Pa ien -D i en Hyb id FES-Exoskele on Con ol
wi h Adap i e Band-Based Assis ance
Jo ge R. Ma ins
Depa men o Indus ial Elec onics (DEI)
Uni e si y o Minho
Guima ˜
aes, Po ugal
0000-0001-6042-3345
Joana F. Almeida
Cen e o Mic oElec oMechanical Sys ems (CMEMS)
Uni e si y o Minho
Guima ˜
aes, Po ugal
0009-0007-7892-582X
Leona do Gizzi
Ins i u e o Modelling and Simula ion o Biomechanical Sys ems
Uni e si y o S u ga
S u ga , Ge many
0000-0003-3009-6261
C is ina P. San os
Cen e o Mic oElec oMechanical Sys ems (CMEMS)
Uni e si y o Minho
Guima ˜
aes, Po ugal
0000-0003-0023-7203
Abs ac —Spinal Co d Inju ies (SCIs) a e a p e alen con-
di ion wo ldwide, leading o a loss o mo o unc ion ha sig-
ni ican ly hinde s Ac i i ies o Daily Li ing (ADLs). Op imized
ehabili a ion is essen ial o mi iga e hese challenges, p omo ing
neu oplas ici y and mo o elea ning. One o he mos p omising
app oaches combines Func ional Elec ical S imula ion (FES)
o muscles wi h o ces gene a ed by exoskele ons. Howe e ,
p e iously de eloped sys ems ha e limi a ions, pa icula ly in
os e ing ac i e pa ien engagemen . This p oposal in oduces
a coope a i e and Assis -As-Needed hyb id sys em (FES-EXO)
designed o ADL-o ien ed ehabili a ion. The p ima y objec-
i e o he con ol s a egy is o enhance pa ien pe o mance
and independence h ough a Human-In-The-Loop (HITL) ap-
p oach. The assis ance is dynamically adjus ed o he use ’s
needs in eal ime, ea u ing h ee ope a ional modes,no assis-
ance, FES-only assis ance, and hyb id assis ance, based on he
use ’s mo o s a e. Addi ionally, he sys em add esses challenges
ela ed o muscle a igue and insu icien ac i e pa icipa ion.
The p oposed con ol amewo k will be in eg a ed in o he
hyb id sys em and, in u u e wo k, will be alida ed wi h heal hy
indi iduals and, subsequen ly, wi h SCI pa ien s in a case s udy
a Guima ˜
aes’ Hospi al.
Index Te ms—Spinal Co d Inju ies, Exoskele on, Func ional
Elec ical S imula ion, Rehabili a ion, Neu oplas ici y, Daily
Ac i i ies, Pa ien -D i en Assis ance.
I. INTRODUCTION
Spinal Co d Inju ies (SCIs) a ec o e 15 million people
wo ldwide [1], o en leading o loss o senso y and/o mo o
unc ion below he inju y le el. This esul s in se e e mobili y
limi a ions and di icul ies pe o ming Ac i i ies o Daily
Li ing (ADLs), o en accompanied by seconda y complica-
ions like spas ici y, p essu e ulce s, and os eopo osis [1].
These limi a ions signi ican ly educe pa ien s’ indepen-
dence and Quali y o Li e (QoL), equen ly equi ing as-
This wo k was suppo ed by FCT na ional unds, unde he na ional
suppo o R&D uni s g an , h ough he e e ence p ojec UIDB/04436/2020,
UIDP/04436/2020 and 2023.13876.PEX [2], and unde he Re e ence Schol-
a ship unde g an s 2022.15668.MIT BI 07 2024 CMEMS and CM/3 23-
02/2025
sis i e de ices such as c u ches o wheelchai s [1]. E ec-
i e ehabili a ion can help o e come hese challenges by
le e aging neu oplas ici y o mo o elea ning [3]. Howe e ,
cu en clinical ehabili a ion is expensi e, ime-consuming,
and o en lacks eal-li e ele ance and pe sonaliza ion [1].
Hyb id FES-Exoskele on sys ems show p omise o p o-
mo ing unc ional eco e y beyond clinical se ings. Ye ,
challenges emain: pa ien in ol emen is o en limi ed, and
exis ing p o ocols a ely accoun o long- e m p og ession o
adap a ion [4]. Chap e II u he explo es limi a ions speci ic
o each sys em.
This p ojec aims o enhance pa ien s’ unc ional pe o -
mance in daily ac i i ies h ough a hyb id FES-Exoskele on
sys em ha deli e s ADL-o ien ed, epe i i e aining. I
adop s an Assis -as-Needed (AAN) s a egy, combining FES-
gene a ed in e nal o ces wi h exoskele on-based ex e nal
suppo . The con ibu ion o each subsys em is dynamically
adjus ed o mee he use ’s e ol ing needs, os e ing au-
onomy in ADL execu ion. A Human-In-The-Loop (HITL)
con ol amewo k is p oposed, placing he use a he cen e
o sys em beha io .
The scien i ic con ibu ion o his wo k lies in a no el,
use -cen e ed hyb id con ol s a egy ha in eg a es oli ional
e o , adap i e FES-Exoskele on coope a ion, and pe sonal-
ized assis ance. By p io i izing use in en ion and minimizing
unnecessa y aid, he sys em encou ages ac i e engagemen
and suppo s mo e e ec i e, na u al ehabili a ion o ADL
asks [5]. This wo k p esen s he cu en s a e-o - he-a and
he challenges iden ied, and desc ibes he p oposed solu ion.
II. STATE-OF-THE-ART ON REHABILITATION
Exoskele ons and o hoses a e wea able de ices ha as-
sis locomo ion, ei he passi ely o ia ac ua o s. Howe e ,
in ehabili a ion, hey may educe pa ien engagemen , a
phenomenon known as ”slacking,” whe e he use becomes
passi e [4].
FES applies elec ical cu en s o igge muscle con ac-
ions, enabling mo emen e en in pa alyzed muscles. I s main
limi a ions a e apid muscle a igue and po en ial discom-
o [6].
Con en ional ehabili a ion elies hea ily on physio he a-
pis s o mobilize limbs [1]. While FES is inc easingly used,
exoskele on adop ion emains limi ed, as clea ad an ages o e
adi ional me hods ha e no ye been i mly es ablished.
In eg a ing FES wi h exoskele ons o o hoses is a g owing
app oach o o e come hei indi idual limi a ions [4]. This
e iew examines he cu en s a e o he a in hyb id eha-
bili a ion sys ems combining bo h echnologies. The li e a u e
sea ch was conduc ed om Oc obe o Decembe 2024 using
Scopus, IEEE Xplo e, and PubMed.
Following he P e e ed Repo ing I ems o Sys ema ic
e iews and Me a-Analyses (PRISMA) guidelines, 12 ele an
a icles we e selec ed o ull- ex eading and analysis.
A. Exoskele on-FES-Use Roles
The use ’s ole a ied widely ac oss s udies. In some cases,
like Lyu e al. (2023) [7], pa ien s we e ins uc ed o elax,
wi h hei mo emen ea ed as a dis u bance. Howe e , mos
sys ems encou aged ac i e pa icipa ion [4], [6], [8]–[12].
Some e en placed he use a he cen e o con ol, p o iding
assis ance only when needed, as in Ch is ou e al. (2024) and
Jung e al. (2024) [6], [8].
In many sys ems, he exoskele on se ed as he main sou ce
o assis ance [7], [10], [11], [13], [14], whe eas in Ch is ou
e al. (2024) [8], i ac ed as a seconda y suppo , used only
when necessa y.
When he exoskele on was he p ima y assis i e de ice, FES
ypically played a suppo ing ole, educing eliance on he
obo and p omo ing muscle ac i a ion [9]. No ably, FES was
o en limi ed o speci ic join s due o channel cons ain s.
B. Senso s
The hyb id sys ems e iewed employed a a ie y o senso s
o suppo bo h he exoskele on and FES con olle s.
Join posi ion senso s, including po en iome e s and op i-
cal/induc i e encode s, a e o en in eg a ed in o exoskele ons
o measu e join angles [7], [8], [10]–[12], [14]. In e ac ion
o ces we e ypically measu ed using load cells and s ain
gauges [4], [6], while g ound eac ion o ces (GRFs) we e
cap u ed using o ce-sensi i e esis o s (FSRs) o insole p es-
su e senso s [4], [6], [9], [13], [15].
Ine ial Measu emen Uni s (IMUs), commonly placed on
he oo o o so, acili a ed gai e en de ec ion and o ien a ion
acking [9], [13]. Addi ionally, elec omyog aphy (EMG) sen-
so s, o en loca ed on muscles such as he hams ings, Tibialis
An e io (TA), and Soleus, we e u ilized o moni o muscle
ac i i y and suppo oli ional con ol [6].
C. Sys em Con ol
1) Exoskele on Con ol: In hyb id sys ems, exoskele on
con ol o en in ol ed adjus ing mo o s i ness and damping
o suppo mo emen [4], [8]. Adap i e s a egies alloca ed
o que based on join posi ion o muscle weakness [8], [12],
while impedance models used eal- ime join eedback o
op imized assis ance [13].
Mos s udies ocused on join ajec o y acking wi h ixed
e e ences [4], [7]–[11], [14]–[16], using low-le el PD o
PID con olle s. A less common app oach by Ch is ou e al.
(2024) [8] applied I e a i e Lea ning Con ol (ILC) o imp o e
pe o mance h ough epe i ion.
2) FES Con ol: FES con ol in hyb id sys ems a ied
widely. Adap i e me hods adjus ed s imula ion based on eal-
ime eedback, including muscle a igue o s eng h es ima es,
while simple app oaches used h esholds o Fini e S a e
Machines (FSM) [9], [13]. Mo e ad anced echniques in ol ed
I e a i e Lea ning Con olle s (ILC) [8] and Deep Neu al
Ne wo ks (DNN) gene a ing FES pa e ns om EMG [16].
Commonly modula ed pa ame e s we e pulse wid h (PW)
and ampli ude (PA), whe eas pulse equency (PF), which can
educe a igue [17], was less explo ed. S imula ion iming was
ypically synch onized wi h mo emen phases like he gai
cycle [6].
3) Exoskele on-FES Collabo a ion: Collabo a ion s a e-
gies be ween FES and exoskele ons included o que alloca ion
ia ixed o dynamic coe icien s [7], [14]. The adap a ion
o he coe icien s could be decided based on muscle a-
igue/s eng h, o mo emen phase [12]. O he me hods in-
ol ed hyb id con ol ma ices and phase-based swi ching [8],
[15].
Real- ime o que dis ibu ion could be op imized using
Model P edic i e Con ol (MPC) [14] o coo dina ed h ough
syne gy-based con ol [12]. Ch is ou e al. (2024) p oposed
a band-based sys em ha dynamically selec s be ween no
assis ance, FES, o hyb id modes based on join angle acking
e o [8].
4) Use In en ion: Mos o he analyzed hyb id sys em
a icles did no inco po a e use in en ion. Howe e , one s udy
by Jung e al. (2024) [6] measu ed oli ional EMG signals o
es ablish a baseline o que, wi h he hyb id sys em supplying
addi ional o que as needed.
5) Muscle Fa igue: Muscle a igue is a majo limi a ion
o FES and is o en add essed in con ol s a egies. Some
s udies assessed a igue e ospec i ely ia ques ionnai es [4]
o EMG me ics [6], while o he s inco po a ed eal- ime
a igue es ima ion o adjus s imula ion o shi e o be ween
FES and exoskele on suppo . Me hods anged om quali a-
i e assessmen s o quan i a i e models and h eshold-based
app oaches [4], [8], [11], [12], [14]. Howe e , no consensus
exis ed on he bes con olle esponse: mos educed s imula-
ion when a igue was de ec ed [8], [11], [12], [14], only one
sys em inc eased i o main ain pe o mance [4].
D. ADLs
Mos s udies ocused on walking assis ance, including ead-
mill [6], [8], [9], [15] and o e g ound walking [4], [11], [12],
while ewe add essed si - o-s and (STS) mo emen s [7], [10],
[13], [14], [16].
Muscle a ge s a ied wi h he ac i i y: STS s udies mainly
s imula ed knee muscles, especially he quad iceps [7], [13],
[14], [16], wi h one including hams ings [10]. Gai s udies
ypically s imula ed bo h quad iceps and hams ings [4], [8],
[9], [11], [12], [15], some imes adding ankle muscles [6],
[9]. Ankle s imula ion was no ably absen in STS p o ocols,
e lec ing di e en muscula demands.
E. Valida ion
Valida ion p o ocols o hyb id sys ems a ied acco ding o
he ADLs a ge ed o assis ance. Pa icipan numbe s we e
gene ally small, anging om 1 o 6 indi iduals. S udy coho s
included exclusi ely heal hy subjec s [4], [6]–[9], [13], [15],
mixed g oups o heal hy and SCI pa ien s [12], [14], [16], o
solely SCI pa ien s [10], [11].
E alua ions ypically ocused on ajec o y acking pe -
o mance, using me ics such as Roo Mean Squa ed E o
(RMSE) be ween desi ed and ac ual join angles [7], [8], [12],
[14]–[16] and a e age join angle alues [4], [6], [9]–[11],
[13]. In e nal sys em pa ame e s, including s imula ion o e
ime [4], [7], [8], [15], o que [6], [7], [10], [11], [13], [15], and
s i ness [4], [8] we e also assessed. Some s udies inco po a ed
use pe cep ion measu es, such as Visual Analog Scales (VAS)
o com o and a igue [4], [6].
III. PROPOSED CONTROL SOLUTION
To add ess he limi a ions o exis ing hyb id s a egies,
his a icle p esen s a con ol amewo k o a hyb id sys em
designed o assis gai . The sys em comp ises an 8-channel
s imula o (Mo ionSTIM8, Medel, Ge many) and a 1-DoF
unila e al ankle o hosis (Ankle-H3, Technaid, Spain). Since
he sys em a ge s he ankle, i p o ides hyb id assis ance o
he Tibialis An e io (do si lexo ) and Soleus (plan a lexo ).
To acili a e eal- ime use eedback and assess indi idual
needs, addi ional senso s aside om hose embedded in he
o hosis a e in eg a ed, including Delsys T igno EMG, FSRs
and IMUs.
All u u e pa icipan s mus p o ide in o med consen
in acco dance wi h he e hical guidelines es ablished by
E hics Commi ee o Guima ˜
aes’ Hospi al (77/2023-CAF and
PIC 57/2023).
Figu e 1 illus a es he con ol diag am, depic ing all he
componen s ha cons i u e he p oposed con ol solu ion.
The solu ion p oposal con ol can be di ided in o 5
main blocks: i) Voli ional Es ima ion; ii) Hyb id Con olle ;
iii) FES Con olle ; i ) Exoskele on/O hosis Con olle ; )
Exoskele on-FES-Human block.
The oli ional EMG is ex ac ed om he measu ed mus-
cula ac i i y wi hin he Voli ional block. Since EMG da a is
acqui ed concu en ly wi h FES s imula ion, he aw signal,
EMG aw, con ains s imula ion a i ac s. The e o e, a i ac
emo al is necessa y o isola e he oli ional componen o
he EMG eco dings.
This p ep ocessing ollows he p ocedu e desc ibed in [6].
A e a i ac elimina ion using a blanking me hod, he il e ed
EMG signal, EMG , is ob ained, enabling he ex ac ion o
he oli ional EMG signal, EMG ol, h ough a comb il e
ollowed by a 2 Hz low-pass il e .
The hyb id block is esponsible o selec ing he app op ia e
e e ence ajec o ies, EMG e and θ e , o guide he assis-
ance o bo h Exoskele on and FES componen s, as well as
iden i ying he equi ed o que o he sys em o complemen
he use ’s exe ed o que, τsys. Fo ha , i ecei es da a om
FSRs and IMUs o iden i y he gai phase and de ec walking
speed. Based on hose inpu s, he adequa e con ol e e ences
a e selec ed.
Wi hin his block, EMG e e ences a e con e ed o es i-
ma ed e e ence o que, τ e using a CNN model as in Mo ei a
e al. [18], simila o [6]. This model is also esponsible o
con e ing he oli ional EMG, EMG ol, ex ac ed om he
measu ed muscle ac i i y o he use in he Voli ional block,
o he oli ional o que o he use , τ ol. Ha ing τ e and τ ol,
τsys can be calcula ed as he di e ence be ween he wo.
The sys em o que, τsys, as a o que e o , de e mines he
le el o assis ance ia a band sys em inspi ed by Ch is ou e
al. [8]. As he e o inc eases, he sys em ansi ions h ough
h ee egions: No assis ance (low e o ), FES-only assis ance
(mode a e e o ), and Hyb id assis ance (high e o ).
A e selec ing he app op ia e assis ance le el based on he
use ’s exe ed o que and he e e ence o que, each compo-
nen , FES and Exoskele on, is con olled using o que e o -
based eedback con olle s.
Rega ding FES con olle , a PID con olle is used o
compu e he PA alue a each momen .
No ably, he eedback con olle also con ains a o que- o -
PA con e e , con e ing he o iginal PID ou pu o a PA alue,
uF ES.
The FES con olle in eg a es a Muscle Fa igue Es ima o
based on he model p oposed by Bao e al. (2020) [14], which
cha ac e izes a igue dynamics. The es ima o elies on pa ien -
speci ic a igue and eco e y ime cons an s, iden i ied h ough
o que o EMG moni o ing, and equi es indi idualized cali-
b a ion o ensu e accu acy. The es ima o ou pu s a muscle
a igue index, µ, anging om 0 ( ull a igue) o 1 (no a igue).
This index scales he o iginal PA alue, so highe a igue
esul s in educed s imula ion.
Conce ning he exoskele on/o hosis con olle , i employs
a eedback con olle , PID, o p o ide a o que con ol signal,
uEXO.
Bo h FES and exoskele on con olle s inco po a e an ILC
o exploi he cyclic na u e o gai by e ining con ol signals
o e epe i ions. Al hough bo h sys ems use ILC, hey adap
di e en pa ame e s. In he exoskele on con olle , s i ness
(k) is upda ed o e gai cycles. In con as , he FES con olle
adjus s an ILC gain pa ame e , λ, as in [8], which anges om
0 o 1. This gain, combined wi h he muscle a igue index,
scales he FES PID ou pu .
The ILC ope a es based on he pa ame e alues om he
p e ious cycle (kk−1and λk−1) as well as eal- ime eedback
om he cu en i e a ion om he cu en cycle, θ and τ .
Fig. 1. P oposed Con ol Diag am o he Hyb id Sys em.
IV. CONCLUSION
This wo k p esen s a no el, pa ien - ailo ed hyb id eha-
bili a ion amewo k ha syne gis ically combines in e nally
gene a ed o ces ia FES wi h ex e nally applied suppo om
a obo ic exoskele on. By dynamically in eg a ing eal- ime
use needs and a iable assis ance le els, he sys em os e s
mo e na u al and e ec i e ehabili a ion.
The coope a i e con ol amewo k con inuously adap s
assis ance acco ding o he use ’s mo o eco e y, aiming
o enhance au onomy in pe o ming ADLs. I s alida ion is
expec ed o o e aluable insigh s in o he cus omiza ion o
ehabili a ion s a egies o indi iduals wi h SCI.
This app oach lays he g oundwo k o a new ehabili a ion
pa adigm, aiming o achie e pa ien eco e y beyond cu en
me hodologies.
REFERENCES
[1] Wo ld Heal h O ganiza ion, Spinal Co d Inju y. [Online]. A ailable:
h ps://www.who.in /news- oom/ ac -shee s/de ail/spinal-co d-inju y
[Accessed: Oc . 17, 2024].
[2] N. F. Ribei o, ”AI-based sma exoskele ons o elde mobili y and
all p e en ion,” Fundac¸˜
ao pa a a Ciˆ
encia e a Tecnologia (FCT),
2025. [Online]. A ailable: h ps://scip oj.p c is.p /en/176625PRJ, doi:
10.54499/2023.13876.PEX.
[3] He Y, Xu Y, Hai M, Feng Y, Liu P, Chen Z, Duan W. Exoskele on-
Assis ed Rehabili a ion and Neu oplas ici y in Spinal Co d Inju y. Wo ld
Neu osu g. 2024 May;185:45-54.
[4] Del-Ama AJ, Gil-Agudo A, Pons JL, Mo eno JC. Hyb id gai aining
wi h an o e g ound obo o people wi h incomple e spinal co d inju y:
a pilo s udy. F on Hum Neu osci. 2014 May 13;8:298
[5] R. M. And ade, S. Sapienza, E. E. Faba a and P. Bona o, ”T ajec o y
T acking Impedance Con olle in 6-DoF Lowe -Limb Exoskele on o
O e -G ound Walking T aining: P elimina y Resul s,” 2021 In e na ional
Symposium on Medical Robo ics (ISMR), A lan a, GA, USA, 2021, pp.
1-6,
[6] Jung S, Bong JH, Kim K and Pa k S, 2024, ’Machine-lea ning-based
coo dina ion o powe ed ankle– oo o hosis and unc ional elec ical
s imula ion o gai con ol’. F on . Bioeng. Bio echnol. 1.
[7] Lyu C, Mo im PT, Penzlin B, R¨
oh en F, Be gmann L, on Pla en P,
Bollheime C, Leonha d S, Ngo C. Closed-Loop FES Con ol o a
Hyb id Exoskele on du ing Si - o-S and Exe cises: Concep and Fi s
E alua ion. Ac ua o s. 2023; 12(8):316.
[8] A. Ch is ou, A. J. Del-Ama, J. C. Mo eno and S. Vijayakuma , ”Adap i e
Con ol o T iadic Human-Robo -FES Collabo a ion in Gai Rehabili-
a ion: A Pilo S udy,” 2024 IEEE In e na ional Con e ence on Robo ics
and Au oma ion (ICRA), Yokohama, Japan, 2024, pp. 5561-5568.
[9] F. Ji, S. Qiu, Y. Liu, J. Wei, Y. Zhang and D. Ming, ”A T ajec o y-
Adap i e Walking Assis ance S a egy Based on Func ional Elec ical
S imula ion o Exoskele on o Help S oke Pa ien s Res o e Na u al
Gai ,” 2022 IEEE/ASME In e na ional Con e ence on Ad anced In el-
ligen Mecha onics (AIM), Sappo o, Japan, 2022,
[10] S. A. Mu ay, R. J. Fa is, M. Gol a b, C. Ha igan, C. Kandilakis and
D. T uex, ”FES Coupled Wi h A Powe ed Exoskele on Fo Coope a i e
Muscle Con ibu ion In Pe sons Wi h Pa aplegia,” 2018 40 h Annual
In e na ional Con e ence o he IEEE Enginee ing in Medicine and
Biology Socie y (EMBC), Honolulu, HI, USA, 2018.
[11] K. H. Ha, S. A. Mu ay and M. Gold a b, ”An App oach o he
Coope a i e Con ol o FES Wi h a Powe ed Exoskele on Du ing Le el
Walking o Pe sons Wi h Pa aplegia,” in IEEE T ansac ions on Neu al
Sys ems and Rehabili a ion Enginee ing, ol. 24, no. 4, pp. 455-466,
Ap il 2016.
[12] N. A. Alibeji, V. Molazadeh, F. Moo e-Clingenpeel and N. Sha ma,
”A Muscle Syne gy-Inspi ed Con ol Design o Coo dina e Func ional
Elec ical S imula ion and a Powe ed Exoskele on: A i icial Gene a ion
o Syne gies o Reduce Inpu Dimensionali y,” in IEEE Con ol Sys ems
Magazine, ol. 38, no. 6, pp. 35-60, Dec. 2018.
[13] M. A. Alouane, W. Huo, H. Ri ai, Y. Ami a , and S. Mohammed,
”Hyb id FES-Exoskele on Con olle o Assis Si -To-S and Mo emen ,”
IFAC-Pape sOnLine, ol. 51, no. 34, pp. 296–301, 2019.
[14] X. Bao, V. Molazadeh, A. Dodson, B. E. Dicianno and N. Sha ma,
”Using Pe son-Speci ic Muscle Fa igue Cha ac e is ics o Op imally
Alloca e Con ol in a Hyb id Exoskele on—P elimina y Resul s,” in
IEEE T ansac ions on Medical Robo ics and Bionics, ol. 2, no. 2, pp.
226-235, May 2020.
[15] Jona han Casas, Chen-Hao Chang, and Vic o H. Duenas. “Swi ched
Concu en Lea ning Adap i e Con ol o T eadmill Walking Using a
Lowe Limb Hyb id Exoskele on”. In: IEEE T ansac ions on Con ol
Sys ems Technology 32,2023.
[16] V. Molazadeh, Q. Zhang, X. Bao and N. Sha ma, ”An I e a i e Lea ning
Con olle o a Swi ched Coope a i e Alloca ion S a egy Du ing Si -
o-S and Tasks wi h a Hyb id Exoskele on,” in IEEE T ansac ions on
Con ol Sys ems Technology, ol. 30, no. 3, pp. 1021-1036, May 2022.
[17] Vahab Nekouka . “Con ol o Func ional Elec ical S imula ion Sys-
ems Using Simul aneous Pulse Wid h, Ampli ude, and F equency
Modula ions: Con ol o FES Using Simul aneous PW, PA, and PF
Modula ions”. In: Neu omodula ion: Technology a he Neu al In e ace
24 (2020).
[18] L. Mo ei a, J. Figuei edo, J. Ce quei a and C. P. San os, ”Assis -As-
Needed Elec omyog aphy-Based Con ol o a Wea able Ankle Robo ic
O hosis,” 2024 10 h IEEE RAS/EMBS In e na ional Con e ence o
Biomedical Robo ics and Biomecha onics (BioRob), Heidelbe g, Ge -
many, 2024, pp. 716-721.
