Design of Low-Cost, Fully Passive Prosthetic Knee
for Persons with Transfemoral Amputation in India
by
Venkata Narayana Murthy, Arelekatti
B.Tech., Indian Institute of Technology Kharagpur, India (2010)
Submitted to the Department of Mechanical Enginering
in partial fulfillment of the requirements for the degree of
Master of Science in Mechanical Enginering
ARCHIVES
MASSACHUSETTS INSTITUTE
OF TECHNOLOLGY
at the
JUL 3 0 2015
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
LIBRARIES
June 2015
Massachusetts Institute of Technology 2015. All rights reserved.
Author...........
Signature redacted
Department of Mechanical Enginering
May 8, 2015
Certified by...........Signature
redacted
inter, V
Am
Assistant Professor of MechafiI Engineering
Thesis Supervisor
Accepted by........Signature
redacted ....
David E. Hardt
Professor of Mechanical Engineering
Graduate Officer
....-
..
=
.e
.......--
Design of Low-Cost, Fully Passive Prosthetic Knee for Persons
with Transfemoral Amputation in India
by
Venkata Narayana Murthy, Arelekatti
Submitted to the Department of Mechanical Enginering
on May 8, 2015, in partial fulfillment of the
requirements for the degree of
Master of Science in Mechanical Enginering
Abstract
An estimated 230,000 above-knee amputees are in need of prosthetic devices in India
with a majority of them facing severe socio-economic constraints in their daily lives.
However, only a few passive prosthetic knee devices in the market have been designed
to enable normative gait and to meet the unique daily life needs of above-knee amputees in the developing world. This thesis builds upon a past study at MIT, which
established optimal mechanical component coefficients in prosthetic knee function required for achieving able-bodied kinematics. A mechanism for the design of a fully
passive, low-cost prosthetic knee device, which aims to facilitate able-bodied kinematics at a low metabolic cost is presented. The mechanism is implemented using
an automatic early stance lock for stability, a linear spring for early stance flexionextension and a differential friction damping system for late stance and swing control.
For preliminary validation of the knee mechanism two field trials were carried out on
five above-knee amputees in India, which showed satisfactory performance of the early
stance lock and enabled smooth stance to swing transition by timely initiation of late
stance flexion.
Thesis Supervisor: Amos G. Winter, V
Title: Assistant Professor of Mechanical Engineering
3
....
...
.
.
..
.
.
.
-...
...
...
....
.
.....
a.
..
..
. .......
.....
.
,,...........
..
.......
..
.
.
. ..
. .
,,...
...
,..........
.
..
...
. ..
.
,-
.
s,,.ss..1.v.....,.m
m.e
-.....
.
. .,
Acknowledgments
I would like to thank the following people and organizations for their contributions
at different stages of my work:
" The five subjects with above-knee amputation, who volunteered to try out my
prototype and gave me very useful feedback based on their insights and experience.
" Professor Amos Winter, for his crucial advice and technical inputs through
out the course of my research effort; and for his creativity, humor, unbridled
enthusiasm for engineering, deep concern towards the well being of his graduate
students, and for personifying all the qualities I would like to emulate in the
future as an engineer, scholar and designer.
" Yashraj Narang, for introducing me to research methodologies in biomechanics,
and for his exhaustive research work on this topic. His genuine curiosity and
rigorous, analytical approach to research and meticulous presentation skills have
taught me a great deal and shaped my research experience in graduate school.
" Dr.
Mathur, Mr.
Pooja Mukul, Rajender, Dr.
Mehta and technicians at
BMVSS (Jaipur Foot) organization, for their timely support to conduct usertrials at Jaipur, India.
Their insights, based on years of experience in the
prosthetics industry, served as crucial reality checks during my visits to Jaipur.
" Dan Dorsch, Ben Peters, Michael Buchman and rest of my fellow graduate students at the MIT GEAR Lab, for their incredible patience and selfless yearning
to share their prototyping hacks with me.
" Nevan Hanumara, Chintan Vaishnav, Jesse Breneman and Rob Stoner of the
Tata Center, for their entrepreneurial push and constant feedback. Funding for
my tuition and research was provided by the Tata Center of Technology and
Design at MIT.
* Friends and family, for their love, support, encouragement, wit and humor.
5
6
Contents
C over
1
2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
A bstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
Acknowledgments. ........
5
Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6
Introduction
9
1.0.1
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
1.1
Biomechanics of human gait . . . . . . . . . . . . . . . . . . . . . . .
11
1.2
Above-Knee Prosthesis: Terminologies
. . . . . . . . . . . . . . . . .
12
1.3
Thesis O utline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13
Design considerations for Prosthetic Knee Function
15
2.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
2.2
User needs in the developing world
16
2.3
Optimal stiffness, damping and engagement parameters of passive pros-
2.4
3
................................
. . . . . . . . . . . . . . . . . . .
thetic knees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
19
Prosthetic knee function
. . . . . . . . . . . . . . . . . . . . . . . . .
21
2.4.1
Prior art: Design strategies for stability . . . . . . . . . . . . .
22
2.4.2
Prior art: Design strategies for achieving normative kinematics
24
Design and Testing of the Prosthesis Mechanism
27
3.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
3.2
Design Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
28
3.3
Mechanism Architecture and Function
29
7
. . . . . . . . . . . . . . . . .
3.4
4
3.3.1
Early stance lock . . . . . . . . . . . . . . . . . . . . . . . . .
30
3.3.2
Early stance flexion-extension
. . . . . . . . . . . . . . . . . .
32
3.3.3
Flexion during late stance and swing . . . . . . . . . . . . . .
33
3.3.4
Differential Damping System . . . . . . . . . . . . . . . . . . .
33
3.3.5
Prototype iterations
. . . . . . . . . . . . . . . . . . . . . . .
37
User Trials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
37
3.4.1
Trial Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . .
39
3.4.2
Trial Results . . . . . . . . . . . . . . . . . . . . . . . . . . . .
40
Discussion and Future Work
45
4.1
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
45
4.2
Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47
A Research Protocol Documents
49
8
Chapter 1
Introduction
The objective of the present thesis was to use the tools of biomechanics, mechanical
design and user-centric design process to develop a fully passive, low-cost prosthetic
knee which can enable able-bodied gait for users in India with above-knee amputation
(termed as transfemoral amputation in biomechanics [1]). The work by Narang and
Winter [2-6] at MIT forms the basis of this thesis, which is briefly summarized in this
chapter for background information.
Relevant parts of their work are discussed in
further detail in the later chapters. This introductory chapter provides a brief review
of human gait and prosthesis technology for transfemoral amputees. An outline of
the thesis is also presented at the end of the chapter.
1.0.1
Background
In 2011, Bhagwan Mahaveer Viklang Sahayata Samiti (BMVSS, a.k.a., "Jaipur Foot")
and the Massachusetts Institute of Technology (MIT) initiated a collaboration to
develop a prosthetic knee for transfemoral amputees in India.
BMVSS, a non-
governmental organization (NGO) based in Jaipur, India, is a major developer, manufacturer and distributor of prosthetic, orthotic, and assistive devices throughout India
and one of the largest organizations in the world serving people in need of prostheses [7]. Since it's inception in 1975, it has helped rehabilitate more than 1.3 million
amputees and polio patients, mostly in India [7]. Using outside funding sources, they
9
distribute all their products free of charge to amputees.
In January 2012, an initial project meeting was held with BMVSS, and the following
design requirements for a new design of the prosthetic knee were given by clinicians
and prosthetists at BMVSS [21:
" Allows normal gait on flat ground
" Provides stability on uneven terrain
" Costs less than $100 to manufacture
After additional meetings at BMVSS and personal interaction with a number of other
organizations across India in 2012 and 2013, many specific design requirements were
documented.
Prosthetists, amputees, technicians, engineers, physicians, professors
and administrators at prosthesis fitment centers, rehabilitation hospitals, and academic institutions across India were interviewed as a part of this interaction.
A
structured survey of Indian amputees was also conducted to quantify the demographics, functional capabilities and functional needs of future end users [2,5]. To explore
biomechanical feasibility of the design of a low-cost prosthetic knee which can enable
transfemoral amputees to walk with reduced energy expenditure and normative gait
kinematics, two-dimensional inverse dynamics simulations were performed. The effects of inertial alterations of a prosthetic leg on the energy expenditure required to
walk with normative gait kinematics were quantified. The effects of inertial properties on the knee moment required for normative gait was also explored. A passive
mechanical model, comprising springs and friction dampers, was formulated to accurately reproduce the required knee moment for able-bodied gait. The work of Narang
and Winter provided a blueprint for the present thesis, which formed the second phase
of the project. The relevant findings of the work of Narang and Winter are described
in further detail in Chapter 2.
10
WEIGHT
%OF CYCLE
INITIAL
CONTACT
1
1
0% 2%
SIGLM
SUPPORT
LOADING
RESPONSE
i
12%
MID
STANCE
-
SWING LIMB ADVANCEMENT
SINGLE LIMB
ACCEPTANCE
TASKS
EVENTS
SWING PHASE
STANCE PHASE
PHASES
TERMINAL
STANCE
PRE
SWING
i
i
50%
31%
62%
DACMN
INITIAL
SWING
MID
SWING
i
i
75%
87%
TERMINAL
SWING
i
100%
Figure 1-1: Phases of the able-bodied human gait cycle. Adapted from [9].
1.1
Biomechanics of human gait
The periodic motion of walking is referred to as the "gait cycle". Qualitatively, the
gait cycle is often divided into phases based on whether one or more legs are in contact
with the ground. "Stance" is when the foot of a specified leg is in contact with the
ground, and "swing" is when the foot of the leg is off the ground. Stance and swing
of one leg alternate with those of the other. "Single limb support" occurs when a
single leg is on the ground, and "double limb support" occurs when both legs are on
the ground. These terms are illustrated in Figure 1-1. The gait cycle can also be
divided into phases based on the forward progression of the body. The Rancho Los
Amigos Gait Analysis Committee [8] proposed a taxonomy that is commonly used in
the literature.
A typical transfemoral amputee gait deviates from able-bodied gait significantly
because of a combination of reasons such as loss of active torque generating muscles,
loss of sensory feedback and proprioception. In addition, the prostheses being used
by above-knee amputees can add additional, undesirable degrees of freedom and may
also cause pain in the residual limb due to the socket-limb interface. And most im-
11
portantly, the inherent limitations of the prosthesis function could be a huge limiting
factor in enabling normative gait. Common gait deviations observed in transfemoral
amputees has been well documented in literature. Some of the common abnormalities
are: Lateral trunk bending, Wide waking base (abduction), circumduction, vaulting,
swing-phase whips, uneven heel rise, terminal impact and uneven step length [10,11.
1.2
Above-Knee Prosthesis: Terminologies
socket
exoskeletal
shank
endcskeletal
shank (pylon)
knee
shank
foot
Figure 1-2: Drawing of a typical above-knee prosthesis. In an exoskeletal shank, load
is borne upon a shell, whereas in an endoskeletal shank, load is borne primarily upon
a pylon. As depicted, an endoskeletal shank may also have a cosmetic cover over the
pylon. Adapted from [12] and [2]
An "above-knee prosthesis" is a prosthetic leg that has been designed for individuals amputated above the knee. Typically, an above-knee prosthesis consists of 5 major
parts: the suspension, the socket, the knee, the shank, and the foot (Figure 1-2). The
present thesis focuses primarily on the design of the knee.
12
In an able-bodied human, the knee allows a large range of motion, and the muscles
of the leg (e.g., quadriceps and hamstrings) control the fiexion of the knee as required
for a given activity. During walking, normal knee kinematics are critical, as deviations
from normal kinematics have been found to increase metabolic energy expenditure
[13]. Unfortunately, above-knee amputees typically have reduced muscle function due
to muscle loss and atrophy, making flexion of a prosthetic knee difficult to control.
The ideal prosthetic knee not only allows a large range of motion, but also replaces
lost muscle function by providing appropriate resistance and/or propulsion to allow
normal kinematics during walking and other activities [14]. Many Different types of
prosthetic knees in the context of developing countries have been designed and their
function are described in detail in chapter 2.
1.3
Thesis Outline
The present thesis, as described at the beginning of this chapter, builds upon the
earlier work done by Narang and Winter [2-6] and is primarily focussed on the mechanism design and testing of a fully passive prosthetic knee that can enable normative
gait for transfemoral amputees in India. Following is an outline of the thesis:
" Chapter 2: This chapter introduces the needs of users in India, biomechanical
modeling and analysis for the design of an idealized prosthetic knee. The design
of existing prosthetic knees are analyzed in detail, in terms of their functions
to enable stability and normative gait for the user. The costs of different prosthetic knees in the market, both in the developed and developing world are also
compared.
* Chapter 3: This chapter lays out the design of a beta prototype of the prosthetic knee with the help of detailed illustrations which describe the function of
the prototype knee at each stage. An earlier, "alpha" version of the prototype is
also illustrated. The process of user-centric testing of the prototype at BMVSS
(Jaipur-foot) is described along with the feedback provided by the subjects and
13
prosthetists involved in the user trials.
* Chapter 4: This chapter discusses the results obtained from testing the prototype and the insights gained from the work completed in the present thesis.
Novel research contributions and limitations of the study are also presented. In
conclusion, future recommendations for further refinement of the prototype and
the mechanism design process are proposed.
14
Chapter 2
Design considerations for Prosthetic
Knee Function
2.1
Introduction
This chapter briefly summarizes the important considerations to be taken into account
for the design of a completely passive prosthetic knee, based on existing literature
and products in the market, both in the developing and developed countries. For an
introduction and definition of terminologies used in above-knee prosthesis technology,
see section 1.2. A crucial first step towards establishing functional requirements for
the design process (section 3.2) is articulating the user needs. This chapter summarizes the results of a detailed user-needs survey conducted by Narang et al [2] followed
by an analysis of gait biomechanics and prior art. We articulate a clear distinction
between conflicting biomechanical requirements of walking with a passive above-knee
prosthesis, vis-A-vis kinematics of walking as against stability (body-weight support)
function of walking. This is followed by a cost-function mapping of existing prosthetic
knees, both in the developed and developing world.
15
2.2
User needs in the developing world
It is estimated that there are currently 30 million people across the world in need of
prosthetic and orthotic devices [15-17]. In India alone, we estimate the total number
of above-knee amputees to be in excess of 230,000 [2]. Other studies have estimated
a number of 6.7 million above-knee amputees in Asia, with a majority living in developing countries of large population such as India and China [17].
According to
an estimate by the World Health Organization, 90-95% of amputees in developing
countries do not receive any prosthetic device [18] and only 20% of amputees are able
to afford currently available prostheses in the market 119].
A majority of Indian amputees belong to economically poor families [21].
In a
past study by Narang et al. [20], 47% of Indian amputees reported changing their
occupation after amputation, as most of the amputees were earlier employed in jobs
that demanded physical exertion such as agriculture and manual labor involving long
hours of standing, walking and lifting heavy weights. In the interviews conducted as
a part of our earlier work [2], amputees reported social discrimination in their families and communities because of their conspicuous disability and unnatural gait. The
severe social consequences and stigma endured by people who undergo lower-limb
amputation in the context of different cultures have been well documented [22-241.
Acute financial constraints coupled with socio-economic considerations project an urgent need for a low-cost product that can deliver high levels of functional performance.
In addition to the biomechanical requirements of walking and standing, a user-centric
approach was used to establish design requirements based on activities of daily living,
fitment, manufacturing, distribution, maintenance, and compliance to international
standards (Table 2.2). There were three important components to this approach:
1. Collaboration and interaction with Bhagwan Mahaveer Viklang Sahayata Samiti
(BMVSS, also known as, the Jaipur Foot organization) based in Jaipur, India.
2. Interviews of Stakeholders: Technicians, engineers, physicians, professors and
administrators at different prosthesis fitment clinics, rehabilitation hospitals,
16
Importance and Difficulty of Various ADLs
100
Bicycle
Wet Mud
/
Stairs
Sit for long
Lie Down
Grass
95
Stand for long
Heavy objects
Sit cross'-legged
Sit-Stand Transition
90
Squat
85
Rocks
0
E
80
Water
75
Kneel
Hills
70
0
10
20
30
40
50
60
70
80
90
100
%
Difficulty
Figure 2-1: Quantitative Survey Results showing the difficulty and importance of
various activities of daily living. The horizontal axis is the percentage of respondents
who rated each activity difficult. For each who rated ADLs difficult, the vertical axis
shows the percentage who answered that their lives would be significantly improved
if they could perform the activity easily with an alternative prosthesis [2,20]
and academic institutions across India.
3. A structured user-needs survey of 19 transfemoral amputees in Jaipur, India
to identify the specific needs with respect to their common activities of daily
living [25, 261.
A wide range of functional requirements was established and ranked in order of importance based on quantitative and qualitative data, which served as the guidepost
for further analysis and design of mechanism (Table 2.1).
Although a number of advanced prosthetic limbs and assistive devices have been
designed in the developed world in the last few decades (Figure 2-3), very few of
17
Functional Requirements
* Able-bodied kinematics
Biomechanical Requirements
9 Stability
9 Energy Conservation
e Ability to stand for long
9 Easy sit-stand transition
* Ability to walk on wet mud
Requirements articulated by the
users
*
Ability to walk carrying heavy objects
9 Sitting cross-legged (important in Indian culture)
* Ability to squat and climb stairs
9 Cost per device <$100
@ Normal looking gait on flat ground
* Stability on uneven terrain
* ISO 10328 compliance
Requirements articulated by the
stakeholders
* Mass-manufacturable
* Ease of fitment, alignment and maintenance
e Appropriate for amputees with long
residual limbs
* Aesthetically pleasing cosmesis
Table 2.1: Functional requirements established based on biomechanical considerations
and user centric approach.
18
them have been suitable for large-scale use in developing countries due to vastly different and complex socio-economic considerations and resource-constrained settings.
Prosthetic knee joints in the United States and Europe cost several thousand dollars
to manufacture and distribute.
Popular active above-knee prostheses that deliver
very high performance can cost up to $50,000 [27]. Even the passive knee joints in
developed countries are too expensive to meet the requirements of amputees in the
developing world.
2.3
Optimal stiffness, damping and engagement parameters of passive prosthetic knees
Designers of prosthetic devices have used components such as springs and dampers
and optimized them with the aim of replicating ideal knee moment required for walk-
ing with able-bodied kinematics [28]. Using a rigid body model of human walking in
the sagittal plane, the work of Narang et al [2] theoretically established the mechanical feasibility of achieving able-bodied kinematics using low-cost, passive mechanical
components such as linear springs and friction dampers (Figure 2-2). Their study also
optimized the mechanical component coefficients by accounting for changes in inertial properties of prosthetic legs, which typically weigh less than physiological legs [1].
Their study showed that using a single linear spring and two friction dampers that
engage and disengage at the prescribed points during the gait cycle, it is possible
to accurately replicate physiological knee moment (adjusted to the change in inertial
properties of prosthetic components compared to the able-bodied leg segments).
A mechanical embodiment of such a knee would need the mechanism to engage
and disengage the spring and dampers at optimal points of time during the gait
cycle (Figure 2-2). Additionally, by tuning the spring stiffness and damper friction
to prescribed values (based on the body weight of the person and weight of the
prosthesis), it should be possible to closely replicate the desired knee-moment to
achieve able-bodied kinematics.
19
trunk
Knee
Angle
clutch
springupelg
-yo
,
(stump + socket)
-{LF- damper
lower leg
(shank)
knee joint
foot
0.6
First
Damper
b1
Spring
k1
E
-
Second
Damper
b2
-
E 0.2
E
o
0
D-0.2 --024
a)-0
-
-Ideal
-Prosthesis
-_0.6
Z -0.8
R 2 =0.90
0
10
20
0
10
20
30 40 50 60 70 80 90 100
% gait cycle (time)
30 40 50 60 70 80 90 100
70 0 Spring engage
60 -
4Spring release
First damper engage
First damper release
~a 50 .0' Second damper engage
4 Second damper release
a)
a
40
C:
<30
c20
10
0
Late stance
flexion
Early stance
flexion-extension
Stance Phase
Swing phase
extension
Swing Phase
Figure 2-2: Determination of optimal mechanical component coefficients for replicating able-bodied knee moment. Narang and Winter [2] used a rigid body model
comprising foot, ankle joint, lower leg, knee joint, and upper leg (top). Using inverse dynamics, they predicted the spring stiffness (ki) and frictional damping (bi
2
and b2) required for replicating able-bodied moment with R = 0.90 (Middle). The
engagement-disengagement points during each gait cycle were also established as a
part of this analysis for one spring and two friction dampers (bottom). The knee angle is the relative angle measured between the upper leg and lower leg (top). Adapted
from [6]
20
2.4
Prosthetic knee function
The primary biomechanical goal of an above-knee prosthesis, as briefly discussed in
chapter 1, is to substitute the function of lost limb as closely as possible. In literature
[10, 291, this is described as achieving reliable stance-phase control and swing-phase
control. While this description of prosthetic knee function is fairly accurate, it is still
incapable of explaining the complete functionality of the prosthesis as a substitute
for the lost limb. As an alternative approach, we break down the prosthetic knee
function as the following:
1. Stability
2. Able-bodied kinematics
3. Metabolic energy expenditure reduction.
Based on biomechanics and physics of walking and past studies [30] that explore the
correlation between achieving able-bodied kinematics and metabolic expenditure reduction, it can be argued that achieving able-bodied gait in unilateral above-knee
amputees leads to a reduction of metabolic energy expenditure. Therefore, the primary requirements of an above-knee prosthesis can be reduced to achieving stability
during walking and standing, and enabling walking with able-bodied kinematics. This
section discusses the prior art in detail, specifically focussing on how each of the existing passive prosthetic knees are designed to meet the aforementioned objectives
of prosthetic knee function.
However, active above-knee prostheses have not been
analyzed as a part of this work for multiple reasons, although they can enable very
high performance in terms of metabolic expenditure and able-bodied gait. Firstly, active or quasi-passive prosthetic knees are electromechanically controlled using modern
electronics (usually, a microprocessor) and are battery driven. Secondly, the actuators used for achieving the required mechanical impedance are extremely expensive
(Table 2.2). For example, the Ossur's Rheo-Knee, which costs about $40,000 [311,
uses a magnetorheological fluid which changes it's damping resistance based on the
magnitude of electric current applied. Thirdly, they are not suitable for use in the
21
socio-economic context of developing countries, not only due to cost, but because of
technological barriers of adaptability. Most active knee prostheses are battery powered and need daily recharging and are not robust to withstand diverse environmental
conditions that users in India would experience [10,18].
2.4.1
Prior art: Design strategies for stability
One of the fundamental design challenges in the design of a passive knee joint is
achieving reliable stance control, which is important for stable locomotion and avoiding falls during early stance [28]. During early stance (Figure 2-4), GRF acting at the
COP is posterior to the physiological location of knee axis and causes a large flexion
moment at the knee. However, despite this large flexion moment, the physiological
knee does not buckle as the extensor muscles in the leg provide an opposite internal
extension moment and limit early stance flexion of the knee to a maximum of about
20 degrees (Figure 2-2). Advanced electromechanical knee joints, counter this large
flexion moment by either providing a counter extension torque using an active powered component or regulate the resistance of the joint based on electro-mechanical
sensing of the center of pressure [32,33]. In a passive knee joint, which does not have
any sensors or battery driven active component, stance control is a serious challenge.
Different designs of passive prosthetic knees tackle this problem of stance control by
compromising on early stance flexion through mechanical means well documented in
literature [18, 28]. For example, single axis knee joints rely on voluntary control of
hip musculature to resist flexion during early stance. Single axis locking knee joints
such as ICRC knee (Figure 2-3) use a mechanical latch engaged by the user to provide extra stability, which leads to a stiff legged gait suited only for new amputees or
low-activity elderly users who demand hyper-stability [28]. Polycentric mechanisms,
using 4-bar mechanism or 6-bar mechanism rely on a moving instantaneous center of
rotation to provide stability. The instantaneous center of rotation starts off posterior
to the GRF vector at the beginning of stance and moves anterior to the GRF vector
just before toe-off enabling some late stance flexion [28]. The LCKnee, recently developed by Andrysek et al [18], uses an automatic stance locking mechanism to lock
22
Full Extension
Maximum Flexion
a. ICRC Single
Axis Knee
b. BMVSS Manual
locking knee
c. BMVSS Single
Axis knee
d. Jaipur-Stanford
polycentric knee
e. LCKnee Stance
locking knee
Figure 2-3: Different low-cost prosthetic knees being used in the developing world.
Adapted and updated from [2]. Costs of these prosthetic knees are summarized in
table 2.2.
23
Early Stance
Mid Stance
Late Stance
Ground Reaction
Force Vector
Figure 2-4: Relative position of the Ground Reaction Force (GRF) vector and the
knee. The GRF vector is posterior in early stance and late stance causing a flexion
moment at the knee. During mid-stance, the vector is anterior to the knee. Green
arrow depicts the direction of the net moment at the knee during each stage because
of the GRF vector, inertial forces and the hip moment. The moment exerted by hip
muscles and the inertial forces are not shown.
the knee during early stance and unlocks it during late stance. A similar automatic
stance locking mechanism was earlier developed by Farber and Jacobson [341. Table
2.2 summarizes the cost and functions of the most commonly used knee prostheses,
both in the developed and the developing world.
2.4.2
Prior art: Design strategies for achieving normative kinematics
Normative kinematics during stance involves an early stance flexion of up to 1525 degrees and a late-stance and swing flexion of up to 70 degrees. The knee also
has to extend back to zero degrees (full extension) before heel strike. None of the
passive prostheses designed for the developing world incorporate any feature to enable
the early-stance flexion [2,18]. However, some of them have built-in mechanisms for
effective swing-phase control. These control mechanisms have been designed primarily
to limit maximum flexion during swing and to control the motion of the lower-leg
assembly for smooth extension at the end of swing phase.
24
Constant friction mechanisms are one of the the simplest ways to limit maximum
flexion angle. Friction control is extremely inexpensive, tunable by changing the tension in a bolt by a locknut. This system is generally incorporated in polycentric knees.
However, friction based damping systems are also highly variable due to wear and
change in environmental conditions, such as moisture and humidity which change the
coefficient of friction between interacting surfaces.
Friction systems also cannot be
tuned dynamically based on walking cadence, as walking faster requires higher levels
of damping. More advanced passive prostheses use a fluid system for achieving variable damping dependent on cadence speed. Pneumatic and hydraulic cylinders are
generally used in fluid based damping systems. Though these systems are highly functional in terms of achieving the required damping, they can be extremely expensive
and may require regular maintenance. Many prosthetic knees are also provided with
an extension assist mechanism designed to reduce the length of swing-phase and to
enable a more symmetrical swing-phase control [10]. These assist mechanisms often
lead to a large terminal impact and may necessitate use of damping pads to reduce
the loud terminal impact at full extension.
25
Name
Organization
Type of knee
Cost
Type of cost
Manuallocking
BMVSS
Manuallocking
< $10
Manufacturing
[35]
Limbs International
BMVSS
Four-bar
$15-20
Manufacturing
136]
Four-bar
$20
Manufacturing
[37]
Single-axis,
auto-locking
Single-axis,
free-swinging
$50-100
Not reported
[38,39]
$147
Retail
1401
Single-axis,
$2,060
Retail
[41]
$54,510
MSRP
[42]
$75,000
MSRP
[43]
knee
LIMBS
Knee
StanfordJaipur
knee
LCKnee
Jan Andrysek
Niagara
Knee joint
Niagara Prosthetics & Orthotics
802 Nylon
Aulie
Knee
vices, Inc.
hydraulic
C-Leg
Otto Bock
Single-axis,
De-
Source(s)
microprocessor, hydraulic
Genium
Otto Bock
Single-axis,
microprocessor, hydraulic
Table 2.2: Costs of various prosthetic knees for the developing and developed world.
Adapted and updated from [2]
26
Chapter 3
Design and Testing of the Prosthesis
Mechanism
3.1
Introduction
For developing a functional knee prosthesis prototype, a structured design process was
employed as discussed in this chapter. Structured design processes have been shown
to facilitate 'deterministic' designs [44-46], as they enable breakdown of a complicated
design challenge into smaller modules, based on the established functional requirements.
A Structured design process can minimize risks and redundancies at each
of the stages of the process.
It can also make it easier to spot inefficiencies and
unexpected outcomes in the functions of the physical embodiment of the proposed
design [44]. The specific process implemented for the design of a low-cost prosthetic
knee prototype in this work is schematically represented in Figure 3.2. The process
is divided into three basic phases: strategy, concept development and division of the
selected concept into smaller modules.
This chapter discusses the design strategy, based on the analysis of gait biomechanics,
employed for building early prototypes. This is followed by detailed explanations of
functions of each of the modules in the selected concept. Iterations of earlier concepts
are briefly discussed, followed by user-centric testing done in India. The feedback col-
27
lected from those user trials served as a preliminary evaluation of each of the modules
in the prototype, which are discussed in detail at the end of this chapter.
Concepts for
Modules
Concepts
Strategy
Modules
j
Functional
Requirements
Prototype
-R-
-
-
I
I
I
I
I
I
I
I
I
-~
%
Iterative prototyping and testing
-
-
I
I
I
I
I
I
I
I
I
I
Figure 3-1: Structured design process used for deterministic design, also referred to as
bottom-up design process in Machine design [441. After functional requirements are
established through user-centric techniques [46], strategy decisions are made based on
choices of technology or approach. Different concepts are then generated, and evaluated through engineering or scientific analysis, brainstorming and prior art analysis.
After a concept is selected, it is broken down into many modules, each of which
might be designed and prototyped independently. Some of these modules may be
more critical than the rest, because failure of those critical modules can lead to a
complete prototype failure. After each module is validated, modules are built together into the prototype. Iterative prototyping, testing and evaluation ensures that
the prototype meets functional requirements of the design problem. Alternative approaches, such as the top-down product design approach may also be employed for
the functional requirements which may not be addressed by a bottom-up machine
design approach [44].
3.2
Design Strategy
The design strategy employed for the mechanism design was to use the movement of
center of pressure (COP), through stance, as an indicator of the state and progress of
28
gait cycle during stance. This approach was particularly useful as a design strategy,
as the COP travel from the heel region to the toe region of the foot is invariant across
different gaits, able-bodied or otherwise. As the COP travels from the heel to the
toes, the Ground Reaction Force (GRF) causes extension or flexion moment at the
knee based on the position of the COP. This strategy has been used by polycentric
knee joints, such as the 4 bar and 6 bar knee joints, albeit it is articulated as "follow-
ing the load line" in literature [18,28,29,47].
Beyond stability, achieving correct kinematics and kinetics during stance involves
early stance flexion (kinetic energy storage) followed by extension (kinetic energy release). This early stance flexion-extension involves energy storage and energy release
in nearly equal proportion [48]. A late stance flexion of up to 45-50 degrees with appropriate damping is also essential for a smooth transition into swing. However, most
passive knee designs, as discussed above, do not facilitate appropriate early stance
flexion-extension and appropriately timed late stance flexion.
As discussed in the
next section, our prototype aims to tackle this tradeoff between kinematics, kinetics
and stability.
3.3
Mechanism Architecture and Function
Figure 3-2 shows the CAD model of the prosthesis assembly and the mechanism of the
knee joint designed to replicate the desired behavior of the prosthetic knee as shown
in Figure 2-2. The three axes crucial for the mechanism's function are the knee axis,
the early-stance flexion (ESF) axis and the locking axis (Figure 3-2). In the following
sections, the mechanism's operation for appropriate prosthetic knee function through
stance and swing phases of level ground walking are described with detailed illustrations.
Figure 3-3(a), which is a sectional view of the disassembled mechanism, shows parts
labeled 1, 2 and 3 mounted on different control axes. Part 1 is connected directly to
29
Socket
"t-i
- AXIS
Prosthetic
Knee Axis
knee
Pylon
Locking Axis
Prosthetic foot
Pylon Connector
Figure 3-2: Prosthesis assembly. The three axes crucial for mechanism's function
shown: The early-stance flexion (ESF) axis, the knee axis and the locking axis.
the socket and is mounted on the ESF axis shared by part 2. The knee axis is housed
within part 2, which is the main rotating element through late stance and swing
phases.
Part 3 is a latch that houses the locking axis and serves as a lock during
early stance and mid stance. Part 3 is directly connected to the lower leg assembly
comprising the pylon and the prosthetic foot. Part 4 is a housing that functions as
the connecting element between part 2 and part 3 and is mounted through both the
knee axis and the locking axis. During late stance and swing phases, part 4 and part
3 rotate about the knee axis along with rest of the lower leg assembly.
3.3.1
Early stance lock
At full swing extension before heel-strike, the mechanism is locked against flexion by
spring-loaded part 3 (Figure 3-3(a)), which jams part 2 (on to which part 1 is also
mounted on the ESF axis). The extension stop bumper prevents the lower leg assembly from hyperextending during standing, stance, or at the end of the swing phase.
30
At heel-strike
3b
Early Stance
Flexion-extension
usoce p
Adjustable spring
position for
variable torsional
Isifess
'
Red dots
represent the
relative
,
Knee
Axis
I
.
location of the
three axes
(Fig. 2)
L4
A
GRF
Posterior to all t ree
axes causing flexion
moment about all three
GRF
remains posterior to all three axes
Lock remains
engaged due
to flexion
moment at
locking axis
initially for early stance flexion about the
ESF axis. As GRF moves anterior to the
ESF axis (orange line) during early
stance, the lower-leg assembly extends
back to zero degrees
axes
Figure 3-3: Operation of the mechanism through the early stance phase. (a) shows the sectional view of the mechanism in the
locked position. Part 1 is connected rigidly to the socket and mounted on part 2 through the ESF axis. Part 3 is mounted on
locking axis shared by part 4. Part 4 is the connecting plate between part 2 and part 3 through the knee axis. The curved red
arrow indicates the flexion moment about the locking axis that keeps the latch engaged due to jamming between part 3 and
part 2. (b) shows early stance flexion-extension made possible between the thigh and the lower leg by the mechanism. Parts
2,3 and 4 (locked together) flex and extend about the ESF axis. Flexion of not more than 20 degrees is allowed. Note that (b)
state is followed instantaneously after (a) during the gait cycle. A different color for each part has been used only for illustrative
clarity.
At the beginning of the gait cycle, i.e.
at heel-strike and during early stance, the
Ground Reaction Force (GRF) is posterior to all the three axes causing a flexion
moment about them.
This flexion moment about the knee axis and locking axis
reinforces the locking action between the latch (part 3) and part 2 (Figure 3-3(a)).
3.3.2
Early stance flexion-extension
During early stance, the flexion moment about the ESF axis causes the combined
assembly of part 2, 3, 4 and the lower leg to flex about the ESF axis (Figure 3-3(b)).
Similar to the normative function of a physiological knee [481, this flexion compresses
the spring between part 1 and part 2, storing and releasing energy during early stance.
As discussed in figure 2-2, based on the Ktance coefficient (2.86
_-, g) [2,3], a spring-
stiffness of 270 kN/m was selected to provide the required torsional stiffness about
the ESF axis. The adjustable lever arm between the ESF axis and the spring (Fig.
3b) is provided by a slot, which facilitates the tuning of torsional stiffness for a range
of equivalent able-bodied body-weight range (50kg to 90kg, including the weight of
the prototype).
The hard-stop between part 1 and part 2 at complete spring compression prevents
early stance flexion of more than 20 degrees (Figure 3-3(b)). As the center of pressure
for the GRF travels anterior during early stance, it then causes an extension moment
about the ESF axis (while still providing flexion moment about both the knee axis
and the locking axis that keeps the lock engaged). During this early stance extension,
the spring releases the stored energy from early stance flexion. This early-stance elastic flexion-extension is crucial for reduced metabolic expenditure as the residual hip
muscles would not need to apply the necessary resistive moment during early stance
flexion and an active moment during early stance extension. Equally important is
the able-bodied kinematics enabled because of this elastic cushioning in the knee at
heel strike (the 20 degree early stance flexion-extension in figure 2-2).
32
3.3.3
Flexion during late stance and swing
The horizontal position of the locking axis is located so that the GRF vector moves
anterior to the locking axis during mid-stance, which causes an extension moment
about the locking axis (as well as the knee axis, Figure 3-4(a)).
This extension
moment causes the latch (part 3) to unlock, freeing up the lower leg assembly for
flexion. During mid-stance, the GRF vector is anterior to the knee axis resulting in
extension moment about the knee axis. As a result, the lower leg assembly cannot flex
yet (unless the hip exerts a large flexion moment). However, during late stance, the
GRF vector acts from the toe region and passes posterior to the knee axis but remains
anterior to the knee axis (Figure 3-4(b)).
The flexion moment during late stance
because of the posterior GRF vector and the flexion moment exerted by the residual
hip musculature initiates late stance flexion about the knee axis, which continues into
swing (Figure 3-4(b)). During flexion, parts 1 and 2 remain rigid together (because of
the stiff spring preventing any flexion about the ESF axis). Parts 3, 4, the pylon and
the prosthetic foot comprise the lower leg assembly, which rotate together about the
knee axis for flexion and extension. Mechanical contact between part 1 and part 3
serves as the hard-stop to prevent any accidental flexion over 90 degrees during swing
phase (normative peak flexion required during swing is around 65 degrees [1, 48]).
The gait cycle is completed by late-swing extension as the lower-leg assembly freely
extends back to zero degrees.
This mechanism needs the user to extend the knee
completely up to zero degrees for the spring-assisted lock to be engaged before heelstrike (Figure 3-4(b)). At full extension, the mechanism reverts to the locked state
(Figure 3-3(a)).
3.3.4
Differential Damping System
To achieve normative kinematics and to minimize metabolic energy expenditure,
dampers are needed to dissipate knee power during late-stance and swing phases
(Figure 2-2). As demonstrated by Narang et al [2,3], the optimal zero-order damping
moment for resisting flexion during late-stance and swing is almost 4 times the value
33
4a
4b
Late-stance and Swing
Mid-stance
\m
\,
hy perextension
lo ;k to counter
th e extension
oment at the
knee axis
Hip moment exerted
through
late stance and
swing for knee flexion
2- -- Latch disengaged
3-
e
as GRF moves
anterior to the
locking axis
4-
Inthe absenc~e of GR
during swing, the latch
is ready to lock at full
extension before heel
strike (Fig. 3a)
GRF
Anterior to all three
axes causing extension
moment about all three
axes
GRF
remains anterior to the locking axis (maintaing the latch disengaged) but
moves posterior to the knee axis before toe-off causing a flexion moment
about the knee axis.
Figure 3-4: Operation of the mechanism through the mid-stance, late stance and swing phases. (a) shows the unlocking of the
latch due to extension moment about the locking axis during mid-stance. As the GRF vector passes posterior to the knee axis
(b), the lower leg assembly flexes about the knee axis. During swing, the spring-bias keeps the latch (part 3) ready for locking
at full extension.
of zero-order damping moment resisting extension during swing (the ratio of B!iex
(0.29
N-m)
to Bext (0.069 N-m) [2, 3]).
In this mechanism, this is achieved by a
differential friction-braking system that uses a one-way bearing as a clutch.
Figure 3-5 shows the implementation of the differential damping system in the mechanism. The two radially spaced dampers with brake-pad components (hereby referred
to as the large damper and the small damper) apply equal normal force (symbol N)
from both sides on part 4. This normal force is equal to the tension in the knee
shaft exerted by the lock-nut fastened on to the external thread on one end of the
shaft. High tension (up to as high as 5 kN) in the shaft is made possible by the use of
Belleville washers compressed to the rated pre-load by the lock-nut. The knee shaft is
coupled to part 3 using a key and part 4 is mounted on the knee shaft with lubricated
bushings, which facilitate smooth rotation and serve as low-cost bearings. When the
mechanism is unlocked and is free to flex during late-stance phase and swing phase,
the friction pads on both the large and small dampers apply total resistive friction
torque, Tfiex, on part 4. The small damper is coupled with a key to the knee shaft and
slips against part 4 in both directions of rotation. The large damper is mounted on a
one-way roller bearing, which resists rotation during flexion but rotates freely during
extension along with part 4. The large damper, therefore, does not exert any resistive
torque during the extension phase of swing. The relative size of the dampers is determined by the ratio of the damping coefficients
(B
) from the following relations
(derived in [49]):
(21LN(R? -r
)
(3.1)
Tflex
BflexW =
2
3(R- r )
Text
-
Bflex
B(R
3(R2
- r2)
2 N(Rr)
13(Rr)f
Text = BextW =
Tflex
2p.N(Ri - r.)
+
(3.2)
1
(R3 - r3)(R - r!) +
- rs)(Ri - r,)
(3.3)
J
where Tflex is the total resistive friction torque applied by dampers during flexion of
35
Knee
Shaft
Lock nut
Large
damper
II
BellevilleJ
washer
7
-3
Small
One-way
roller dutch
dampe
-4
Knee shaft
coupled to Part 3
Small damper
coupled to knee
shaft
-,
Flanged bushing
for free rotation
large damper and
Braking
pads on
small
damper
exert
frictional torque
about knee axis
Figure 3-5: Differential Damping System- exploded view and cross sectional view of
the mechanism. (Bottom) The dampers carry radially spaced brake pads, inner and
outer diameters are labeled (see equations 3.1, 3.2 and 3.3).
36
late-stance and swing, Text is the total resistive friction torque during swing extension
(applied only by the small damper as the larger damper does not slip on part 4 and
rotates along with part 4), N is the normal force between the damper and part 4,
R, and r, are the outer and inner diameters respectively of the large damper, R,
and r, are the outer and inner diameters respectively of the smaller damper (Figure
3-5), Bflez (0.29 N-m) and Be.t (0.069
N-m)
are the damping coefficients based on
a previous study (Figure 2-2 and [2,31), W is the body weight of the user in kg, p
(0.55) is the kinetic coefficient of friction between part 4 (made of Aluminum 7075
alloy for this prototype) and brake pads (which are made of a composite material).
The mechanism is tuned to the desired damping values based on the weight of the
user. This is achieved by increasing or decreasing tension N in the knee shaft by
loosening or tightening the lock-nut against the Belleville washer [49] on the knee
shaft (Figure 3-5).
3.3.5
Prototype iterations
An earlier embodiment of the latch mechanism was prototyped and tested, demonstrated in Figure 3-6 (hereby referred to as the alpha prototype [46]) . The differential damping mechanism in the alpha prototype was prototyped by a braking surface
mounted on a spring loaded one-way roller clutch. The braking action was implemented by the interaction between the braking surface and the rotating joint. The
magnitude of damping, however, was not tunable and wasn't quantified for the alpha
iteration.
3.4
User Trials
During the course of prototype development, early stage user trials were conducted
twice at the Jaipur-Foot clinic (in India) in order to validate the major functional
features of the mechanism.
Two subjects were tested with the most recent proto-
type (Section 3.3), with the help of trained prosthetists at the Jaipur-Foot clinic in
February 2015. Earlier in August 2014, three subjects were also tested with the alpha
37
CAD Model
Assembled
prototype
Pylon and
prosthetic foot
Sectional views
(disassembled)
Thigh Socket
connector
Fricton
Knee axis
-
Damper
mounted on
one-way
clutch
Lock Engaged,
-
Friction tuner
using
Flexion moment
at the locking
axis for lock
adjustable
spring preload
engagement
Locking axis
Ready to
flex
LockJ
disengaged
The cam surface
engages with
Extension
moment
disengages the
the preloaded
damper and
slips against the
damper during
lock
flexion and rolls
during
extension,
thereby
providing
Spring bias for
the lock. At the
end of swing
phase the
differential
spring bias will
lock the knee at
full extension
before heel
damping
Adjustable
strike of the
horizontal
position of the
pylon and foot
next stance
phase
Figure 3-6: Architecture and function of the earlier iteration of the prosthesis mechanism. During early stance, GRF direction causes flexion moment at the locking axis
while the lock is engaged. During mid stance, the extension moment at the locking
axis disengages the lock. During late stance, when the GRF vector passes posterior
to the knee axis, late stance flexion at the unlocked knee joint can take place.
38
prototype (Figure 3-6) in an identical trial.
3.4.1
Trial Protocol
The primary objective of the two trials was to establish the level of functionality of
the mechanism's locking feature and the differential damping system and whether
that enabled a smooth stance to swing transition, as observed by the prosthetists and
as experienced by the subjects themselves. The objective of the trial was not to carry
out a conventional kinematic gait-analysis study but to establish early-stage feasibility of the mechanism architecture based on user-feedback. The MIT Committee on
the Use of Humans as Experimental Subjects approved this field validation study [50].
Subjects were selected from the sample of above-knee prostheses users available at
the Jaipur-foot clinic based on the following criteria:
1. Willingness to participate in the study: voluntary consent was sought by communicating the process of the study in complete detail in the local language
(Hindi). The consent form is included in the first appendix.
2. Experience of at least one year of using an above-knee prosthesis after amputation.
3. Users with only unilateral amputations were considered.
4. Medium length of the stump of the residual limb: users with very short or very
long residual limb were not considered for the purpose of the study.
This is
because a very short or long residual limb demands a more complicated socket
and prosthetic knee layout.
5. Only male users were selected as female prosthetists culturally trained to fit the
prostheses for female users were were not available during trials.
After the selected subjects were fitted with the prototype prosthesis with the help
of prosthetists at Jaipur-foot clinic, the study began by an extended period (about
39
30 minutes) of acclimatization, which involved walking using training rails for safety.
After the subjects felt confident, the two-minute walk test [301 was conducted multiple
times (typically five to six iterations) with rest periods between each of the iterations.
Each two-minute walk test involved walking for about two minutes at self selected
cadence without the use of training rails. Subjective, visual gait observations were
made by prosthetists during these tests. Subjects were also video-recorded during
tests using a hand-held video camera for detailed observations later.
The verbal
comments made both by the prosthetists and the users during the tests were recorded
in written notes for later analysis. At the end of two-minute walk tests, subjects were
asked open-ended questions about their walking experience for detailed qualitative
feedback which was also recorded. Subjects were then asked to answer the following
open-ended questions:
1. Please describe your experience of walking with the prototype prosthesis.
2. How would you compare the performance of this prototype prosthesis with the
prosthesis currently being used by you?
3. What features of this prototype prosthesis do you like? What features do you
not like?
4. How can this prototype be improved in future? If those improvements were
made, would you permanently replace your existing prosthesis with the improved prototype prosthesis?
3.4.2
Trial Results
All five subjects were able to walk in the 2-minute walk test after a period of acclimatization and learning to use the prototype knee in both the trials. Three of the five
subjects were also able to disengage the lock easily midway through stance (Figures
3-7, 3-8), which enabled late-stance flexion and a smooth transition from stance to
swing. This was ascertained through visual observation by the prosthetists and oral
40
Figure 3-7: Preliminary user trial with the alpha prototype (Figure 3-6) to validate
the early stance latch design. (a) Subject 1 using the alpha prototype for the 2-minute
walk test. (b) Subject 2 during the 2-minute walk test, late stance flexion of up to
40 degrees can be seen. (c) Subject 2 walking outdoors on uneven terrain.
41
Figure 3-8: User trials with the second prototype (Section 3.3) to further test the
early stance latch design along with the additional differential damping system. (a)
and (b) show the prosthesis assembly as fitted on to the subject. Late stance flexion
after disengagement of the latch in the knee mechanism shown in (c) and (d).
42
feedback from the subjects. All subjects also provided useful feedback for future improvements of the mechanism design. One such important observation was that the
disengagement of the lock (Figure 3-4(a)) led to a small amount of hyperextension of
the lower-leg assembly, which was deemed uncomfortable by both the subjects. The
engagement of the lock before stance was found to be loud and was reported as an
undesirable feature by each subject. None of the subjects felt the prosthesis to be
heavy in comparison to their current prosthetic devices. The differential damping
effect was observed during the late stance and swing phases. However, the spring
meant for early stance flexion-extension (Figure 3-4(b)) was not effectively engaged
by the subjects during trials. A possible reason for this was that both the subjects
were not confident to shift their body weight forward (as compared to the able-bodied
individuals) during the transition from double support to single support phase of early
stance [1]. In the future trials, this may be addressed by further training and acclimatization focused on a more natural transition from double support to single support
of the amputated foot after heel-strike.
43
44
Chapter 4
Discussion and Future Work
4.1
Discussion
An important design strategy for this mechanism was using the movement of the center of pressure from foot's heel region to the toe region (and the resulting change in
direction of the GRF vector) as an indicator of the phase of the gait cycle. Polycentric knees also similarly utilize the GRF vector direction to enable late-stance flexion
but do so at the cost of achieving ideal kinematics of stance-swing transition [28,29].
Polycentric knees with extremely posterior instantaneous centers of rotation make
it difficult for the users to transition from stance into swing because of delayed late
stance flexion. With the instantaneous center of rotation set closer to the knee axis,
the prosthesis becomes unstable with a higher risk of a fall, in case of accidental flexion. In our design of the early-stance lock, it was possible to precisely position the
locking axis (with respect to the knee axis and the foot) by using the center of pressure data and the GRF data. By locating the locking axis in the correct horizontal
position, we ensured that the lock disengages only after the early flexion-extension
phase of stance but before the engagement of the damper during late stance flexion
phase. Similarly, the ESF axis was located horizontally based on the GRF vector so
that early stance flexion-extension was made possible. The locking mechanism used
in the design is similar in function to the one used by Andrysek [291 in the LCKnee.
However, our mechanism implements a simpler architecture of the lock by using only
45
a single lever for the latch acting posterior to the knee axis (part 3), which may lend
itself for lower cost manufacturing.
Based on the earlier analysis of the swing phase by Narang and Winter [2,4], which
requires two damping torques of different magnitudes, we argue that an extensionassist spring may not be necessary for accurate swing phase control. This is demonstrated by the differential damping mechanism in our prototype.
Extension-assist
springs have been used widely in many passive above-knee prostheses [10, 51] for
achieving high-resistance flexion during late stance and early swing (when the spring
stores energy) and resistance-free extension during swing (when the spring releases
the stored energy). Use of extension springs without sufficient damping leads to terminal impact at the end of swing phase [10, 51]. This is also far from the ideal in
terms of normative kinetics, as springs do not dissipate energy from the system. Prosthetic knee designs with extension springs commonly use viscoelastic dampers [51] to
cushion the impact at the end of swing extension, further adding to the cost and
functional complexity of the product. Basing our design on a theoretical analysis,
we implemented a novel differential damping system in our prototype with the aim
of achieving relatively resistance-free extension and negligible terminal impact. This
design feature, however, was not conspicuous to the two subjects in the field trial
(Figure 3-8) and must be validated through quantitative gait analysis in the future.
There are a few functional limitations in the current design of the mechanism. The
current design necessitates full extension of the knee at the end of swing phase (Figure
latestance(b)). Failure to lock the knee before stance can lead to unstable stance and
possible buckling of the joint. This may further increase the risk of accidental falls
for some users. Use of friction based braking pads for damping may not be robust,
as past designs of knees with friction dampers have been reported to have variable
damping due to wear, changes in humidity and exposure to outdoor dust and rain [18].
During the field validation trial, both the subjects deemed the mild clicking sound
of the latch as undesirable as they felt it made their disability conspicuous to others.
46
Secondary user needs of Indian transfemoral amputees such as squatting, cross-legged
sitting [21 were also not addressed by this prototype.
4.2
Future work
Our future work to develop this design further will take the limitations of the design
into account, as discussed in the previous section of this chapter. Clinical gait analysis
of subjects fitted with the prototype, with appropriately tuned spring stiffness (for
early stance flexion-extension) and damping friction (for late-stance and swing), will
be carried out for kinematic evaluation of the mechanism. These experimental results
would then be benchmarked against the predictions of theoretical analysis, based
on which our prototype was designed.
Additional design features critical for the
robust performance of the prosthesis will need to be incorporated in future iterations,
such as enabling activities of running, sitting cross-legged and squatting. One of the
assumptions of the present work is that by using simple mechanical components such
as springs and friction pads, it is possible bring down the manufacturing cost of the
prosthetic knee. Future efforts in designing a low-cost knee would need to validate
this assumption with a detailed manufacturing process analysis, sourcing costs of
materials, labor costs of assembly and distribution costs.
47
48
Appendix A
Research Protocol Documents
49
50
Prof. Amos Winter
Murthy Arelekatti
Permission is requested to receive informed consent orally, ratherthan via a signature. There is a concern that asking
the subjects to sign informed consent documents may be offputting, since signaturesare often associatedwith
government documents andsome users have limited writingability and may be embarrassed. Hence, the investigator
proposes readinginformed consent information to subjects, askingfor oral consent, and only proceeding in conducting
interviews if oral consent is received.
CONSENT TO PARTICIPATE IN STUDY (SCRIPT FOR INDIAN PARTICIPANTS)
I am a graduate student from the Massachusetts Institute of Technology in the US working with BMVSS to design a
new and improved prosthetic knee. We have a new prosthetic knee here that we would like to test with your help. If
you choose to participate, I would like to have you try out a second prosthetic limb in addition to the limb that you will
be receiving today. I'd have you walk in both prostheses for a while to get used to them, then walk in the gait lab with
them to record information about how you walk in them. Lastly I would ask you what you like and what you do not like
about both of the prostheses.
Please listen to the following information and ask questions about anything you do not understand, before deciding
whether or not to participate. You will be given a written copy of what was just read to you.
- The study is entirely voluntary. You can stop at any time for any reason. The study will be done while you are here,
and will take about two hours total.
- You will not be compensated for participation in this study, and your treatment at BMVSS will not be affected at all
by anything you say or do in this interview.
- Unless you give us permission to use your name, title, and/or quote you in any publications or presentations that may
result from this research, the information you tell us will only be shared along with all of the other information we get
from other participants.
- There is an organization in the United States called the FDA which monitors studies that involve medical devices,
such as prosthetic feet. At some point in the future, they may review this study. If that is the case, authorized
representatives from the FDA may see your name and information you provide during this study. These representatives
are required by law to keep this information confidential. It will only be used to review the study and will in no way be
available to anyone else, either in the US or in India.
- In addition to the video that we take in the gait lab, we may take pictures and/or videos of you using the prosthetic
knee if you allow us to do so. The pictures and videos would be used in presentations and publications. This is
voluntary. You can participate in the study even if you do not want pictures or videos taken other than the video
required in the gait lab.
EMERGENCY CARE AND COMPENSATION FOR INJURY
If you feel you have suffered an injury, which may include emotional trauma, as a result of participating in this study,
please contact the person in charge of the study as soon as possible.
In the event you suffer such an injury, M.I.T. may provide itself, or arrange for the provision of, emergency transport or
medical treatment, including emergency treatment and follow-up care, as needed, or reimbursement for such medical
services. M.I.T. does not provide any other form of compensation for injury. In any case, neither the offer to provide
medical assistance, nor the actual provision of medical services shall be considered an admission of fault or acceptance
of liability. Questions regarding this policy may be directed to MIT's Insurance Office, (617) 253-2823. Your
insurance carrier may be billed for the cost of emergency transport or medical treatment, if such services are
determined not to be directly related to your participation in this study.
Subject consents to
Participation
Use of direct quotes in publications and presentations
Use of photographs and/or video in publications and presentation
(yes / no)
(yes / no)
(yes / yes (masked) / no)
This testing will be completed by Aug 20th, 2014
Please contact Murthy Arelekatti at murthya@mit.edu, phone: +91-9880388224 or +1-(617)-417-6434 with any
questions or concerns.
If you feel you have been treated unfairly, or you have questions regarding your rights as a research subject, you may
contact the Chairman of the Committee on the Use of Humans as Experimental Subjects, M.I.T., Room E25-143b, 77
Massachusetts Ave, Cambridge, MA 02139, phone +1-617-253-6787.
Prof. Amos Winter
Murthy Arelekatti
Permission is requested to receive informed consent orally, rather than via a signature. There is
a concern that asking the subjects to sign informed consent documents may be off putting,
since signatures are often associated with government documents and some users have limited
writing ability and may be embarrassed. Hence, the investigator proposes reading informed
consent information to subjects, asking for oral consent, and only proceeding in conducting
interviews if oral consent is received.
CONSENT TO PARTICIPATE IN STUDY (Translated Script)
-+7 TfT
af 7-r 97 7 (-iwrd Te9T) %*319 Pr , AT TA-;iT V, k i\T BMVVSS * M Tw
Tq7 -f\T
-T * wr
#
sft -mTr
7wrf -t 'M
#
1t
e 79T twr
t iirrrTiri# -r aiI
r ftbiTIf 4 1f4vv
afrq-* TiK -frw
-;r 3T7
* 3Tafti
-4 -r
wq 3Tff Tt-uq 4
wr wr
T
rr T frftit4 T-+ T NT w
rWr t i wf
1 W ff9T
wq
7g *-#vrtIA--4
-wqrsr ' tTrgh
II
~ff~T1 --44':1te-44
:' I'
wq * zft9 3% w
fk7T #T4
# FDAe (reqleg) rrwm abiT rf/w ftft
|
-fl 7'; r7 ery 7 9-T wT zf 7T
g~ift n r 3wu-h#
ET fi p
o fkbrE -ift 4
,TR ftr aff, 7f 7sKrt *r -,17
q* 9Tw i ai1[- i u k qifi q f
kk * wr;7r WT9* BMVSS T7
7 fk*
-cT7
Tgl 7Taqwl Tft fkzr wrlm i
wcSry 4 -Mr e Trdf \'n I I - rfi1 3TNq fhft
<51 MqIWl4 MII--If 47 -T TWWT +- I - qft se~i
7 TTrgrftM # - r Wqr* Mrsq, qft 3TP-r/A 7-97 TT ma 4 1r a a
47 R Tft ft I4
*
alilftRn -4 FDA (\T
)
qTTN/ 4199 1 *7-r n fft~M l \6 ' 1T f14KA~
,i-ir aTrm* -,rq
7Ttf vftffir Tf aTTzR zf *wr *;
T4
-W* I -+Tft 4 # T
A il1(Iw TErw NtW I -T-q Vftff * ~KrM 3114- -7 A 1 11 1 44f
q TMTr Treg Fr 7-crW
TqW 3TtszR
Tq -g
zTrat -r wr fzfh
WTq%q~4r aleG~ f ft MREh *T < It
r- r r q rirr 1 RTM I - T W T-M 3Wrq -asf
4 7
r7T : -7f -1
-4g~i # rTffRm tT
I Tf
9
'M -1+ R * R-T - R M19* tI Ft ' Vr M R 41 -ft 4 1
;qI * P17g - 311 11, 1;iffe %rtq;[T IJWTi
7w 'TF
.
zmSrg aTr(:r
Tf ;3Trqf lfw4 -qT \cTw-4ft a i 14- * -, i
ue wq-, wm- r qTtfr f t
-T mT~rr -Trc-T -r
9-7wr t, DtW -c-i 7; tW FTsvvff 4 wa7 4 wRFf *r #qi-t-
qzT # r- T fz 4-, M.I.T. Efi PI ;g
rr T
,T WI ;31 hI
td1 fi r 7ri*rFftmrr
A 114 47 fW7 4 1TqT -b w
3th 3TT47fi , 9 q I * 4, -9T srf3f
~r f ftwr M qw,
Mq3r;T~ * fk7
Wc-+W1T7 rTk t. M. I.T. q-t * 917g gTrqw * ft* aT-e F-7rr w ;T-7ru
t. ft*
* H N @1i, -r *r VFcIT f
ft- M I Fia T Wrfrf Wk * f4\7, at ; ft
ft MI
Ik e Eft C I a Ni-
VMTfrf 4rr 1 Tr q1< R,7 zft -zft
7 FlT -b i 1 ft 77 N i irC fimr wrg. Tf *ff Tr' ;r -4w
L a~( 1, H31t 94t9T -l ie i "q, (6 17 ) 2 5 3-2 82 3 * f 4r7 fH ffrK ft -zT W t. i* * NTi * 7 q 4 Wzr 4
aTq-q4 iqwfl * ftr 4 -ft -tN I% #k fk7\tR9Vft-I TTT -# W T-F4T rT,
S(1-
9_
T M ft Mr Is
_
_
-91<T
_
qrri
_
(WjTIf)
__f__T
r W 4T 1a r -q-'TrT
VFXR it<3 T4
3ft / -4T ~Efilt R IT d9
(Wr / Mt)
_
)
_h
(-T /4T(119)T
-47
A_-I_ o _T, _o _-TT
(Ii(w 7T7T
-Tt- wr vr ff -4 "7,ff w witl" (Murthy Arelekatti )W
4_
311 4 1-ti 4;-r
* i:\T iffw *RT Wr 9-.r t.
__I_
w.TrRTI 3fT w;fr#
[TrTfT
Zf
#19-0
:
t4 r murthya@mit.edu
RT %Tq Trf
4rb wdt Ti-4w Trm t zrr wr
T RTi-tT - 3Tyk a1ftyR
-
zft aiT~ -Wmt -cf 3;F
-qfT - vqN f, -er #-4
0
Chairman of the Committee on the Use of Humans as Experimental Subjects, M.I.T.,
Room E25-143b, 77 Massachusetts Ave, Cambridge, MA 02139
-frM 1-617-253-6787 1iten414
W
Bibliography
[1] Jacquelin Perry and Judith M. Burnfield. Gait Analysis: Normal and Pathological Function. SLACK Incorporated, 2nd edition, 2010.
[21 Yashraj S. Narang. Identification of Design Requirements for a High-Performance
, Low-Cost , Passive Prosthetic Knee Through User Analysis and Dynamic Simulation. Master's thesis, Massachusetts Institute of Technology, Cambridge MA,
May 2013.
[3] Y. S. Narang, V. N. M. Arelekatti, and A. G. Winter V. The Effects of Prosthesis
Inertial Properties on Prosthetic Knee Moment and Hip Energetics Required to
Achieve Able- bodied Kinematics. (In review), 2014.
[4] Yashraj S Narang and Amos G Winter. The Effects of the Inertial Properties
of Above-Knee Prostheses on Optimal Stiffness , Damping , and Engagement
Parameters of Passive Prosthetic Knees. In Review, pages 1-20, 2014.
[5] Yashraj S Narang and Amos G Winter. Using Biomechanical and HumanCentered Analysis to Determine Design Requirements for a Prosthetic Knee for
Use in India. In Review, pages 1-20, 2014.
[6] Yashraj S Narang and Amos G Winter. Effects of prosthesis mass on hip energetics, prosthetic knee torque, and prosthetic knee stiffness and damping parameters required for transfemoral amputees to walk with normative kinematics. In ASME 2014 InternationalDesign Engineering Technical Conferences and
Computers and Information in Engineering Conference, pages V05AT08A017V05AT08AO17. American Society of Mechanical Engineers, 2014.
What we do: Above-Knee
[7] Bhagwan Mahaveer Viklang Sahayata Samiti.
Prosthesis.
http://jaipurfoot.org/what-we-do/prosthesis/above-knee_
prosthesis.html (Accessed 5/19/14).
[8] Observational Gait Analysis. Los Amigos Research and Education Institute,
Inc., Pathokinesiology Service and Physical Therapy Department, Rancho Los
Amigos National Rehabilitation Center, 4th edition, 2001.
[9] Jessica Rose and James G. Gamble, editors.
Williams and Wilkins, 3rd edition, 2005.
51
Human Walking.
Lippincott
[101 Dominik Wyss. Evaluation and design of a globally applicable rear-locking prosthetic knee mechanism. Master's thesis, University of Toronto, 2012.
[11] Jason M. Wilken and Raul Marin. Care of the Combat Amputee, chapter 19.
Dept. of the Army, 1st edition, 2010.
[12] Alvin L. Muilenburg and A. Bennet Wilson Jr. A manual for above-knee
(trans-femoral) amputees. http: //www. oandp. com/resources/patientinfo/
manuals/akindex. htm (Accessed 5/9/13).
[13] Arthur D. Kuo and J. Maxwell Donelan. Dynamic principles of gait and their
clinical implications. Physical Therapy, 90(2):157-74, 2010.
[14] Dale Berry. Microprocessor prosthetic knees. Physical Medicine and Rehabilita-
tion Clinics of North America, 17:91-113, 2006.
[15] World Report on Disability. Technical report, World Health Organization, 2011.
[16] Guidelines for Training Personnel in Developing Countries for Prosthetics and
Orthotics Services. Technical report, World Health Organization, 2005.
[17] Samuel R. Hamner, Vinesh G. Narayan, and Krista M. Donaldson. Designing
for Scale: Development of the ReMotion Knee for Global Emerging Markets.
Annals of Biomedical Engineering, 41(9):1851-9, September 2013.
[18] Jan Andrysek. Lower limb prosthetic technologies in the developing world: a
review of literature from 1994-2010. Prosthetics and Orthotics International,
34(4):378-98, 2010.
[19] D Cummings. Prosthetics in the developing world: a review of the literature.
Prosthetics and orthotics international,20(1):51-60, April 1996.
[20] I.C. Narang, B.P. Mathur, P. Singh, and V.S. Jape. Functional capabilities of
lower limb amputees. Prosthetics and Orthotics International, 8:43-61, 1984.
[21] Dinesh Mohan.
A report on amputees in India.
Orthotics and Prosthetics,
40(1):16-32, 1986.
[22] W Yinusa and ME Ugbeye. Problems of amputation surgery in a developing
country. Internationalorthopaedics, 27(2):121-124, 2003.
[23] Bruce Rybarczyk, David L Nyenhuis, John J Nicholas, Susan M Cash, and James
Kaiser. Body image, perceived social stigma, and the prediction of psychosocial
adjustment to leg amputation. Rehabilitation psychology, 40(2):95, 1995.
[24] Olga Horgan and Malcolm MacLachlan. Psychosocial adjustment to lower-limb
amputation: a review. Disability & Rehabilitation, 26(14-15):837-850, 2004.
52
[251 Susan J. Mulholland and Urs P. Wyss. Activities of daily living in non-Western
cultures: Range of motion requirements for hip and knee joint implants. International Journal of Rehabilitation Research, 24(3):191-8, 2001.
[261 Janet Fricke. Activities of daily living. In J.H. Stone and M. Blouin, editors,
InternationalEncyclopedia of Rehabilitation. 2013. http: //cirrie. buf falo.
edu/encyclopedia/en/article/37/ (Accessed 4/29/13).
[27] Ottobock. Reimbursement by product. http: //professionals. ottobockus.
com/cps/rde/xchg/ob-us-en/hs.xsl/48354.html?id=48372
(Accessed on
05/19/14).
[28] C.W. Radcliffe. Four-bar linkage prosthetic knee mechanisms: kinematics, alignment and prescription criteria. Prosthetics and Orthotics International, 18:159-
73, 1994.
[29] Jan Andrysek, Susan Klejman, Ricardo Torres-Moreno, Winfried Heim, Bryan
Steinnagel, and Shane Glasford. Mobility function of a prosthetic knee joint
with an automatic stance phase lock. Prosthetics and Orthotics International,
35(2):163-70, 2011.
[30] Richard W. Baker. Measuring Walking: A Handbook of Clinical Gait Analysis.
Mac Keith Press, 1 edition, 5 2013.
[311 ZDNet. Ossur rheo knee. http: //www.zdnet.com/pictures/photos-hightech-prosthetics-1-1-70/ (Accessed 05/19/15).
[32] Ernesto C. Martinez-Villalpando and Hugh Herr. Agonist-antagonist active knee
prosthesis: A preliminary study in level-ground walking. Journal of Rehabilitation Research & Development, 46(3):361-74, 2009.
[33] Frank Sup, Amit Bohara, and Michael Goldfarb. Design and control of a powered transfemoral prosthesis. The International Journal of Robotics Research,
27(2):263-73, 2008.
[34] B S Farber and J S Jacobson. An above-knee prosthesis with a system of energy
recovery: a technical note. Journal of rehabilitationresearch and development,
32(4):337-48, November 1995.
[35] Conversation with BMVSS leadership in Jaipur, India in January 2012.
[36] Rob Goodier. A durable, cheap prosthetic knee is tested for developing countries,
March 2010. https: //www. engineeringf orchange. org/news/2010/03/13/a_
durable.cheapprosthetic-kneeistestedfor-developing-countries.
html (Accessed 5/1/13).
[37] Rob Goodier. A low-cost prosthetic gives legs to amputees in the developing
world, March 2011. https: //www. engineeringf orchange. org/news/2011/03/
06/a-low-cost-prosthetic.giveslegs-to-amputees-inthedeveloping_
world.html (Accessed 5/2/13).
53
[38] Jane Langille. $50 artificial knee wins global health innovation award, November 2012.
http: //janelangille. com/50-artificial-knee-wins-globalhealth- innovation-award/ (Accessed 5/2/13).
[39] Bloorview Research Institute. Commercialization: Low-cost prosthetic knee
joint.
http://www.hollandbloorview.ca/research/commercialization/
kneejoint.php (Accessed 5/2/13).
[40] Niagara Prosthetics & Orthotics. Niagara knee joint kits.
products/howtoorder . html (Accessed 5/2/13).
http://npoi.ca/
[41] Phone conversation with Aulie Devices, Inc. in February 2012.
[42] Ottobock.
C-Leg microprocessor-controlled prosthetic knee.
http:
//professionals.ottobockus.com/cps/rde/xbcr/ob-us-en/C-LegReimbursementReferenceGuide_12072401.2.pdf (Accessed 5/2/13).
http://professionals.
Genium bionic prosthetic system.
[43] Ottobock.
ottobockus.com/cps/rde/xbcr/ob-us-en/GeniumReimbursement_
ReferenceGuide_12072401.1. pdf (Accessed 5/2/13).
[441 Alexhander H. Slocum. FUNdaMENTALS of Design. Open Access, 2008.
[45] Robert R Donaldson. Deterministic approach to machining accuracy. Technical
report, California Univ., Livermore. Lawrence Livermore Lab., 1972.
[46] Karl Ulrich and Steven Eppinger. Product Design and Development, 5th Edition.
McGraw-Hill/Irwin, 5 edition, 5 2011.
[47] C.W. Radcliffe. Functional considerations in the fitting of above-knee prostheses.
Artificial Limbs, 2:35-60, 1955.
[48] David A. Winter. Biomechanics and Motor Control of Human Movement. John
Wiley & Sons, Inc., 4th edition, 2009.
[49] Richard Budynas and Keith Nisbett. Shigley's Mechanical Engineering Design
(McGraw-Hill Series in Mechanical Engineering). McGraw-Hill Science/Engi-
neering/Math, 9 edition, 1 2010.
[50] The committee on the use of humans as experimental subjects, 2015.
[51] Alex Furse, William Cleghorn, and Jan Andrysek. Development of a Lowtechnology Prosthetic Swing-phase Mechanism. 31(2):145-150, 2011.
54