Motion Capture and MultiBody Simulations to Determine Actuation

Requirements for an Assistive Exoskeleton

Daniel S

a Pina

1,2

, Joaquim Gabriel

and Renato Natal

Faculty of Engineering, University of Porto, Porto, Portugal

INEGI - Institute of Science and Innovation in Mechanical and Industrial Engineering, Porto, Portugal

Keywords:

Biomechanics, OpenSim, Exoskeleton.

Abstract:

EUROSTAT’s projections show that, by 2040, the people aged 65 or more will account to almost one fourth of

the population. These statistics raise concerns over the sustainability of the society, so technological solutions

have been emerging to prolong the active age of European citizens. One of the main impairments for elders

is an increasing difﬁculty in performing daily lower-limb activities (i.e. walking, climbing stairs) due to

Sarcopenia, among other issues. Therefore, the authors are developing an active exoskeleton whose sole

purpose is to assist the gait of an elderly person. The proposed system is based on a low-proﬁle design,

allowing a smaller frame that allows the device to be worn beneath loose clothing, making it more desirable

to wear in public by reducing social awkwardness. This article shows the methodology used to determine

the actuation requirements for the exoskeleton. Two subjects performed a number of trials depicting daily

life activities in a biomechanics laboratory that acquires motion sensor and force-plate data. Each activity

was performed with additional weights to emulate the presence of an exoskeleton. The data was used in

a multibody simulation program (OpenSim) to determine the requirements (angular speed, torque) for the

actuation system in the exoskeleton.

1 INTRODUCTION

The ageing process in the human being results in sev-

eral changes in the musculoskeletal system. Among

other effects, the muscles shrink and lose mass i.e.

Sarcopenia, the number and size of muscle ﬁbers de-

crease, the tendons and cartilages become less tolera-

ble to stress, the heart lowers the speed at which it can

pump blood and the bones lose mass, becoming more

prone to fractures (des, 2017). One of the ﬁrst major

symptoms that appear with ageing is an irregular gait

and decreased gait speed (Riley et al., 2001).

The appearance of these symptoms usually results

from an increasingly sedentary life. Consequently,

a sedentary life will aggravate the aforementioned

changes in muscle and bone mass reductions. An ab-

sence of muscle stimulation results in loss of mus-

cle mass (Fiatarone and Evans, 1993) and the lack of

stress applied to human bones will prevent the piezo-

electric effect and mechanotransduction that main-

tains their density (Muscolino, 2016).

This results in a continuous self-feeding cycle

where the symptoms contribute directly to worsening

the conditions, as shown in Figure 1.

Figure 1: Cycle of symptoms and consequences during ag-

ing.

Additionally, adopting a sedentary life and moving

less frequently also decreases a person’s conﬁdence

and motor control, increasing the chances of suffer-

ing from falls (Steadman et al., 2003). The lack of

an active life is also associated with social isolation

(Shankar et al., 2011). The end result is a lower qual-

ity of life that can result in depression and other psy-

chological disorders (Steptoe et al., 2013).

There are classical solutions for medium mobility

impairment, such as crutches and walkers. However

they occupy the upper limbs which results in a drastic

Pina, D., Gabriel, J. and Natal, R.

Motion Capture and MultiBody Simulations to Determine Actuation Requirements for an Assistive Exoskeleton.

DOI: 10.5220/0007403601830191

In Proceedings of the 12th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2019), pages 183-191

ISBN: 978-989-758-353-7

183

change of lifestyle and the misuse of these solutions

can cause injuries in the long term. Moreover, the

mobility walkers can be difﬁcult to use within tight

spaces such as house interiors and bathrooms.

Due to an increase in life expectancy and decrease

of birth-rate, many developed countries are suffering

from an ageing population. Europe is the continent

with the oldest population in the world. According

to a recent EUROSTAT report (EUROSTAT, 2014),

close to 20% of the European population is aged 65

or above. This results in almost 85 Million people in

that age bracket. Within these, 23% show a moderate

difﬁculty while walking and 21% show severe difﬁ-

culty, as seen in Figure 2.

Figure 2: Proportions and number of people (in millions) in

the European Union aged 65 years or older with moderate

or severe mobility impairment, according to a 2014 EURO-

STAT study (EUROSTAT, 2014).

EUROSTAT’s projections also show that, by 2040,

the people aged 65 or more will account to almost

one fourth of the population (EUROSTAT, 2017).

Therefore, the ratio between working age (18-

65 years old) and non-working age people would be

close to 2 to 1, which raises questions regarding the

sustainability of the European society (Commission,

2015). Due to this growing concern, many stud-

ies and initiatives (Walker, 2010), (HARTLAPP and

SCHMID, 2008), (Peine et al., 2014) have emerged in

order to extend the active age for European citizens.

Moreover, the age evolution projections would in-

crease the number of people aged 65 and over with

moderate mobility difﬁculties from 20 Million to over

35 Million.

It is therefore of the utmost importance to develop

solutions that contribute to solving the several prob-

lems of mobility within the 3rd and 4th ages, which

are aggravated by the ageing population in the Euro-

pean Union.

The proposal for the project is to develop a low-

proﬁle active exoskeleton with a smaller range of uti-

lization than other active exoskeletons currently in the

market (such as the HAL (Sankai, 2006)), which is

uniquely to assist an elderly person’s gait and other

daily life activities such as climbing stairs.

By focusing on these activities, it is possible that

an exoskeleton that assists the lower limb activities

can be achieved with a small frame, which could go

unnoticed if the user is wearing loose clothes.

This paper refers to determining the actuation re-

quirements to assist the user of the exoskeleton being

developed.

An explanation for the methodology used to de-

sign the mechanical frame can be seen in (Pina et al.,

2018). Because the exoskeleton is aimed at helping

the movement in the lower limbs in people with a re-

duced degree of strength and mobility, the system pro-

poses to assist the lower limbs’ biomechanical forces

by 50%. The exoskeleton is planned to wear a total

of 20 kg, though it is designed to support itself by ex-

tending into the ground.

2 THEORETICAL FRAMEWORK

The exoskeleton is designed to support its own

weight. However, when the exoskeleton is equipped,

the combined system composed of human body and

exoskeleton is bound to demand additional torque

than the human body alone. This happens because the

total weight is higher, so the torque levels required to

perform daily life activities will also be greater.

The formula for the torque is expressed in equa-

tion 1:

τ = I × α (1)

Where τ is the torque in N.m, I is the moment of

inertia in kg.m

and α is the angular acceleration.

Given I as the moment of Inertia dependent on the

mass m and distance r to the pivot:

I = m × r

(2)

therefore, the torque τ for a given movement de-

pends on the mass m, distance r and angular acceler-

ation α:

τ = m × r

× α (3)

If the purpose of the exoskeleton is to provide

the same natural movements that are performed on a

daily basis, then the angular acceleration α is ideally

the same, and r is constant because the exoskeleton

adapts its length to the user’s limbs. Therefore, an

BIODEVICES 2019 - 12th International Conference on Biomedical Electronics and Devices

184

increment on the mass m of the bodies being moved

results in an increment on the torque τ.

So the sum of torques required to move a system

composed of human body and exoskeleton is greater

than the sum of torques required to move the human

body:

∑

(humanwithexoskeleton)

∑

(human)

(4)

The system proposes to assist the human move-

ment in daily activities in close to 50%, so the torque

provided by the exoskeleton actuators must be de-

signed to perform 50% of the system combined of

human with exoskeleton.

To achieve these values, a number of trials with

human subjects was made with no weights and then

with attached weights to emulate the weight of the

exoskeleton.

3 MOTION CAPTURE DATA

ACQUISISTION

The trials were performed in a laboratory capable of

doing motion capture through cameras and markers

placed on the subjects. There are also four force

plates that exist to determine the ground reaction

forces for gait trials, and are arranged according to

Figure 3.

Figure 3: Arrangement of the force plates in the biomechan-

ics laboratory.

During a gait trial, if the ﬁrst force plate to be stepped

on is plate ”L”, then it is assumed that plate A will

be stepped on by the right foot and plate B will be

stepped on by the left foot. This arrangement allows

the system to acquire the ground reaction forces for a

full gait cycle.

The trials were performed by two subjects: a fe-

male with a weight of 52.4 kg and height of 1.52 m,

and a male with a weight of 63.8 kg and a height of

1.73 m.

As seen in (Pina et al., 2018), the mechanical

frame of the exoskeleton weights close to 8 kg. In

order to include the weight of the backpack with ac-

tuators, batteries and other subsystems, the exoskele-

ton simulation trials were performed with the subjects

carrying an additional weight of 20 kg. The trials

were performed with a weight distribution similar to

the exoskeleton frame, in order to approach the dy-

namic behavior of the exoskeleton. The subjects at-

tached different weights throughout the lower limbs

and its distribution can be seen in Figure 4.

Figure 4: Distribution of the attached weights to approach

the weight dynamics of the exoskeleton developed.

Pictures of the female and male subjects with and

without the attached weights can be seen in Figure

Figure 5: Pictures of the female and male subjects. On top,

with normal clothing and no added weights. On the bottom,

with the attached weights and backpack.

The subjects performed the trials described in Figure

6. The sitting and standing results are not considered

in this work because they did not contribute to the end

results.

Motion Capture and MultiBody Simulations to Determine Actuation Requirements for an Assistive Exoskeleton

185

Figure 6: Diagram of the trials performed for each subject.

Figure 7 shows a gait trial performed by the female

subject, with and without the additional weights.

Figure 7: Pictures of the female subject performing gait tri-

als with normal clothing on the left and attached weights on

the right.

To perform the stair trials, a set of mock-up stairs was

built with brick and wood to provide a safe method

to simulate stair ascend and descend. The stair climb

and descend cycles have the same description as gait

cycles. Each leg performs a stance phase (between

the foot making contact with one step and lifting off

to the next step) and swing phase (between the foot

lifting off one step and reaching the next). There

are two mock-up steps: with a height of 40 cm and

a second with a height of 80 cm. Although these

mock-up stairs are higher than usual stairs in modern

houses, this way they present a ”worst-case scenario”,

as higher stairs are prevalent in older houses or pub-

lic buildings. The lower set of stairs was placed in

force plate A (from Figure 3) and the taller set was

placed in force place B. The acquisition software for

the force plates was then tuned to subtract the weight

of the mock-up steps. To complete the cycle after the

taller step, there was a wooden block to complete the

the ascend or initiate the descend movements. The

wooden block was not placed over force plates.

In order to organize the ﬁnal simulation results,

the subjects were instructed to climb the mock-up

stairs with the right leg on the lower step, and then

to descend the stairs with the right leg ﬁrst.

Figure 8 shows pictures of the tests for stair ascend

and stair descend.

Figure 8: Pictures of the female subject performing stair

trials with normal clothing on the left and attached weights

on the right. The top two pictures show stair climbing while

the bottom pictures show stair descending.

4 OpenSIM WORKFLOW

The software used to perform the multibody simula-

tions is the OpenSim 3.3 Simulation Toolkit devel-

oped by Scott Delp (Delp et al., 2007). The software

is open source and is capable of performing multi-

body simulations for biomechanical applications. The

muskuloskeletal tridimensional model used for the

simulations is the reference Gait2392 (Delp et al.,

1990), which is based on anatomical data provided

by several studies (Friederich and Brand, 1990), (Hoy

et al., 1990). Gait2392 has 23 degrees of motion and

was designed to perform lower limb simulations.

OpenSim takes the marker data and performs In-

verse Kinematics to translate into motion data for the

Gait2392 model, in the form of angular position per

joint through time. There is also a Scaling func-

tion, which uses marker data and the subject’s weight

to customize the Gait2392 model. With the scaled

model, motion data and ground reaction forces from

the force plates, it is possible to perform Inverse Dy-

namics which provides the torque values for each

joint.

However, the Inverse Dynamics procedure in-

duces virtual residual actuators (consisted of torques

and forces applied to the pelvis) to help stabilize the

model due to data acquisition errors in the motion

BIODEVICES 2019 - 12th International Conference on Biomedical Electronics and Devices

186

capture system and force plates. In complex models

like the Gait2392 using complex motions, the resid-

ual actuators often assume large values, which signif-

icantly reduces the validity and precision of the cal-

culated torques.

The Residual Reduction Algorithm (RRA) uses

the Computer Muscle Control (CMC) algorithm to

control the torque applied by point actuators placed

in each joint. The CMC algorithm was developed

to calculate muscle activations for any given move-

ment through power-saving paths, as a biomimetic ap-

proach (Seth et al., 2011). At the same time, the RRA

makes slight modiﬁcations the motion data and pro-

poses mass changes to each body in order to reduce

the values of the residual actuators.

Figure 9 shows the regular workﬂow for perform-

ing simulations using RRA:

Figure 9: Workﬂow for the OpenSim Simulation Toolkit to

achieve precise values for joint torques using RRA.

According to the software developers, the results are

only valid if the values for residual forces and er-

ror displacements are below pre-deﬁned thresholds.

Therefore, the results obtained in this article were

achieved by tuning the simulation timings to ﬁt the

thresholds for the residual actuators.

For performing the simulations with the additional

weights, the exoskeleton components described in

(Pina et al., 2018) were adapted to the body measure-

ments of the male and female subjects. Afterwards,

the exoskeleton components were ported and adapted

into the Gait2392 model with the correct values for

geometry, mass, center of mass and inertia. The

weight distribution is different between the weights

attached to the subjects in the trials and the virtual

exoskeleton. To compensate for this difference, the

RRA procedure for mass changes was repeated for

each virtual model until the proposed changes were

minimal. At the moment of writing, the mass change

procedure is only possible through command line and

is not available through the software’s GUI.

The models used the OpenSim simulations can be

seen in Figure 10.

Figure 10: From left to right: male subject with exoskele-

ton, female subject with exoskeleton, male subject without

exoskeleton, female subject without exoskeleton.

5 OpenSIM RESULTS

The results presented in this section show the torque

values for the hip and knee joints for each activity and

for each subject. These joints are planned to be actu-

ated and assisted by the exoskeleton. For some tri-

als, the ankle torque is also shown. The ankle joint is

not actuated by the system, but the torque differences

may be a concern. However, a large proportion of

the ankle plantar ﬂexion torque originates from pas-

sive elastic forces from extending the Soleus muscle.

Also, many types of footwear limit the plantar ﬂexion

so its importance in this system is considered partially

reduced. Each result graphic shows a joint torque

from a trial with weight and another joint torque from

the equivalent trial without weight, to provide better

means of comparison.

The OpenSim’s graphics creator utility deﬁnes a

positive torque for hip ﬂexion and negative torque for

hip extension. For the knee, ﬂexion is negative and

extension is positive.

In some results, the same trial for the same joint

is split into two simulations. The reason for this is

because in some simulations the values for the resid-

ual actuators and displacement errors would start to

increase.

In the following torque graphics, the blue and

green lines show a result without additional weight

Motion Capture and MultiBody Simulations to Determine Actuation Requirements for an Assistive Exoskeleton

187

and in red a result with additional weight.

5.1 Gait Results

The following results were obtained through the

OpenSIM RRA simulations with data taken from the

gait trials. A screenshot of the male model with and

without the exoskeleton can be observed in Figure 11

Figure 11: Screenshot of the simulation with the male

model with the exoskeleton on the left and without the ex-

oskeleton on the right.

Hip Joint

Figure 12 shows the torque results for the hip joint

ﬂexion/extension in the female subject, during the

gait trials. The higher positive torque values in the

hip ﬂexion correspond to the beginning of the swing

phase.

Figure 12: Hip Extension/Flexion torque values for the gait

trials with the female subject.

With the female subject, the additional weight causes

a difference of 12 N.m in the left hip ﬂexion. Hip

extension (negative values) is showing larger torque

values without weight. A possible explanation for this

is the subject assuming a faster walking pace.

Figure 13 shows the same trials for the male sub-

ject.

The male subject shows a 12 N.m difference in the

right hip ﬂexion.

Figure 13: Hip Extension/Flexion torque values for the gait

trials with the male subject.

Knee Joint

The torque results for the female subject’s knee joints

during gait are shown in Figure 14. The higher pos-

itive torque values in the knee correspond to the mo-

ment of heel contact.

Figure 14: Knee Extension/Flexion torque values for the

gait trials with the female subject.

The knee results in the female trial do not show signif-

icant torque differences, and the same was observed

with the male trials.

5.2 Stair Ascend and Descend Results

The stair ascend and descend simulations could not

be resolved for a full ”stair” cycle. Due to the gener-

ally increased complexity of the movements, the sim-

ulation periods are reduced to keep the residual actu-

ator values within the pre-established thresholds de-

termined by OpenSim’s developers. Therefore, it is

expected for the joint results in one side to match half

a cycle and the other side to match the other half (e.g.

left leg joints show the swing phase and the right leg

joints show the stance phase). The following graph-

ics show the results with a larger difference observed

between trials with and without weight.

Screenshots of the male model with and without

BIODEVICES 2019 - 12th International Conference on Biomedical Electronics and Devices

188

the exoskeleton during the simulations with stairs can

be observed in Figure 15

Figure 15: Screenshot of the simulation with the male

model with and without the exoskeleton while climbing

stairs on the left and descending stairs on the right.

Hip Stair Ascend

Figure 16 shows the hip results for the male subject.

Figure 16: Hip Extension/Flexion torque values for the stair

ascend trials with the male subject.

The maximum torque values for the male subject’s

right hip shows a 10 N.m difference in extension dur-

ing the stance phase.

Knee Stair Ascend

Figure 17 shows the knee torque values for stair as-

cend with the female subject.

Figure 17: Knee Extension/Flexion torque values for the

stair ascend trials with the female subject.

The weight trials show a 22 N.m increase in max-

imum torque for knee extension during the swing

phase with the left knee.

Hip Stair Descend

Figure 18 shows the hip joint torque values for the

male subject during the stair descend trial.

Figure 18: Hip Extension/Flexion torque values for the stair

descend trials with the male subject.

The male subject results show a 20 N.m difference in

the right hip extension.

Knee Stair Descend

The stair descend trial with the knee joint shows the

largest torque differences between the results with and

without weight. The knee extension peak values cor-

respond to the moment when the subjects’ feet make

contact with the lower step. This moment matches

the highest torque observed in any of the movements

studied in the trials.

Figure 19 shows the same trials for the male sub-

ject. The knee torque differences with and without

weight are similar between the male and female sub-

ject. The male subject shows a difference of 50 N.m

in the left knee.

Figure 19: Left Knee Extension/Flexion torque values for

the stair descend trials with the male subject.

Motion Capture and MultiBody Simulations to Determine Actuation Requirements for an Assistive Exoskeleton

189

6 CONCLUSIONS

Although the subjects were instructed to perform the

same movements, the additional weights created a

difference in the body dynamics, and therefore the

movements themselves. For this reason, some tests

show large differences in torque values between trials

with and without additional weight.

The Exoskeleton proposes to assist the move-

ment in the daily activities by 50%. To achieve

this, the maximum torque values from the actua-

tion system should be able to perform 50% of the

maximum torque values for a given movement with-

out the weights (τ(t

max

)

normal bod y

), plus the differ-

ence in torque for the same movement with the

weights(∆τ(t

max

)

body with exoskeleton

Therefore:

τ(max)

actuationsystem

= τ(t

max

)

normal bod y

× 50%)+∆τ(t

max

)

body with exoskeleton

(5)

The aforementioned torque values and torque differ-

ences are observed in the following trials:

Hip Extension:

Stair ascend with male subject: 60 N.m without

weight and ∆τ 10 N.m.

τ(max)

actuationsystem

= (0.5 × 60) + 10 = 40 N.m

Hip Flexion:

Stair Descend with male subject: 50 N.m without

weight and ∆τ 20 N.m.

τ(max)

actuationsystem

= (0.5 × 50) + 20 = 45 N.m

Knee Extension:

Stair Descend with male subject: 100 N.m without

weight and ∆τ 50 N.m.

τ(max)

actuationsystem

= (0.5 × 100) + 50 = 100 N.m

Knee Flexion:

Gait with male subject: 27 N.m without weight and

∆τ 3 N.m.

τ(max)

actuationsystem

= (0.5 × 27) + 3 = 16.5 N.m

The largest requirement for torque assistance from the

exoskeleton is the knee extension with a value of 100

N.m. Given the large difference between the torque

requirement for this speciﬁc movement (knee joint

in stair descend) and the torque requirements for the

other movements, the 50% assistance value may not

be met in this case, or additional types of solutions

may be studied. For example, since the knee exten-

sion during stair descend is performing negative work,

the same torque can be obtained through a control-

lable brake.

With this data, it will be possible to develop an ac-

tuation system that is neither under or overengineered

for the exoskeleton. This process can therefore be

able to save time and costs during the exoskeleton de-

velopment.

ACKNOWLEDGEMENTS

The authors would like to thank the “Minist

erio da

encia, Tecnologia e Ensino Superior – Fundac¸

para a Ci

encia e a Tecnologia, Portugal” for the

funding provided by the research project “LAETA –

UID/EMS/50022/2013”.

REFERENCES

American academy of orthopaedic surgeons, “ef-

fects of aging”. as of 2017 available at

http://orthoinfo.aaos.org/topic.cfm?topic=A00191.

Commission, E. (2015). The 2015 ageing report: Economic

and budgetary projections for the 28 eu member states

(2013-2060). http://ec.europa.eu.

Delp, S., Loan, J., Hoy, M., Zajac, F., Topp, E., and Rosen,

J. (1990). An interactive graphics-based model of the

lower extremity to study orthopaedic surgical proce-

dures. IEEE Transactions on Biomedical Engineering,

37(8):757–767.

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.

EUROSTAT (2014). Physical and sensory functional limi-

tations by sex, age and educational attainment level.

EUROSTAT (2017). Population structure and ageing. ISSN

2443-8219.

Fiatarone, M. A. and Evans, W. J. (1993). 11 the etiology

and reversibility of muscle dysfunction in the aged.

Journal of Gerontology, 48(Special):77–83.

Friederich, J. A. and Brand, R. A. (1990). Muscle ﬁber ar-

chitecture in the human lower limb. Journal of Biome-

chanics, 23(1):91–95.

HARTLAPP, M. and SCHMID, G. (2008). Labour market

policy for ‘active ageing’ in europe: Expanding the

options for retirement transitions. Journal of Social

Policy, 37(03).

Hoy, M. G., Zajac, F. E., and Gordon, M. E. (1990). A

musculoskeletal model of the human lower extremity:

the effect of muscle, tendon, and moment ann on the

moment-angle relationship of musculotendon actua-

tors at the hip, knee, and ankle. Journal of Biome-

chanics.

BIODEVICES 2019 - 12th International Conference on Biomedical Electronics and Devices

190

Muscolino, J. E. (2016). Kinesiology: The Skeletal System

and Muscle Function, volume Chapter 3.9 “Effects of

Physical Stress on Bone”, pp. 47-49. C V MOSBY

CO.

Peine, A., Rollwagen, I., and Neven, L. (2014). The

rise of the “innosumer”—rethinking older technology

users. Technological Forecasting and Social Change,

82:199–214.

Pina, D. S., Fernandes, A. A., Jorge, R. N., and Gabriel, J.

(2018). Designing the mechanical frame of an active

exoskeleton for gait assistance. Advances in Mechan-

ical Engineering, 10(2):168781401774366.

Riley, P. O., Croce, U. D., and Kerrigan, D. C. (2001). Ef-

fect of age on lower extremity joint moment contribu-

tions to gait speed. Gait & Posture, 14(3):264–270.

Sankai, Y. (2006). Leading edge of cybernics: Robot suit

HAL. In SICE-ICASE International Joint Conference.

IEEE.

Seth, A., Sherman, M., Reinbolt, J. A., and Delp, S. L.

(2011). OpenSim: a musculoskeletal modeling and

simulation framework for in silico investigations and

exchange. Procedia IUTAM, 2:212–232.

Shankar, A., McMunn, A., Banks, J., and Steptoe, A.

(2011). Loneliness, social isolation, and behavioral

and biological health indicators in older adults. Health

Psychology, 30(4):377–385.

Steadman, J., Donaldson, N., and Kalra, L. (2003). A ran-

domized controlled trial of an enhanced balance train-

ing program to improve mobility and reduce falls in

elderly patients. Journal of the American Geriatrics

Society, 51(6):847–852.

Steptoe, A., Shankar, A., Demakakos, P., and Wardle, J.

(2013). Social isolation, loneliness, and all-cause mor-

tality in older men and women. Proceedings of the

National Academy of Sciences, 110(15):5797–5801.

Walker, A. (2010). The emergence and application of active

aging in europe. In Soziale Lebenslaufpolitik, pages

585–601. VS Verlag f

ur Sozialwissenschaften.

Motion Capture and MultiBody Simulations to Determine Actuation Requirements for an Assistive Exoskeleton

191