grasp and force based taxonomy of split-hook prosthetic ... · pdf filegrasp and force based...
TRANSCRIPT
Grasp and Force Based Taxonomy of Split-Hook Prosthetic Terminal Devices
Joseph T. Belter, Bo C. Reynolds, and Aaron M. Dollar, IEEE Senior Member
Abstract — In this paper, we analyze the use of the body-powered split-hook prosthetic terminal device, which is the most commonly used upper-limb prosthesis. We developed two taxonomies of split-hook use, one on grasp shape and one on force exertion, illustrating the functional capabilities and use cases of the device. Video captured from an amputee using a body-powered split-hook during a number of common activities was used to lend weight to the completeness of the classifications. These taxonomies serve to establish a common language and means of comparing the types of grasps achievable by simple terminal devices to those of advanced myoelectric terminal devices or even human hands. The first taxonomy categorizes the grasp type based on the contacts with the environment while the second is categorized by the method and limitation of force exertion. We discuss the difference between grasps capable of holding objects compared to those that are capable of acquiring objects and the importance of non-prehensile uses of the split-hook. The classification schemes lay the groundwork for further detailed study of split-hook use, and the discussion of the use cases described may help guide terminal device developers to create improved prostheses.
Keywords—split-hook; terminal device; prosthetic hand
I. INTRODUCTION
Although hundreds of prosthetic terminal devices are available, ranging from task specific devices to almost indiscernible cosmetic replicas of human hands [1], the voluntary opening split-hook (shown in Fig. 1) is widely accepted as the most commonly utilized functional terminal device [2]. Even with advanced multi-fingered and multi-DOF myoelectric terminal devices now available on the market, many prosthetists and amputees still turn to the body-powered split-hook due to its proven robustness, performance, low cost, and light weight [3], and many amputees who have learned to utilize the split hook without any major difficulties are reluctant to switch to a newer technology with a long learning curve.
Researchers have long categorized and organized human hand grasp types to better understand and emulate the capabilities of human manipulation. An early taxonomy divides human grasps into six main categories: cylindrical, tip, hook, palmer, spherical, and lateral based on the approximate shape of the hand while grasping an object [4].
This was later extended to a more complex human hand grasp taxonomy by Cutkosky based on the grasp types required in a manufacturing environment [5]. Manipulation taxonomies have also been proposed that classify the relative motion between the object and hand rather than the “shape” of the grasp [6]. The information on human hand use is commonly collected through observation and through the analysis of film collected in a natural use setting. A recent study by Zheng et al. analyzed video footage of human hand use taken during non-structured tasks for machinists and housemaids [7]. By recording each grasping task, the relative distribution of grasp types was measured.
In this paper, we present two sub-classifications of split-hook prosthesis use to enable a better understanding of the functional capabilities and usage of this class of terminal devices. While the specific pose or shape of the hook with relation to the object may help to define the grasp type, it was also observed that within different grasp types there are multiple limitations to the forces that can be exerted on the object. For this reason, we separated our look at the split-hook into two taxonomies. The first is based on the nature of contact with an object, while the second is based on the types of force exertion that can be utilized.
There has been relatively little work related to thorough classifications of split-hook usage that focuses on the function of the device instead of the task. While the majority of studies on prosthesis use to date has primarily involved written surveys of users (e.g. [8]), others have supplemented with follow up visits to the amputee’s home to observe the use of the device in the normal and unstructured environment [9]. As far as the authors are aware, all of these studies focus on high-level function (e.g. usage frequency, classes of tasks) without detail on the specific ways in which the device is used. The most relevant study that we were able to identify was performed by Fraser in 1998, in which 66 amputees were videotaped at their homes using their terminal device while performing a set of common tasks [10]. The video was then analyzed and each action involving the terminal device was categorized into a manipulative or non-manipulative category. The secondary categories included, grip, release, hold, transfer, support, steady, etc. While generally a very thorough study, it did not focus on specifics of the device usage and the categorization was broad enough to cover the wide range of devices used, including purely cosmetic terminal devices.
This work was supported in part by the Gustavus and Louise PfeifferResearch Foundation.
J. T. Belter, B. C. Reynolds, and A. M. Dollar are with the department of Mechanical Engineering and Material Science, Yale University, NewHaven CT 06511, USA (email: [email protected], [email protected], and [email protected])
978-1-4244-7929-0/14/$26.00 ©2014 IEEE 6613
tysimsdtainud
A
th1th[
tiFNfmfgcgcnthgAth
toacsthlc
We are inteype terminal d
specific desigmproved, and
scheme. We bdesign of the saxonomies. Wncluding ways
utilize it to imdesigns.
A. Description
The split-hohe original pa
1912 [11]. Desphe most popul2].
In the most itanium or stai
Fig. 1). The inNitrile or Neofriction and prmetal hooks. Tfeature (sometigrasp slender, lcigarette. At thgrip surfaces ccutout in the naturally centerhe wrist. The i
gripping surfacAlso, the clamphe contact poin
The top fingeo the plane of
and supports tcontrol cable asimilar to the whe back side ength of the
control cable a
Figure 1. Standarand the method of
erested in studydevice usage w
gn shortcomintherefore wou
begin this papesplit hook and
We then presens to extend it tomprove and
II. M
n of Split-Hook
ook terminal datent was awarpite its crude alar terminal de
basic form, tinless steel fingnside of the finoprene grip marevent objects The grip matimes called thlong cylindrica
he base of the called the clam
center to prr cylindrical obintended purpoce on anotherp can provide nts are closer toer has a long fethe finger. Thi
the attachmentand serves as way the humanof a pencil (sthumb determ
about the hinge
rd features of the bf operating the hoo
ying the specifwith enough d
ngs and areauld like a detaier with a detad methods usednt and discuss o other devicerefine future
METHODS
k
device has charded to David aesthetic appeaevice for uppe
the split hook gers shaped to ngers is coatedaterial to incr
from being sterial has an e “cigarette noal objects simifingers there a
mp. These platovide additionbjects along anose of the clamr plane to thata much largero the pivot poieature protrudiis feature is cat point for th
a contact pon thumb base asee Fig 1, cen
mines the mome at the base o
body powered splitok using a body-po
fics of split-hodetail to identas that can led classificatiailed look at td to develop tthe taxonomi
e types as well terminal devi
anged little sinW. Dorrance
arance, it remaier limb ampute
consists of twform a hook (s
d with a layer rease the surfascratched by tadditional bulotch”) that hellar to a pencil are two opposites have a smnal support a
n axis in line wmp is to provide
t of the finger grip force sinnt of the devicing perpendicualled the “thumhe body-poweroint for graspiand palm suppnter image). Tment arm of tof the fingers.
t-hook terminal deowered control cab
ook ify be
ion the the es, as
ice
nce in
ins ees
wo see of
ace the lge lps or
ing mall and
with e a ers. nce e.
ular mb” red ing ort
The the In
the vol1, the felastic usuallyuser’s elastic strengthan objeto havinfatigue
The weight split-hograms fthe simabout eWhen mfingers importahands.Dorransplit-hodifferendesign results,
B. Ter
In othat wil
evice. Here, the thble.
luntary-openingfingers are conbands. A pul
y accomplishedshoulders or ubands and o
h of the elasticect when it is bng a higher nuassociated witmajor benefitsand durability
ook weighs onfor modern my
mple metal conenvironmental manipulating sdoes not obst
ant factor whWhile we fo
nce 5x style hooook type tences such as la
features may, as described l
rminology
order to removell be used withPrehensile –supported wirequirement oContact – Podevice and thGrip Securityexternal distuExternal Forworld that wothe terminal include pushiInternal Forcfingers of theinclude pincfashion.
hree views show th
g style terminanstantly being ll of the bodyd through rel
upper arm, oveopens the finc band determbeing gripped umber of elasticth pulling the cs of the split hy. The standar
nly 213 grams yoelectric handnstruction userfactors such a
smaller objecttruct the view
hen comparingocus the taxooks, the resultserminal deviack of the “thuy require sliglater.
e ambiguity, whin the taxonom– The object ithin the termof gravity to hooint of interactie object or envy – Ability to urbances applierce – A force ould require adevice. These
ing, pulling, or ce – A force the terminal deviching and gra
he classic Hosmer
al device, as sheld together
y-powered conlative movemeercomes the fongers. The nu
mines the forcein the device. c bands is the control cable. hook are attribrd Hosmer Doas compared t
ds [12]. Also, rs do not needas heat and abts, the thin proof the object.
g hooks withonomies on ths apply to mostices. Howevumb” or additioht modificatio
we have definedmy. is intended to
minal device wold the object inion between thvironment.
maintain the ed to the object applied to th
an equal force e types of lolifting.
hat is created bce. These typeasping in a
Dorrance 5XTi ho
een in Fig. by a set of
ntrol cable, ent of the orce of the umber and e placed on
A tradeoff increase in
buted to its orrance 5X to 400-600 because of d to worry brasion [2]. ofile of the
This is an functional
he Hosmer t classes of
ver, small onal finger ons to the
d the terms
o be fully without the n place. he terminal
grasp with t. he external to support ads would
between the es of forces
prehensile
ook shape
6614
C
dupfncsiwruvvgdainr
A
thfpmnloaintyapimhty
C. Method of E
The presentediscussions betuser of a Hosmpowered split-footage was anumerous acticooking/eating shows four sams being used to
was gathered recording systeuser was recorvideo data wavalidate the taxgrasp or interadeveloped claadditional categn Fig. 3 and
recorded video
A. Grasp-base
The grasp taxhe split-hook i
first divided prehensile catmanipulative anon-prehensile ocations of co
all surfaces of nner surface wypes of interac
and often in presentations omportant aspe
hook and the sype interactionThe prehensi
Figure 2. Videounilateral body-po
Establishing Ta
ed taxonomies tween the auth
mer Dorrance 5-hook. In addacquired of thivities, includmeals, and sta
mple frames froo accomplish susing a head
em similar to trded while in as studied by xonomies, withaction seen inassification scgories were reFig. 4 are ba
.
III.
ed Split-hook T
xonomy (Fig. 3is used to intethe interactiotegory, in a
and manipulativcategory w
ontact on the hthe hook inclu
were used in a ctions were use
assistance to of terminal devcts are often oshape help it ns more predictile category i
o frames showing owered split-hook
axonomies
were generatehors, one of whX-Ti, voluntar
dition, over 5 hat author’s hding hygienic andard office som the video insimple tasks. T
mounted webthe system usehis natural hoall the autho
h the authors cn the video achemes to equired. The ex
ased on those
RESULTS
Taxonomy
3) attempts to ract with the ens into a no
similar wayve category us
was then subdhook. It was shuding the tip, f
non-prehensiled to provide st
the able havice grasping coverlooked. Thto perform thtably. is first sub-di
different grasp tuser.
ed after thorouhich is a 10-yery-opening bod
hours of vidhook use duri
tasks, laundstyle work. Fign which the ho
The video footabcam and vid
ed in [7], and tome setting. Thors and used categorizing eaaccording to tensure that xamples depict
seen within t
capture any wenvironment. Won-prehensile y to the nosed by [10]. Tdivided by thown that almofront, thumb, ale manner. Thetabilizing actio
and. In standacapabilities, thehe rigidity of te non-prehens
ivided based
types exhibited by
ugh ear dy-deo ing
dry, g. 2 ook age deo the his to
ach the no ted the
way We
or on-The the ost
and ese ons ard ese the sile
on
whethemain pwith ththe “finwithin occurribased osplit-ho“precislong, than addiwhere ain contcombinorganizgrasp scontact
Note in Fig. contralacomplethese gwith retrying tFor exahold apositionwithout
B. For
The sdifferenenvironexertionvery simfunctioThe selimits environsystem,
The used toprosthetradeofattacheduring loads wthe user
The ihook anumberaccordialso bepowereis graspthe mabands, size of
y a
er the fingers aprehensile featuhe outer surfacenger/exterior” the “clamp”.
ing between thon the number ook. These aresion grasps” onhin objects, anitional feature a large portiontact with the
nation with thzed from left tsecurity which ts between the that a few par3, typically caateral hand or
exity and the ggrasps generallespect to the spto pick up an ample, the coman eating orning of the utt placing it in t
rce-based Split
split-hook forcnt methods annment. The fin based on extmilar to the di
ons since all precond sub-catethe amount o
nment. This , the elastic bansuspension sy
o fixate the pretic socket anff exists betwed and the lev
pushing, pullinwithin this cater. internal grippinare generally r and stiffnesing to their pree modulated bed control cableped and all tenaximum force the location o
f the object. F
are used. Althure of the splie of the fingergrasp, with a For prehensi
he fingers, theyof contacts be
e limited to twn small objects)nd often involvsuch as the “th
n of the entire gobject, which
he “clamp”. Tto right based is strongly corhook and the o
rticular grasps,annot be achievr terminal devgeometry of thly require propplit-hook and wobject directly
mmon “Thumbr writing utetensil that is nthe grip using t
t-hook Taxonom
e exertion taxond limitation irst categorizatternal or internivision of nonrehensile graspegorization is bof force that
includes the ands, and the coystem consists rosthesis to thend body powween how tigvel of comfortng, or lifting is
egory are also l
ng forces of thdetermined by
ss of which aeferences, needby a balance e and the elast
nsion is releaseexertion is d
of the object wFor a standard
hough the fingit-hook, graspis was also obsroll of tape),
ile grasps wity were again setween the objwo contacts (u), four contactsving both the fhumb”), or aregrip material sh is sometimeThe entire taxon an increasrrelated to the object. , marked with ved without thevice. Partly duhe objects useper position ofwere not achievy from a table b/Fingers” grasensil, requirenot generally the other limb.
my
onomy (Fig. 4)of exerting lotion separatesal forces. This
n-prehensile or ps require interbased on the can be exert prosthesis ontrol cable.
of the entiree amputee, incwered harnessghtly the pro, the amount s limited. The limited by the
e voluntary-opy the elastic bare chosen byds, and abilitieof tension in ic bands. When
ed from the condependent on within the hoo
amount of el
ers are the ing objects served (e.g.
as well as th contacts sub-divided ect and the usually for s (often for fingers and ea contacts, surface was es used in xonomy is e in object number of
an asterisk e use of the ue to their ed in them, f the object vable when or drawer. sp, used to s specific achievable
) shows the oad on the s the force
division is prehensile
rnal forces. factor that
ted on the suspension
e apparatus cluding the s. Since a osthesis is of loading achievable strength of
pening split bands (the y the user es), but can
the body-n an object ntrol cable, the elastic
ok, and the lastic band
6615
t1incalothcthb
rffadtcs(cm
ension, the gri1-3 cm sized on the clamp ca
crushing a peaalso overpoweroad with the ohis type of gra
cable, minus thhe elastic band
band. Often more d
required. In thforce on the cforce is placedalone. Ultimatedifference betwension stored
concentration, tsmall spring, o(far right). Hocable, it can bemoving their a
Figure 3. The Splharness. *Grasps
ip force at the objects. The grian reach 140 Nanut within ther the strength ooutside of the asp is the tensiohe force requirds, as shown in
delicate and finhis case, the ucontrol cable sd on the objectely, the grip foween the loadin the elastic the user is ableor a compliantowever, due toe difficult for tharm in space.
it-Hook grasp taxoonly observed by p
fingers is betwip force of a 1
N. In Fig. 4, the clamp featurof the elastic bfingers. The l
on in the body red to open th
n Fig. 4 separat
ne movements ouser can alter such that a mut than that of trce exerted on
d in the controbands. With a
e to grasp delict drink lid, as o the requiredhe user to keepThe force-bas
onomy shows all oplacing the object
ween 45-70 N f cm sized obje
he hook is showre. The user cbands and exerlimiting factor powered contr
he fingers due ting a stiff rubb
of the fingers athe position auch smaller grthe elastic ban
n the object is tol cable and ta high degree cate objects likeshown in Fig
d tension on tp that pose whsed taxonomy
observed uses for twithin the hook u
for ect wn can rt a
in rol to
ber
are and rip
nds the the of
e a . 4 the
hile is
sorted require
The tthe spliinteractbased oachievacommoof achiorientatintende
A. Obj
A mdifferentype. SpreposiFor exaFig. 2 (hand pomost gr
the voluntary-openusing the able hand
from left to ried to perform e
two taxonomieit-hook is usedtion types. Alon the objects able prehensilon object sizes ievable grasp tions, dependi
ed to be perform
ject Acquisition
major feature nce in acquisit
Some of the gritioning of theample, the four(bottom) securositions the forasp types invo
ning split-hook as d.
ight by the leveach category o
IV. DISCUSS
es presented ind to perform a lthough the gbeing grasped
le grasps ecoand geometrietypes hold m
ing on the remed.
n versus Holdi
of grasp uttion methods frasps are unace object prior tur-contact, thumring a fork, is rk within the holving multiple
controlled by a bo
vel of user conof force exertio
SION
n Fig. 3 and 4 wide array of
grasp taxonomd, it can be seompass a lares. Furthermore
many objects iequirements o
ing
tilization relatfor each partic
chievable withoto closing the mb/finger grip,only achieved hook. This is the contacts with
ody-powered contr
ncentration on.
4 show that f grasp and
my was not en that the
rge set of e, the array in multiple of the task
tes to the cular grasp out careful split-hook. , as seen in if the able
he case for h the thumb
rol cable and
6616
ohthowbpdsopd
andowSatehthc
B
hmthptyvacin
or clamp. Prephand manipulahe object befo
object into a mwithin hand mabe placed in thplaces a large ddevice for mucshows that not of acquiring ipractical usagedoing unilatera
The necessitya useful grasp non-prehensile devices. In factof the split-howithin-finger tSimilar results amputee requirerminal device
easier to simplhand. Perhaps shis additiona
concentration, t
B. Modificatio
Designers hahook design anmethods [1]. Whe hook could
presented in thypes including
voluntary closappropriately ncylindrical opentended to gra
Figure 4. The spthe maximum for
osition of objetion when oneore within-han
more stable or anipulation in he hook in thedemand on thech of the pract
only is the grit in order to e to both unill and bilateral
y for careful prposition may uses of the
t, the most comook user weretwo-contact gr
were found bres the use oe in holding or ly perform the since non-prehal effort andthey are perfor
ons of Hooks/ N
ave seen limitnd proposed al
We can easily d expand, or
his paper. Fig. g the “Work Hing hook (rignamed based ening between sp the handle o
plit-hook force taxrce exerted and the
ects is observee grasp type isnd manipulatio
useful grasp. the split-hook
e exact positioe contralateral ical function orasp important
determine itlateral and bitasks. repositioning p
lead to the hsplit-hook or
mmon grasps se non-prehensrasps the nextby Fraser [10f the able hausing an objectask complete
hensile functiond require vrmed more freq
New Hook Desi
ations to the lternative featusee how the ffill in, the g5 shows two
Hook” (middleght) [1]. The
on the additithe fingers. T
of a broom or s
xonomy shows thee amount of user c
ed in able hums used to pick on transitions tWith the lack , the object mu
on for use whihand or termin
of the hook. Tht, but the meths relevance alateral ampute
prior to achievihigh incidence r other terminseen in the vidsile grasps, wt most commo]. If a unilate
and to assist tct, then it may ely with the abns do not requ
very little usquently.
igns
traditional splures or actuatifeatures added rasp taxonomadditional ho
e) and the Sierwork hook w
ion of the larThis opening wshovel. When w
e methods of forceconcentration requi
man up the of
ust ich nal his
hod and ees
ing of
nal deo
with on. eral the be
ble uire ser
lit-ion
to ies
ook rra
was rge
was we
look at a moregrasp ty
Whillimitatirobust m“wrap betweepower hands,standarmade uoften saor prost
Whensplit-honature importaobject the devcould bpowereneed tovoluntaoffer ththe hooas deschook incontrolcumber
In adnumeroThese space particulrod [1].
e exertion utilized ired to achieve tha
t the grasp taxoe robust alternype. le the work hion to the capamedium and lagrasps” as the
en grasper andgrasps make uyet are not c
rd split-hook. up for by utilizandwiching larthesis socket an looking at took, it is clear of any voluntant of which than the limitavice were a vbe directly sp
ed cable. One o be maintaineary closing hohe ability to exok. The voluntacribed above buin position whl cable (althoursome and somddition to thesous task speciftypes of devicof required glar aspect such.
by split-hook userat particular task.
onomy in Fig. native to the
hook partially abilities of the sarge diameter pey involve a
d object to supup a large portcapable of beiThe lack of azing the split-rge cylindrical
and their chest the force exethat there are
tary opening teis the inabilityation imposedvoluntary clospecified by th
drawback is ed while holdiook lacks the xert larger forary-closing oput also featureshen the tensiough this lockin
metimes unreliase changes in fic hook-type tces do not attgrasp types, h as holding a
rs. The organizatio
3, we see thatfour-contact f
addresses it, split hook is thpower grasp (olarge amount
pport large loation of grasps ing performeda large power -hook wearer’sl items betweeor stomach. rtion capabilitlimitations ba
erminal devicey to close tigh
d by the elasticsing split-hookhe tension in that this tens
ng an object. clamp feature
ces within theeration gives ts a lock that won is releasedng feature can
able). the split hoo
terminal devicetempt to span but focus sobaseball glove
on is based on the
t this offers finger-span
one major he lack of a often called
of contact ads). These for human
d using the grasp was
s body and en the hook
ties of the ased on the e, the most hter on an c bands. If k, the load
the body sion would The Sierra e but does e fingers of the benefits ill hold the
d from the n be a bit
ok designs, es are sold.
the entire olely on a e or fishing
limitation to
6617
fptotohothafcwimoo[cwnosaginpr
P
[
[2
[3
V.
Using the splfunctional use prostheic hand o perform a wo acquire all
hand and the aof the voluntarhe presented ta
analysis in orfunction of numclassified the twe would likmportance of e
or the types of others. This an5,6,7], but non
classified and rwould like to enumber of subjof grasp typestatistical distriadd an additiongrasp listed in nformative de
prosthetic handrequired by the
The authorsPfeiffer Resear
1] RSL Steeper 2009. http://bebioni_limb_catalog
2] C.M. Fryer, JComponents”RehabilitationSurgeons, edi
3] G. H, Kejlaprotheses to u163, 1993
Figure 5. Basic (middle) and the S
CONCLUSION
it-hook taxonoof the hook wuse. Although
ide variety of the grasps wit
ability to exert ry opening conaxonomies cander to compamerous terminypes of grasps
ke to get a each grasp typegrasped objec
nalysis has been-prehensile furecorded. To dextend our vidjects and taskse and duratioibution of grasnal element relthe taxonomy
ecision makingds depending o end user.
ACKNOW
s would like to ch Foundation
REFER
“Upper Limb ProRetrieved
ic.com/uploads/filegue.pdf J.W. Michael, “U” Atlas of Limb Pn Principles, Ch. 6ition 2, 1992, repri
aa, “Consumer coupper limb ampute
Hosmer 5X hookSierra Hook (right)
NS AND FUTURE
omy, we can eaith both humanh the split-hoofunctions, it lathout assistancforces is limit
ntrol strategy. n allow for moare and betternal devices. Als achievable bbetter unders
e based on the ts that are mor
en performed functions of the do so with termdeo analysis fos, as well as toon to enable sp type utilizatlated to the imand therefore
g process for on the usage a
WLEDGMENTS
thank the Gusn for supporting
RENCES osthetic Componen
Jan es/upper
pper-Limb ProsthProsthetics: Surgi6A, American Acinted 2002 oncerns and the ees,” Prosthet Orth
k (left) compared t) [1].
E WORK
asily compare tn hand and oth
ok has the abilacks in the abilce from the abted by the natuWe believe th
ore detailed grar understand tlthough we hay the split-hoostanding of tfrequency of u
re important thfor able hands hand were rare
minal devices, wr a much grea
o include loggicalculations
tion. This woumportance of ea
allow for a mothe designer
and functional
tavus and Loug this work.
nts Catalogue” 202013 fr
hetics: Body-Poweical, Prosthetic, aademy of Orthope
functional value hot Int, vol.7, p. 1
to the “Work Hoo
the her lity lity ble ure hat asp the ave ok, the use han
in ely we
ater ing of
uld ach ore of
lity
ise
08-rom
ered and edic
of 57-
[4] G. in EGer
[5] M.RHan(19
[6] I.MClaTra
[7] J.ZGraTas9-1
[8] C. PmyDec
[9] A. Zuiamp
[10] C.MfunOrt
[11] D.W191
[12] J.TDesProDev
ok”
Schlesinger, “DerErsatzglieder und rmany: Springer, 1R. Cutkosky, “On nds for Manufactu
989) 269-279. M. Bullock, R.Rassification of Humansactions on Hapt. Zheng, S. De Lasp Type and Fresks” Int. Conf. on 3, 2011. Pylatiuk, S. Schul
yoelectric prosthetic. 2007 Van Lunteren, GHithoff, “A field putees” Prosthet O
M. Fraser, “An enctional prosthesesthot Int, 22:216, 19W. Dorrance, “Ar12 . Belter, J. Segil,sign and Perfo
osthetic Hands” velopment, Aug. 2
r Mechanische AuArbeitshilfen, M
1919. n Grasp Choice, Gruring Tasks,” IEEE
R. Ma, and A.Muman and Robot Dtics, 2013 vol. 6(3
La Rosa, and A.Mequency in Daily Robotics and Auto
lz, L. Doderlein, “Ric hand users” Pro
HM Van Lunterenevaluation of a
Orthot Int. 7, 141-evaluation of thes by unilateral up998 rtificial Hand”, U
A.M. Dollar, anormance Specific
Journal of R2013 50(5):599-61
ufbau der Kunstlic. Borchardt et al.
rasp Models, and E Trans. Robot. A
M. Dollar “A Dexterous Manipu), pp. 129-308.
M. Dollar “An InvHousehold and Momation, Shanghai
Results of an Interosthet Orthot Int.,
n Gerritsen, H.G. arm prostheses f-151, 1983 e use made of c
upper limb amput
US Patent #1,042,4
nd R.F. Weir “Thecations of AnthRehabilitation Re18
hen Glieder,” , Eds. Berlin,
the Design of Automat., 5(3),
Hand-Centric ulation” IEEE
vestigation of Machine Shop i, China, May
rnet survey of 31(4):362-70
Stassen, M.J. for unilateral
cosmetic and ees” Prosthet
413. Oct. 29,
e Mechanical hropomorphic esearch and
6618