A Mechatronics-twin Framework based on Stewart Platform for

Effective Exploration of Operational Behaviors of Prosthetic Sockets

with Amputees

Dejiu Chen

, Suranjan Ram Ottikkutti

and Kaveh Nazem Tahmasebi

Unit of Mechatronics and Embedded Control Systems, KTH Royal Institute of Technology, 10044 Stockholm, Sweden

Keywords:

Mechatronics-twin, Stewart Manipulator, Transfemoral Amputee, Prosthetics, Human-in-the-Loop,

Cyber-physical System, Biomechanical Modeling, Force-control.

Abstract:

A Stewart platform is a six-degree-of-freedom parallel manipulator widely used as the motion base for

ﬂight simulators, antenna positioning systems, machine tool technology, etc. This work presents a novel

mechatronics-twin framework that integrates such a manipulator with advanced biomechanical models and

simulations for effective exploration of operational behaviors of prosthetic sockets with amputees. By means

of the biomechanical models and simulations, the framework allows the users to ﬁrst analyze the fundamental

operational characteristics of individual amputees according to their speciﬁc body geometries, pelvis-femur

structures, sizes of transfemoral sockets, etc. Such operational characteristics are then fed to one Stewart plat-

form as the reference control signals for the generation of dynamic loads and behaviors of prosthetic sockets

that are otherwise difﬁcult to observe or realize with the real amputees. Experiments in form of integration

testing show that the proposed control strategy is capable of generating expected dynamic operational condi-

tions. Currently, the mechatronics-twin framework supports a wide range of biomechanical conﬁgurations and

the quantiﬁcation of the respective intra-socket load conditions for socket design optimization and anomaly

detection.

1 INTRODUCTION

Limb amputations cause serious physical disabilities

that compromise the quality of life of many people

around the globe. Limb prostheses offer a solution

to reduce the negative impact of such disabilities, at-

tempting to restore a normal functionality and am-

putee autonomy, as much as possible. It is estimated

that 90% of amputees will wear a prosthetic limb for

the rest of their lives. At present, despite some im-

portant recent advances in prosthetics, 40 to 60% of

amputees exhibit a rather low satisfaction level due to

comfort issues (Baars et al., 2018).

As a critical interface between the amputee (nat-

ural) stump and the prosthetic (artiﬁcial) device, a

suitable prosthetic socket must ensure efﬁcient ﬁtting,

appropriate load transmission, stability, and control.

The performance often constitutes a key factor for the

success or failure of the prosthesis itself. The opti-

mization of prosthetic sockets is however a difﬁcult

https://orcid.org/0000-0001-7048-0108

https://orcid.org/0000-0002-0699-3889

https://orcid.org/0000-0002-1685-5586

task as each solution is inherently individual, while

suffering from the fact that a wide range of opera-

tional conditions can only be partially observable or

quantiﬁable. Such conditions are typically related to

the dynamic load distribution, stump volume ﬂuctua-

tion and tissue evolvement, etc. For example, in cur-

rent practices, the load bearing capability of the stump

can only be checked by prosthetists using “touch and

feel” technique. The socket-related issues that are of

concern range from reduced bio-mechanical ﬁtness,

hampered dynamic control, to poor comfort and med-

ical complications (e.g. skin lesions).

This paper presents a novel mechatronics-twin

framework that addresses the above-mentioned chal-

lenge by serving as an analytical replica for revealing

the complex operational interplay of amputee, pros-

thetic device and prosthetic socket. The overall ap-

proach is characterized by an integration of (a) virtual

behaviors based on a combination of advanced biome-

chanical modelling, simulation, and FEA (Finite Ele-

ment Analysis); and (2) physical behaviors based on

well-controlled motions of a six-degree-of-freedom

parallel manipulator referred to as Stewart platform.

Chen, D., Ottikkutti, S. and Tahmasebi, K.

A Mechatronics-twin Framework based on Stewart Platform for Effective Exploration of Operational Behaviors of Prosthetic Sockets with Amputees.

DOI: 10.5220/0010838600003123

In Proceedings of the 15th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2022) - Volume 1: BIODEVICES, pages 74-83

ISBN: 978-989-758-552-4; ISSN: 2184-4305

Figure 1: Overview of the mechatronics-twin that allows both virtual and physical replications of prosthetic device. The virtual

replication is supported by (1) biomechanical modelling and simulation; (2) FEA; The physical replication is supported by

(3) 3-D printing and integrating; (4) physical testing by Stewart platform.

In particular, while the virtual behaviors are useful for

establishing basic understanding of fundamental bio-

mechanical operational conditions and interactions of

individual amputees, the physical behaviors allow a

reﬁned investigation of such operational conditions

and interactions that are otherwise difﬁcult to ob-

serve or realize with the real amputees. Moreover, the

framework also aims to constitute an important ba-

sis for successful deployment of next generation ﬂex-

ible wearable sensors inside bionic prosthetic sockets

for real-time monitoring of dynamic load conditions.

Due to their inherent physical ﬂexibility, such sensors

could suffer from some performance concerns, relat-

ing typically to the measurement sensitivity, inaccu-

racy and drift (Dejke et al., 2021). By generating op-

erational data with well deﬁned ﬁdelity, such virtual

and physical behaviors allow the training and testing

of data-driven algorithms for the sensor functions.

Currently, the mechatronics-twin framework sup-

ports a wide range of bio-mechanical conﬁgurations

and the quantiﬁcation of respective dynamic opera-

tional situations including the intra-socket load con-

ditions. This is done by the following technical steps:

(1) biomechanical modelling and simulation of over-

all gait dynamics; (2) FEA for basic analysis and vir-

tualization of possible intra-socket load conditions;

(3) 3-D printing and prototyping speciﬁc test stumps

and sockets conﬁgurations for physical tests; (4) phys-

ical testing by Stewart platform with the test stumps

and sockets. An overview of this mechatronic-twin

framework can be seen in Figure 1. The rest of this

paper elaborates these technical steps. It is struc-

tured into the following sections: Section 2 discusses

related concepts and technologies in the domains of

prosthetic design, operation perception and analysis.

Section 3, 4, 5, and 6 elaborate the support for mod-

elling, simulation, prototyping and testing. The re-

sults from a case study is presented in Section 7. Fi-

nally, the conclusion is given in Section 8.

2 RELATED WORK

Transfemoral (above knee) amputation is a surgical

procedure performed to remove the lower limb above

the knee joint when that limb has been severely dam-

aged via trauma, disease, or congenital defect. Es-

sentially, the usage of prosthetic device aims to re-

store the ambulation and self-esteem of amputee to

the maximum extent. The major components of a

transfemoral prosthesis include socket, suspension,

knee joint, pylon and feet (Geng et al., 2012). Among

these prosthetic components, the socket serves as the

interface between the residual limb and the prosthe-

sis. It must protect the residual limb and appropri-

ately transmit the forces associated with standing and

ambulation.

A perception of the intra-socket load conditions

would help the prosthetists optimize the design of

prosthetic devices for improved operational ﬁtness

and comfort. Due to the inherent complexity and

variability of human body and prosthesis operation,

such a perception would also constitute the most im-

portant basis for enabling data-driven analyses where

machine-learning and artiﬁcial intelligence methods

are employed for the modeling and exploration of

complex operational conditions. The state of the art

approaches to modern limb prosthesis are therefore

searching for an integration of mechanical, electronic,

and computing technologies to support novel sen-

sory and data analytic capabilities. This is, however,

A Mechatronics-twin Framework based on Stewart Platform for Effective Exploration of Operational Behaviors of Prosthetic Sockets with

Amputees

always a challenging task. Normally, direct sens-

ing of the pressure conditions is restricted by spe-

ciﬁc requirements on sensor deployment and perfor-

mance. For example, all modern pressure sensor tech-

nologies, including resistive transducer, piezoelectric

transducer, optical pressure transducer, and capac-

itive transducer (Bao., 2000), have their respective

disadvantages. For successful usage of such technolo-

gies, sensor testing and calibration become therefore

important. For the calibration of sensors, it is always

important to take the actual operational condition into

the consideration. One reason for this is that the

speciﬁc structural conditions of each individual pros-

thetic socket, relating for example to the curvature and

surface hardness, could lead to the problems of sen-

sor performance (Khodasevych et al., 2017). Differ-

ent bio-mechanical conditions and operational behav-

iors also lead to varying dynamic conditions that in

turn may interfere negatively with the sensor drift and

hysteresis, as well as other frequency response char-

acteristics (Buis and Convery, 1997). Moreover, sen-

sor re-calibration could also be necessary in order to

compensate for the drift over life cycle.

This paper presents an approach to a novel

mechatronics-twin framework for prosthetic devices

is similar to the notion of digital-twin in regard to the

replication of a physical target system on the basis

of measurements (Boschert and Rosen, 2016). The

approach addresses however in particular the chal-

lenge of complex biomechanical dynamics of pros-

thesis operation as well as the need of physical replica

for data generation and sensor calibration. Clearly,

any approaches that rely on repeated experimentation

on the amputees will not be preferable. This would

in the worst case cause further trauma to the am-

putees. Therefore, in the mechatronics-twin frame-

work, a six-degree-of-freedom parallel manipulator

referred to as SP (Stewart Platform) has been adopted

for a reﬁnement of the digital virtual replication given

by modeling and simulation.

Since being ﬁrstly introduced in 1949, various SP

based solutions are widely used as the motion base

for antenna positioning systems, machine tool tech-

nology, ﬂight simulators, etc. Today, there are over

1400 research articles about manipulators analysis

and design. See e.g. (Furqan et al., 2017), (Wapler

et al., 2003), (Grace et al., 1993), (Brandt et al.,

1999) and (Fichter et al., 2008). Our approach adopts

a PID (Proportional–Integral–Derivative) strategy to

the motion control of SP, with the control refer-

ence generated by biomechanical simulation and real-

time feedback from the operation of SP. Similar ap-

proaches to SP control can be found in (S¸umnu et al.,

2017) and (Rossell et al., 2015).

3 BIOMECHANICAL

MODELLING AND

SIMULATION

This technical step aims at eliciting the most fun-

damental operational characteristics of a prosthetic

device as an integral part of amputee. Within the

mechatronics-twin framework, it provides the support

for estimating the piston-forces and moments within

the amputee stump-socket assembly during walking.

Knowledge of such physical interactions is essential

for more detailed analysis of stump and prosthesis dy-

namics. All tasks are based on OpenSim, which is an

open source tool for the modeling, simulating and an-

alyzing of neuromusculoskeletal systems (Delp et al.,

2007).

The work starts with a quantiﬁcation of amputee

body geometries, sizes of the pelvis-femur structure

and prosthetic socket based on a combination of mea-

surement and estimation. These geometries are not

only essential for visual representation of the bod-

ies, but are also important in the estimation of piston-

forces with inverse dynamics when combined with a

conﬁguration of associated body masses. To achieve

a good scaling in biomechanical models, proper mea-

surement of the limbs are required to determine the

joints connecting the various bodies. The masses of

the residual limb are initially approximated by equat-

ing the volumetric density of the default healthy limb

to the residual limb. The Static Optimisation Tool of

OpenSim is then used for a further adjustment of these

masses regarding the kinematics and ground reaction

forces of test-subject. A similar approach is used

to develop a transfemoral biomechanical model in

Figure 2: A snapshot of biomechanical model and related

gait phases.

Figure 3: Positioning femur in a transfemoral socket model.

BIODEVICES 2022 - 15th International Conference on Biomedical Electronics and Devices

Figure 4: (a) Basic force tensors determining the intra-socket loads condition; (b) Basic conﬁguration of an amputee leg model

integrating transfemoral socket and stump (Right Leg).

OpenSim to estimate joint torques (Mohamed, 2018).

A biomechanical model for identifying the piston-

forces of transfemoral or transtibial prosthesis within

dynamic gait cycles is then constructed with Open-

Sim as shown in Figure 2. This model includes de-

tailed stump and socket models as shown in Figure

3. The current modeling of transfemoral prosthesis

is based on a reﬁnement of a well deﬁned transtibial

model(Willson, 2017). In particular, numerous adap-

tations are enabled to match the requirement of esti-

mating speciﬁc piston-forces from a transfemoral am-

putee. This includes replacing the transtibial socket,

tibia pylon, remaining tibia, and with speciﬁc trans-

femoral socket and femur conﬁgurations.

During gait cycles, the piston-forces and moments

are related to the contact forces between stump and

socket as shown in Figure 4 (a). The conjunction

of all force vectors F

over the discrete regions of

stump surface t at each speciﬁc gait phase is equiv-

alent to the piston force F

and moments of the same

stump. The contact force F

at each region is a com-

position of normal force experienced from the pres-

sure F

and shear force F

of the same region. For the

mechatronics-twin framework, an amputee leg model,

shown in Figure 4 (b), is used to stipulate the related

multi-body parameters of particular concern. These

include:

• Interface r: representing the imaginary rigid joint

between the femur and the socket located at the

COM (Center of Mass) of the stump.

• Interface SP: representing the joint between the

socket (S) and the femur pylon (P).

The corresponding piston-forces and moments at

these interfaces are calculated according to the sim-

ulated ground reaction forces during gait cycles and

the corresponding multi-body transformation through

the Inverse Dynamics Tool of OpenSim. In Figure

5 and Figure 6, one example of the piston forces

and moments over 7 gait cycles is shown. They

are represented in terms of percentage of gait cycle.

The sampling frequency for the simulation is 100Hz.

These internal joint reaction data are then exported for

the analysis of socket-stump interface conditions (see

Section 4 below).

Figure 5: One example of piston force along X-Axis

(lateral-medial axis) of 7 consecutive gait cycles.

Figure 6: One example of rotation moment around X-Axis

(lateral-medial axis) of 7 consecutive gait cycles.

A Mechatronics-twin Framework based on Stewart Platform for Effective Exploration of Operational Behaviors of Prosthetic Sockets with

Amputees

4 FINITE ELEMENT ANALYSIS

OF LOAD CONDITIONS

The goal of this technical step is to provide an effec-

tive characterization of possible intra-socket load con-

ditions of concern before any further physical tests.

Within the mechatronics-twin framework, it provides

the support for establishing the virtual behaviors relat-

ing to the contact forces on stump surface, based on

the internal joint reaction data from the biomechanical

modelling and simulation. This joint reaction can be

normalized as seen in Figure 5 in an effort to identify

relationships in the gait by comparing metrics such as

standard deviation between various test subjects (Mo-

hamed, 2018). In addition, the FEA also provides

useful insights about contact forces and displacement

on the interface of stump and socket. The analysis is

based on ANSYS workbench, which is a commercial

software tool providing support for advanced FEA.

Due to the complex ﬁne-grained interplay of related

surface conditions, modeling with conventional soft-

ware such as MATLAB Simscape multibody would be

a less preferred option. The work begins with a setup

of the FEA model by importing the geometries of the

stump, socket and femur to the DesignModeler of AN-

SYS workbench. The simulation setup also includes

deﬁning the material properties of the various bodies

such as density, Young’s modulus and Poisson’s ratio.

The calculated internal joint reaction data by Open-

Sim are thereafter used as the inputs to the FEA soft-

ware. See Figure 7 for one example of the derived

pressure load conditions by FEA.

Figure 7: One example of average pressure load of 7 con-

secutive gait cycles (medial posterior view).

5 3D PRINTING AND

PROTOTYPING

This technical step is responsible for producing and

prototyping the femur-stump assemblies and sockets.

Within the mechatronics-twin framework, it provides

Figure 8: The shell mold casting process of silicone stump.

the physical replica of femur-stump assemblies and

prosthetic sockets for further physical testing by the

test-rig. 3D printing is used mainly due to the excel-

lent lead times. The approach also allows a geograph-

ical distribution of activities, e.g. having the patient

measurement in one region of the world and the test-

ing in anther region.

The printing and prototyping of physical replicas

use the measurements of related femurs, stump as-

semblies and sockets. In particular, a physical replica

of socket could be dimensioned by evenly extruding

the geometry of target stump by 3 millimetres ac-

cording to a widely used socket thickness (Jamaludin

et al., 2018). The design could also be based on the

scanning of a socket being used. The stump repli-

cas are produced by a shell mold casting process, for

which the shell is ﬁrst created based on the geom-

etry of stump as seen in Figure 8. Similarly, a fe-

mur prototype is also generated by printing and in-

tegrated within the stump. The related design tasks

are supported by Meshmixer (Schmidt and Singh,

2010), which is a composition tool for arbitrary sur-

face meshes.

In practices, the sizes of these physical replicas

are often larger than the maximum printing volumes

of commercially available 3D printers. Under the cir-

cumstances, the parts are printed separately and as-

sembled together thereafter. By diving the process

into multiple sections, the lead time could also be also

reduced signiﬁcantly.

6 PHYSICAL TESTING BY

STEWART PLATFORM

Within the mechatronics-twin framework, the phys-

ical testing by Stewart platform allows a more de-

tailed investigation of operational behaviors of pros-

thetic devices as an integral part of amputee. The goal

is also to reﬁne, verify and validate the correspond-

ing virtual behaviors using the physical prototypes

of femur-stump assembly and prosthetic socket. The

physical test-rig consists of the prototypes of femur-

stump assembly and prosthetic socket, and a Stewart

BIODEVICES 2022 - 15th International Conference on Biomedical Electronics and Devices

Figure 9: The conﬁguration of physical test-rig with femur-

stump assembly, prosthetic socket, and Stewart platform.

platform, as shown in Figure 9. Each leg of the Stew-

art platform is mounted with a load cell and a linear

actuator for the motion control.

A schematic representation of the Stewart plat-

form is shown in Figure 11 (a). The platform has two

coordinate systems, located at the geometry center of

the moving-platform, P(X

, Y

, Z

), and the geome-

try center at the base-platform (i.e. the ﬁxed plat-

form), B(X , Y, Z). Points b

and p

are the connect-

ing joints to the base and moving platforms be the

leg i, respectively. The key control parameters are

shown in Figure 11 (b) and (c). The big (R-radius)

and small (r-radius) circles represent the base- and

moving-platform respectively. The leg joints are de-

noted by the b

and p

. The ﬁgures also show the the

length of a leg and the relating angle to Z-plane.

As one key step in the process of motion control,

a lumped-parameter model is used to specify the tar-

get plant given by femur-stump assembly and socket.

This method allows the complex operational behav-

iors of prosthetic devices to be captured by parame-

terized spring-mass-damper models. Especially, the

joint movement along each DoF (Degree of Freedom)

will be expressed as follows:

F = k ×x + d × ˙x + m × ¨x (1)

where k is the spring stiffness, d is the damping coef-

Figure 10: The comparison of estimated force by a lumped

parameter model vs. actual force given a ﬁxed displace-

ment.

ﬁcient, m is the mass, and x is the displacement. The

preferred values for these parameters are estimated

according to the MSE (Mean Square Error) between

the estimated forces in Equation 1 and the estimated

forces from the simulations introduced in Section 3

and 4. The performance of a lumped-parameter model

identiﬁed with a pattern search algorithm (Zanetti,

2021) is shown in Figure 10. Currently, our frame-

work supports the estimation of lumped parameters

for each DoF including polar coordinates.

For the motion control of testing, the current

framework adopts a cascaded force-position control

approach, as shown in Figure 12. The design consists

of the following major function blocks:

• Force Control: This is a PID control function for

deriving the desired displacements of the central

moving plate of platform.

• Position Control: This is a PID control function

associated to each leg of the platform for deriving

the required force for the linear actuator based on

the discrepancy between the desired and measured

leg conditions.

• Inverse Kinematics: This is an analysis function

for generating the reference length of each leg of

the platform based on desired position of the cen-

tral moving plate of platform.

• Forward Kinematic: This is an analysis func-

tion for deriving the kinematic conditions of the

geometry center of the moving-platform based on

the measured leg lengths.

As shown in Figure 12, the control starts with

a speciﬁc force reference (i.e. a trajectory of de-

sired reference forces F

re f

to be implemented), de-

rived from the biomechanical modeling and simula-

tion. The controller ﬁrst calculates the differences be-

tween such reference forces with the force feedback

(i.e. a trajectory of actual measured forces F

f eedback

by the Stewart manipulator) as follows:

= F

re f

− F

f eedback

(2)

where the force feedback is given as the sum of mea-

sured force feedback of each individual leg f

, as de-

ﬁned below:

f eedback

∑

i=1

∗ sin θ

(i = 1, 2, ..., 6) (3)

where the force feedback of each leg, f

, is measured

by the load cell mounted on each leg. The angle

denotes the angle of each leg with respect to the

XY plane on the base platform coordinate system as

shown in Figure 11. The value is calculated with the

position sensor feedback.

A Mechatronics-twin Framework based on Stewart Platform for Effective Exploration of Operational Behaviors of Prosthetic Sockets with

Amputees

Figure 11: (a) A schematic view of Stewart Platform(S¸umnu et al., 2017) ; The key leg parameters with (b) side view, and (c)

upper view.

Figure 12: Overview of cascaded force-position control.

The Force Control function derives the desired

displacements of the moving-platform dis

for the ac-

tuation of the desired trajectory as follows:

dis

= (K

+ (K

)

(τ)d(τ) + (K

)

(τ)

(4)

This desired displacement is then combined with

the observed Z-position for deﬁning newly desired Z-

position. The Inverse Kinematics function takes then

the desired Z-position as inputs and derives the corre-

sponding reference length of each leg.

The observation of operational conditions is sup-

ported by the Forward Kinematics function, which

derives the current kinematics conditions regarding

the current positions of each leg. The design follows

the concept presented in (Harib and SrinivasanHood,

2003). With this approach, the moving-platform posi-

tion after each iteration is given based on a Jacobian

matrices. To quickly ﬁnd a good approximation, a nu-

merical solution based on Newton-Raphson method

(Wilson and Sadler, 1993) is used.

The Position Control function takes the desired leg

length references as well as the actual measured leg

positions by the load cell as the input signal. Simi-

lar to force control strategy, the computed errors are

multiplied by PID constants, which have been esti-

mated by trial and error, at each time step. The needed

forces (actuator efforts), F

act

, for each leg are the out-

put of this PID controller which are used to generate

the needed length of each leg to perform the desired

movements of the moving platform of Stewart plat-

form.

BIODEVICES 2022 - 15th International Conference on Biomedical Electronics and Devices

act

= (K

+ (K

)

(τ)d(τ) + (K

)

(τ)

(5)

7 CASE STUDY AND RESULTS

In order to validate the proposed approach, a case

study is carried out using a stump replica casted ac-

cording to the conﬁguration of an amputee. The ge-

ometry and additional data such as weight and height

are identiﬁed from a test led by Ossur Inc. as shown

in Table 1.

Table 1: Test-subject information.

Patient info. Value

Age 42 Years

Body mass 80Kg

Height 180cm

A scan of the amputee’s residual limb, shown in

Figure 13 (a), is performed to generate the 3D geom-

etry data for the 3D printing and prototyping of the

stump and socket replicas. For the case study, a cyclic

gait load is applied to the stump as a well-deﬁned gait

behavior. Through the biomechanical modeling and

simulation (Section 3), the corresponding normalized

piston forces are identiﬁed, shown in Figure 13 (b).

Figure 13: (a) Scanned residual limb of test-subject; (b) The

related piston force during operation.

The entire assembly is 3D printed and casted

within 72 hours by using six Ultimaker 3D Print-

ers, and thereafter integrated with the Stewart Plat-

form. The Stewart Platform takes the normalised pis-

ton forces as the control references and conducts the

testing. Within the Stewart Platform, the load cells

of leg are Vetek VZ-101BH, whereas the actuators are

Transmotec DLA series linear actuators rated for 250

N of nominal dynamic load. The position control

function is implemented with an embedded computer

node based on Arduino Mega Microcontroller. A cus-

tom package is developed in ROS (Robot Operating

System) to support the control of overall behaviors.

The feedback signals from the position sensors and

load cells are collected with this embedded computer

node and passed the other control functions located

at a host PC (running with Linux, a Ryzen 9 mobile

processor and 32GB RAM).

A comparison of the resulting force trajectory

from the test is compared with the reference trajectory

as shown in Figure 14. In the test, the weight of the

platform, femur and stump assembly are subtracted

from the net force vector to effectively apply the gait

piston force load. The result shows that the proposed

control strategy is capable of achieving the required

magnitude and wave form of the dynamic load cy-

cle. Although a phase shift is observed, mainly due to

the computation delay, it would not affect the overall

quality of test regarding the the replication of targeted

mechanical process. The damping of the signal can

be attributed to the weak of knowledge of the linear

actuators. The inertia of the gearbox can also cause

a damping of the system thus preventing a faster re-

sponse. Further tuning and development of the PID

controllers may alleviate these issues.

8 CONCLUSION AND FUTURE

WORK

In this paper, a novel simulation and testing frame-

work for effective exploration of complex opera-

tional behaviors of prosthetic devices in terms of a

mechatronics-twin was presented. It serves as the an-

alytical replica of prosthetic devices with both virtual

and physical behaviors for effective data-driven anal-

ysis and sensor calculation. The approach provides

also a platform for effective optimization of prosthetic

devices by revealing undesired load conditions with-

out real tests by amputees. This would for example

avoid unnecessary trauma to the amputees.

The results by case study show that the proposed

solutions can fulﬁll the expected goals. Additionally,

the current design of this framework opens up many

perspectives for future research. In further studies,

the modeling support can be enhanced with a mix-

ture of heterogeneous friction coefﬁcients as well as

more complex hyperelastic material models for the

stump. The simulation and testing cases can also be

automated for different load or gait conditions. A re-

ﬁnement of the controller design based on for exam-

ple optimal control would also improve the control

performance. Various software and hardware tech-

nologies would also need be explored to support the

implantation of the proposed framework in industrial

scale.

A Mechatronics-twin Framework based on Stewart Platform for Effective Exploration of Operational Behaviors of Prosthetic Sockets with

Amputees

Figure 14: A result of the force control of Stewart Platform.

ACKNOWLEDGEMENTS

This work was supported by the research project

SocketSense (https://www.socketsense.eu/), funded

by the European Union’s Horizon 2020 research and

innovation programme under grant agreement No

825429.

Test-subject data such as residual limb geome-

try, weight, height and age were provided by

Ossur

(https://www.ossur.com) based on a pilot study.

REFERENCES

Baars, E. C., Schrier, E., Dijkstra, P. U., and Geertzen,

J. H. (2018). Prosthesis satisfaction in lower limb am-

putees: A systematic review of associated factors and

questionnaires. Medicine, 97(39).

Bao., M. (2000). Micro mechanical transducers: pressure

sensors, mechanical transducers: pressure sensors, ac-

celerometers and gyroscopes. Handbook of Sensors

and Actuators, Elsevier.

Boschert, S. and Rosen, R. (2016). Digital twin—the sim-

ulation aspect. In Mechatronic futures, pages 59–74.

Springer.

Brandt, G., Zimolong, A., Carrat, L., Merloz, P., Staudte,

H., Lavall

ee, S., Radermacher, K., and Rau, G. (1999).

Crigos: a compact robot for image-guided orthopedic

surgery. IEEE Trans Inf Technol Biomed.

Buis, W. and Convery, P. (1997). Calibration problems

encountered while monitoring stump/socket interface

pressures with force sensing resistors: techniques

adopted to minimise inaccuracies. Prosthetics and or-

thotics international, 21(3).

Dejke, V., Eng, M. P., Brinkfeldt, K., Charnley, J., Lussey,

D., and Lussey, C. (2021). Development of proto-

type low-cost qtss™ wearable ﬂexible more enviro-

friendly pressure, shear, and friction sensors for dy-

namic prosthetic ﬁt monitoring. Sensors, 21(11):3764.

Delp, S. L., Anderson, F. C., Arnold, A. S., Loan, P.,

Habib, A., John, C. T., Guendelman, E., and The-

len, D. G. (2007). Opensim: open-source software

to create and analyze dynamic simulations of move-

ment. IEEE transactions on biomedical engineering,

54(11):1940–1950.

Fichter, E., Kerr, D., and Rees-Jones, J. (2008). The

gough—stewart platform parallel manipulator: A ret-

rospective appreciation. Journal of Mechanical Engi-

neering Science, 223:243–281.

Furqan, M., Suhaib, M., and Ahmad, N. (2017). Studies on

stewart platform manipulator: A review. Journal of

Mechanical Engineering Science, 31:4459–4470.

Geng, Y., Yang, P., Xu, X., and Chen, L. (2012). Design

and simulation of active transfemoral prosthesis. In

2012 24th Chinese Control and Decision Conference

(CCDC), pages 3724–3728.

Grace, K., Colgate, J., Glucksberg, M., and Chun, J.

(1993). Studies on stewart platform manipulator: A

review. Proceedings IEEE International Conference

on Robotics and Automation.

BIODEVICES 2022 - 15th International Conference on Biomedical Electronics and Devices

Harib, K. and SrinivasanHood, K. (2003). Kinematic and

dynamic analysis of stewart platform-based machine

tool structures. Robotica, 21:541–554.

Jamaludin, M. S., Hanafusa, A., Yamamoto, S.-I., Agarie,

Y., Otsuka, H., and Onishi, K. (2018). Evaluation

of the effects of geometrical changes in prosthetic

socket towards transfemoral residuum via ﬁnite el-

ement method. In 2018 IEEE-EMBS Conference

on Biomedical Engineering and Sciences (IECBES),

pages 314–319. IEEE.

Khodasevych, I., Parmar, S., and Troynikov, O. (2017).

Flexible sensors for pressure therapy: Effect of sub-

strate curvature533and stiffness on sensor perfor-

mance. Sensors, 17(10).

Mohamed, A. (2018). Modeling and simulation of trans-

femoral amputee gait. PhD thesis, University of New

Brunswick.

Rossell, J., Palacios-Quinonero, F., Rubio-Massegu, J., and

Vicente-Rodrigo, J. (2015). Tracking control for a

stewart platform prototype. 2015 International Con-

ference on Advanced Mechatronics, Intelligent Man-

ufacture, and Industrial Automation (ICAMIMIA),

0:58–63.

Schmidt, R. and Singh, K. (2010). Meshmixer: an inter-

face for rapid mesh composition. In ACM SIGGRAPH

2010 Talks, pages 1–1.

S¸ umnu, A., G

uzelbey,

I. H., and C¸ akir, M. V. (2017). Simu-

lation and pid control of a stewart platform with linear

motor. Journal of Mechanical Science and Technol-

ogy, 31(1):345–356.

Wapler, M., Urban, V., Weisener, T., Stallkamp, J., D

urr,

M., and Hiller, A. (2003). Studies on stewart platform

manipulator: A review. Transactions of the Institute

of Measurement and Control, 25:279–280.

Willson, A. M. (2017). A Quasi-Passive Biarticular Pros-

thesis and Novel Musculoskeletal Model for Transtib-

ial Amputees. PhD thesis.

Wilson, C. and Sadler, J. (1993). Kinematics and Dynam-

ics of Machinery. Harper Collins College Publishers,

New York, 3rd edition.

Zanetti, L. R. (2021). Lumped parameter and modal models

to simulate ground reaction forces due to running.

A Mechatronics-twin Framework based on Stewart Platform for Effective Exploration of Operational Behaviors of Prosthetic Sockets with

Amputees