Diagram of external shank forces and reaction forces and moments for kneeling near full flexion with respect to the GCS Figure 9: Diagram of external shank forces and reaction forces and moments for kneeling near full flexion with respect to the ASCS 30 .Figure 9 shows the orientation of the forces and moments as presented in this research. with respect to the ASCS. External force diagrams for all postures are shown in Figure 10.
Figure 10: External force diagrams with respect to the anatomical shank coordinate system 31 .
therefore the sum of all forces and moments at the knee were calculated and summed to equal zero. R. y. and z directions were all summed to zero to determine the reaction force.Joint equilibrium was assumed. The sum of the external forces in the x. necessary to stabilize the knee due to the application of the net external forces (Fknee). ∑F x =0 F1x + F2 x + Rx = 0 Rx = − F1x − F2 x ∑F y =0 F1 y + F2 y + R y = 0 R y = − F1 y − F2 y ∑ Fz = 0 F1z + F2 z + R z − W LL − Ft / c − Fh / g = 0 R z = W LL − F1x − F2 x − Ft / c − Fh / g Fknee = − ( Rx + R y + Rz ) 32 .
M F 1 = KJC1x − COP1x M F2 x [ = [KJC KJC y − COP1 y KJC y − COP2 y KJC z − COP1z × F1x KJC z − COP2 z 2x − COP2 x ] [ ] × [F F1 y F2 y F1z ] F2 z ] M LL = KJC x − COM LL x M t / c = KJC x − COPF t / c x M h / g = KJC x − COPF h / g x [ KJC y − COM LL y KJC y − COPF t / c y KJC y − COPFh / g y KJC z − COM LL z × [0 0 − WLL ] KJC z − COPF t / cz × [0 0 − Ft / c ] KJC z − COPF h / g z × 0 0 − Fh / g ] [ [ ] ] [ ] ∑M KJC =0 M KJC + M t / c + M h / g + M F2 + M F1 + M LL = 0 M KJC = − (M t / c + M h / g + M F2 + M F1 + M LL ) M knee = − M KJC = M t / c + M h / g + M F2 + M F1 + M LL 33 . The net external knee moment applied to the knee joint (Mknee) was also determined. z directions were also summed to equal zero. and. y.The sum of the external moments at the knee joint in the x.
4.5 N and 164 N.1 THIGHCALF AND HEELGLUTEUS CONTACT FORCES 4.17 m along the long axis of the tibia from the medial epicondyle of the femur. All forces and moments presented are in reference to the anatomical shank coordinate system. Subject 1 was a 1. 720 N. Pressure distributions. This pressure distribution is shown in Figure 12. 59 year old female. 34 .43 m along the long axis of the tibia from the medial epicondyles of the femur. Neither subject had a history of knee injuries or pathologies.1 Near Full Flexion The mean Ft/c in the kneeling near full flexion posture for subjects 1 and 2 were 127.0 RESULTS Data from two subjects were analyzed using the developed computation model. respectively. Subject 2 was a 1. The centers of pressure were 0. 675 N.14 m. Subject 1 had a mean Fh/g of 56. and 0.1. Subject 2’s posture did not cause any contact between the heel and gluteal muscles. showing the centers of pressure for this posture are shown in Figure 11.4. 19 year old male.56 m.8 m.1 N with a center of pressure 0.
a) b) Proximal Shank Figure 11: Thighcalf contact pressure distributions for kneeling near full flexion a) Subject 1 b) Subject 2 Heel Figure 12: Heelgluteus pressure distribution for kneeling near full flexion. Subject 1 35 .
15 m.2 Squat The mean Ft/c in the squatting posture for subjects 1 and 2 were 173 N and 195 N.14 m along the long axis of the tibia from the medial epicondyle of the femur. showing the centers of pressure for this posture are shown in Figure 13. a) b) Proximal Shank Figure 13: Thighcalf contact pressure distributions for squat a) Subject 1 b) Subject 2 36 . and 0. respectively.1. The centers of pressure were located 0.4. Pressure distributions.
Both subjects had very small standard deviations in the included joint angles over time. Subject 1 Included Near 90° Flexion One Knee Near Full Flexion Squat Standing 83.2 6.3 7.251.3 5.42) 1.7 10 (0.35 4.7 7.29) (0. creating a standard deviation of 2.0 2.69° for varus/valgus.221.47) 0.21) 24. .16) (0.Varus .1 (1. Table 1: Average knee angle (degrees).9 5.2 (0.4.57) (.368.28) 75. and included knee angles for Subjects 1 and 2 are shown in Table 1 and Table 2.6 (0.73) 1.Ext rot 82.22) (.2 KNEE ANGLES Mean varus/valgus. This 30° range in angles caused a standard deviation of 8.41) (0.44° for the included angles.6 (2.47) (1.2 1.0 14.55) Without Kneepads + Valgus + Int rot Included . Varus/valgus deviations were the largest for both subjects. Subject 2 ranged from 3° varus to 3° valgus when wearing kneepads and kneeling on one knee.55) (0. Subject 1 ranged from 15° varus to 15° valgus when wearing kneepads and kneeling near 90° flexion.1 5.07) 173. and .7 (8.25) (3.78 (.9 8.43) (1. Standard deviations across 5 second portions of the trials are shown in parentheses.0 (0.6 .44) 23.6 2.0 (0. respectively. internal/external rotation.60 (0.1°.0 (3.32) 22.Ext rot 0.7 0.5 (0.87) (1.98° for int/ext rotation.07) 37 .3 (.35) 80.Varus . The angles varied over time with standard deviations ranging from .30) (3.55°.39) 23.45) (0. A typical time series plot of the joint angles is shown in Figure 14.0) (0.78) Subject 1 With Kneepads + Valgus + Int rot .
34) 2.32) (1.9 (.27) 35.48) (2.8 4 .1 5.4) 33.7 . Subject 2 Near 90° Flexion One Knee Near Full Flexion Squat Standing Subject 2 With Kneepads + Valgus + Int rot Included .9 (.6 .9 .28 (.55) Without Kneepads + Valgus + Int rot Included .9 .Ext rot 84.3 .81 5.40) (.1) (1.5) (1) (.26) (.Varus .09 12.47) (1.05 (.Varus .9 (.7 (.40 4.25) 36.02) Figure 14: Joint angles for Subject 1 kneeling near full flexion without kneepads 38 .7 (.9 9.9 1.5 (.69) (.35 (.6) (.7) (.1 (.1) (1.63 7.31) 40.Table 2: Average knee angles (degrees).31) 37.77) 179.28 (.69) 31.25) (.51) (.35) (.Ext rot 38.46) (.9 1.58 13.14) .
57) 5.65 (.54) (.00 43.67 7.94 14.63 (.61 17.21 (.34) (. Average net forces for Subjects 1 and 2 are shown in Table 3 and Table 4.38) (1.87) (1. Time series of the net forces for Subject 1 kneeling near full flexion without kneepads are shown in Figure 15.07) (.61 41.53 (.74) 16.10) Without Kneepads.95 16. and multiplying by 100% yielding values in percentage of body weight (%BW).41) (1.78) .Medial . Subject 1 Near 90° Flexion One Knee Near Full Flexion Squat Standing Subject 1 With Kneepads.23 27.41) (.50 (.27) (1. Table 3: Average net external forces normalized by body weight.4.57 11.Distal 6.62) (1.80 13.62 13.57 8.70 13.78) (. % BW + Lateral + Anterior + Proximal .78) (. by the body weight in N.3 KNEE FORCES All forces are with respect to the anatomical shank coordinate system and normalized to body weight.78) (1. respectively.70 74. The largest lateral forces were found in Subject 1 when kneeling on one knee and the largest medial forces were shown in Subject 2 when squatting without kneepads.20) 7.Posterior .58 (. by dividing the force in N.23 (.58 39.86 27.Medial . The largest posterior forces were shown when kneeling on one knee in both subjects.08 6.51) 39 .26) 5.14) (.33) 5.Posterior .29 6.Distal 8.68) 2.89 (.86 (.74) (. % BW + Lateral + Anterior + Proximal .33 67.61) (. The largest proximal forces occurred in Subject 1 when squatting and in Subject 2 when kneeling on one knee.56) (.30) (.69) (.
85 (.99 (.19) 6.80 8.74 16.06 (.30 (.14 (.88 28.81 45.18 43.40 12.31) (.29) (.Distal 8.29) (.79) (.39) (.12 (.28) 3.75 (.Distal 1.Table 4: Average net external forces normalized by body weight.21) (.08) (.60) (.22 18.41) .51) 4. Subject 2 Subject 2 With Kneepads + Lateral + Anterior + Proximal .52 11.72) (.61) (.12 42.93 16.80) (.Posterior .42 16.Posterior .16) (.86 (.29) (.24 22.49 19.79 23.Medial .69) Near 90° Flexion One Knee Near Full Flexion Squat Standing Figure 15: Net external forces normalized by body weight for Subject 1 kneeling near full flexion without kneepads 40 .31 55.37) 6.19) (.46) (.Medial .73) Without Kneepads + Lateral + Anterior + Proximal .66 5.51) (.56 15.92) (.28) (.20 (.26) 10.07 53.18) 8.
Kneeling near 90° flexion. and kneeling near full flexion resulted in internal rotation moments. All sagittal moments imposed on the knee due to static kneeling were flexion moments. yielding a value in Nm. All moments were divided by this body weight*ht value. Abduction moments were shown in all postures without kneepads and when kneeling near full flexion and squatting with kneepads. Subject 2 showed internal rotation moments for all postures. yielding moments with units in %BW*Ht. Time series showing the sagittal moment contributions for Subject 1 kneeling near full flexion without kneepads are shown in Figure 16. 41 . Subject 1 showed external rotation moments when kneeling near full flexion without kneepads and when squatting. Adduction moments occurred in Subject 1 for all postures. Average net moments for each posture with and without kneepads are shown in Table 5 and Table 6. kneeling on one knee.4. [41] The body weight in Newtons was multiplied by the height in meters. Adduction moments were created in Subject 2 when kneeling near 90° flexion and kneeling on one knee with kneepads.4 KNEE MOMENTS All moments are in respect to the anatomical shank coordinate system and normalized to a percentage of the body weight*height to reduce the differences in moments due to gender.
23) 5.52 (.21 . Subject 1 Subject 1 With Kneepads Flexion Near 90° Flexion One Knee Near Full Flexion Squat Standing .92 (.01) .34 (.60 (.05) (.81 (.09) .07 (.10) (.32 (.91 .45 .19) 4.16 (.External Rotation .00 (.02) 5.26 (.33 (.27 (.23) 4.06 (. Subject 2 Subject 2 With Kneepads Flexion + Abduction Moment .08 (.31) Adduction .54 (.86 .08) + Internal .01) Near 90° Flexion One Knee Near Full Flexion Squat Standing 42 .04) .93 (3.07) 2.25 (.1) .06) Abduction .16) (.Adduction .05) (.01) 6.03) .57 (.12) .19 (.04) .42 (.28) 1.63 .79 .06) (.46 (.05 (.07 (.06) 6.03) Flexion 3.11) .25 (.08) Without Kneepads + Internal Adduction .04) 1.32 (.17 (.08) .02) .21) 3.03) Flexion .58) 1.05 (.03) (.71 (.14 (.04) .02) Internal Rotation .Table 5: Average net external knee moments normalized by Body weight*Height.03) .33) .17 (.85 (.39 (.04 (.30 (.71 (.03) .04) .12 .12) Without Kneepads Internal Rotation .24 (.38) 1.28 (.69 (.03) .49 (.01) 4.42 (.51 1.03) .01) .76 (.53 (.32 (.External Rotation .03) .11) (.09) 1.08 .01) Table 6: Average net external knee moments normalized by Body weight*Height.05) (.15) 5.19) (.04) .90 (.21 (.
Figure 16: Moment contributions normalized by Bodyweight*Height for Subject 1 kneeling near full flexion without kneepads 43 .
applying a force of 8 %BW. Internal rotation moments were larger with kneepads (11 Nm) than without (3 Nm). 75 %BW. This increase in internal rotation moment was also due to the kneepad changing the location of the COP of the force at the knee from 1.3% BW*Ht.9 cm medial to the KJC.27 %BW was placed on the toes. External rotation moments were higher with kneepads (10 Nm) than without (3. The moments generated by these forces were almost balanced by the weight of the lower leg. When kneeling near 90° flexion about 47 %BW was placed on the right knee and another 0.8 % BW*Ht.87 Nm). 44 .4 cm to 3.54 %BW. kneeling near full flexion. The force in the zdirection moved from 1. . Adduction moments in this posture were lower with kneepads (9 Nm) than without (10 Nm). These forces created a flexion moment which was reduced by the extension moment generated by the weight of the lower leg. The weight at the toes was higher than any other kneeling posture. and squatting.5 cm medial to the KJC.5 SUMMARY SUBJECT 1 Kneeling near 90° flexion showed the smallest flexion moments.4. increasing the abduction moment by nearly 7 Nm.15 Nm). When kneeling near full flexion the weight at the knee was 28 %BW. creating a mean net flexion moment of 0. was placed on the knee and a small amount of weight. When kneeling on one knee a predominant amount of weight. The contact between the thigh and the calf was 18 %BW. The higher rotation moment with kneepads was almost entirely due to the moment created by the force at the knee (9. followed by kneeling on one knee.4 cm to 2. This increase in moment was due to the kneepad changing the COP of the force at the knee. due to the subject sitting on their heels. creating a net flexion moment of 0. This posture had the smallest abduction moments: 3 Nm without kneepads and 1 Nm with kneepads. was placed on the toes. 21 %BW.
and created a thighcalf contact force of 25 %BW. Squatting applied a force of 44 %BW to the foot. The source of this difference was the xcomponent of the force at the foot which increased from 20 N without kneepads to 40 N with kneepads.5 Nm) and without (5. 4. 51 Nm when squatting and the 45 .6 SUMMARY SUBJECT 2 Subject 2 showed high variability in knee forces. and angles between kneepad states. External rotation moments were similar for squatting with (5. with a net flexion moment of 6 %BW*Ht. This difference in transverse moments may also be attributed to the COP of the force at the knee moving from 1. This difference is most likely due to the subjects kneeling posture which may or may not have been directly related to the kneepad.The thighcalf and heelgluteus contact forces along with the weight of the lower leg created extension moments which acted to stabilize the knee joint in this high flexion posture. moments.2 cm medial to the KJC. When not wearing kneepads the smallest flexion moment was 34 Nm when kneeling near full flexion. The flexion moment generated by the weight at the toes and the knees were still sufficient to create a net flexion moment at the knee of 4 %BW*Ht.3 cm to 4. followed by 42 Nm when kneeling near 90° flexion. Adduction moments were larger with kneepads (14 Nm) than without (11 Nm). When wearing kneepads a 5 Nm internal rotation moment was created and without kneepads a 1 Nm external rotation moment was created. Adduction moments were higher with kneepads (16 Nm) than without (10 Nm).9 Nm) kneepads. The thighcalf contact force along with the weight of the lower leg created an extension moment in opposition to the flexion moment created by the ground reaction force at the foot.
62 Nm when squatting. but still higher than that during full flexion (35 Nm). followed by 35 Nm when kneeling near full flexion. followed by kneeling near full flexing. The largest abduction moments were created when squatting. followed by keeling near full flexion.highest flexion moment was 68 Nm when kneeling on one knee. This created thighcalf contact which was not accounted for in the analysis of their data and may have caused an overestimation of the flexion moments. kneeling near full flexion. Some 60 %BW was distributed to the right side with 40 %BW at the right knee and another 20 %BW at the right toes. and kneeling on one knee. This new moment is less than squatting (62 Nm). Internal rotation moments were largest when kneeling near 90° flexion. When wearing kneepads kneeling near 90° flexion had the smallest flexion moment of 7 Nm. and 73 Nm when kneeling on one knee. and squatting. Kneeling on the right knee showed the largest flexion moments because most of their body weight was distributed to their right side. kneeling near 90° flexion. Squatting and kneeling on one knee had the largest flexion moments for both knee pad states. the thighcalf contact moment created in the squatting posture was applied to this posture. Squatting and kneeling near full flexion created adduction moments and kneeling one knee and kneeling near 90° flexion created adduction moments. The left leg may have acted to provide balance and stability. 46 . followed by kneeling on one knee. Internal rotation moments were largest when kneeling near 90° flexion. To determine the effect of thighcalf contact on this moment. When including this contact the flexion moment decreased from 73 Nm to 46 Nm. squatting. These postures were found to have similar flexion angles: 37° for squatting and 38° for kneeling on one knee. and kneeling on one knee.
Subject 1 showed significant differences in frontal moment (p = . and internal/external rotation angle (p = . This difference is not thought to be a result of the kneepad itself. The mean values of the forces. To determine the effect of the kneepads on the forces and moments at the knee. but of Subject 2 not repeating the postures correctly. their included angle was 85°.001). When later instructed to kneel their included angle decreased to 39°. This was expected as Subject 1’s posture did not change much (04. Internal/external rotation angles also varied between postures and these varied joint angles may have largely contributed to the differences in joint forces and moments between postures.021).4. Subject 1 showed a significant difference in sagittal moment (p <. Kneeling near 90° flexion created flexion angles of 83°.001).001).037).05 was used to demonstrate statistical significance. No significant differences were found between kneepad states for Subject 1.000). multivariate ANOVA was performed for each subject.040). posterior/anterior force (p = .7 STATISTICS In this study two subjects were asked to simulate 4 postures with and without kneepads while measurements were recorded.001) between postures. When originally instructed to kneel near 90° flexion. Table 7 and Table 8 show the resulting pvalues from ANOVA where significant values are in bold font. This was also expected due to the differences in joint angles. Subject 2 showed a significant difference in medial/lateral force magnitude between kneepad states (p = . and angles were compared between kneepad states. posterior/anterior force (p <. frontal moment (p = . This 46° increase in knee flexion.001). 47 . moments. compared to the 23° of kneeling near full flexion. included knee angle (p <.3° change in joint angles) between kneepad states. A pvalue of 0. distal/proximal force (p <. along with the 8° changes in valgus angles may have largely contributed to the significant difference in medial/lateral forces.
001 <.982 . and squatting created proximal forces and abduction moments compared to the distal forces and adduction moments created in some other postures. Subject 1 Subject 1 Kneepad Moments Sagittal Frontal Transverse Medial/Lateral Posterior/Anterior Distal/Proximal Included Varus/Valgus Int/Ext Rotation .221 .934 .908 .485 .288 <.984 Posture <.001 Angles Forces 48 .001 <.027) between postures. Table 7: Resulting pvalues for ANOVA.and distal/proximal force (p = .001 .037 .148 .948 . These differences were also expected as kneeling near 90° flexion and kneeling on one knee created increased posterior forces.443 .727 .801 .001 .
021 . Table 9: Resulting pvalues for ANOVA of Kneepad*Posture Interaction.521 Data from Subject 1 and Subject 2 were combined to see if there was a significant difference in forces.970 .153 .805 .795 .065 . Subject 2 Subject 2 Kneepad Moments Sagittal Frontal Transverse Medial/Lateral Posterior/Anterior Distal/Proximal Angles Included Varus/Valgus Int/Ext Rotation . and joint angles due to the interaction of kneepad and posture.776 .692 <.993 .546 .276 .758 .074 .802 .001 . Subjects 1&2 Forces Kneepad*Posture Interaction Moments Sagittal Frontal Transverse Medial/Lateral Posterior/Anterior Distal/Proximal Included Varus/Valgus Int/Ext Rotation 49 .564 . Table 9.964 .929 Angles Forces .392 .387 .339 .903 . No significant differences were found.027 . moments.Table 8: Resulting pvalues for ANOVA.967 .125 Posture .040 .
and included angles were shown to exist. Significant differences in all moments. Table 10.65 <.16 <.06 . Subject comparison pvalue Moments Sagittal Frontal Transverse Medial/Lateral Posterior/Anterior Distal/Proximal Angles Included Varus/Valgus Int/Ext Rotation <.001 <.ANOVA was also performed to determine if significant differences existed between Subject 1 and Subject 2.52 Forces 50 .001 .001 . posterior forces. Table 10: Resulting pvalues for ANOVA.001 <.001 .
External force diagrams for all postures are shown in Figure 16. The motion of the markers was shown to be less than 1. Figures 19 and 20. 51 .5 cm over the course of a squatting trial. and Figure 21. sensitivity analyses were performed. the forces were varied by ± 26 N in 2 N increments. varying the location of the knee joint center and measuring its effect on the knee moments. Although static postures were studied there was some expected motion of the markers due to motion artifact and the subject maintaining balance.1 VARYING MODEL PARAMETERS The mean values of the zcomponents of the moment arms and the ycomponents of the forces were varied to determine their effect on the sagittal knee moments. and squatting are shown in Figure 17. All mean values used in this analysis were from Subject 1 without kneepads. One force plate in the laboratory had a threshold of approximately 3 N. kneeling near full flexion. To account for this error. Another analysis was performed. To include this source of error the moment arms were varied by ± 13 cm in 1 cm increments.5. Figure 18. 5. kneeling on one knee. An analysis of the model parameters was performed for each posture to determine the effect of varying the moment arms and forces on the sagittal moments.0 SENSITIVITY ANALYSIS To determine the major sources of error in the moment calculations. Plots of the varied parameters and resulting moments for kneeling near 90° flexion.
Varied moment arms resulted in sensitivities with units of Nm/cm. Parameters with the least change in moment are shown in italics. Also shown are the percentages of change in sagittal moment per centimeter or per Newton. The pink squares represent the sagittal moments resulting from varying the parameter. Table 10 shows the sensitivity of the sagittal moments to the model parameters. Varied moment arm results in change in moment per centimeter and varied forces results in change in moment per Newton. This illustrates the amount of variability of the sagittal moment that is expected due to the model parameter. The green line connecting the green triangles represents the standard deviation of the model parameter. Parameters creating the largest changes in moments are shown in bold. 52 . In all figures the blue diamond represents the mean value of the model parameter and its corresponding sagittal moment.respectively.
N 5 Sagittal Moment.5 Sagittal Moment. cm 54 0 5 10 15 20 15 12 9 6 3 0 ycomponent of F1 force. Varying the zcomponent of the F1 moment arm by 3 cm changed the sagittal knee moment by . the moment changed by 3. Varying the ycomponent of the low leg weight changed the moment by 1. When F1 was varied by 6 N. cm 0 5 10 15 20 30 35 40 45 ycomponent low leg weight. and the weight of the foot+shank (low leg). all forces are shown in N. 53 . The forces and moment arms were varied for the ground reaction force at the toes (F1).2 Nm. Varying the zcomponent of the COM of the lower leg by 3 cm changed the moment by 1. N Mean Value Varied SD Figure 17: Varied forces and moment arms and corresponding sagittal knee moments for kneeling near 90° flexion. Varying the zcomponent of the F2 moment arm by 3 cm was sufficient to more than triple the magnitude of the sagittal moment and change its interpretation. Nm Sag it t al Mo m en t . changed the sagittal moment by .1 Nm.1 Nm. Nm 0 5 10 15 20 64 62 60 58 56 zcomponent of F1 moment arm. cm 2 0 5 10 15 20 318 315 312 309 306 303 300 ycomponent of F2. Nm 5 Sagittal Moment. Nm 0 5 10 15 20 6 4 2 0 zcomponent of F2 moment arm. the ground reaction force at the right knee (F2). Nm 5 Sagittal Moment. and all moments are not normalized and are shown in Nm. Varying the ycomponent of the F2 force by 6 N.5 Nm.2 Nm. N 5 5 Sagittal Moment. Note that all moment arm values are shown in cm. Nm 0 5 10 15 20 25 23 21 19 17 15 zcomponent of low leg moment arm.
Nm 0 10 20 30 60 58 56 54 52 50 zcomponent of F1 moment arm. the ground reaction force at the right knee (F2).10 SagittalMoment.3 Nm. Varying the ycomponent of F2 by 6 N changed the moment by .2 Nm. Nm 0 10 20 30 6 2 0 4 zcomponent of F2 moment arm. cm 10 Sagittal Moment. cm 2 0 10 20 30 490 485 480 475 470 465 ycomponent of F2. an extension moment was created. 54 . Varying the ycomponent of F1 by 6 N. N 45 Mean Value Varied SD Figure 18: Varied forces and moment arms and corresponding sagittal knee moments for kneeling on the right knee. N 10 0 10 20 30 Sagittal Moment. When this moment arm was increased by 3 cm. Nm 30 33 zcomponent of low leg moment arm. Decreasing the zcomponent of the F2 moment arm by 3 cm more than doubled the sagittal moment. changed the sagittal moment by 3. Varying the zcomponent of the F1 moment arm by 3 cm. cm 36 39 42 ycomponent low leg weight. The forces and moment arms were varied for the ground reaction force at the toes (F1). Varying the zcomponent of the low leg COM by 3 cm and the ycomponent of the low leg weight by 6 N changed the moment by 1.1 Nm. Nm 10 0 10 20 30 25 23 21 19 17 15 Sagittal Moment. N 10 SagittalMoment.5 Nm. Nm Sagittal Moment. changed the sagittal moment by . Nm 10 0 10 20 30 23 20 17 14 11 8 ycomponent of F1. and the weight of the foot+shank (low leg).
and the heelgluteus contact force. cm 58 155 150 145 140 ycomponent of F1.2 Nm.40 Sagittal Moment. Varying the zcomponent of the low leg COM by 3 cm changed the moment by 1. the thighcalf contact force. Nm 45 50 55 60 25 23 21 19 17 15 zcomponent of low leg moment arm. N 135 45 50 55 60 40 Sagittal Moment. Nm 45 50 55 60 30 35 40 ycomponent of low leg weight. and the weight of the foot+shank (low leg).3 N. Varying the ycomponent of the low leg weight by 6 N changed the moment by 1. 55 . Varying the zcomponent of the F2 moment arm by 3 cm changed the moment by 5. Nm 40 45 50 55 60 200 195 190 185 180 ycomponent of F2 .7 Nm. N 45 Mean Value Varied SD Figure 19: Varied forces and moment arms and corresponding sagittal knee moments for kneeling near full flexion.2 Nm. The forces and moment arms were varied for the ground reaction force at the toes (F1). Varying the ycomponent of F2 by 6 N changed the moment by . Varying the ycomponent of F1 by 6 N changed the moment by 3. Nm 45 50 55 60 9 7 5 3 1 zcomponent of F2 moment arm.4 Nm. N 40 40 Sagittal Moment.7 Nm. cm Sagittal Moment. Nm Sagittal Moment. Varying the zcomponent of the F1 moment arm by 3 cm changed the moment by 4. the ground reaction force at the right knee (F2). Nm 40 45 50 55 60 65 64 63 62 61 60 59 zcomponent of F1 moment arm. cm Sagittal Moment.
m Sagittal Moment. 56 . Varying the heelgluteus moment arm by 3 cm changed the moment by 1. N 40 Sagittal Moment. N 60 Mean Value Varied SD Figure 20: Varied forces and moment arms and corresponding sagittal knee moments for kneeling near full flexion. the ground reaction force at the right knee (F2). cm 40 45 50 55 60 110 115 120 125 130 135 y component Thighcalf Force. The forces and moment arms were varied for the ground reaction force at the toes (F1). and the weight of the foot+shank (low leg).8 Nm. Varying the heelgluteus contact force by 6 N changed the moment by 2. Nm 45 50 55 60 48 46 44 42 40 38 z component of Heelgluteus moment arm. continued. Nm 45 50 55 60 20 18 16 14 12 10 8 zcomponent of Thighcalf moment arm. Nm 45 50 55 60 45 50 55 ycomponent of Heelgluteus Force.5 Nm.7 Nm. Varying the zcomponent of the thighcalf moment arm by 3 cm changed the moment by 3. the thighcalf contact force. Varying the thighcalf contact force by 6 N changed the moment by .40 Sagittal Moment. and the heelgluteus contact force.5 Nm. Nm 40 Sagittal Moment.
Nm 60 SagittalMoment. 57 . changed the moment by . and the thighcalf contact force. m 41 90 315 310 305 300 295 290 285 y component of F2. changed the moment by 1. Varying the zcomponent of the F2 moment arm by 3 cm. changed the moment by 9 Nm. Varying the ycomponent of the low leg weight by 6 N.2 Nm. Nm Sagittal Moment. N 60 Sagittal Moment. N Mean Value Varied SD Figure 21: Varied forces and moment arms and corresponding sagittal knee moments for squatting. the weight of the foot+shank (low leg).9 Nm.1 Nm.7 Nm. cm 10 90 160 165 170 175 ycomponent of Thighcalf force . Varying the zcomponent of the thighcalf moment arm by 3 cm. The forces and moment arms were varied for the ground reaction force at the foot (F1). Nm 70 70 80 80 90 18 16 14 12 zcomponent of Thighcalf moment arm. Varying the zcomponent of the low leg COM by 3 cm. Varying the ycomponent of F2 by 6 N. Varying the ycomponent of the thighcalf contact force by 6 N. N 60 Sagittal Moment. Nm 60 Sagittal Moment. changed the moment by 5. Nm 60 70 70 80 80 90 48 47 46 45 44 43 42 z component of F2 moment arm. changed the moment by . changed the moment by 2.60 Sagittal Moment. Nm 70 70 80 80 90 25 19 23 21 17 z component of low leg moment arm.9 Nm. m 15 90 24 26 28 30 32 34 36 38 y component low leg weight.
8 %/N 4.1 Nm/cm 55 %/cm .2 Nm/N 3.44 %/N .2 %/cm .7 Nm/cm 2.2 Nm/N 1.55 Nm/N 10.2 Nm/N .13 %/cm .6 %/N .Table 11: Sensitivity of sagittal moments to changes in model parameters for all postures Varied Parameter F1 moment arm F1 Near 90° Flexion F2 moment arm F2 low leg moment arm low leg weight F1 moment arm F1 F2 moment arm One Knee F2 low leg moment arm low leg weight F2 moment arm F2 Squat low leg moment arm low leg weight Thighcalf moment arm Thighcalf Force Model Sensitivity .31 Nm/cm .55 %/N .2 %/cm .13 %/N .8 Nm/cm 37 %/cm .5 %/N 3.03 Nm/N .02 Nm/N .6 %/N 3 Nm/cm 4.55 Nm/N 4.4 %/cm .44 %/cm .35 Nm/cm 6.39 Nm/cm 3 %/cm .16 Nm/cm 1.4 Nm/N .06 Nm/cm 1.2 %/cm .27 %/N 1.2 %/N 58 .15 Nm/N .
5 Nm/cm 2.8 %/cm .4 %/cm . The sagittal moment was the least sensitive to the thighcalf contact force.38 %/N 1. increasing by 3 Nm for every 1 cm this force was moved distally.1 %/N . In the squatting model. decreasing by .2 Nm/cm 2.8 Nm with every 1 cm this force was moved distally.15 Nm with every 1 N this force 59 .9 %/N In the kneeling near 90° flexion model.27 %/N .9 Nm/cm 3.75 %/cm . the location of the force at the knee (F2) increasing the sagittal knee moment by 3.39 Nm/cm . increasing by. the kneeling on one knee model was the most sensitive to the location of the force at the knee. Similarly to the kneeling near 90 flexion model. increasing by .48 Nm/N .05 Nm/N .03 Nm for every 1 N increase in force.84 %/cm . the sagittal moment was the most sensitive to the location of the force at the foot.61 Nm/N 1.02 Nm with a 1 N increase in force. The sagittal moment was the least sensitive to the force at the knee. The sagittal moment was the least sensitive to the force at the knee.14 Nm/N . F1 moment arm F1 F2 moment arm F2 Near Full Flexion low leg moment arm low leg weight Thighcalf moment arm Thighcalf force Heelgluteus moment arm Heelgluteus force 1. increasing by 4.1 Nm per every centimeter this force was moved distally.2 Nm/N .44 Nm/cm .7 %/N 1.Table 11 (continued).7 %/cm .
the sagittal moment was the most sensitive to the location of the force at the knee increasing by 1. and Table 11.9 Nm for every 1 cm this force was moved distally. kneeling on one knee. Percent change in moments resulting from variation in KJC locations for these postures are shown in Table 8. In the kneeling near full flexion model.05 Nm for every 1 N increase in force. some variation in the placement of these markers was expected. and Figure 25.was increased. increasing by . Figure 24. This variation in marker placement will affect the location of the KJC. 5. Moving the location of the thighcalf contact force 1 cm distally decreased sagittal moments by more than 1 Nm in the kneeling near full flexion and squatting models. Table 10.2 VARYING KNEE JOINT CENTER LOCATION Knee joint center (KJC) locations are dependent on the placement of surface markers by laboratory researchers. respectively. Table 9. The sagittal moment was least sensitive to the force at the knee. Changes in the KJC locations resulted in linear changes in the joint moments. 60 . Although one researcher was responsible for palpating the medial and lateral epicondyles of the femur. A sensitivity analysis was performed on the KJC location to determine the effect of its variation on the knee moments. The KJC was varied by ±3 cm along the anatomical shank axes. Figure 23. respectively. kneeling near full flexion and squatting are shown in Figure 22. Although no moments changed by an order of magnitude. Plots showing varied KJC locations and resulting moments for kneeling near 90° flexion. thereby affecting all moment calculations. some variations were sufficient to change the interpretation of the moments.
3 Nm. and transverse moments. Varying the xcomponent by 3 cm changed the frontal moment by 1.Figure 22: Varied KJC locations and resulting knee moments for kneeling near 90° flexion. Varying the x. Varying the ycomponent by 3 cm changed the sagittal moment by 1. respectively. Varying the xcomponent by 3 cm and varying the ycomponent by 3cm more than doubled the transverse moment and changed the interpretation of this moment. 61 .7 Nm. y. and zcomponents of the KJC location had no effect on the sagittal. frontal. Varying the zcomponent by 3 cm changed the adduction moment by 1.3 Nm. Varying the zcomponent by 3 cm more than doubled the sagittal moment.
59 21. 0.01. 0] [.5 Nm to 4.73 254.60 84.65 68. Adduction moments were most sensitive to the zlocation of the KJC.6 Nm extension.02] [0.00 0. .02.02. 0.00 7. .56 48. .01. When the KJC was moved 3 cm posterior the moment decreased to a 2.87 %/cm 169.83 45.3 Nm when the KJC was moved 3 cm posterior.00 0.3 Nm to 11.00 0.87 169.37 16. 0.01.91 49.01] [0.02. .70 35.00 22.39 53. . 0.29 14. .00 0. 0.6 Nm when the KJC was moved 3 cm lateral.7 Nm to a 14 Nm flexion moment when the KJC was moved 3 cm anterior.Sagittal moments were most sensitive to the zlocation of the KJC when kneeling near 90° flexion.66 68.83 45.00 0.57 Change in Moment.03] Sagittal 0.48 22.03] [0.88 48.48 Transverse 84. 0] [0. 0] [0. 0] [0.59 21.00 0. 0] [.00 0.01.02. 0.39 53. 0.00 0.00 0. This moment increased from a 5.37 0. This moment became an external rotation moment of 5. 0. 0.00 0.00 0.00 0. 0. .74 254. increasing from 2.09 17.00 62 .04 145.88 7.09 0. 0] [0. . 0] [0.52 97.29 14.00 0. % Frontal 16. .1 Nm when the KJC was moved 3 cm medial.46 %/cm 32.70 35.04 145.60 17. Table 12: Percent change in moments due to varying KJC for kneeling near 90 degrees flexion Varied by. . . 0] [0. 0] [0.46 32.00 0.03.91 49. . 0] [. 0. Transverse moments were most sensitive to the xposition of the KJC increasing from 3. 0] [.03. 0.03. m [. and decreasing to .8 Nm when the KJC was moved 3 cm anterior.02] [0.01] [0.52 97.00 0. 0] [.03.
Figure 23: Varied KJC locations and resulting knee moments for kneeling on right knee. Varying the x, y, and zcomponents of the KJC location had no effect on the sagittal, frontal, and transverse moments, respectively. Varying the ycomponent by 3 cm changed the flexion moment by 2.7 Nm. Varying the zcomponent by 3 cm changed the sagittal moment by 13.6 Nm and changed it interpretation. Varying the xcomponent by 3 cm changed the adduction moment by 2.7 Nm. Varying the zcomponent changed the frontal moment by 1.5 Nm. Varying the xcomponent by 3 cm changed the transverse moment by 9.1 Nm. Varying the ycomponent by 3 cm changed the transverse moment by 1.5 Nm.
63
Flexion moments were most sensitive to the zlocation of the KJC when kneeling on one knee. This moment increased from 12.8 Nm to 26 Nm when the KJC was moved 3 cm anterior. When the KJC was moved 3 cm posterior the moment became a .78 Nm extension moment. Adduction moments were most sensitive to the xlocation of the KJC. Increasing from 10.6 Nm to 13.3 Nm when the KJC was moved 3 cm anterior, and decreasing to 7.8 Nm when the KJC was moved 3 cm posterior. Transverse moments were most sensitive to the xposition of the KJC. The internal rotation moment changed from 3.5 Nm to 17.1 Nm when the KJC was moved 3 cm medial. This moment became an external rotation moment of 10.1 Nm when the KJC was moved 3 cm lateral.
Table 13: Percent change in moments due to varying KJC for kneeling on one knee
Varied by, m [.01, 0, 0] [.02, 0, 0] [.03, 0, 0] [.01, 0, 0] [.02, 0, 0] [.03, 0, 0] [0, .01, 0] [0, .02, 0] [0, .03, 0] [0, .01, 0] [0, .02, 0] [0, .03, 0] [0, 0, .01] [0, 0, .02] [0, 0, .03] [0, 0, .01] [0, 0, .02] [0, 0, .03] Sagittal 0.00 0.00 0.00 0.00 0.00 0.00 7.13 14.27 21.40 7.13 14.27 21.40 35.36 70.71 106.07 35.36 70.71 106.07
Change in Moment, % Frontal 8.64 17.28 25.91 8.64 17.28 25.91 0.00 0.00 0.00 0.00 0.00 0.00 4.73 9.46 14.19 4.73 9.46 14.19 Transverse 129.27 258.54 387.81 129.27 258.54 387.81 14.28 28.56 42.84 14.28 28.56 42.84 0.00 0.00 0.00 0.00 0.00 0.00
64
Figure 24: Varied KJC locations and resulting knee moments for kneeling near full flexion. Varying the x, y, and zcomponents of the KJC location had no effect on the sagittal, frontal, and transverse moments, respectively. Varying the ycomponent by 3 cm changed the flexion moment by 1.6 Nm. Varying the zcomponent by 3 cm changed the flexion moment by 3.6 Nm. Varying the xcomponent by 3 cm changed the adduction moment by 1.6 Nm. Varying the zcomponent by 3 cm changed the adduction moment by 1.1 Nm. Varying the xcomponent by 3 cm changed the transverse moment by 3.6 Nm and changed its interpretation. Varying the ycomponent by 3 cm changed the transverse moment by 1.4 Nm and changed its interpretation.
65
00 3.03. . .00 0.02] [0. 0.28 Transverse 184. 0] [. increasing from 11.03.60 58. . 0] [0.03. 0] [0.00 2.53 369.93 Nm when the KJC was moved 3 cm lateral. .00 0.02.00 0.02.67 9. .35 14.07 553. 0.91 Change in Moment.85 10. Table 14: Percent change in moments due to varying KJC for kneeling near full flexion Varied by.59 117.01] [0.00 0.61 6.00 3. 0] [0.03. .00 0.00 0. . 0] [. .19 175.00 0.00 0.00 66 . 0.02] [0. 0.6 Nm when the KJC was moved 3 cm lateral. 0.85 10. 0.31 4. 0] [0.01.Flexion moments were most sensitive to the zlocation of the KJC when kneeling near full flexion.66 9.5Nm when the KJC was moved 3 cm medial.2 Nm when the KJC was moved 3 cm medial. 0.00 0.00 0.99 1.55 369.02. 0] [.43 6. 0] [0.61 6. This moment increased from 51.2 Nm when the KJC was moved 3 cm posterior.01] [0.65 Nm to 4.06 553.76 58. Transverse moments were most sensitive to the xposition of the KJC. . . .00 0.02. This moment became an internal rotation moment of 2.17 175.00 0.75 Nm to 55 Nm when the KJC was moved 3 cm anterior and decreased to 48.03] Sagittal 0.00 0.77 0. 0.00 0.30 4.28 3. % Frontal 4.99 1. 0.01. 0] [0.43 6. and decreasing to 9. .59 184. 0] [.01. Adduction moments were most sensitive to the xlocation of the KJC.02 4. 0.02 0.00 0.1 Nm to 12.92 2.01.03] [0. 0] [. The external rotation moment increased from .99 2.00 2. 0. 0] [0. m [.00 0.60 117.35 14. 0.
67 . Varying the ycomponent by 3 cm changed the flexion moment 5. and zcomponents of the KJC location had no effect on the sagittal. frontal.6 Nm. Varying the x.6 Nm.6 Nm.6 Nm. Varying the xcomponent by 3 cm changed the adduction moment by 5. Varying the xcomponent by 3 cm changed the transverse moment by 3 Nm. respectively. Varying the zcomponent by 3 cm changed the flexion moment by 3 Nm. and transverse moments. Varying the ycomponent by 3 cm changed the external rotational moment by .Figure 25: Varied KJC locations and resulting knee moments for squatting. Varying the zcomponent by 3 cm changed the adduction moment by 0. y.
00 0. and decreasing to 3. .02.35 5.00 2. 0.00 0. m [. Transverse moments were most sensitive to the xposition of the KJC.8 Nm when the KJC was moved 3 cm medial and decreased to .2 Nm to 76.82 19. 0.00 0. 0.61 39. When the KJC was moved 3 cm inferior the moment decreased to 65. .32 15. This moment increased from 71.82 4.20 Transverse 26.12 52.22 Change in Moment.20 2. .9 Nm when the KJC was moved 3 cm superior. . 0. .92 2.28 7.06 4.00 0.82 4.00 0.03. 0] [.00 68 .00 0. 0.2 Nm when the KJC was moved 3 cm medial.00 0. 0] [.00 2.00 0.03. . 0] [.23 78.00 0. 0] [0.6 Nm.02.Flexion moments were most sensitive to the ylocation of the KJC when squatting.48 5.64 5.61 39. The external rotation moment increased from 3.00 0. .01. 0] [0. . Increasing from 9.8 Nm when the KJC was moved 3 cm lateral.41 2.01] [0.01.00 0. 0. 0. 0.03.28 7.32 15. . 0] [0.00 0. .22 58.13 6. 0. 0] [.03] [0.16 10.07 4.02.22 1.12 52. 0. 0. Table 15: Percent change in moments due to varying KJC for squatting Varied by.16 10.00 0.01] [0. % Frontal 19.01.00 0.02] [0.01. 0] [0.35 26.64 5.82 0. Adduction moments were most sensitive to the xlocation of the KJC.41 2.23 78.92 1.59 Nm to 15.8 Nm to 6.03] Sagittal 0. 0] [0. 0] [0.13 6. 0] [0. . 0] [.48 0.03.9 Nm when the KJC was moved 3 cm lateral. .21 58.02.02] [0.00 0. 0.
4 %/cm Transverse 2.5 %/cm no change .20 Nm/cm 5.7 %/cm no change .4 %/cm no change 1 Nm/cm 1 %/cm 1.2 Nm/cm 184.4 %/cm no change 1.5 %/cm .8 %/cm .8 Nm/cm 48.6 %/cm no change 69 .92 Nm/cm 7.42 Nm/cm 7. anterior.3 %/cm .1 %/cm .6 %/cm no change .5 Nm/cm 14.3 %/cm no change 1 Nm/cm 26.9 Nm/cm 2.7 %/cm 1.5 Nm/cm 129.58 Nm/cm 17. Greatest changes in moments are shown in bold.3 %/cm 2.6 %/cm no change .89 Nm/cm 19.5 %/cm no change .5 Nm/cm 35.38 Nm/cm 3.2 Nm/cm 2.2 Nm/cm 2.52 Nm/cm 4.38 Nm/cm 58.9 Nm/cm 8.6 %/cm 1 Nm/cm 1.1 %/cm .8 Nm/cm 84. and proximal directions.2 %/cm no change 1. The largest sensitivities where shown in the transverse moments due to moving the KJC 1 cm laterally in all postures.1 %/cm 4. Table 16: Sensitivity of moments to KJC location for all postures Sensitivity (Nm/cm & %/cm) of Moments Posture Varied direction Sagittal lateral Kneeling Near 90° Flexion anterior proximal lateral Kneeling on One Knee anterior proximal lateral Squatting anterior proximal lateral Kneeling Near Full Flexion anterior proximal no change .42 Nm/cm 16.3 %/cm Frontal .5 Nm/cm 4. Values are shown due to varying KJC locations by 1 cm in the lateral.58 Nm/cm 22.Summaries of model sensitivities to KJC locations for all postures are shown in Table 16.7 %/cm no change 4.9 %/cm . Sensitivity values are shown in Nm changes per cm and percentage of moment change per cm.
due to the complexity of using segment coordinate systems. The result was point [0 0 0]. thereby validating the KJC as the origin of the ASCS. and 68°. thereby validating the model. These values were also comparable to published research. the model was validated by several means. Hemmerich reported flexion. and adduction angles between of 153157°. and 710°. respectively for squatting and 144155°. Instead the force and moment calculations were validated by determining the forces and moments during quiet standing and comparing these values to published research. 1112°. external rotation. no direct comparisons could be made to force and moment values reported in literature. a mean adduction moment of .27 Nm/kg. [44] Although flexion angles agree with those in their research. 911°. Due to the lack of research on kneeling knee forces and moments. To ensure accuracy of the anatomical coordinate system calculations. 43] Joint angles were compared to values reported in published research. respectively for kneeling near full flexion.0 MODEL VALIDATION The developed computational model is based on techniques which have been previously used to determine the 3D joint moments in gait analysis. However.04 Nm/kg applied to the knee joint.15 Nm/kg. The KJC location in the GCS was multiplied by [TGAshank]1. the adduction and 70 . the origin of the anatomical coordinate system was verified to equal to the KJC.6. [42. and a mean internal rotational moment of . Quiet standing produced a mean flexion moment of .
but did kneel with their buttocks over their heels.6° valgus and . Knee rotations also agree with the 11. This may have resulted in increased abduction moments.external rotation angles are somewhat larger than those in this study. In this study.42. subjects performed kneeling in full flexion without placing any weight on their heels and in a more upright posture. [45] The knee joint angles when standing were also compared to published research. The right knee had an included angle of 174179° with 2. Subject 1 performed kneeling near full flexion with their buttocks resting on their heels.8° of internal rotation. In Hemmerich’s study.1 ± 6. and internal rotation in Subjects 1 and 2 of this study.7° of internal rotation found in passive knee flexion to 150°. Also the orientation of the feet may have created external rotation in their subjects. Subject 2 did not rest their buttocks on their heels. This is mostly due to the differences in kneeling postures. [46] 71 .1. which is comparable to published research.
When wearing kneepads this moment arm changed to [.1 KNEEPADS Kneepads were not expected to significantly affect the knee joint angles and no statistically significant differences in joint angles were found within subjects.014. contributing 2. Kneepads were also not expected to significantly affect the forces at the knee. The knee pad was expected to change the location of the center of pressure at the knee.036.14 Nm to the internal rotational moment of the knee. Also the limitations of this study which may restrict the extent to which these findings may be generalized are discussed.04.0 DISCUSSION This discussion section includes the sensitivity of the computational model to the model parameters and knee joint center location which are important factors when applying this model to future research.021].025]. .7. 72 . However. which would change the moment arm of that force. . they were expected to affect the moments at the knee joint. In Subject 1 when kneeling near 90° flexion without kneepads. and the advancements are explained. subject variability. thereby affecting the moment. The effect of kneepads on the force and moments. . . the moment arm of the groundreaction force at the knee was [. the implications of findings.031. Although no statistically significant differences were found between kneepad states. there were differences in the transverse moments which may have been caused by the kneepads. 7.
2 MODEL SENSITIVITY 7.73] N and [51. This magnitude of torque may introduce significant changes in the stresses and strains experienced by the internal stabilizing structures of the knee. The design of the kneepad (articulating with a hard.contributing 9.2 mm medially when wearing kneepads. 308. however some variations were sufficient to change the interpretation of the moment from flexion to extension.2. the model was most sensitive to the moment arm of the 73 .07 Nm of the 7.1 Model Parameters The developed computational model showed variations in sagittal moments as a result of changes in moment arms and forces. 327. However. 47.87. No variation in moment arm or force changed the moment by an order of magnitude.86 Nm. When analyzing the kneeling on one knee and kneeling near 90° flexion postures.25. Forces were also different between kneepad states [66. The other 7.52. 40.91.72 Nm increase in internal rotational moment. contoured outer shell) is thought to have contributed to this increase in transverse moment. Similar changes in moments were seen in the kneeling on one knee and kneeling near full flexion postures for this subject. the model was most sensitive to the location of the ground reaction force at the knee. In kneeling near full flexion and squatting. 7. respectively.65 Nm were due to the proximal force and its medial distance from the KJC. doubling or tripling the moment when varying the location by 3 cm.90] N without and with kneepads. This implies that the COP of the groundreaction force at the knee was shifted 2. the differences in lateral forces and distal distances accounted for merely .
especially the moment arms of the thighcalf contact force and the COP of the force at the knee to reduce errors associated with these model parameters. 74 . One subject was responsible for palpating and marking these anatomic landmarks.2. All efforts should be made to ensure proper positioning of all moment arms.thighcalf contact force. Sagittal moments changed from flexion to extension when the zcomponent of the COP at the knee was increased by 3 cm for kneeling near 90° flexion and kneeling on one knee. In this study the location of the medial and lateral epicondyle was marked while the subject sat with their knee at 90°. However the error in marker placement is expected to be less than the 3 cm variations studied. Adduction moments were increased when the KJC was moved laterally or proximally. It is possible that changes in marker placement may create error in moment calculations and joint angles. in some cases doubling the knee moment due to varying the COP by 3 cm.2 Knee Joint Center Location Variations in knee joint center location had large effects on the moment calculations and varied between postures. Due to the high sensitivity of moments to KJC location care must be taken to ensure proper placement of surface markers on the medial and lateral epicondyles. 7. Internal rotational moments were increased when the KJC was moved laterally or posteriorly. Flexion moments were decreased when the KJC was moved posteriorly and increased when the KJC was moved proximally.
When kneeling near 90° flexion both subjects had similar forces and moments with kneepads. This caused increased ground reaction forces at the foot and decreased ground reaction forces at the knee leading to increased flexion moments in Subject 2. Subject 2 was in a more flexed posture creating increased flexion moments. Without kneepads. When kneeling on one knee Subject 1 was in a more upright posture. Subject 2 was in a much more crouched posture with joint angles closer to that in full flexion and squatting. Subject 2 did not sit on their heel and kept their feet in a vertical position with minimal rotation. Subject 1 and Subject 2 showed significantly different joint angles. A comparison of the two subjects yields some insights into potential intersubject variability.7. Subject 1 had a wider stance which may have accounted for the adduction moments created compared to the abduction moments in Subject 2. there did appear to be differences between them. When kneeling near full flexion. This accounted for the decreased varus angles and frontal and transverse moments of Subject 2. Inclusion of more subjects will be needed to determine any anthropometric effects on the results with and without kneepads. Subject 1 sat on their heels with their feet rotated laterally. 75 . and moments during kneeling. kneeling close to 90° with the right leg and the supporting left leg. forces.3 SUBJECT VARIABILITY While the model was only tested on two subjects. When squatting. This variability may be due to their difference in stature as well as kneeling postures.
9 and 1. the distribution of mass in any body segment is not uniform nor is it concentrated at one location.4 LIMITATIONS The limitations of the developed model are associated with the complexity of kneeling near the end range of flexion. A Tekscan ClinSeat® system was used to measure the thighcalf and heelgluteus contact forces. Therefore the relative motion of these markers with respect to the anatomical landmarks is not expected to introduce error in the knee joint center estimation. The use of this type of system. a system which uses resistive technology. [47] Markers were placed on the medial and lateral epicondyles of the femur while the subject was in a standing Tpose for 5 seconds.8 mm for Subject 1. Motion of these markers was minimal with standard deviations between . as degenerative cartilage increases the friction at the articulating surfaces. as well as with the use of inverse dynamics. These markers were then removed and recreated in data analysis. The external applied forces are not applied directly to the tibia. The patellar tendon and tibial tubercle may change the center of pressure location and magnitudes of forces applied to the tibia during kneeling thereby affecting the force and moment calculations. also known as soft tissue artifact. has been found to introduce creep in the measurement of static forces varying the force measurements by 10 % 76 . Also. thereby causing forces which may not be linearly related to the resulting moments due to tissue deformation. The use of reflective markers to track the motion of palpable landmarks may have introduced some sources of error. Motion of the skin to which reflective markers are attached. strongly affects the estimation of joint angles characterized by small range of motion. The assumptions necessary to use the computational model may not be valid in all circumstances. The state of the joint cartilage may null the assumption of a frictionless joint.7.
Thighcalf contact creates a pressure distribution which has higher contact pressures closer to the popliteal region. Along with the issues associated with measuring the contact forces. Heelgluteus contact creates a pressure distribution with pressures highest in the middle of the contact area. However the lack of pressure measurements in these areas decreased the measured total force. however these contact forces are more complex. This decrease in force may have been significant to increase the calculated flexion moments. One researcher calibrated the system by sitting on the sensor atop a table with their left leg on the portion of the sensor that is placed under the subject’s lower leg. It is possible that the values reported from this system are inaccurate due to the method of calibration. Future studies on thighcalf contact should use measurement tools capable of quantifying the entire contact area and measuring 77 . thereby creating an overestimation in sagittal moments. [48] Also calibration of this system may reduce the accuracy and repeatability of its measurements. the system used to measure these contact force was only capable of measuring axial pressure. This distance was accounted for when determining the center of pressures of these contact forces. the full contact surface could not be measured. Both contact pressures create tissue deformation which may act to distribute the axial load in multiple directions. Therefore shear forces resulting from this contact were neglected from the model. Also due to the design of the pressure sensor. The active sensing units on the ClinSeat® system are located approximately 1.5 cm away from the top of the sensor. thereby calibrating the sensor to bodyweight. The thighcalf and heelgluteus contact forces were represented as single resultant forces. there are limitations associated with the representation of these forces in the model. However. Every attempt was made to ensure repeatability of the calibrated values. [49] The recommended calibration is performed by applying a constant pressure to the sensor by sitting on the sensor.to +20 %.
Future studies on knee biomechanics in which imaging techniques can be engaged should account for this change in joint center location. when modeling this contact tissue deformation should be accounted for to reduce error in frontal and transverse moments. In the medial compartment there is little to no movement until 120° of flexion when there is a posterior displacement of 9 mm. it neglects the possibility that similar contact may be present in the other postures. prior to collection of motion capture and force data. [51] This movement accounts for the increase in varus/valgus and internal/external rotations at higher knee flexion. Since subjects were not given specific instructions on kneeling postures. imaging techniques were not employed which would have been necessary to determine the location of the knee joint center for all joint postures studied. These angles introduced thighcalf contact which was not measured nor accounted for in the computational model.forces in multiple axes. Subject 2 had included angles similar to those during their squatting postures. there are limitations associated with the KJC location used in this model. When kneeling on one knee and kneeling near 90° flexion without kneepads. Although this was necessary to determine the moments for these postures. In the lateral compartment the femur moves 20 mm posterior from 0 to 120° of flexion and an additional 10 mm when kneeling into a deep squat. Thighcalf and heelgluteus contact force measurements were made for kneeling near full flexion and squatting. In this study. Also. Neglecting this contact resulted in 78 . it is possible that they may assume the kneeling near 90° flexion and kneeling on one knee posture with higher flexion angles than anticipated. Knee joint center location estimates have been shown to affect the interpretation of joint moments in gait studies with small moment magnitudes. [50] Although the sagittal joint moments in this model are of a higher magnitude. Femoral “rollback” causes the KJC to move during knee flexion.
meniscus. concrete conclusions cannot be drawn on the forces and moments created in restricted postures. Internal and external rotational moments will increase the loading to the medial and lateral compartments of the knee. Adduction moments open the medial joint space. the goal of this master’s thesis project was the development of the model to determine the external forces and moments. In the future this model will be applied to data from a much larger dataset and resulting forces and moments will be published. respectively. stressing the medial soft tissues and applying increased loads to the lateral compartment of the tibia. 79 .overestimations of the flexion moments for these postures. increasing the loads to the posterior tibia. Due to the small sample size. increasing stresses in lateral tissues and loads transmitted to the medial compartment.5 IMPLICATIONS OF FINDINGS Kneeling and squatting create tibial loading conditions which differ substantially from standing. The small sample size creates a very small power and statistics may not have been appropriate to show significant differences between postures or subjects. Increased flexion moments open the anterior aspects of the joint. 7. These complex loading conditions were created in the kneeling postures in this study. N=2. Abduction moments open the lateral joint space. and not to characterize the forces and moments seen in kneeling and squatting. The implications of these increased joint torques on the muscles. In future studies it may be necessary to instruct subjects to assume very specific postures such that all contact forces may be accounted for and comparisons may be made between kneepads states. Although data from two subjects were presented. ligaments. and articular cartilage are presented.
Internal rotation recruits the anterior bundles of the anterior cruciate ligament (ACL) along with the anterior and posterior bundles of the posterior cruciate ligament (PCL) with primary stability from the posterior 80 . EMG data of the quadriceps and hamstrings muscles were collected.2 Ligament Recruitment The passive stability due to the ligaments and soft tissues will depend on knee orientation and loading. Studies of passive knee motion reveal interesting findings of the recruitment patterns of the passive knee stabilizers when kneeling near 90° flexion. Results showed minimal activity.1 Muscles Activity To resist the externally applied forces and moments and reduce tibial translations. The transference of the results of this model to the FEM model being developed as part of the larger effort within the project will give an understanding of the magnitude of these passive tissue loads. However. it is assumed that the quadriceps muscle group is responsible for resisting the externally applied flexion moments. This implies that the knee stability is mostly achieved by the passive stabilizers. Active knee stabilizers include the muscles and tendons and passive stabilizers include the ligaments and soft tissues. the implications of this study are that the high moments and forces calculated in the model may be transferring to the passive stabilizers of the knee. Thus. although not included in the results. active and passive internal knee stabilizers are used.5.7. in these static postures. < 5% MVC. When considering the forces applied to the knee. 7.5. In this study muscle activity was not used as a model input and it is expected that the results would change significantly if muscle activity were included in the force and moment calculations.
To restrict tibial translations and rotations when in these postures. the anterior bundles of the PCL and MCL. The application of a posterior force recruits. It is believed that the posterior soft tissues of the knee (posterior horns of the meniscus. hamstrings muscles. [56] If findings from these insitu studies are correlated to the forces and moments from this study. fat. when kneeling near 90° flexion the MCL is the primary stabilizer with secondary restraint from the PCL. 57] However.5 N when subjected to a 10 Nm adduction moment at 90° flexion. A maximum 500 N posterior force. [53. in high flexion (above 120°) the PCL does not contribute substantially to knee stability. posterior capsule. it is not expected that kneeling postures requiring approximately 90° of knee flexion (kneeling near 90° flexion and kneeling on one knee) will cause detriment to the ligaments. and 11 Nm external rotational moment were applied to the tibia when kneeling in these postures. 10 Nm adduction moment. [52. [52] Although adduction moments are generally stabilized primarily by the ACL and the MCL. 54] Kneeling near 90° flexion and kneeling on one knee created posterior forces along with adduction and internal rotational moments in both subjects.bundles of the PCL. [58] 81 . Fukuda and colleagues found an insitu ACL force of 19. the PCL is expected to be the primary stabilizer with assistance from the ACL and MCL. and skin) not the PCL provide stability in high flexion. [55] Carlin and colleagues found an insitu PCL force of 95 N when subjected to a 100 N posterior force at 90° flexion. The stability of the knee in the fully flexed postures of kneeling near full flexion and squatting is more complicated. In vivo studies of weight bearing knees in high flexion show that up to 120° the PCL plays a major role in providing knee stability from posterior translation.
This translation is also accompanied by the posterior translation of the meniscus with the lateral horn translating more than the medial horn. [60] Compared to extension. adduction moments were created when kneeling near full flexion and squatting.5. The adduction moments which were as much as 16 Nm in conjunction with the posterior forces of up to 120 N. [61] In Subject 1. the contact between the tibia and the femur occur at the posterior aspects of the knee. These loads are then transferred to the articular cartilage with 82 . The moments which were up to 21 Nm in combination with 160 N posterior forces. In the kneeling near 90° flexion and kneeling on one knee postures. [59] This increases the load on the medial meniscsus. increased the shear load to the meniscus.3 Meniscal Loading During walking 70% of the total knee load is distributed to the medial compartment of the knee joint. decreasing the contact area resulting in increased stresses in the posterior meniscus. increase the shear loads transmitted to the meniscus.7. Along with knee flexion playing a major role in the loads and contact areas of the meniscus. However. abduction moments increased the loading to the medial meniscus. the contact area between the tibia and the meniscus is thought to be sufficient to withstand the forces created under these conditions without damage. the load increases. In Subject 2. in high flexion. When an abduction moment is added to the tibia. This posterior translation functions to increase the contact area between the tibia and the meniscus and may play a crucial role in distributing compressive loads in full flexion. as in the kneeling near full flexion and squatting postures. adduction and abduction moments contribute to the loading of the meniscus. there is posterior translation of the tibia. with increasing joint flexion. Adduction and abduction moments increase loading to the lateral and medial compartments of the knee. The adduction moments in postures with high knee flexion may cause detrimental effects on the lateral meniscus.
and principal strain have been suggested as mechanisms of articular cartilage damage. 32] Stair climbing produces forces and moments greater than that during 83 . [63] Excessive shear stress. Type 1 articular cartilage damage is characterized by damage without disruption of the underlying bone or calcified cartilage layer. and sitting crossedlegged. the implications are that the high moments and forces associated with kneeling near full flexion and squatting may cause excessive stress on the articular cartilage of the tibiofemoral joint thereby increasing the risk for knee osteoarthritis.maximal shear stress occurring at the cartilagebone interface away from the center of contact.65] Thambyah and colleagues found articular cartilage stresses of 14 MPa during gait and these stresses increased by more than 80% when in deep flexion. sideknee bending. stair climbing. [17. 7. [66] Thus. 7. [62] This creates areas of high stress on the articular cartilage and when sustained for prolonged periods could lead to the deterioration of the meniscus and articular cartilage damage. kneeling. This increased stress reached the damage limit of the cartilage and was thought to increase the risk of articular cartilage damage. Two forms of articular cartilage damage occur from loading.5.4 Osteoarthritis Progression Articular cartilage damage may occur as a result of biological and mechanical factors.5 Postures Associated with Osteoarthritis Epidemiologic studies have determined postures associated with increased prevalence of knee osteoarthritis including squatting.5. Type 2 damage is characterized by bone fracture with or without damage to the overlying cartilage. tensile stress. [63 .
68] Stair descent creates flexion. [68] In this study all postures created frontal moments and posterior forces greater than standing.6 ADVANCEMENTS The developed computational model accounts for the contact between the heel and the gluteal muscles which until now. In this study the heelgluteus contact force was 54 N which was almost half of the 124 N thighcalf contact force. varus/valgus. 7. Early interventions to reduce the stresses and strains at the knee joint 84 . Thighcalf contact has been shown to significantly affect the forces in the quadriceps and it is expected that the heelgluteus contact will have similar effects on muscle force estimations. increased posterior force. and high knee flexion. If a force of 124 N is significant to effect muscle force estimations than an additional force of 54 N should also be significant to affect these forces. it is expected that postures with higher joint flexion will. Although when using this model this may be a limitation. Although kneeling with joint flexion close to 90° is not expected to cause damage to the articular cartilage. create high loads on the medial and lateral posterior compartments of the knee and may create the biomechanical stresses necessary to initiate the developments and progression of medial and lateral compartment knee osteoarthritis. The combination of increased frontal moments. Results of the sensitivity analysis revealed that the location of the thighcalf contact has a large effect on the knee moments when kneeling near full flexion and squatting.walking. [67. had been neglected from other models. when considering the design of interventions this may be an important factor. making it a more stressful activity. and external rotational moments greater than level walking or stair ascent.
i. Newer devices should consider maximizing the thighcalf and heelgluteus contact forces in hopes of distributing more body weight along the lower leg. It was also found that the kneepad acted to change the COP of the force at the knee by as much as 2. slightly changing postures or the distribution of body weight may be useful.e.have focused on the outside of the knee. It therefore becomes important to consider the effect of new devices on the transverse and frontal moments of the knee which are often neglected due to their smaller magnitudes when compared to sagittal moments. knee pads. However.9 cm in the medial/lateral direction which created an 8 Nm increase in transverse moments. thereby creating an extensor moment to stabilize the knee joint in full flexion. 85 . This increase in rotational moment may be detrimental to the internal knee structures.
and crawling. patellar tendon. and stooping. will allow development of better solutions to the complex knee problems facing many lowseam coal miners. force data. soft outer shell) was used. 86 . crawling. Gaining a better understanding of the pressures. Subjects were also instrumented with a custom designed pressure sensor on their right knee under their kneepad.0 FUTURE WORK In the near future the forces and moments determined using this model will be used as inputs into a finite element model of the knee. Also a second commonly used kneepad with a different design (nonarticulating. and tibial tubercle with and without kneepads. stooping. and moments at the knee joint and in the internal stabilizing structures. forces. electromyography and knee surface pressure data were collected as part of this project. In addition to motion capture. Electromyography was used to determine which muscles were active during different stages of kneeling. The endproduct of this project will be the design of a new intervention which will be effective in reducing the stress and strains in the internal knee structures caused by frequent and prolonged kneeling. This pressure sensor was used to measure the pressure distributions at the patella. and thighcalf and heelgluteus contact measurements.8.
data from two subjects were presented in this paper. up to 500N. There were also large variations in the moments between postures and kneepads states.9. However. In some cases the interpretation of the joint moment changed due to varying the KJC by 3 cm. This model was custom fitted to each subject and based off their individual anthropometry. Also ankle positioning appeared to play a large role in the increased frontal moments associated with kneeling. and transverse moments in all postures. ground reaction forces. Postures requiring larger flexion angles generally created larger flexion moments at the knee. One subject internally rotated their ankles when kneeling which caused large knee adduction moments. up to 16 Nm. This model was developed for use on a much larger dataset currently being collected.0 CONCLUSION In this work a 3D computational model based on inverse dynamics and segment coordinate systems was developed in MATLAB to determine the joint angles and net force and moments applied to the knee in static kneeling postures. This increased adduction moment in combination with increased posterior forces. These subjects revealed that kneeling creates high external knee joint moments which may be as much as 10 to 40 times that during quiet standing. frontal. thighcalf contact force. and heelgluteus contact force measurements. Sensitivity analysis revealed that varying the location of the knee joint center affected the sagittal. increases the forces applied to the medial compartment of the knee and may be significant to create cartilage 87 . The more upright postures of kneeling near 90° flexion and kneeling on one knee created increased posterior forces.
21 Nm. However. A larger dataset will be necessary to determine the effect of these kneeling postures on the internal structures of the knee. this study suggests that kneeling increases the flexion moments applied to the knee as well as the adduction moments and posterior forces and may be significant to initiate meniscal and articular cartilage damage. The other subject had large abduction moments. when squatting which increased loading to the lateral aspects of the meniscus.damage. It was also found that the kneepad used in this study changed the center of pressure of the force at the knee. in some cases increasing the transverse moment by as much as 8 Nm. 88 .
A.m. knee_angle. and joint angles at the knee consists of eight MATLAB® files. moments.1 TRANSFORM_M2A_SHANK_THIGH. perform all necessary calculations and are included in this appendix. The three files: transform_m2a_shank_thigh.M Transform_m2a_shank_thigh.m. 89 .m is used to calculate the transformation matrix from the global coordinate system to the anatomical coordinate system. and forces_moments. The other files are necessary for importing data and exporting results. the transformation matrix from the global coordinate system to the measured coordinate system. Three files are used to calculate the forces.m. and angles. net moments.APPENDIX MATLAB SOFTWARE The MATLAB® software developed to determine the net forces. and the transformation matrix from the measured coordinate system to the anatomical coordinate system.
78:80). left_thigh=anatomical_data(:.93:95). 8/19/08 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Load Anatomical Data cd(results_directory).87:89). Determine the anatomical coordinate system for shank and %%%% calculate the transformation matrix from global to anatomical %%%% coordinates %%%% 3.Shank %%%% 2.102:104). right_shank=anatomical_data(:. left_shank=anatomical_data(:.96:98).mat %% Determine length of data [n o]=size(anatomical_data).81:83). left_knee_medial=anatomical_data(:. right_thigh_rear=anatomical_data(:. Assign data to specific markers right_knee=anatomical_data(:. left_thigh_rear=anatomical_data(:.48:50).33:35). left_shank_rear=anatomical_data(:.60:62). rasis=anatomical_data(:.Thigh %%%% 24 for thigh %%%% 5.function [filename]=transform_m2a(results_directory) %%%%%%%%%%%%%%%%%%%%%%%%%%% transform_m2a. Determine the transformation matrix from measured to anatomical %%%% and get the mean value of this matrix %%%% Part 2 . left_ankle_medial=anatomical_data(:.51:53).45:47). %% 1.36:38). 90 . right_ankle=anatomical_data(:. right_thigh_front=anatomical_data(:.120:122). right_ankle_medial=anatomical_data(:.111:113). Assign data to specific markers %%%% Part 1 .114:116).84:86).42:44).105:107).66:68). left_shank_front=anatomical_data(:. right_knee_medial=anatomical_data(:. right_shank_rear=anatomical_data(:. load anatomical. Save transformation matrix TMAshank to mat file %%%% %%%% Jonisha Pollard. right_shank_front=anatomical_data(:. left_ankle=anatomical_data(:. JPollard@cdc. This function uses %%%% only the anatomical (calibration) data %%%% 1. right_thigh=anatomical_data(:. Determine the measured/testing coordinate system and calculate %%%% the transformation matrix from global to measured %%%% coordinates %%%% 4.63:65).123:125). left_knee=anatomical_data(:. left_thigh_front=anatomical_data(:.gov.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%% %%%% This function is designed to produce the transformation matrix from %%%% testing coordinates to anatomical coordinates. lasis=anatomical_data(:.69:71).
%% Determine location of joint centers in global coordinates AJC=(right_ankle + right_ankle_medial)/2; % Global Right AJC KJC=(right_knee + right_knee_medial)/2; % Global Right KJC AJC_L=(left_ankle + left_ankle_medial)/2; KJC_L=(left_knee + left_knee_medial)/2; % Global Left AJC % Global Left KJC
% Global HJC Calculation. Bells method % Leardini et al. 1999, Bell et al. 1990 HJC_origin=(lasis+rasis)/2; PW=abs(lasis(:,1)rasis(:,1)); % inter Asis distance HJC(:,1)=HJC_origin(:,1)+.36*PW; HJC(:,2)=HJC_origin(:,2).19*PW; HJC(:,3)=HJC_origin(:,3).3*PW; HJC_L(:,1)=HJC_origin(:,1).36*PW; HJC_L(:,2)=HJC_origin(:,2).19*PW; HJC_L(:,3)=HJC_origin(:,3).3*PW;
%%%%%%%%%%%%%%%%%%%%%% Part 1  Shank %%%%%%%%%%%%%%%%%%%%%%%%%% %% 2. Determine anatomical coordinate system for shank and calculate %% transformation matrix from global to anatomical coordinates % All coordinate systems were chosen such that hey align with the % Global coordinate system at Standard anatomical position % % % % Global & Anatomical Coordinate Systems x is medial/lateral y is anterior/posterior z is superior/inferior
% Right for i=1:n r1=KJCAJC; r2=right_knee  right_knee_medial; r3=cross(r1,r2); r4=cross(r3,r1); x(i,:)=r4(i,:)/norm(r4(i,:)); y(i,:)=r3(i,:)/norm(r3(i,:)); z(i,:)=r1(i,:)/norm(r1(i,:)); end TGAshank=zeros(4,4,n); for i=1:n TGAshank(1,:,i)=[1 0 0 0]; TGAshank(2:4,1,i)=KJC(i,:); TGAshank(2:4,2,i)=x(i,:); TGAshank(2:4,3,i)=y(i,:); TGAshank(2:4,4,i)=z(i,:); end
% % % %
local zaxis in xdirection local yaxis local xaxis
% rotation matrix
%right KJC is origin
91
clear r1 r2 r3 r4 x y z
% clear variables
% calculate AJC and KJC in anatomical shank coordinates for i=1:n G_ajc(i,2:4)=AJC(i,:); G_ajc(:,1)=1; G_kjc(i,2:4)=KJC(i,:); G_kjc(:,1)=1; asAJC_all(i,:)=inv(TGAshank(:,:,i))*G_ajc(i,:)'; %Shank AJC asKJC_all(i,:)=inv(TGAshank(:,:,i))*G_kjc(i,:)'; %Shank KJC end asAJC=mean(asAJC_all(:,2:4)); asKJC=mean(asKJC_all(:,2:4));
% verified to equal [0 0 0]
% Left for i=1:n r1=KJC_LAJC_L; r2=left_knee_medial  left_knee; r3=cross(r1,r2); r4=cross(r3,r1); x(i,:)=r4(i,:)/norm(r4(i,:)); y(i,:)=r3(i,:)/norm(r3(i,:)); z(i,:)=r1(i,:)/norm(r1(i,:)); end TGAshank_L=zeros(4,4,n); for i=1:n TGAshank_L(1,:,i)=[1 0 0 0]; TGAshank_L(2:4,1,i)=KJC_L(i,:); TGAshank_L(2:4,2,i)=x(i,:); TGAshank_L(2:4,3,i)=y(i,:); TGAshank_L(2:4,4,i)=z(i,:); end clear r1 r2 r3 r4 x y z
% % % %
local zaxis in xdirection local yaxis local xaxis
% rotation matrix
% left KJC is origin
% clear variables
% calculate AJC and KJC in anatomical shank coordinates for i=1:n G_ajc_L(i,2:4)=AJC_L(i,:); G_ajc_L(:,1)=1; G_kjc_L(i,2:4)=KJC_L(i,:); G_kjc_L(:,1)=1; asAJC_all_L(i,:)=inv(TGAshank_L(:,:,i))*G_ajc_L(i,:)'; %Shank AJC asKJC_all_L(i,:)=inv(TGAshank_L(:,:,i))*G_kjc_L(i,:)'; %Shank KJC end asAJC_L=mean(asAJC_all_L(:,2:4)); asKJC_L=mean(asKJC_all_L(:,2:4)); %% 3. Determine the measured coordinate system for shank and calculate the %% transformation matrix from global to measured %% Global & Anatomical Coordinate Systems
92
%% x is medial/lateral %% y is anterior/posterior %% z is superior/inferior % Right shank for i=1:n r1=right_shankright_shank_front; r2=right_shank_front  right_shank_rear; r3=cross(r2,r1); r4=cross(r1,r3); x(i,:)=r3(i,:)/norm(r3(i,:)); y(i,:)=r4(i,:)/norm(r4(i,:)); z(i,:)=r1(i,:)/norm(r1(i,:)); end
% % % %
local zaxis in y direction local xaxis local yaxis
% rotation matrix
TGMshank=zeros(4,4,n); for i=1:n TGMshank(1,:,i)=[1 0 0 0]; TGMshank(2:4,1,i)=right_shank_front(i,:); TGMshank(2:4,2,i)=x(i,:); TGMshank(2:4,3,i)=y(i,:); TGMshank(2:4,4,i)=z(i,:); end clear r1 r2 r3 r4 x y z % clear variables % Left shank for i=1:n r1=left_shankleft_shank_front; r2=left_shank_front  left_shank_rear; r3=cross(r2,r1); r4=cross(r1,r3); x(i,:)=r3(i,:)/norm(r3(i,:)); y(i,:)=r4(i,:)/norm(r4(i,:)); z(i,:)=r1(i,:)/norm(r1(i,:)); end TGMshank_L=zeros(4,4,n); for i=1:n TGMshank_L(1,:,i)=[1 0 0 0]; TGMshank_L(2:4,1,i)=left_shank_front(i,:); TGMshank_L(2:4,2,i)=x(i,:); TGMshank_L(2:4,3,i)=y(i,:); TGMshank_L(2:4,4,i)=z(i,:); end clear r1 r2 r3 r4 x y z % clear variables
% origin
% % % %
local zaxis in y direction local xaxis local yaxis
% rotation matrix
% origin
%% 4. Calculate transformation matrix from measured to anatomical coords % Right TMAs_alldata=zeros(4,4,n); % TMAshank for all data points for i=1:n TMAs_alldata(:,:,i)=inv(TGMshank(:,:,i))*TGAshank(:,:,i);
93
Thigh %%%%%%%%%%%%%%%%%%%%%%%%%% %% 2.i).:).:.2:4)=0.:).:)/norm(r1(i.1. TMAshank_L(1. TGAthigh(2:4. z(i.i)=inv(TGMshank_L(:.:)).r2).:. end % calculate the mean of TMAshank_L_alldata TMAshank_L=zeros(4. x(i.4.:)=r1(i. r3=cross(r1. for i=1:n TGAthigh(1.:)).i)=HJC(i. TGAthigh(2:4.j.i))*TGAshank_L(:.j)=mean(TMAs_alldata(i.:.right_knee_medial.i)=z(i. % % % % local zaxis in xdirection local yaxis local xaxis % rotation matrix %right HJC is origin 94 . r4=cross(r3.:).2:4)=0.4). y(i.:)).i)=x(i.4). TMAshank_L(1.r1). r2=right_knee .i)=[1 0 0 0].:)=r4(i.:.:)/norm(r3(i. TMAshank(1.4.n).1)=1. end end % Left TMAs_alldata_L=zeros(4.n).:).3.j.4.:)). end end %%%%%%%%%%%%%%%%%%%%%% Part 2 .2. end TGAthigh=zeros(4. TMAshank(1.end % calculate the mean of TMAshank_alldata TMAshank=zeros(4.i)=y(i. TGAthigh(2:4. % TMAshank_L for all data points for i=1:n TMAs_alldata_L(:. for i=2:4 for j=1:4 TMAshank_L(i.:)).:)/norm(r4(i.1)=1.:)=r3(i. for i=2:4 for j=1:4 TMAshank(i.j)=mean(TMAs_alldata_L(i. TGAthigh(2:4. Determine anatomical coordinate system for thigh and calculate %% transformation matrix from global to anatomical coords %% Global & Anatomical Coordinate Systems %% x is medial/lateral %% y is anterior/posterior %% z is superior/inferior % Right Thigh for i=1:n r1=HJCKJC.
r4=cross(r3.:.:).2:4)).:).i))*G_hjc(i. % verified to equal [0 0 0] atKJC=mean(atKJC_all(:.:)/norm(r1(i.:)=inv(TGAthigh_L(:.1)=1. 95 .i))*G_kjc(i.r2).2:4)=HJC_L(i. %Thigh KJC end atHJC_L=mean(atHJC_all_L(:.i)=HJC_L(i.:)).1)=1.:)=inv(TGAthigh(:.2.:)=inv(TGAthigh_L(:.:)'. G_kjc_L(:.n).:).:).:. % Left Thigh for i=1:n r1=HJC_LKJC_L. G_hjc_L(:.:).:). % verified to equal [0 0 0] atKJC_L=mean(atKJC_all_L(:.:.i)=[1 0 0 0]. end clear r1 r2 r3 r4 x y z % % % % local zaxis in xdirection local yaxis local xaxis % rotation matrix %left HJC is origin % clear variables % calculate HJC and KJC in anatomical thigh coordinates for i=1:n G_hjc_L(i. TGAthigh_L(2:4. z(i.i))*G_hjc_L(i. G_kjc_L(i. TGAthigh_L(2:4.2:4)=HJC(i. atHJC_all_L(i.:. y(i.2:4)=KJC_L(i.:)/norm(r4(i.:.r1).:)=inv(TGAthigh(:.:). G_kjc(:.:)/norm(r3(i. r3=cross(r1. G_kjc(i.i)=z(i.:)'.4.i))*G_kjc_L(i. G_hjc(:.2:4)).1)=1.:)=r4(i. r2=left_knee_medial . for i=1:n TGAthigh_L(1. %Thigh KJC end atHJC=mean(atHJC_all(:.left_knee.1)=1.4.end clear r1 r2 r3 r4 x y z % clear variables % calculate HJC and KJC in anatomical thigh coordinates for i=1:n G_hjc(i. %Thigh HJC atKJC_all(i. TGAthigh_L(2:4.:)'.:).:)).:)=r3(i. x(i. TGAthigh_L(2:4.2:4)).1.:)=r1(i.:)'.i)=x(i. end TGAthigh_L=zeros(4.2:4)). atHJC_all(i.i)=y(i.2:4)=KJC(i.:)). %Thigh HJC atKJC_all_L(i.3.
:). end TGMthigh_L=zeros(4. y(i.:)/norm(r3(i.3.:). for i=1:n TGMthigh_L(1.4.4.:)).:). x(i.i)=x(i. for i=1:n TGMthigh(1. z(i. r2=left_thigh_frontleft_thigh_rear.r1). r2=right_thigh_frontright_thigh_rear.4.n). x(i. TGMthigh_L(2:4.3.:.:)).:)=r1(i. TGMthigh(2:4.i)=z(i.:)/norm(r4(i. r3=cross(r2.n).i)=right_thigh_front(i.i)=[1 0 0 0]. TGMthigh_L(2:4. end TGMthigh=zeros(4. z(i.:)=r4(i. r4=cross(r1.:)/norm(r4(i. %origin TGMthigh_L(2:4. TGMthigh(2:4.1.i)=[1 0 0 0].r1).i)=z(i. r4=cross(r1.:)). end clear r1 r2 r3 r4 x y z % Left for i=1:n r1=left_thighleft_thigh_front.:)). Calculate transformation matrix from measured to anatomical coords 96 .:)/norm(r3(i.:).:)=r3(i. TGMthigh(2:4.:)=r4(i.:).4.:)/norm(r1(i.2.:).:).r3).:)).:)/norm(r1(i.:)=r1(i.1.i)=x(i. y(i.r3).:)=r3(i.:)).i)=left_thigh_front(i.%% 3. r3=cross(r2.2. %origin TGMthigh(2:4. end clear r1 r2 r3 r4 x y z % clear variables % % % % local zaxis in y direction local xaxis local yaxis % clear variables % % % % local zaxis in y direction local xaxis local yaxis % rotation matrix % rotation matrix %% 4.:).i)=y(i. Determine the measured coordinate system for thigh and calculate %% transformation from global to measured %% Global & Anatomical Coordinate Systems %% x is medial/lateral %% y is anterior/posterior %% z is superior/inferior % Right for i=1:n r1=right_thighright_thigh_front.:.i)=y(i. TGMthigh_L(2:4.
m 97 . end % calculate the mean of TMAthigh_alldata TMAthigh=zeros(4.:.1)=1. Save transformation matrixes KJC. AJC and HJC mat file save anatomical.% Right TMAt_alldata=zeros(4.mat TMAshank TMAthigh asAJC TMAshank_L TMAthigh_L .. TMAthigh(1.j)=mean(TMAt_alldata_L(i.n).i)=inv(TGMthigh_L(:.i).:.4). end % calculate the mean of TMAthigh_alldata TMAthigh_L=zeros(4. asAJC_L append end % end of function transform_m2a. for i=2:4 for j=1:4 TMAthigh(i. for i=2:4 for j=1:4 TMAthigh_L(i. TMAthigh_L(1.i))*TGAthigh_L(:. TMAthigh_L(1.i))*TGAthigh(:. end end %% 5.i)=inv(TGMthigh(:.j)=mean(TMAt_alldata(i.j.2:4)=0.:))..:)).4. TMAthigh(1. % TMAthigh for all data point for i=1:n TMAt_alldata_L(:.2:4)=0.:.1)=1.4.n).:.4). end end % Left TMAt_alldata_L=zeros(4.j. % TMAthigh for all data point for i=1:n TMAt_alldata(:.:.:.i).
75:77). left_thigh=testing_data(:. 98 . and the angle of internal/external rotation are calculated.2 KNEE ANGLE.M Knee_angle. left_shank=testing_data(:.mat %%%% %%%% Jonisha Pollard. left_thigh_rear=testing_data(:. right_shank_front=testing_data(:. function knee_angle(posture. right_thigh_rear=testing_data(:. 7/23/08 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% cd(results_directory) load(posture) load anatomical %% %load mat file with data for specific postures %load mat file with anatomical markers and TMAshank 1. Load testing marker data and assign data to specific markers %%%% 2.81:83). Save joint angles to *posture*.m is used to calculate the orientation of the knee joint. the angle of varus/valgus. %%%% 1.36:38). Load testing marker data and assign data to specific markers [n a]=size(testing_data). the included joint angle. JPollard@cdc. right_thigh_front=testing_data(:. left_shank_rear=testing_data(:. with a Euler order of X’y’z”. Calculate anatomical coordinate system from measured system and %%%% determine transformation from thigh to shank Tts %%%% 4.gov.33:35). right_shank=testing_data(:.27:29). left_shank_front=testing_data(:.90:92). Determine measured coordinate system for shank and thigh %%%% 3. right_thigh=testing_data(:. results_directory) %%%%%%%%%%%%%%%%%%%%%%%%%%% knee_angle.A.45:47).69:71).m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%% This function is designed to determine the knee angle by calculating %%%% the angle between the thigh coordinate system and the shank coordinate %%%% system.39:41). Using Euler angle decomposition.87:89). right_shank_rear=testing_data(:. Decompose transformation matrix for Euler angles %%%% 5.78:80). left_thigh_front=testing_data(:.48:50).
i)=right_shank_front(i. r2=right_shank_front .:).i)=[1 0 0 0].:)/norm(r1(i. z(i.:. for i=1:n TGMthigh(1.r3).:)=r1(i.:). end TGMthigh=zeros(4. y(i.r1).4.:)/norm(r1(i.:).:)=r4(i. TGMshank(2:4.2.:)).:)/norm(r3(i.1.3. x(i. for i=1:n TGMshank(1. r2=right_thigh_frontright_thigh_rear.i)=x(i. TGMthigh(2:4.:)=r3(i. TGMthigh(2:4.i)=right_thigh_front(i. end clear r1 r2 r3 r4 x y z % Right thigh for i=1:n r1=right_thighright_thigh_front. end clear r1 r2 r3 r4 x y z % Left shank for i=1:n r1=left_shankleft_shank_front.3.:)=r1(i.:).:)=r4(i.:)).2. % % % % local zaxis in y direction local xaxis local yaxis % rotation matrix % origin % local zaxis % in y direction 99 .:)/norm(r4(i.left_shank_rear.i)=x(i. r3=cross(r2. TGMthigh(2:4. y(i. x(i. r2=left_shank_front .:)=r3(i.4.i)=z(i.r3). z(i.i)=[1 0 0 0].:)). end % % % % local zaxis in y direction local xaxis local yaxis % rotation matrix TGMshank=zeros(4.:)/norm(r4(i.r1).i)=y(i.:)/norm(r3(i. r3=cross(r2. Determine measured coordinate system for shank and thigh %% Global Coordinate System %% x is medial/lateral %% y is anterior/posterior %% z is superior/inferior % Right shank for i=1:n r1=right_shankright_shank_front. TGMshank(2:4.right_shank_rear.:)).n). r4=cross(r1.4. TGMthigh(2:4. TGMshank(2:4.i)=z(i.:).:.i)=y(i.:)). r4=cross(r1.%% 2.:).:).4.n).:)).:). % origin TGMshank(2:4.1.
r3=cross(r2,r1); r4=cross(r1,r3); x(i,:)=r3(i,:)/norm(r3(i,:)); y(i,:)=r4(i,:)/norm(r4(i,:)); z(i,:)=r1(i,:)/norm(r1(i,:)); end
% local xaxis % local yaxis % rotation matrix
TGMshank_L=zeros(4,4,n); for i=1:n TGMshank_L(1,:,i)=[1 0 0 0]; TGMshank_L(2:4,1,i)=left_shank_front(i,:); % origin TGMshank_L(2:4,2,i)=x(i,:); TGMshank_L(2:4,3,i)=y(i,:); TGMshank_L(2:4,4,i)=z(i,:); end clear r1 r2 r3 r4 x y z
% Left thigh for i=1:n r1=left_thighleft_thigh_front; r2=left_thigh_frontleft_thigh_rear; r3=cross(r2,r1); r4=cross(r1,r3); x(i,:)=r3(i,:)/norm(r3(i,:)); y(i,:)=r4(i,:)/norm(r4(i,:)); z(i,:)=r1(i,:)/norm(r1(i,:)); end TGMthigh_L=zeros(4,4,n); for i=1:n TGMthigh_L(1,:,i)=[1 0 0 0]; TGMthigh_L(2:4,1,i)=left_thigh_front(i,:); TGMthigh_L(2:4,2,i)=x(i,:); TGMthigh_L(2:4,3,i)=y(i,:); TGMthigh_L(2:4,4,i)=z(i,:); end clear r1 r2 r3 r4 x y z
% % % %
local zaxis in y direction local xaxis local yaxis
% rotation matrix
% origin
%% %%
3. Calculate anatomical coordinate system from measured system and determine transformation from thigh to shank Tts for i=1:n % Right TGAshank(:,:,i)=TGMshank(:,:,i)*TMAshank; TGAthigh(:,:,i)=TGMthigh(:,:,i)*TMAthigh; Tts(:,:,i)=inv(TGAthigh(:,:,i))*TGAshank(:,:,i); Rts(:,:,i)=Tts(2:4,2:4,i); % rotation matrix % Left TGAshank_L(:,:,i)=TGMshank_L(:,:,i)*TMAshank_L; TGAthigh_L(:,:,i)=TGMthigh_L(:,:,i)*TMAthigh_L;
100
Tts_L(:,:,i)=inv(TGAthigh_L(:,:,i))*TGAshank_L(:,:,i); Rts_L(:,:,i)=Tts_L(2:4,2:4,i); % rotation matrix end %% 4. Decompose transformation matrix for Euler angle
% Global Coordinate System % 1st axis  flexion/ext  X % 2nd axis  varus/valgus  Y % 3rd axis  inversion/eversion  Z % Euler order Xy'z''
for i=1:n z_angle(i,:)=((180/pi) * atan2(Rts(1,2,i),Rts(1,1,i))); x_angle(i,:)=180 + ((180/pi) * atan2(Rts(2,3,i),Rts(3,3,i))); y_angle(i,:)=((180/pi) * atan2(cos(z_angle(i,:))*Rts(1,3,i),... Rts(1,1,i))); z_angle_L(i,:)=((180/pi) * atan2(Rts_L(1,2,i),Rts_L(1,1,i))); x_angle_L(i,:)=180 + ((180/pi) * atan2(Rts_L(2,3,i), ... Rts_L(3,3,i))); y_angle_L(i,:)=((180/pi) * atan2(cos(z_angle_L(i,:)) ... *Rts_L(1,3,i),Rts_L(1,1,i))); end %% 5. Save angles save(posture,'x_angle','y_angle','z_angle','Tts','TGAthigh',... 'TGAshank','x_angle_L','y_angle_L','z_angle_L','Tts_L', ... 'TGAthigh_L','TGAshank_L','append') end % end of function knee_angle.m
101
A.3
FORCES_MOMENTS.M
Forces_moments.m is used to calculate the net external forces, moments, and moment contributions at the knee joint in the global coordinate system. The forces and moments are then transformed into the anatomical coordinate system.
function forces_moments(weight, height, posture, results_directory) %%%%%%%%%%%%%%%%%%%%%%%%%%% forces_moments.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%% This function is designed to calculate the net forces and moments at %%%% the knee joint for squatting and full flexion %%%% 1. Load testing marker data and assign data to specific markers %%%% 2. Load force plate data %%%% 3. Determine measured coordinate system for shank %%%% 4. Get anatomical shank coordinates from testing coordinates %%%% 5. Determine COP and weight of segments and joint rotation centers %%%% 6. Calculate the joint reactive forces at the right knee in Global %%%% coordinates %%%% 7. Calculate the moment at the right knee in Global coordinates %%%% 8. Calculate forces and moments in Anatomical coordinates %%%% 9. Normalize forces and moments %%%% 10. Saves forces and moments to *posture*.mat %%%% %%%% Jonisha Pollard, JPollard@cdc.gov, 7/23/08 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% cd(results_directory) load(posture) %load mat file with data for specific postures load anatomical %load mat file with anatomical markers and TMAshank %% 1. Load testing marker data and assign data to specific markers [n a]=size(testing_data); right_shank=testing_data(:,39:41); right_shank_rear=testing_data(:,48:50); right_shank_front=testing_data(:,45:47);
%%
2. Load force plate data F1=forces_data(:,2:4); COP1=forces_data(:,5:7); F2=forces_data(:,9:11); COP2=forces_data(:,12:14);
102
:)=RGAshank(:.:)).3.2.:.:.n). for i=1:n TGMshank(1.:)/norm(r3(i. r4=cross(r1. TGMshank(2:4.:)=TGAshank(:. g_KJC(i.i)*[1 0 0 tc_dist]'.i)=TGAshank(2:4.:. end % origin %% 4. COP_hg(i.:. end 103 .i)*[1 0 0 0]'. TGMshank(2:4. % Thighcalf and Heelgluteus contact forces in global coordinate system gtc_cont(i.i)*[0 hg_cont 0]'. and contact forces in global coordinates for i=1:n % calculate location of AJC and KJC in global coordinates g_AJC(i. Determine COM. COP.:)/norm(r1(i.:).:)).:)=r1(i. ghg_cont(i.:)/norm(r4(i.% KJC is origin of ACS [0 0 0] % calculate thighcalf and heelgluteus COP in global coordinates % these contact forces are located along the z axis of the shank coord COP_tc(i.i)*[1 0 0 hg_dist]'. x(i.:.2:4.:.i)*TMAshank. z(i.4.%% 3. RGAshank(:.:)).i)=TGMshank(:. TGMshank(2:4.:.:)=r3(i.r3).:)=RGAshank(:.:)=TGAshank(:.:). Determine measured coordinate system for shank %% %% %% %% Global Coordinate System x is medial/lateral y is anterior/posterior z is superior/inferior for i=1:n r1=right_shankright_shank_front.:.1.:)=TGAshank(:.:. r2=right_shank_front .4.i)*[0 tc_cont 0]'.i)=z(i. end % % % % local zaxis in y direction local xaxis local yaxis % rotation matrix TGMshank=zeros(4.:). end % transformation matrix % rotation matrix %% %% 5.r1).i)*[1 asAJC]'. TGMshank(2:4. joint rotation centers.:)=r4(i. y(i.i).i)=[1 0 0 0].:). r3=cross(r2.right_shank_rear.i)=right_shank_front(i.i)=y(i.:)=TGAshank(:. Get anatomical shank coordinates from measured coordinates for i=1:n TGAshank(:.:.i)=x(i.
MF1(i. %% 7.5% of distance from proximal to distal COM_lowleg=gAJC+. %eliminate '1' in 1st col gKJC=g_KJC(:. 104 . W_lowleg=zeros(n.:)). %% 6.:)=cross(r_lowleg(i.ghg_cont(i.:). r_tc=(COPtcgKJC)/1000.:)=cross(r_FP1(i.F2(i. COPtc=COP_tc(:. MF2(i. COPhg=COP_hg(:. Mhg(i.2:4). Muscle moment will be () Mknee= M_lowleg + MF1 + MF2 + Mtc + Mhg.525*(gKJC . Mtc(i.:). Calculate the joint reactive forces at the right knee in Global %% Coordinates % Sum of all forces at knee % Sum forces = Force_r_knee % Knee reaction % Force imposed on knee =F1 should equal zero + F1 + F2 + W_lowleg + tc + hg = 0 force = (F1 + F2 + W_lowleg + tc + hg) + F2 + W_lowleg + tc + hg % Force imposed on Knee (reaction force will be ) Fknee= F1 + F2 + W_lowleg + gtc_cont + ghg_cont.2:4).:)=cross(r_FP2(i.:)=cross(r_tc(i.3)=.:). Calculate moment at the right knee joint in Global Coordinates % Sum of all moments at knee should equal zero % Moments will be calculated using the vector cross product % Sum of Moments at knee= Knee_reaction_Moment+ MF1 + MF2 + Mlowleg + Mtc % + Mhg= 0 % Knee_reaction_Moment = (MF1 + MF2 + Mlowleg + Mtc + Mhg) % Moment imposed on knee= MF1 + MF2 + Mlowleg + Mtc + Mhg % moment arm vectors in m r_lowleg=(COM_lowleggKJC)/1000.W_lowleg(i. r_FP2=(COP2gKJC)/1000. % Lower leg COM and weight are modified from Clauser et al 1969 and % Hinrichs 1990 % 47.2:4).2:4).gtc_cont(i.gAJC=g_AJC(:. r_hg=(COPhggKJC)/1000. % calculate moments in Nm for i=1:n M_lowleg(i.:)=cross(r_hg(i.:).:)).:)).gAJC).:)).058*weight. end % Moment imposed on knee.3). r_FP1=(COP1gKJC)/1000.:)). W_lowleg(:.:).F1(i.
moments.[ 0 tc_cont 0]).aF2(i.i))*[1 COM_lowleg(i.:)=inv(RGAshank(:.3). ar_FP1=aCOP1(:. % Moment arms ar_lowleg=aCOPlowleg(:.:)=(inv(TGAshank(:. ar_hg=[0 0 hg_dist]/1000. end % Moment imposed on knee aMknee=aM_lowleg + aMF1 + aMF2 + aMtc + aMhg. and moment arms in Anatomical shank coordinates % Forces % FA = [RGA]^1 * FG for i=1:n aF1(i. aW_lowleg(i.:)'.:)=cross(ar_hg. aMtc(i.:).i))*[1 COP2(i.:.i))*W_lowleg(i.:)=inv(RGAshank(:.2)=hg_cont. aMF1(i.:)).aW_lowleg(i.3). aF2(i. % Moments for i=1:n aM_lowleg(i.:.aF1(i.:)]'). aCOP2(i.i))*F2(i.:)'.[0 hg_cont 0]).:).:)=cross(ar_tc.%% 8. aCOP1(i.:)]').:)).:)=cross(ar_lowleg(i. aMhg(i.:)=cross(ar_FP2(i.:. aMF2(i.:)=(inv(TGAshank(:. ar_tc=[0 0 tc_dist]/1000.2:4)/1000. % COP locations aCOPlowleg(i. ahg_cont=zeros(n.:)'. atc_cont(:. Calculate forces.:)]').:)=inv(RGAshank(:. end atc_cont=zeros(n.:. ar_FP2=aCOP2(:.:)).:.:.:)=cross(ar_FP1(i. 105 .2:4)/1000. ahg_cont(:.:).2)=tc_cont.2:4)/1000. % Total Force imposed on Knee aFknee= aF1+ aF2 + aW_lowleg + atc_cont + ahg_cont.:)=(inv(TGAshank(:.i))*F1(i.i))*[1 COP1(i.
'aM_lowleg'. 'nghg_cont'. 'nMF1'... 'MF2'. 'r_hg'. 'nF2'. Normalize forces and moments % forces will be normalized to body weight % moments will be normalized to bodyweight*height % Global forces nF1=(100/weight)*F1. 'M_lowleg'. 'natc_cont'. naW_lowleg=(100/weight)*aW_lowleg.. 'ar_FP2'. 'nMF2'. 'aW_lowleg'.'aF2'. 'aMF1'. 'r_FP2'.m 106 . naFknee=(100/weight)*aFknee. Save forces and moments to *posture*. nMhg=(100/(weight*height))*Mhg.'nFknee'. %% 10. nghg_cont=(100/weight)*ghg_cont. 'Fknee'. naMF1=(100/(weight*height))*aMF1.. 'append') end % end of function forces_moments. nF2=(100/weight)*F2. 'aF1'.'naW_lowleg'.. naMF2=(100/(weight*height))*aMF2. nMknee=(100/(weight*height))*Mknee. 'naFknee'. 'ngtc_cont'. 'Mhg'. 'gtc_cont'. nahg_cont=(100/weight)*ahg_cont. naMknee=(100/(weight*height))*aMknee. 'naMhg'. 'nMknee'. 'aFknee'.. .'F2'. nMF1=(100/(weight*height))*MF1.'r_lowleg'. naM_lowleg=(100/(weight*height))*aM_lowleg. 'nMtc'. % Anatomical Forces naF1=(100/weight)*aF1. nM_lowleg=(100/(weight*height))*M_lowleg. 'ahg_cont'. 'ar_tc'. 'naMF1'. . ngtc_cont=(100/weight)*gtc_cont. 'naF2'.. 'W_lowleg'. . nMF2=(100/(weight*height))*MF2.. 'ar_FP1'. 'r_tc'..'F1'. 'ar_lowleg'. 'r_FP1'.%% 9. 'aMknee'... nFknee=(100/weight)*Fknee. natc_cont=(100/weight)*atc_cont... 'aMhg'. 'MF1'.. 'naF1'. 'naMknee'..'nW_lowleg'. 'naMtc'. naMtc=(100/(weight*height))*aMtc. 'nahg_cont'. 'naMF2'.. 'nM_lowleg'. naMhg=(100/(weight*height))*aMhg. naF2=(100/weight)*aF2. . .. 'ar_hg'.. 'Mtc'. nW_lowleg=(100/weight)*W_lowleg. nMtc=(100/(weight*height))*Mtc. 'nMhg'. 'naM_lowleg'.. 'aMF2'. 'atc_cont'. . 'ghg_cont'. .. 'aMtc'. 'nF1'.. 'Mknee'...mat cd(results_directory) save(posture.
1980. Estimates of the Prevalence of Arthritis and Other Rheumatic Conditions in the United States. Verdonschot N. and Coggon D. 2007. 2004. Reading I. Occup Environ med.128(1):179189 Kivimäki J. Ann R Coll Surg Engl. 2008.147. Tomlinson F. Loeffen R. Helmich CG. De Waal Malefijt M. Factors Associated With Osteoarthritis of the Knee in the First National Health and Nutrition Examination Survey (HANES I). and hypertension: A comparative study. American Journal of Epidemiology. Li LC. Knee disorders in the general population and their relation to occupation.794797 Jensen. Brit J. Scand J Work Environ Health. Lilli K.58(1):2635 Maetzel A. Pencharz J. Arthritis & Rheumatism.tricountyortho. Clinical Orthopaedics and Related Research. Industr.31 suppl 2:6874 Sharrard. Barink M. Thighcalf contact force measurements in deep knee flexion. Riihimäki H. WJW.60.19. rheumatoid arthritis.401 Anderson JJ & Felson DT. The economic burden associated with osteoarthritis. Pressure Effects on the Knee in Kneeling Miners. Kneestraining work activities. Liddell FDK.BIBLIOGRAPHY [1] Lawrence RC. 1965. et al.195202 http://www. Knee disorders in carpet and floor layers and painters. Injuries to the Semilunar cartilage of the knee in miners. Med. Bombardier C. Cooper C. 2005.jpg Retrieved April 17. Clinical Biomechanics. Ann Rheum Dis. Knee Joint Structure and Function. Felson DT.714 Baker P. 1962.36.18(5):310316 Zelle J.63(4):395. selfreported knee disorders and radiographically determined knee osteoarthritis. 1988.30924 Sharrard WJW. Scand J Work Environ Health.22:821826 Welsh RP. Hänninen K. Part 2. 2003.com/mmg/pated/knee_problems/anatomy/quads. 1992. 2008 [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] 107 .
A Comparison of Histologic Differences in Meniscal Tissue Between Tears in Anterior Cruciate LigamentIntact and Anterior Cruciate LigamentDeficient Knees. Indelicato PA.807811 Tangtrakulwanich B.niams. 1975. McNicholas MJ. (pp 883893). Edinburgh. www.gov Lavender SA and Andersson GBJ.62:567575 Drosos GI. 71149 Baratz ME. Spanier S. Batts J.32(6):14791483 Hoshino A. Ergonomic principles applied to prevention of injuries to the lower extremity.69B. London and New York: Churchill Livingstone. American Journal of Sports Medicine. Am J Sports Med 1986:14(4):270 Hawkins A. Habitual Floor Activities Increase Risk of Knee Osteoarthritis. Meniscectomy and arthritis. Medial and Lateral Meniscal Tear Patterns in Anterior Cruciate LigamentDeficient Knees. Meniscal tears: the effect of meniscectomy and of repair on intrarticular contact areas and stress in the human knee: a preliminary report. Injuries to the knee joint. Geater AF. The causes and mechanisms of meniscal injuries in the sporting and nonsporting environment in an unselected population. pp. Clinical Orthopaedics and Related Research.000 fulltime workers by part of body and selected sources of injury or illness. Fu FH. 2004. FL. Chongsuvivatwong V. Wright. Barrett GR. CRC Press United States Bureau of Labor Statistics. 1987. 2006. The Journal of Bone and Joint Surgey. The American Journal of Sports Medicine. Nichols L.). The Journal of Bone and Joint Surgery. Wallace WA.2(2):5156 [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] 108 . In W Karwowski & WS Marras (Eds. Pozo JL. The Occupational Ergonomics Handbook. Simonet Jacques C. Ann Rheum Dis 1974:33(1):111 Macquet Paul G. Boca Raton.11:143149 Smillie IS. Mengato R. V. Dowson D. Osteoarthritis and meniscus disorders of the knee as occupational diseases of miners. Proceedings: Functions of the menisci: a preliminary work. 2001. 2006. The Knee. CME Orthopaedics 2001. 2005. Van de Berg André J. ImpactAbsorbing Properties of the Human Knee. Occup Environ Med.57A(6):766771 Smith JP. Femorotibial WeightBearing Areas.[12] [13] [14] [15] Seedhom BB. 2005.454:147154 National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS).nih. Franklin J. 1978.29(4):415419 Meister K. Incidence rates for nonfatal occupational injuries and illnesses involving days away from work per 10. McMillan G.
Analysis of kneejoint forces during flexedknee stance. Risk Factors for Progressive Cartilage Loss in the Knee. Ford W. JM Young (1969) Weight. 1994. Mayo P. et al. Hinrichs. Eng Med 1982. Mechanical loads at the knee joint during deep flexion. An Update on the Epidemiology of Knee and Hip Osteoarthritis with a View to Prevention. Arthritis & Rheumatism. Key Biscayne. The Association of Meniscal Pathologic Changes with Cartilage Loss in Symptomatic Knee Osteoarthritis. Occupational activity and osteoarthritis of the knee. Agrawal V. Niu JB. Biomechanics and motor control of human movement. Xu L et al. Wiley.23(9):949951 [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] 109 . The Beijing osteoarthritis study. Dyrby CO. et al. Andriacchi TP. A comparison of the accuracy of several hip center location prediction methods.53:9093 Zhang Y. Coggon D. Journal of Biomechanics. 2002. Association of squatting with increased prevalence of radiographic Tibiofemoral knee osteoarthritis. J. Industr. Summer Bioengineering Conference. 1984. J Biomechanics. Nevitt MC. Journal of Orthopaedic Research. Arthritis & Rheumatism. J Bone Joint Surg Am. Adjustments to the segment center of mass proportions of Clauser et al. Pedersen DR. Scand J Work Environ Health.[26] [27] Biswal S. Seedhom BB. Dieppe P. Zhang YQ. June 2529. Bergman GA. 1990. Alexander EJ. Arthritis & Rheumatism. RN.A. Andriacchi TP. Hunter DJ. Brit. McAlindon T. Ann Rheum Dis. Ohio. Rheumatism In Miners Part II: XRay Study.20:881886 Caruntu DI. 2003. Florida Winter. 1990.50(4):11871192 Dahlvist NJ.617621 Clauser CE.57:961967 Nagura T.volume and center of mass of segments of the human body.54(3):795801 Felson DT & Zhang Y. Arthritis & Rheumatism. 1998.10(6):429433 Kellgren JH & Lawrence JS. 2006. JT McConville. Hastie T. Goel VK. (1969).41(8):13431355 Anderson JA.9:197207 Cooper C. Athrosis and its relation to work. Antonelli D. Brand RA. D. 2002. Modeling the Knee Joint in Deep Flexion: ”Thigh and Calf” Contact. Dennis MJ. WrightPatterson Air Force Base. Goitz HT. Hefzy MS.23.11:6976 Perry J. Egger P. AMRLTR6970. New York Bell AL. 1975. Med. Forces during squatting and rising from a deep squat.46(11):28842892 Hunter DJ. 1952. 2004. (1990).
Comparison of capacitive versus resistive joint contact stress sensors. de Lange A. The effect of variation in knee center location estimates on net knee joint moments.28(6):753758 Hemmerich A.7. Suda Y. Journal of Orthopaedic Research. Hurschler C. Canada. Zayontz S. Aissaoui R. DeFrate LE. June 2729. Function of the knee cruciate ligaments as restraints on varusvalgus torque. Normalization of joint moments during gait: a comparison of two techniques. Bryant JT. Journal of Biomechanical Engineering. Ramakrishnan HK. 2004.11. 1999. Kinetic analysis of the lower limbs during walking: what information can be gained from a threedimensional model? Journal of Biomechanics. Role of the anterior cruciate ligament and the medial collateral ligament in preventing valgus instability. Understanding TibioFemoral motion.24:137141 Martinelli L. Shott S. 1991. Proceedings of Annual Meeting of Japanese Society for Orthopaedic Biomechanics. Simultaneous bilateral 3D ablebodied gait. Calibration Issues of Tekscan Systems for Human Pressure Assessment. Brown H. 2001.36:599603 Allard P. Clinical and Orthopaedics and related research. Lachance R. Duhaime M. and ankle kinematics of high range of motion activities of daily living. Recruitment of Knee Joint Ligaments. Reid SA.[41] [42] [43] Moisio KC. Presented at RTO HFM Specialists Meeting on “Soldier Mobility: Innovations in Load Carriage Systems Design and Evaluation”. Niki Y et al. Hideo M. Rosenbaum D. Measurement of lower extremity kinematics during level walking. 8:383392 Stagni R.113:94103 Matsumoto H. Capello A. Hurwitz DE. et al. 1996. Journal of Biomechanics. Marthandam SSK. 2006. Whiteside RA. Huiskes R. Journal of Orthopaedic Research. Hip. 2006. Human Movement Science. Simultaneous measurement of tension along the ACL and PCL. 1995. Journal of Orthopaedic Research. Wootten ME. Otani T. Journal of Orthopaedic Sciences. 2000 Holden JP.6(1):2832 Toshiyuki M.8188 Blankevoort L. The Knee.24(4):770781 Li G. knee.16 Williams A. Smith S. and Wyss UP. Kingston. Yoshihiro S et al.20:151156 110 [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] .1998. 2004. Summer DR.22(1):9095 Kadaba MP.15:327346 Eng JJ & Winter DA. 2003.447:214220 Morin EI. Gait & Posture. Propagation of anatomical landmark misplacement to knee kinematics: Performance of single and double calibration. Yasumfumi S. Gait & Posture. Fantozzi S. Stanhope SJ. 1990. Logan M. Kinematics of the knee at high flexion angles: an in vitro investigation. 2006.
Goh JCH. Deluzio KJ.[55] [56] Fukuda Y.27:329335 Costigan PA. Hovinga KR. Rietbergen B.24(2):193197 Papannagari R. DeFrate LE.21(6):11071112 Carlin GJ. and Huiskes R. Das De S. Effect of posterior cruciate ligament on posterior stability of the knee in high flexion. Nha KW et al. 2008. Lee J. Knee and hip kinetics during normal stair climbing.36:845851 Atkinson TS. Wyss UP. Dynamic knee loads during gait predict proximal tibial bone distribution. 2007.37(5):779783 Hurwitz DE. Hughes C. 1998. 2002.26(5):673684 Schipplein OD and Adriacchi TP. Journal of Biomechanical Engineering. Interaction between active and passive knee stabilizers during level walking. Journal of Biomechanics. 1996. Journal of Biomechanical Engineering. Gait characteristics of patients with knee osteoarthritis. DeFrate LE. Livesay GA.210:2737 Thambyah A. Journal of Orthopaedic Research. Impactinduced fissuring of articular cartilage: An investigation of failure criteria. Insitu Forces in the Human Posterior Cruciate Ligament in Response to Posterior Tibial Loading. Annals of Biomedical Engineering. Gait and Posture. Lewis J. Normal contact of elastic spheres with two elastic layers as a model of joint articulation. Journal of Orthopaedic Research. Most E. Contact stresses in the knee joint in deep flexion. Woo S. 2003. Donkelaar CC. 2003. Magnetic resonance image analysis of meniscal and tibiomeniscofemoral contact in deep knee flexion. Journal of Biomechanics.. 1996. Transmission of Rapidly Applied Loads Through Articular Cartilage Part 1: Uncracked Cartilage. 2001. Andriachhi TP. 2005. Medical Engineering and Physics. Pathways of loadinduced cartilage damage causing cartilage degeneration in the knee after meniscectomy. Lancianese SL. et al. Harner CD et al. Sumner DR. 1991. American Journal of Sports Medicine. 1998:120:181187 Eberhardt A.9:113119 Wilson W. Journal of Biomechanics.31:423430 Yao J.38:15071512 Li G.34:907915 [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] 111 . and Keer L. Tsuda E et al. and Altiero NJ. An K.16:3137 Kaufman KR. Journal of Orthopaedic Research. Function of posterior cruciate ligament bundles during in vivo knee flexion. 2004. Haut RC. Loh J. Morrey M. and Sugar DA. A quantitative analysis of valgus torque on the ACL: a human cadaveric study. Morrey BF. Lerner AL. Journal of Biomechanics. Jounral of Engineering Medicine. 1991:113:410417 Kelly PA and O’Connor JJ.
This action might not be possible to undo. Are you sure you want to continue?
Use one of your book credits to continue reading from where you left off, or restart the preview.