Kinematic Analysis of Powered Lower Limb Orthoses for Gait

Kinematic Analysis of Powered Lower Limb Orthoses for Gait Rehabilitation of Hemiplegic

and Hemiparetic Patients

M. R. Safizadeh, M. Hussein, M. S. Yaacob, M. Z. Md Zain, M. R. Abdullah, M. S. Che Kob, K. Samat

Abstract - In this paper, the kinematic analysis of constructed assistive robotic leg for rehabilitation of patients who encounter the neurological injury is presented. In order to design an efficient new mechanism, studies were carried out to distinguish the human architecture and dynamics. In the study, the motion of a healthy physical subject in walking situation of 1 km/h speed was recorded. Thereafter, a novel robotic leg mechanism was developed to produce similar motion. The robotic leg is driven by a single actuator to drive both the hip and the knee joints mechanism. In order to verify the robot motion with respect to human motion, kinematic analysis of all robot’s joints and links are formulated and are simulated in MATLAB software. The results obtained from the kinematic analysis of the developed assistive robotic system show that its motion conforms to the motion and dynamics of a healthy human. Keywords - Kinematic analysis, assistive robotic leg, lower extremity exoskeleton, hemiplegic and hemiparetic patient.

I. INTRODUCTION The third and seventh causes of death in the world are stroke and accident, respectively. In USA for example, according to the American Heart Association and the National Heart, Lung, and Blood Institute (NHLBI), the cost of tending to people who suffer from cardiovascular diseases and stroke in 2009 is estimated to be $475.3 billion, making stroke a major financial burden to society. This cost includes both direct and indirect costs; direct costs include the cost of physiotherapists and other professionals, hospital and nursing home services, the cost of medications, home health care and other medical durables and indirect costs include lost productivity that results from illness and death. Patients with hemiplegia and palsy may not able to carry out the daily activities such as talking, walking, crouching and grasping; therefore, they need to improve their abilities by active and passive rehabilitation therapy iteratively and regularly. In passive exercises, the patient receive the rehabilitation exercises by physiotherapist; whereas, the

active exercises are done by the patients. Unfortunately, although many attempts and researches have been done to prevent the people from accident and death, the number of stroke and injured patients who survive from these events which require rehabilitation services increase with time and also the number of specialists, physiotherapists and rehabilitation centers is not sufficient to respond to a large number of patients efficiently. Thus, after successful presence of robot in industries, the robotic knowledge has been applied in order to carry out the rehabilitant tasks efficiently and less costly. However, medicine and medical doctor must be involved in the development to assure the robot is employed in an ethical way [1].

In recent years, the attention on robotic rehabilitation has been increasing in order to train the patient base on their daily activities. However, because of some disadvantages of medical robots such as neglecting the social and psychological needs of patients, suppression the individuality and uniqueness of services, rejection of autonomous services, and increase of cost, they are not widely applied in medical and rehabilitation centers these days [1]. Hence, researchers have been trying to improve the interaction of robots and their environments and patients in order to carry out some beneficial repetitive exercises.

Indeed, psychological feedback as one of significant step in medical robot design should be taken into account to assess the patients and their performance. Medical robot should be able to record and quantify patient’s performance and provide score, statistics, comparisons between normal human and tested patient as well as comparison between new and previous performance to show the patient’s progress. In addition, the feedback should able to motivate patients to utilize all their efforts in improvement; also it can inform the user and physiotherapist about movement errors and patients’ motion conditions.

As the number of survivors from incidents such as accident, wars and strokes grows, the attention of researchers on developing rehabilitation robots for lower and upper extremity exoskeletons to help hemiplegic patients has been augmented, [2-7, 8]. Professor Sankai in University of Tsukuba for example, has developed a hybrid assistive leg (HAL-3) which assists the disabled in developing normal walking motion. He uses two DC motors

INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

Issue 3, Volume 5, 2011 490

onenmthsymSufloesefcaextostnohisccoarco

ham(AanelantoknorthofortomhiacanorFopr[2potorespkncocobeal

ofpapaitsligtaDvecaon

n knee and hincoder to me

measure floor rhe required toystem [5-6]. H

myoelectricity uzuki K. et aoor reaction stimation in ffectively [9]. arried out extextremity exosko carry signiaircases and rot able to accighlighted foucheme, high-pommunicationrchitecture toonsumption [1

Besides, sevave been prop

motion. ArtificiAFO) [11-15]nd electro-pnectrical actua

nd 18] and hyo provide the nee and anklrthosis with trhan 10 years af patients. Lorthosis, utilizeo handle patien

movement usinip and knee pctuator that atnd exoskeletorthosis have bor instance, troposed as a n20]. The conceotential energorque on the esistant duringpring potentianee flexion anonsists of muontrol rotationetween shankigned centrica

This paper wf one DOF larallel to humatient with reps ability to cghtweight, sim

arget of the deC motor thatelocity of the am-follower mn the human m

ip joints, senseasure the joreaction and morques for knHowever, sinc

is not estimal. proposed t

force (FRForder to

Professor Kaensive researckeleton (BLEificant loads rocky slopes, complish thesur features wpowered com

n protocol ano decrease 10]. veral other de

posed to providial pneumatic , was propos

neumatic gaiators for diffeydraulic actuat

required force joints correreadmill train

and has been pokomat Gait es treadmill annt balance. T

ng exoskeletonpatient’s jointttach to the exon with activbeen designedthe Spring Bnew techniqueept of Spring

gy that beinghip flexion. T

g knee extensal energy, a bnd maintain knultiple-brake n of the kneek segments anally with the jwill present thlower extrem

man hip and kpetitive physiclimb or wample and accsign. It is an et can synchrohip and knee

mechanism. Thmotion data a

sory systems oints angles, myoelectricitynee and hip jce the biologimated for Htwo new algo) estimation estimate pa

azerooni and h on developi

EEX) in order in some s

in which whse tasks. The which includempact power nd electronic

the comple

signs and metde comfort tomuscles prot

sed for knee-ait orthosis [rent numberstors were alsoce and torqueectly [10]. Tning has beenproven to cureOrthosis [19

nd body weighis orthosis an structure suts are supportxoskeleton. Bve drivers, sod in order torake Orthosise in order to pBrake Orthos stored in thThe spring alsion. Other thbrake is designee extensionand brake-ac

e joint. The brnd thigh segmoint axis.

he design and mity exoskeletknee joints incal therapy w

alk on the stcurate mechaexoskeleton ponize the ange joints by mehe cam profilecquired from

in terms of rforce senso

sensor to estijoints and coical signal su

HAL-3 accuraorithms in ca

and torso atient’s intenhis research ging Berkeley lto help the pituations suc

heeled vehicledeveloped sy

e: a novel cosupplies, sp

s, and a speexity and p

thods of actuapatient in wa

otype (KAFOankle-foot ort16]. In addof DOF [5,

o designed in e to move theherapy in ca

n applied for e and improve9], one of roght support syassists patient’upport; in addted by linear esides the ort

ome other pa exercise pats (SBO) has provide hip flesis, generally,

he spring to lso acts as a

han the functiogned to countn. This exosketuator in ordrake was mou

ments then ex

kinematic anaton that movn order to traiwithout considtairs. A low anism is the owered by a sgular positioneans of a geae is obtained bjoint angles o

rotary ors to imate ontrol

uch as rately, ase of angle

ntions group lower eople

ch as es are ystem ontrol pecial ecific

power

ations alking

O) and thosis

dition, 7, 17 order

e hip, ase of

more e gait

obotic ystem ’s leg

dition, drive

thosis assive tients.

been exion , uses assist knee

on of teract eleton der to unted xactly

alysis ves in in the dering

cost, main

single n and ar and based of the

lowecond

F

II

Ianthrtwo anthrvariajointinjurmajomotidriveof lidesigtwo be fkinemDC mkneesyste

er limbs. Thducted in orde

Fig. 1: Capt

Fig. 2: Experim

DATA ACQU

n order to ropomorphic

major coropomorphic able to be usedts should rotary or pain toor concern is ion. This is ae mechanism nks and lowegn an anthropmain links wifastened to tmatics of hummotor, gears ae joint motionem a single DO

he necessary r to obtain the

tured video in

mental Hip (a)

UISITION AN

design an adesign contex

onsiderations. design is to md for several pte in the sam the patient to match the

achieved throwith lightwei

er energy conpomorphic anith variable lehe thigh andman motion, aand cam-follon follows the OF mechanism

experimentse essential dat

n MATLAB M

) and Knee (b

ND MECHAN

appropriate roxts were follo

The firstmake the lengpatients. As th

me axis with hshould be ave leg kinema

ough the desiight componennsumptions. Tnd economic aength are propd shin. To ha mechanism

ower is propoship joint ang

m as shown in

s (Fig. 1) ata (Fig. 2).

MPlay GUI

) data angles

NISM DESIGN

obotic leg, towed, producit concern gth of each limhe exoskeletonhuman joint, avoided. Anothatic with humign of a propnts, less numb

Thus, in order architecture, tposed which chave a suitab

m consisting ofsed such that tgle to make tn Fig.3.

are

N

the ing of

mb n’s any her

man per ber to

the can ble f a the the



A. prexDbyexMprsekmara

(ascoillm11boprlinFi

ancoacvadewbereswsit

Ashth

. Cam Desig

Since the oroduce vivid lxperiment, huVD 605 with y extracting xtracted fram

MATLAB imrocessing at equential framm/h. The jointre later used tosingle cycle dIn the design

1 ), whereas,

s well as cam pontributes to tlustrates that

movement for 15° and 126° oth hip and kractical data tnear data whiig. 7). After design

nd software (ontrolling thccommodate varies from onesigned to be

with simple gue set for patehabilitation witch is alstuations.

As mentioned ehould follow he kinematic

Fig. 3: B

ign

objective of lifelike humanuman motion

ability of 25 fframes out

mes (Fig. 1) amage processi

the stage of mes are proct coordinates o derive joint data. n, the hip anglthe knee angl

profile. The athe accuracy ot the range walking speeto 178°, resp

knee profiles thereafter befich are used t

ning the robo(driver, transdhe robot mvarieties of pae to another. e user-friendl

uides can opertients of diffexercise spe

so included

III KINEMA

earlier, since tthe anthropomof lower ex

Block Diagram

designing assn motion [21]is captured u

fps frame rateof the video

are processeding toolbox.

f data acquisessed for waobtained fromangles and ca

le is determinele depends on

accuracy of caof the knee anof angles f

ed of 1 km/h ectively. The are revised t

fore conversioto produce th

t, the controlducers, user

motions weratient needs; a

The control sly as such anrate the systemferent limb sieed requirem

to accomm

ATIC ANALY

the powered lmorphic and tremity exosk

m

sistive robot], in the conduusing Sony D

e and are proco; thereafter. d as images u

Through imition, a totalalking speed

m image proceam profile base

ed by gear rotn the link 1 m

m profile therngle motion. Ffor hip and

is between 7angular data

to produce a on into follow

he cam profile

l system hardinterface etc.

re developedas patient dynasystem softwany physiotherm. The systemizes and diff

ments. Emergmodate emerg

SIS

ower limb ortergonomic dekeleton shoul

is to ducted DCR-essed

The using

image l 400 of 1

essing ed on

tation motion

refore Fig. 2

knee 72° to

from more wer’s e (see

dware .) for d to amics are is rapist

m can ferent gency gency

thosis esign, ld be

propobjeis caMATsystethreerelat

A. Sobtalowethigh

desig

acquof hifoun

Fig.

Iwere

Bis inpoin

pV

VV

VV

Orelatdisplon (4

2

wher

relat

1L

si1L

per for humanctive, the kinarried out baTLAB prograem are acquiree links (link 1tive motion pr

Kinematic An

Since the deined accordiner extremity eh’s angle with

gn, all meas

uired from patip angle; ther

nd appropriate

3.

n the kineme made:

The gear m(d1). Point P is a pThe rotating

Based on the rtroduced to fo

nt P) on link 1.

/P G GV V

cos(

sin 1

GP

GP

VVVV

sin(

sin

/

/

GGP

GGP

VVVV

Once the velotionship betwelacement of th4) below.

12

LVP

re, 1 is the a

tionship betwe

( 22

21 dd

sinin 12 d

n motion. In oematic analys

ased on simplamming. The ed on the basi1 to link 3; seinciples [22].

nalysis of ‘lin

sign and meng to the adapexoskeleton k

h pendulous m

urements suc

tient’s thigh sirefore, the size for 1d and lin

atic analysis,

ovement invo

point on the slaxes are assum

relative motioormulate the v

)

cosin

21

2

GV

)(

cos

21

22

GV

ocity of poineen angular vhe gear (see F

cos( 11

LVG

angular displa

een 1L and 2

)cos2 121 dd

s21

order to achiesis of the powle dynamic rkinematic relis of kinematiee Fig. 3), ca

nk 1’

easurement option of humankinematic, lin

motion around

ch as 1d and

ize and the exze of 3.1 cm ank 1 respective

, the followin

olves the mo

lot closest to Gumed to be at l

on of rotatingvelocity of on

cosos 21

cossin 21 G

nt P has beevelocity of linFig. 5) can be

)2

acement of th

2 can be expre

)

sin(sin1

11 Ld

ve the desirabwer assistive lrelationship alationship of tc analysis of t

am-follower a

of all links an kinematic ak 1 follows tpoint 2O . In th

link 1 slot a

xperimental daand 39.5 cm aely, as shown

ng assumptio

otion of the r

G. link 1.

axes, (1) to (ne point (such

(1)

0 (2)

0 (3)

n acquired, tnk 1 and angue obtained bas

(4)

he gear, and t

essed as follow

(5)

)1 (6)

ble leg and the the and

are and the his

are

ata are

n in

ons

rod

(3) as

)

)

)

the ular sed

)

the

ws:

)

)



vege

V

veis of3.

Thereafter, elocity of theear as shown i

22LVKnee

Fig. 4: Ang

Fig.5: Ang

Fig.6: A Because of

elocity of linkin accordance

f 0.26 Hz. Me28 rad/s.

the relatioe knee and thin Fig. 6 is ob

gular displacem

gular velocity

Absolute veloc

f pendulous mk 1 as well as te with the peneanwhile, the

onship betwehe angular ditained from fo

ment of link 1

of link 1 ( 2 )

city of knee fo

motion of linthe absolute vndulous motioangular veloc

een the absisplacement oollowing equa

( 2 ) for one c

) for one cycle

or one cycle

nk 1, the anvelocity of the on with a frequcity of DC mo

solute of the ation:

(7)

cycle

e

ngular e knee uency otor is

B. IveloTo tlink rolle

velo

obtafixedAs inprod

Ithe flink that the mobtacomprelatprodsimu

IMATshou(Figsrolle

as sh

Fig

Kinematic An

n order to fincity of roller this end, since1 directly af

er velocity ca

cities named,

ined by assumd and the link ndicated earlie

duce cam profit can be seenfollower for rcondition in there are dwe

middle of the ining smoothparison is dotion to rotationduces a smoulation. n order to derTLAB prograuld be acquires. 8 and Fig.

er while the lin

hown in Fig.10

Fig.7: Cam pr

. 8: Interpolat

nalysis of ‘lin

nd the angulaA; a point one the angular

ffected the vean be found

, 1AV and AVming the camis rotating wh

er, the linear mile are shown n from the grarotational linkvertical direc

ells at the begcycle. These

h motion of one to illustranal link is cor

ooth motion

rive the kinemamming, the md by interpola9). As a res

nk is assumed

0.

rofiles in case

ed follower dione cy

nk 2’

ar velocity ofn follower hasr velocities foelocity of rolld by dividing

2A . Velocity

m is rotating while cam is fixmotion data win Fig. 7.

raph that the k is smootherction. The proginning and eare importanthe follower

ate that the drrect and applfor the fol

matic relationshmost correct ating the real csult, the abso

d fixed ( 1AV ) a

e of fix and rot

isplacement aycle

f the link 2, ts to be obtaineor both cam aler, the absolug AV into tw

1AV and 2AV a

while the linkxed respectivewhich are used

displacement r than the fixofile also shond as well as

nt parameters f. Actually, thdata for cam licable. Hencelower and t

hip based on tand proper dacam profile da

olute velocity are demonstrat

tational link

and real data fo

the ed.

and ute wo

are

k is ely. d to

of xed ws in for his in

e it the

the ata ata of

ted

for



Thca

taca

ve

anro

th(8ab

V

F

Fig. 10: Th

The most c

he exact positam ( 3 ) and

angential vectalculate 2AV w

elocity of AVnalyzing the oller to 2O , an

he absolute ve8) below. Figbsolute velocit

AA VVV 21

Fig.9: Interpo

he absolute ve

complicated p

tion of the rolthe rod ( 2 )

tor of cam awith its exact

2AV is acquired

angular velocnd the angle b

elocity of the g. 12 shows ty V and the

AAA VVV 2 212

2

olated cam pro

elocity of 1AV

part is finding

ller based on ) are needed.

at this positit direction. O

d (as shown

city of link between 1AV a

roller, V , is the relation

gear angular d

AA cos

ofile

for one cycle

g the velocity

the rotation oSubsequently

on is requireOnce the abs

in Fig. 11)

1 and distancand 2AV (i.e.

obtained basenship betweendisplacement.

(8

e

y 2AV .

of the y, the

ed to solute

), by

ce of

A ),

ed on n the .

8)

Sand anguabsocombConsrespeveloanaly

ShV

JV

JV

S

/J ShVshowcaseare a(13)

KV

Fig. 11: The

Fig. 12: The

Since the lowrotates by m

ular velocity olute velocitybination of sequently, theect to the kncity of joint ysis with resp

KShK VV /

AJA VV /

ShShJ VV /

Since link J-Shh is zero; there

wn in vector ds that follow. acquired basethrough (22).

AKSh VV /

absolute veloc

absolute velo

er part of exomeans of caof link 2 is

y of roller both algeb

e magnitude ee joint. On

J can be fect to roller A

K

h is fixed to lefore, (12) is

diagram in FiSubsequentlyd on Figs. 13

AJA V /

city of 2AV fo

ocity of AV fo

oskeleton has am-follower ms acquired bA and the

ebraically anof joint Sh the other han

found using A.

link 2, the relas derived from

ig. 13 to Fig.y, the value of3 to Fig. 17 a

or one cycle

r one cycle

general motimechanism, tby utilizing t

knee, and tnd graphicalis generated

nd, the absolurelative moti

(9)

(10)

(11)

ative velocity m (11) which

17 for the fif / and as well as usi

(12)

ion the the the lly.

in ute ion

of h is

ive /

ing



Case I (0° ):

Fig: 13: Velocity vector diagram Based on the vector diagram in Fig. 13, the vector

components in the x and y directions are:

/ 2: sin cos cos

cos 0A A J A JA K

ShK ShK

x V V VV

(13)

/ 2: cos sin sin

sin 0A A J A JA K

ShK ShK

y V V VV

(14)

Case II ( 180°):

Fig: 14: Velocity vector diagram

: / 0 (15) : / 0

(16) Case III (180° 180° ):


: / 0 (17) : / 0

(18)

Case IV (180° ):


: / 0 (19) : / 0

(20)

Case V ( 360°):


: / 0 (21) : / 0

(22) Since the magnitudes of KV and AV are computed from

(7) and (8), the values and direction of /Sh KV and /J AV are

obtained using (12) and the known angular position of link 1 ( 2 ) and . The angular velocities of link 2 and link 3 can

be found by analyzing the relative velocities of /Sh KV and

/J AV . Once the angular velocity of link 2 is calculated, the

relative velocity /Bot KV can be obtained using (23) below.

/ /

/ /

/

/

/ /



4/ /

2bot K Sh K

LV Vr

(23)

Since of KBotV / can be found, the magnitude as well as

the direction of ankle velocity are obtained based on (24) through (27).

/B B K KV V V (24)

2 2

/ / 2 /2 cos( )A K B K K B K Sh KV V V V V (25)

In addition, the angle between and for different

cases is obtained, using (18) with respect to Fig. 18 to Fig. 22 in which also provide the direction of BV for each cases.

2 2 2

1 /cos ( )2

B K B KB K

B K

V V VV V

(26)

It can be seen from the figures that the value of depends on the configuration of , and / for the respective cases.

Case I (0° ):

Fig. 18: Velocity vector diagram

2[90 ( )]B B K (27)

Case II (θ θ 180°):


90 | | (28)

Case III (180° 180° ):


| | 90 (29)

Case IV (180° ):


| | 90 (30) Case V ( 360°):


90 | | (31)

/

/

/

/



2(anob

asgr(1

onm

F

Comparing (b), the simunalysis show btained from t

In order to cssumed that iround, and re17) (see Fig.

n the basis of maximum angle

Fig. 23: Defini

Fig. 24: Defi

Fig. 25: Defin

Fig. 24 with ulation data good agreem

the MATLABcalculate the sin each step ear foot hip 26). Thus, th

f known valuee.

itive absolute

initive angular

nitive angular

Fig. 2(a), andacquired bas

ment with the B programmingstep size for ththe fore footangle is at m

he step size (

s for the link

velocity of V

r position of h

r position of kn

d Fig. 25 withsed on kinemsimulation re

g. he robotic legt is vertical tomaximum pos

Stepd ) is calcu

length and th

AnkleV for one c

hip for one cyc

nee for one cy

h Fig. matic esults

g, it is to the sition ulated

he hip

cycle

cle

ycle

F

Tthe f

1

3

L

Sd

Stepd

T2 dhum

Textrebasehelp kneeusingangudesigsigniprevloweDC kneesimuThe data obtawill

Tof SUnivsuppsuppgran

Fig. 26: Huma

Trigonometricfollowing equa

2 4cos

24.69

L

1 2sinL sp Sd F

Therefore, th106Stepd cm

man step’s leng

This paper pemity exoskeled on acquire

to train the se angles for g MATLAB ular data wasgn a proper cificant defere

vious designs er extremity emotor to cont

e joints. Inulation were ckinematic resto prove its

ined from thbe applied in

The authors wScience, Tecversiti Teknolport in the rported in partnt number Vot

an step size du

and geometrations:

3cos S

4 3sinL sin F d

he patient gin each cycl

gth and within

IV. CONCL

presents an aleton) which wd experimentstroke sufferer1 km/h walkiimage proces

s converted tocam-follower ence between

is that the xoskeleton mtrol the move

the end, thcarried out basults were coms validity. The kinematic afuture studies

ACKNOWLE

would like to thnology andlogi Malaysia esearch workt by the Mal. 79338.

uring each stag

ric relationshi

1 4L L

31Sd cm 53Stepd cm

gait through le which is a

n the patient’s

CLUSION

assistive robowas designedtal walking d

ers. The experking speed wessing toolbox.to linear mot

mechanism. n the present

design and dmechanism usinement of both he kinematicased on the pmpared with the validated analysis are vs.

EDGMENT

thank the Mald Innovation a (UTM) for tk. This worklaysia eScienc

ge of the gait

ps are shown

(32)

m (33)

(34)

a distance accordance w

ability.

otic leg (lowd and developdata in order rimental hip aere obtained . Thereafter, tion in order In this work

t prototype adevelopment ng only a singthe hip and t

c analysis aproposed desigthe experimensimulation davery useful a

laysian Minis(MOSTI) a

their continuok was financces Fund und

in

)

is with

wer ped

to and by the to

, a and

of gle the and gn.

ntal ata

and

try and ous cial der



REFERENCES

[1] Liliana Rogozea, Florin Leasu, Angelu Repanovici, Michaela Baritz, Ethics, robotics and medicine development, Proceedings of the 9th WSEAS international conference on Signal processing, robotics and automation, Robotics and Automation, 2010, pp. 264-268, ISSN: 1790-5117.

[2] Daniela Mariana Barbu, Ion Barbu, "Dynamical Model for an Original Mechatronical Rehabilitation System, Proceedings of the 14th WSEAS International Conference on Applied mathematics, 2009, pp. 23–28, ISSN: 1790-276.

[3] Kao, P.C, Motor adaptation during dorsiflexion-assisted walking with a powered orthosis, Journal of Gait and Posture, vol. 29, pp. 230–236, 2009.

[4] Sawicki, G. S., Domingo, A. & Ferris, D. P, The Effects of Powered Ankle-Foot Orthoses on Joint Kinematics and Muscle Activation During Walking in Individuals With Incomplete spinal cord injury, Journal of Neuro Engineering and Rehabilitation, vol. 3, pp. 1-17, 2006.

[5] Kawamoto, H. & Sankai, Y., Power Assist System HAL-3 for Gait Disorder Person, in Computers Helping People with Special Needs. Vol. 2398/2002, ed Heidelberg: Springer Berlin, 2002, pp. 19-29.

[6] Kawainoto, H., Lee, S., Kanbe, S. & Sankai, Y., Power Assist Method for HAL-3 Using EMG-Based Feedback Controller, in Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics, 2003, pp. 1648–1653.

[7] Banala, S. K., Kulpe, A. & Agrawal, S. K., A Powered Leg Orthosis for Gait Rehabilitation of Motor-Impaired Patients, in IEEE International Conference on Robotics and Automation, Roma, Italy, 2007, pp. 4140-4145.

[8] Hussein A. Abduallah, Cole Tarry, Mohamed Abderrahim, Robotic Techniques for Upper Limb Rehabilitation and Evaluation, Proceedings of the 2nd WSEAS International Conference on Dynamical Systems and Control, Bucharest, Romania, 2006, pp. 78-83.

[9] Kenta, S., Gouji, M., Hiroaki, K., Yasuhisa, H. & Yoshiyuki, S., Intention-Based Walking Support for Paraplegia Patients with Robot Suit HAL, Journal of Advanced Robotics, vol. 29, pp. 1441-1469, 2007.

[10] Zoss, A., Kazerooni & Chu, A., On the Mechanical Design of the Berkeley Lower Extremity Exoskeleton (BLEEX), in IEEE/RSJ International Conference on Intelligent Robots and Systems, 2005, pp. 3132-3139.

[11] Ferris, D. P., Gordon, K. E., Sawicki, G. S. & Peethambaran, A., An Improved Powered Ankle–Foot Orthosis Using Proportional Myoelectric Control, Journal of Gait and Posture, vol. 23, pp. 425-428, 2005.

[12] Ferris, D. P., Sawicki, G. S. & Domingo, A., Powered lower limb orthoses for gait rehabilitation, Top Spinal Cord International journal of Rehabilitation, vol. 11, pp. 34-39, 2005.

[13] Gordona, K. E., Sawicki, G. S. & Ferrisa, D. P., Mechanical Performance of Artificial Pneumatic Muscles to Power an Ankle–Foot Orthosis, Journal of Biomechanics, vol. 39, pp. 1832-1841, 2006.

[14] G. S. Sawicki and D. P. Ferris, A Pneumatically Powered Knee-Ankle-Foot orthosis (KAFO) With Myoelectric Activation and Inhibition, Journal of NeuroEngineering and Rehabilitation, vol. 6, pp. 1-16, 2009.

[15] Sawicki, G. S., Gordon, K. E. & Ferris, D. P., Powered Lower Limb Orthoses: Applications in Motor Adaptation and Rehabilitation, in 9th International Conference on Rehabilitation Robotics, Chicago, IL, USA, 2005, pp. 206-211.

[16] Belforte, G., Gastaldi, L. & Sorli, M., "Pneumatic Active Gait Orthosis," Journal of Mechatronics, vol. 11, pp. 301-323, 2001.

[17] Ruthenberg, B. J., Wasylewski, N. A. & Beard, J. E., An Experimental Device for Investigating the Force and Power Requirements of a Powered Gait Orthosis, Journal of Rehabilitation Research and Development, vol. 34, pp. 203-213, 1997.

[18] Jerry, E. P., Benjamin, T. K. & Christopher, J. M., The Roboknee: An Exoskleton for Enhancing Strength and Enharance During Walking, in International Conference on Robotic & Automation, New Orleans, 2004, pp. 2430-2435.

[19] Gery Colombo, Matthias Joerg, Reinhard Schreier, BNIE, Volker Dietz, Treadmill training of paraplegic patients using a robotic orthosis, Journal of Rehabilitation Research and Development, Vol. 37, No. 6, 2000, pp. 693-700.

[20] S. Gharooni, B. Heller, and M. O. Tokhi, A New Hybrid Spring Brake Orthosis for Controlling Hip and Knee Flexion in the Swing

Phase, IEEE Trans Neural System and Rehabilitation Engineering, Vol. 9, No.1, 2001, pp: 106-107.

[21] J. Kang, B. Badi, Y. Zhao and D. K. Wright, Human Motion Modeling and Simulation, Proceedings of the 6th WSEAS International Conference on Robotics, Control and Manufacturing Technology, Hangzhou, China, April 16-18, 2006, pp: 62-67.

[22] Meriam, J. L., Kraige, L. G. & Palm, W. J., Engineering Mechanics: Dynamics, 5 ed.: John Wiley, 2001.



Documents

Kinematic Analysis of Powered Lower Limb Orthoses for Gait