A Low Cost Design of Powered Ankle-Knee Prosthesis for Lower
Limb Amputees
Preliminary Results
Ruben C. Martinez, Roberto L. Avitia, Miguel E. Bravo and Marco A. Reyna
Department of Bioengineering and Environmental Health , Autonomous University of Baja California,
A. Obregon Ave., Mexicali, B.C., Mexico
Keywords: Powered Prosthesis, Series Elastic Actuator, Clutch, Amputee, Lower Limb.
Abstract: In this paper we described a new kind of a powered knee and powered ankle prosthesis for individuals who
have suffered a complete or partial lower limb amputation. Our prototype prosthesis consist of two modules,
the ankle module and the knee module. The first contains a unidirectional spring configured in parallel with
a force-controlled actuator. This spring is intended to store energy in dorsiflexion, and then released it to
assist power plantar flexion. The knee modules consist of a series of elastic clutch actuators and a
unidirectional spring positioned in parallel to the motor. Preliminary results show two modules designs and
ankle module prosthesis prototype almost complete. Two modules working together will enhance the
performance of amputee individual, producing more natural gait and reducing the metabolic cost at walking
in level-ground.
1 INTRODUCTION
Individuals with lower limb amputation have shown
to expend more metabolic energy than an individual
with a healthy leg during normal walking.
Transfemoral amputees expend up to 60% more
metabolic energy compared with healthy subjects
(Waters, 1976). A transtibial amputee tends to
expend 20-30% more metabolic energy in normal
walking (Colborne, 1992). Thus, both cases of
amputation tend to walk more slowly than an
individual with healthy lower limbs. In addition
amputees exhibit asymmetric gait patterns compared
to non-amputees (Winter, 1991).
Currently most of the commercial prostheses
available are passive prostheses. These are not able
to bring positive work at phase stance, also have
increased the risk of joint and back pains. Some
researchers have shown that powered prostheses for
lower limb are able of mimic human gait. They can
provide negative and positive work in the stance
phase, as wells as to improve amputees performance
in a more natural gait and normal walking (Au,
2009; Martinez-Villalpando).
Ideally, a good design of prosthesis needs to
have some characteristics described as: (1) be able to
produce net power to the gait; (2) the lowest possible
energy consumption; and (3) should not exceed the
weight and the height of the missing limb.
Improvements in elastic elements to prosthetic
devices have been shown several advantages. These
include, increasing tolerance to the load impact,
stored and released energy, as wells as reducing
energy requirements to the actuation and increasing
peak power output (Grimmer, 2012). Therefore
many applications have been developed as result of
these advantages such as it has been implemented
successfully in many applications, for example in
exoskeletons devices, active ortheses and robotic
legs (Dollar, 2008; Zoss, 2005).
Currently most of the developed power
prostheses are still on development stage. One of
them is the MIT powered ankle-foot prosthesis that
has been developed using series elastic actuators and
parallel springs. This has shown to reduce the
metabolic cost of walking in transtibial amputees
(Au, 2009). F. Sup et al. have developed a powered
transfemoral prosthesis which has incorporate a
spring in parallel to the ankle joint to reduce peak
motor torque requirements and to increase the
bandwidth (F. Sup, 2009). Another example is the
power knee developed Also Bellman et al proposed
a prosthetic foot that it used successfully elasticity
253
C. Martinez R., L. Avitia R., E. Bravo M. and A. Reyna M..
A Low Cost Design of Powered Ankle-Knee Prosthesis for Lower Limb Amputees - Preliminary Results.
DOI: 10.5220/0004914402530258
In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2014), pages 253-258
ISBN: 978-989-758-013-0
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
elements, they called it The AMP-Foot prosthesis
and it has a mechanism that permit storage and
release of the energy spring and later releasing it
when they needed (Bellman, 2008). This will permit
the use of smaller actuators and less energy
requirements. Therefore this will reduce the overall
prostheses weight, a crucial issue in the development
of a powered prosthesis.
In this paper we proposed to develop a power
knee and ankle prosthesis divided in two modules
that can work together or separately to be used in
individuals with a complete or partial amputation.
This prosthesis should be able to produce net
positive and net negative work during the stance
phase of normal walking, thus it will satisfy the
requirements of normal gait at level-ground walking.
We described design procedures, as well as
preliminary results, presenting power ankle
prototype module completely.
2 MATERIALS AND METHODS
Prosthesis proposed has two modules, the knee and
the ankle module. The purpose of designing the
prosthesis in two modules is for a wider range of
amputees have the opportunity to use this prosthetic
device. It would be used by transtibial amputee
individuals, as well as transfemoral amputee
individuals. Therefore a person with a transtibial
amputation should use only the ankle module, and in
the case of a transfemoral amputee it can use both
modules or choose only one.
Both modules are able to work together in
coordination and each one can work independently
on of the other. Both designs were made in
SolidWorks ® (see Figure 3 and Figure 4).A simple
description of two modules is described in Figure 1,
where we can appreciate springs and clutch needed
for both modules that are considered in design as
crucial elements working as storage-release energy
jointly with actuators.
The chassis was designed to contain the
actuators, on the other hand, the electronic and the
battery sections are placed on the right side of the
prosthesis. The structure of the prosthesis except to
the foot was made with aluminum 6061 T6.
The prosthesis range motion measured in degrees
was established as close as possible to the leg-
human range motion during level-ground walking,
as we can see in Table 1.
Figure 1: A simple description of springs and clutch used
in prosthesis proposed. Two modules are presented
separately by ankle module and knee module.
Table 1: Maximum joint motion ranges used in power
prosthesis design.
Phase / Condition
Human
Walking
Max.
Powered
Prosthesis
Max.
Ankle dorsiflexion 14.1
0
15
0
Ankle plantar flexion 20.6
0
25
0
Knee flexion 73.5
0
120
0
2.1 Knee Design
The actuator features transmissions comprises of a
3:1 timing belt drive coupled to a ball-screw (NSK
10x3 mm) in series with a spring. The translational
movement of the ballscrew converts an angular
rotation motion of the knee via the series spring with
a moment arm 0.045m.
During the leg-human motion as described in
Figure 2 we can appreciate early stance phase of the
knee flexion and extension, the knee torque-angle
relationship behaves like a spring and it takes
approximately 40% of the gait cycle. Thus this
period is linear and consists of the greatest positive
and negative mechanical power phase required
during the gait cycle. The series elasticity was
chosen to mimic this linear region. Thus, choosing
the correct spring knee prosthesis can mimics the
same behaviour of a biological knee at stance phase.
Spring stiffness that we have chosen satisfies
requirements of 74 kN/m and provides a rotational
stiffness of 150 Nm/rad to the powered knee
module. The series elastic actuator (SEA) uses a
brushless DC motor (Maxon® EC-Max 30). A
unidirectional spring was placed in parallel to the
actuator to assist to motor when the knee is more
than 74
0
degrees in flexion. The stiffness of the
tension spring is 86 kN/m.
Adding a clutch in parallel to the SEA can be
reduced the electrical energy consumption in the
actuator; therefore we opted to add a clutch to our
BIODEVICES2014-InternationalConferenceonBiomedicalElectronicsandDevices
254
Figure 2: Walking phases described by prosthesis. Clutch is activated during heel strike to mid-stance phase emulating a
passive spring.
design (Rouse, 2013). When the clutch is disengage
of the transmission the powered knee it behaves like
a SEA, and when the clutch is engage the prosthesis
behaves like a passive spring. The clutch is
disengaged automatically when the spring release all
stored energy.
On the top and in the bottom of the chassis was
placed a pyramidal connector which is connected to
the socket and the extension tube respectively of the
amputee (See Figure 1 and Figure 1).
Figure 3: Internal schematic of power knee prosthesis and
its corresponding associate sections.
2.2 Ankle Design
The Ankle prosthesis is composed of an actuator in
parallel with a unidirectional spring. For driving
train system, we used a brushless DC motor
(Maxon®, EC-max 30, 60 Watts) that drives a ball-
screw (NSK®, 10x3 mm) via timing belt that drives
the transmission with a 3:1 ratio. The ball-screw is
coupled in series with low profile prosthetic foot,
(Ossur®, Flex foot). This elastic leaf spring was
used to emulate the function of a human foot. It
provides shock absorption, stores energy during
early stance, delivers energy during late stance, as
well as to minimize the ground reaction shock to the
transmission. At the top of the structure we placed a
pyramidal connector (See Figure 3 y Figure 4).
During the stance period the human ankle can
reach a velocity of 5 rad/s. In the table 2 it can be
seen that the ankle module can achieve a velocity of
3.45 rad/s, therefore it cannot satisfy that
requirement. However this problem can be solved
changing the motor for another one that be more
faster. A good solution is the brushless DC motor
Maxon EC-Powermax 30 of 200 W. This motor is
similar in size but with almost than thrice the
nominal speed and torque, compared with the motor
that we are using. Therefore doesn’t affect the
design of the actuator. This change will enhance the
performance of the actuation, increase nominal
velocity and the output torque of the actuator.
The ankle actuator incorporates two springs with
stiffness of 162 KN/m in parallel with the motor.
The purpose is to supplement power output during
plantar flexion. The parallel spring is unidirectional,
and it is used only to provide a rotational stiffness
value of 518 Nm/rad when the ankle angle is greater
ALowCostDesignofPoweredAnkle-KneeProsthesisforLowerLimbAmputees-PreliminaryResults
255
than zero.
Table 2: Physical parameter specifications of the powered
ankle prosthesis.
Variable Value
Max. dorsiflexion 15
0
Max. plantar flexion 25
0
Height
Weight
Peak output torque
Peak velocity
Parallel offset stiffness
27 cm
2.5 kg
135 Nm
3.45 rad/s
324 kN/m
Figure 4: Internal schematic of power ankle prosthesis and
its corresponding associate sections.
2.3 Sensing System
Sensing system is a crucial section required to get
kinematic and kinetic information related with
prosthesis movements that should be processed on
real-time by controller algorithm.
There are several methods for measuring force
and torque. We choose three different types of
sensors for this application. For measuring the
ground reaction force and the sagittal plane moment
we opted to use load cells, due to their simplicity in
design, long lasting, and easy to implement in the
control architecture.
The load cells are strategically placed on
prosthesis, and each module includes a uniaxial load
cell (Omega®). Since impedance changes in a strain
gages are very small, strain gages were connected
using a Wheatstone bridge circuit configuration.
The sagittal plane moment was measured above
the knee joint at the socket Interface. For measuring
the force ground reaction we placed the load cell at
the top of the ankle module inside the socket of the
pyramidal connector (see Figure 3 and Figure 4).
Once we have measured force ground
reaction
in the
load cell we proceed to calculate the torque
that would be on the ankle in a static situation using
equation 1:
(1)
Where
is the force ground reaction vector,
is
the force ground reaction to ankle vector, and is
the ankle torque calculated.
In this prosthesis we used a linear potentiometer
on the series spring of the knee and on the parallel
spring of the ankle to measure the spring deflection
and then estimate the joint torque.
We used a position sensor to determine the angle
between knee and ankle joint. Since each joint is
actuated along a single axis in the sagittal plane, a
single position sensor can be placed on each joint. A
rotary encoder was attached directly on each joint; it
was possible to have a more accurate measurement
of the angle for each joint. In order to have a better
control for each actuator, we placed an optical
encoder on the shaft of both actuators.
For reading inclination angle of the residual limb
of the amputee it needs a system that provide us
measurements in real-time that can be later utilized
by the control system to execute movement
commands of the prosthesis. Therefore we opted to
utilize the combination of two sensors; a 3-axis
gyroscope (L3GD20) and a 3-axis accelerometer
(MMA7341L).
One gyroscope and one accelerometer were used
for each joint. This method provides flexion and
extension angles by estimating acceleration of the
joint centre of rotation. Another application in which
we used the accelerometer is in the estimation of the
ground slope. In order to estimate the ground slope,
we used the accelerometer in tangential direction.
Assuming that the foot is positioned in parallel with
the ground, the only component of acceleration
present is the gravity; the gravity direction vector
was calculated and computed by monitoring the
variation in the acceleration component orthogonal
to the long axis of the foot.
3 RESULTS
In this paper, we presented preliminary results of a
design and manufacturing of the powered knee-
ankle prosthesis prototype. We made some trials of
prototype functionality of the powered ankle
separately as indicated in Figure 5.
It has been performed studies of finite element
analysis by computer for all the structural elements
that are part of the prosthesis, this with the purpose
BIODEVICES2014-InternationalConferenceonBiomedicalElectronicsandDevices
256
of ensure that the device can withstand dynamic
loads for subjects with a weight less than 85 kg.
These studies were not addressed in this paper. The
mechanism of the ankle module has demonstrated to
provide a full range of motion of 40 degrees, more
than necessary for represent a normal walking. On
the other hand the ankle module is capable to
generate a torque of 135 Nm, enough to mimic the
biological ankle torque at level-ground walking. The
springs that are attached in parallel with the actuator
on the ankle module can be replaced by composite
materials and this change will permit reduce the
overall weight of the ankle module.
The addition of elastic elements and a clutch
mechanism to our design has been demonstrated
theoretically to reduce significantly the electrical
energy requirements required by the motor.
Simultaneously we amplified the force bandwidth of
the actuator. Thus it was permitted that the
prosthesis performs more complex tasks where we
need more power on the actuation, such as walking
upstairs and down stairs, as well as standing from a
seated posture.
Figure 5: Physical prototype of the powered ankle module.
4 CONCLUSIONS
In this paper we describe the design of a prototype
of a powered ankle- knee prosthesis that is human-
like in weight, size and functionality. The
architecture that comprises this prosthetic device
permits to mimic the behaviour of human leg in
normal walking. Consequently it would be capable
to satisfy the necessary requirements such as, torque
and movements range, performing these tasks with
economical electrical-energy consumption. The
prototype of the ankle module is fully assembled as
can be seen in Figure 5. The weight and the size of
the ankle module can be reduced optimizing the
design and replace some elements that it would
permit enhance the torque bandwidth. The knee
module is on manufacturing process, once
completed this process it will start to test the control
system algorithm that will control the actuation of
the prosthesis
ACKNOWLEDGEMENTS
We would like to thanks to sponsors of Bio-
Prosthesis project UABC and Idalia Martinez. Thank
God for giving us strength and knowledge to make
this project.
REFERENCES
A. M. Dollar and H. Herr., 2008. Design of a quassi-
passive knee exoskeleton to assist running, IEEE/RSJ
International Conference on Intelligent Robots and
System, Nice, France.
Au, S. K., Weber, J., and Herr, H., 2009. Powered Ankle-
Foot Prosthesis Improves Walking Metabolic
Economy, Robotics, IEEE Transactions. 25, (1), pp.
51-66.
A. Zoss, H. Kazerooni, and A. Chu, 2005. On the
Mechanical Design of the Berkeley Lower Extremety
Exoskeleton (BLEEX), IEEE/RSJ International
Conference on Intelligent Robots and System.
D. A. Winter, 1991. The Biomechanics and motor control
of Human gait: Normal, elderly and pathological,
Waterloo Biomechanics, vol. 2.
E. J. Rouse, L. M. Mooney, E. C. Martinez-Villalpando
and H. M. Herr, 2013. A clutchable series-elastic
actuator: design of a robotic knee prosthesis for
minimum energy consumption. IEEE International
Conference on Rehabilitation Robotics.
E. Martinez-Villalpando, and Herr, H. M. Herr., 2009.
Agonist-antagonist active knee prosthesis: A
preliminary study in level-ground walking. Journal of
Rehabilitacion Research & Development. Vol.46, no.
3, pp. 361-73.
F. Sup, H. Varol, J. Mitchell, T. Withrow, and M.
Goldfarb, 2009. Self-contained powered knee and
ankle prosthesis: Initial evaluation on a transfemoral
amputee, Rehabilitation Robotics, ICORR, IEEE
International Conference. Pp. 638-644.
G. R. Colborne, S. Naumann, P.E. Longmuir, and D.
Berbrayer, 1992. Analysis of mechanical and
metabolic factors in the gait of congenital below knee
amputees: A comparsion of the SACH and Seattle
feet, Ammer. J. Phys. Med. Rehabil.,vol. 71 no. 5, pp.
272-278.
ALowCostDesignofPoweredAnkle-KneeProsthesisforLowerLimbAmputees-PreliminaryResults
257
M. Grimmer, M. Eslamy, S. Gliech, and A. Seyfarth,
2012. A Comparison of Parallel- and Series Elastic
Elements in an actuator for Mimicking Human Ankle
Joint in Walking and Running, IEEE International
Conference on Robotics and Automation, RiverCentre,
Saint Paul, Minnesota USA.
R. Bellman, A. Holgate, T. Sugar, 2008. SPARKy 3:
Design of an active robotic ankle prosthesis with two
actuated degrees of freedom using regenerative
kinetic, in Proc. IEEE/RAS-EMBS Int. Conf. Biomed.
Robotics Biomechatron., pp. 511-516.
BIODEVICES2014-InternationalConferenceonBiomedicalElectronicsandDevices
258
