Assessment of the Suitability of the Motorized Ankle-Foot Orthosis as a

Diagnostic and Rehabilitation Tool for Gait

Guillermo As

ın

, Filipe A. Barroso

1,2

, Juan C. Moreno

and Jos

e L. Pons

Bioengineering Group, Spanish National Research Council (CSIC),

Crtra. Campo Real km. 0,200 28500, Arganda del Rey (Madrid), Spain

Adaptive System Behaviour Group - Industrial Electronics Department, University of Minho, Gimar

aes, Portugal

Keywords:

Motorized Ankle-foot Orthosis, Gait, TOP-DOWN, EMG, Neuroengineering, Kinematics.

Abstract:

A unilateral powered exoskeleton (Motorized ankle-foot orthosis, MAFO) is presented in this work, with the

aim of studying muscle and kinematics short-term adaptations of the ankle during rehabilitation tasks. For this

purpose, we conducted this study during gait over a treadmill, measuring surface electromyography activation

and biomechanical data, in different conditions of assistance. This pilot study also aims to demonstrate that the

tool is suitable for measuring biomechanical data while allowing EMG measurements, proving it as a useful

tool during gait assessment and rehabilitation. Gastrocnemius Medialis activation presents slightly higher

amplitude with higher assistances, so the subjects performed a higher range of motion gait pattern. Tibialis

Anterior EMG activation presents consistent data with previous studies. Ankle angle at lower assistances

makes the robot force less the subject to reach the imposed gait pattern, and so the range of motion diminishes.

Regarding ankle angular velocity, at higher assistances, higher velocities are reached. The torque between the

subjects foot and the robot. For lower assistances, the imposed reference pattern is less restrictive, and so the

force the user exerts against the robot is lower.

1 INTRODUCTION

Stroke is the principal neurological disease in the de-

veloped world that culminates in physical disability.

Most promising interventions for the rehabilitation

of locomotor function are based on robotic systems

which are focused on the rehabilitation of the function

by acting at the periphery of the body (BOTTOM-UP

approach). It is unclear how effective these treatments

are, and one of the biggest problems they have is the

non-adherence of the patient to the therapy.

The opposite approach, which focuses on neu-

rological interventions that are based on the state

of the brain after the pathology to alter periph-

eral behavioural outcomes is known as TOP-DOWN

approach(Belda-Lois et al., 2011). Iosa and col-

leagues applied this approach in the framework of the

European Project BETTER with a new tool in which

a speciﬁcally designed ankle-foot orthosis (AFO) is

combined with sEMG (surface electromyography)

and kinematics sensors to provide the user a contin-

uous online feedback of his/her performance (Iosa

et al., 2012).

Several motorized devices for training during

overground gait have been described in the litera-

ture. The WalkTrainer is intended for a patient to re-

learn gait by combining a hybrid orthosis with func-

tional electrical stimulation (Stauffer et al., 2009)

with a BWS (body-weight-support) portable mech-

anism. The IHMC (Institute for Human and Ma-

chine Cognition) Mobility Assist Exoskeleton (Kwa

et al., 2009), the externally powered lower limb or-

thosis (Saito et al., 2005), and the Lower Body Ex-

oskeleton (Costa and Caldwell, 2006) are other simi-

lar devices that allow over ground and treadmill gait

rehabilitation. Focusing on exoskeletons that target

single joints like the device used in this work, the

literature describe apparatuses such as the powered

KAFO (knee-ankle-foot orthosis), a unilateral KAFO

that actuates proportionally to surface EMG signals

from the patient (Sawicki and Ferris, 2009) by actu-

ating artiﬁcial pneumatic muscles. GAIT is a quasi-

passive KAFO that was developed as a low-power de-

vice (Moreno et al., 2008), where the knee is actively

powered, but the ankle relies on a springed passive

actuator to avoid foot drop while providing mobility.

The variable impedance AFO described by Blaya and

Herr (Blaya and Herr, 2004), an ambulatory version

161

Asín G., Barroso F., Moreno J. and Pons J..

Assessment of the Suitability of the Motorized Ankle-Foot Orthosis as a Diagnostic and Rehabilitation Tool for Gait.

DOI: 10.5220/0004652101610166

In Proceedings of the International Congress on Neurotechnology, Electronics and Informatics (SensoryFusion-2013), pages 161-166

ISBN: 978-989-8565-80-8

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

of the AnkleBot (Krebs and Hogan, 2006; Wheeler

et al., 2004), is an AFO that impedes foot drop by

regulating the impedance via a series elastic actua-

tor. Galle and colleagues propose an assistive bilat-

eral AFO exoskeleton (Galle et al., 2013) based on an

adaptive sEMG controller and pneumatic actuators, to

reduce the metabolic cost of walking. Ferris and col-

leagues in several papers presented the use of an AFO

of EMG activation of dorsiﬂexors (Tibialis Anterior)

and/or plantar ﬂexors (Soleus, Gastrocnemius) while

gait (Ferris et al., 2005; Ferris et al., 2006; Sawicki

and Ferris, 2008; Gordon and Ferris, 2007; Kao and

Ferris, 2009; Kao et al., 2010a; Kao et al., 2010b); but

highlighted the main limitation of this kind of actua-

tors: they make the exoskeleton not readily portable.

A motorized AFO developed in the framework of

the BETTER project (MAFO, motorized AFO) (As

ın

et al., 2012) has been proposed as a controllable or-

thosis that can alter the ankle joint neuromuscular

control and therefore be applied to assist in locomo-

tion training after neurological injury.

Our goal is to study muscle and kinematics short-

term adaptations of the ankle with the use of a unilat-

eral powered exoskeleton. For this purpose, we con-

ducted this study during gait over a treadmill, measur-

ing EMG activation and biomechanical data, in dif-

ferent conditions of assistance (ratio between “robot

in charge” and “patient in charge” (Van der Kooij

et al., 2006) concepts). This pilot study also aims

to demonstrate that the tool is suitable for measur-

ing biomechanical data while allowing EMG mea-

surements, proving it as a useful tool during gait re-

habilitation.

2 MATERIALS AND METHODS

2.1 Subjects

Three healthy right-handed subjects (2 male, 1 fe-

male, age 25 ± 1.73 years, body mass 75.67 ± 16 kg)

were enrolled in this study. A left MAFO (see Fig-

ure 1) weighting 1.1 kg was used. It was adjusted to

match each of the subjects lower leg length and malle-

oli position. The insole is located inside the sports

footwear of the subject as a means of tight attachment

to the foot, to reliably transmit the movement to the

joint. A rubber insole of the same height as the robots

insole has been located inside the contralateral shoe

to compensate the height difference.

2.2 Procedure

The subjects underwent robot-aided walking on a

treadmill (TC-450 by Domyos) at 1 m/s with the

MAFO under these conditions:

1. Full Assistance (refer to 2.2.1 for an explanation

of this parameter) (FA): robot in charge, i.e. the

robot performs all the movement with no need of

any contribution from the subject.

2. Medium Assistance (MA): the robot performs

50% of the movement, i.e., the robot is only able

to reach 50% of the targeted trajectory. The re-

maining 50% is accomplished by the subject.

3. Low Assistance (LA): the same as MA condition,

being the ratio 10% from the robot and 90% from

the subject.

As a partial aim of the study was to study angu-

lar position while walking with the orthosis, subjects

were asked to walk as if they were not wearing the

robot, but forcing the robot if needed in order to try

to perform a normal gait pattern. They had no vi-

sual reference so they were just focused on walking.

They had to walk without resting the hands on the

bars of the treadmill, to emulate as close as possible

overground walking.

Figure 1: MAFO (motorized ankle foot orthosis), worn by

one of the subjects of the study.

The trials (summarized in Table 1) had a duration

of 3 minutes and were separated in time by at least

half an hour to ensure independence between trials

and let muscles rest to eliminate fatigue due to the

exercise.

NEUROTECHNIX2013-InternationalCongressonNeurotechnology,ElectronicsandInformatics

162

Table 1: Study conditions.

Trial Velocity Assistance

FA 1 m/s 100%

30 minutes rest

MA 1 m/s 50%

30 minutes rest

LA 1 m/s 10%

2.2.1 Hybrid Controller

The device is controlled using a hybrid controller

where the value of contribution of a position con-

troller and an admittance controller (based on the

measured torque) is selected prior to the beginning of

the exercise (see Figure 2). A parameter called “as-

sistance”, with values from 0 to 100 per cent, modu-

lates the output from the controller. 100 per cent of

assistance indicates that the controller output depends

only on the position controller; at 0 per cent, the con-

troller output comes from the admittance controller;

and so at values in between the controller output is

distributed with “assistance” percentage from the po-

sition controller, and 100 minus “assistance” percent-

age from the admittance controller. This parameter

was calibrated by measuring the percentage of ankle

range of motion the robot was able to reach while per-

forming a dorsi-plantar ﬂexion continuous movement

at different velocities.

The gait pattern is pre-recorded from a healthy

subject as in (Hitt et al., 2009) and modulated as a

function of the stride time as in (Ward et al., 2006).

The admittance controller is dependent on the gait

phase, with a lower stiffness in the swing phase than

in the stance phase, i.e. a higher proportional gain in

the swing than in the stance phase, to avoid the effect

of the interaction torque against the ground.

Figure 2: Hybrid controller conceptual scheme.

2.3 Data Acquisition and Analysis

The MAFO records ankle joint angular position, an-

gular velocity, and torque resulting from the in-

teraction between the foot and the orthosis. Two

footswitches were located at the heel and the toe to

record heel strike and toe-off conditions respectively,

for the identiﬁcation of the gait phase. This allows the

robot to synchronize to users gait phase.

Two muscles have been measured in the study:

Tibialis Anterior (dorsiﬂexion) and Gastrocnemius

Medialis (plantar ﬂexion). These distal muscles ex-

ert the more powerful EMG signals involved in the

control of the ankle joint. Furthermore, as the ankle

joint produces the majority of the positive mechani-

cal work during stance in human walking (Kao et al.,

2010b), further studies with this hardware could lead

to mechanical analysis of the whole limb. They have

been recorded with the equipment EMG-USB by OT

Bioelettronica, in a bipolar conﬁguration according to

SENIAM recommendations (Hermens et al., 1999),

with both DRL and reference at the subjects left wrist.

Kinematic signals were sampled at 1200 Hz, and

EMG signals at 2048 Hz. Raw EMG data were high-

pass ﬁltered (3rd order Butterworth digital) at 20 Hz

and envelopes for each signal were extracted. Indi-

vidual stride cycles were separated with gait events

determined using the footswitches data, obtaining for

the three-minute sessions 90 ± 5 steps. Kinematic

signals were resampled to 2048 Hz sampling fre-

quency to match EMG data sample frequency. The

smoothed EMG and kinematic signals were then av-

eraged (stride-by-stride and for all the subjects) to ob-

tain averaged time-normalized gait cycles for all con-

ditions and data for the comparison between condi-

tions.

3 RESULTS

3.1 EMG Activation

The subjects were asked to walk as normally as pos-

sible, so the hypothesis was that the activation pat-

terns may remain almost unchanged in such a short

exercise. Figure 3 shows the activation of the Tibialis

anterior (TA) and Gastrocnemius Medialis (GM) for

the three subjects. GM presents slightly higher ampli-

tude with higher assistances (FA), which responds to

the fact that at higher assistances, the subjects per-

formed a higher range of motion gait pattern (see

section 3.2 for further details on biomechanical sig-

nals). TA presents, according to gait phases in Table 2

(Perry, 1992), a less difference between the values of

muscle activation at the endpoints in the ﬁgure (heel

strike moment, i.e. initial loading and terminal-swing

moments) and EMG values from mid-stance to mid-

AssessmentoftheSuitabilityoftheMotorizedAnkle-FootOrthosisasaDiagnosticandRehabilitationToolforGait

163

Figure 3: Mean RMS TA and GM activation; per gait phase percentage.

Table 2: Phases of the gait cycle.

Phase Percent of gait cycle

Initial loading 0 – 12

Mid-stance 12 – 30

Terminal-stance 30 – 50

Pre-swing 50 – 62

Initial-swing 62 – 75

Mid-swing 75 – 87

Terminal-swing 87 – 100

swing. These signals are consistent with data pre-

sented by Hidler and Wall (Hidler and Wall, 2005).

3.2 Biomechanical Data

Figure 4(a) presents the ankle angle. These data

shows that for lower assistances, the robot forces less

the subject to reach the imposed gait pattern, and so

the range of motion diminishes. Although the user

was told to perform a normal gait, for very low as-

sistances (LA condition) the lack of a gait reference

imposition maybe the cause of the lower range of mo-

tion.

Figure 4(b) presents the ankle velocity. These data

is a direct consequence of the angle, so the data ex-

tracted is consistent to the data for the ankle angle:

higher assistances lead to higher range of motion and

so higher velocities.

Figure 4(c) presents the torque between the sub-

ject’s foot and the robot. For lower assistances, the

imposed reference pattern is less restrictive, and so

the force the user exerts against the robot is lower, as

the robot tries to follow subject’s movements.

Mean values for the range of motion and maxi-

mum values for the velocity and torque are presented

in Table 3 for the three subjects and the three condi-

tions.

Table 3: Mean range of motion and maximum velocity and

torque for the three subjects.

Subject Trial ROM Top speed Max torque

[

◦

] [

◦

/ s] [N· m]

1 FA 24.40 84.84 10.36

MA 18.29 72.21 6.34

LA 7.15 31.32 3.68

2 FA 27.40 88.25 7.92

MA 21.38 89.20 8.25

LA 6.15 29.62 3.30

3 FA 26.50 87.87 6.61

MA 22.84 79.80 5.11

LA 14.31 57.54 4.56

4 DISCUSSION AND FUTURE

WORK

This work presented a tool to assist physiotherapists

in the rehabilitation tasks, providing not only the

movement task, but measurements to assess func-

tional improvements. The tool proved to be suitable

for these rehabilitation and assessment tasks, being

much more portable than artiﬁcial-muscle-powered

devices.

A deeper study is to be carried out enrolling more

subjects, and modifying the protocol of the current

study to perform a long term experiment, to be able to

observe the evolution on the measurements due to the

rehabilitation exercises. This will lead to the obser-

vation of short-term adaptations in muscle activation

and kinematic signals. Due to its portability, another

future could be performed overground and not over a

treadmill, thus eliminating the reﬂex to walk when the

“ground” moves. This approach makes the presented

experiment set-up prone to be combined with elec-

NEUROTECHNIX2013-InternationalCongressonNeurotechnology,ElectronicsandInformatics

164

(a)

(b)

(c)

Figure 4: Biomechanical signals: angle (a), velocity (b),

and torque (c).

troencephalography to study walking intention, and

further combination of EEG and EMG data could lead

the device to be an assistive device based on a brain-

neural-computer interface.

ACKNOWLEDGEMENTS

This study has been funded by grant from the

European Commission, within the Seventh Frame-

work Programme (FP7-ICT-2009-247935: BETTER

BNCI-driven Robotic Physical Therapies in Stroke

Rehabilitation of Gait Disorders). It also has been

partially funded by grant from the Spanish Ministry of

Science and Innovation CONSOLIDER INGENIO,

project HYPER (Hybrid NeuroProsthetic and Neuro-

Robotic Devices for Functional Compensation and

Rehabilitation of Motor Disorders, CSD2009-00067).

REFERENCES

ın, G., Collantes, I., Moreno, J. C., and L, P. J. (2012).

Dise

no de una

ortesis motorizada de tobillo para reha-

bilitaci

on de ictus con un enfoque top-down. In Sum-

mer School on Neurorehabilitation, 2012.

Belda-Lois, J.-M., Mena-del Horno, S., Bermejo-Bosch, I.,

Moreno, J. C., Pons, J. L., Farina, D., Iosa, M., Moli-

nari, M., Tamburella, F., Ramos, A., et al. (2011). Re-

habilitation of gait after stroke: a review towards a

top-down approach. Journal of neuroengineering and

rehabilitation, 8(1):66.

Blaya, J. A. and Herr, H. (2004). Adaptive control of a

variable-impedance ankle-foot orthosis to assist drop-

foot gait. Neural Systems and Rehabilitation Engi-

neering, IEEE Transactions on, 12(1):24–31.

Costa, N. R. and Caldwell, D. G. (2006). Control of a

biomimetic soft-actuated lower body 10dof exoskele-

ton. Proceedings of the 8th International IFAC Sym-

posiumon Robot Control SYROCO, pages 6–8.

Ferris, D. P., Czerniecki, J. M., Hannaford, B., et al. (2005).

An ankle-foot orthosis powered by artiﬁcial pneu-

matic muscles. Journal of Applied Biomechanics,

21(2):189.

Ferris, D. P., Gordon, K. E., Sawicki, G. S., and Peetham-

baran, A. (2006). An improved powered ankle–foot

orthosis using proportional myoelectric control. Gait

& posture, 23(4):425–428.

Galle, S., Malcolm, P., Derave, W., and De Clercq, D.

(2013). Adaptation to walking with an exoskeleton

that assists ankle extension. Gait & Posture.

Gordon, K. E. and Ferris, D. P. (2007). Learning to walk

with a robotic ankle exoskeleton. Journal of biome-

chanics, 40(12):2636–2644.

Hermens, H. J., Freriks, B., Merletti, R., Stegeman, D.,

Blok, J., Rau, G., Disselhorst-Klug, C., and H

agg, G.

(1999). European recommendations for surface elec-

tromyography. Roessingh Research and Development

The Netherlands.

Hidler, J. M. and Wall, A. E. (2005). Alterations in muscle

activation patterns during robotic-assisted walking.

Clinical biomechanics (Bristol, Avon), 20(2):184–

193.

Hitt, J., Sugar, T., Holgate, M., Bellman, R., and Hollander,

K. (2009). Robotic transtibial prosthesis with biome-

chanical energy regeneration. Industrial Robot: An

International Journal, 36(5):441–447.

AssessmentoftheSuitabilityoftheMotorizedAnkle-FootOrthosisasaDiagnosticandRehabilitationToolforGait

165

Iosa, M., Tamburella, F., Moreno, J. C., Collantes, I., As

ın,

G., Aloise, F., Pisotta, I., Muzzioli, L., Mattia, D.,

Molinari, M., Pons, J. L., and Cincotti, F. (2012). Neu-

rorehabilitation after stroke: a new tool for a top-down

approach. In Terzo Congresso Gruppo Nazionale

Bioingegneria (GNB 2012).

Kao, P.-C. and Ferris, D. P. (2009). Motor adaptation during

dorsiﬂexion-assisted walking with a powered orthosis.

Gait & posture, 29(2):230–236.

Kao, P.-C., Lewis, C. L., and Ferris, D. P. (2010a). Invariant

ankle moment patterns when walking with and with-

out a robotic ankle exoskeleton. Journal of biome-

chanics, 43(2):203–209.

Kao, P.-C., Lewis, C. L., and Ferris, D. P. (2010b). Short-

term locomotor adaptation to a robotic ankle ex-

oskeleton does not alter soleus hoffmann reﬂex ampli-

tude. Journal of neuroengineering and rehabilitation,

7(1):33.

Krebs, H. I. and Hogan, N. (2006). Therapeutic robotics:

A technology push. Proceedings of the IEEE,

94(9):1727–1738.

Kwa, H. K., Noorden, J. H., Missel, M., Craig, T., Pratt,

J. E., and Neuhaus, P. D. (2009). Development of the

ihmc mobility assist exoskeleton. In Robotics and Au-

tomation, 2009. ICRA’09. IEEE International Confer-

ence on, pages 2556–2562. IEEE.

Moreno, J. C., Brunetti, F., Rocon, E., and Pons, J. L.

(2008). Immediate effects of a controllable knee an-

kle foot orthosis for functional compensation of gait

in patients with proximal leg weakness. Medical &

biological engineering & computing, 46(1):43–53.

Perry, J. (1992). Phases of gait. In Gait Analysis: Normal

and Pathological Function, pages 9–16. Slack, Tho-

rofare, NJ.

Saito, Y., Kikuchi, K., Negoto, H., Oshima, T., and

Haneyoshi, T. (2005). Development of externally

powered lower limb orthosis with bilateral-servo ac-

tuator. In Rehabilitation Robotics, 2005. ICORR

2005. 9th International Conference on, pages 394–

399. IEEE.

Sawicki, G. S. and Ferris, D. P. (2008). Mechanics and ener-

getics of level walking with powered ankle exoskele-

tons. Journal of Experimental Biology, 211(9):1402–

1413.

Sawicki, G. S. and Ferris, D. P. (2009). A pneumatically

powered knee-ankle-foot orthosis (kafo) with myo-

electric activation and inhibition. Journal of neuro-

engineering and rehabilitation, 6(1):23.

Stauffer, Y., Allemand, Y., Bouri, M., Fournier, J., Clavel,

R., Metrailler, P., Brodard, R., and Reynard, F. (2009).

The walktrainer-a new generation of walking reeduca-

tion device combining orthoses and muscle stimula-

tion. Neural Systems and Rehabilitation Engineering,

IEEE Transactions on, 17(1):38–45.

Van der Kooij, H., Veneman, J., and Ekkelenkamp, R.

(2006). Compliant actuation of exoskeletons. Mo-

bile robotics-towards new applications, Mammendorf.

ISBN, pages 978–3.

Ward, J. A., Hitt, J., Sugar, T., and Bharadwaj, K. (2006).

Dynamic pace controller for the robotic gait trainer.

ASME.

Wheeler, J. W., Krebs, H. I., and Hogan, N. (2004). An

ankle robot for a modular gait rehabilitation system.

In Intelligent Robots and Systems, 2004.(IROS 2004).

Proceedings. 2004 IEEE/RSJ International Confer-

ence on, volume 2, pages 1680–1684. IEEE.

NEUROTECHNIX2013-InternationalCongressonNeurotechnology,ElectronicsandInformatics

166