sensors

Communication
Human Hand Anatomy-Based Prosthetic Hand
Larisa Dunai 1, * , Martin Novak 2 and Carmen García Espert 3

1 Centro de Investigación en Tecnologías Gráficas, Universitat Politècnica de València, camino de Vera s/n,
46022 Valencia, Spain
2 Faculty of Mechanical Engineering, Czech Technical University in Prague, Technická 4, Praha 6,
166 00 Prague, Czech Republic; [email protected]
3 Hospital La Fe, Avinguda de Fernando Abril Martorell, 106, 46026 Valencia, Spain; [email protected]
* Correspondence: [email protected]; Tel.: +34-689-215-528

Abstract: The present paper describes the development of a prosthetic hand based on human hand
anatomy. The hand phalanges are printed with 3D printing with Polylactic Acid material. One of the
main contributions is the investigation on the prosthetic hand joins; the proposed design enables one
to create personalized joins that provide the prosthetic hand a high level of movement by increasing
the degrees of freedom of the fingers. Moreover, the driven wire tendons show a progressive grasping
movement, being the friction of the tendons with the phalanges very low. Another important point is
the use of force sensitive resistors (FSR) for simulating the hand touch pressure. These are used for
the grasping stop simulating touch pressure of the fingers. Surface Electromyogram (EMG) sensors
allow the user to control the prosthetic hand-grasping start. Their use may provide the prosthetic
hand the possibility of the classification of the hand movements. The practical results included in the
paper prove the importance of the soft joins for the object manipulation and to get adapted to the
object surface. Finally, the force sensitive sensors allow the prosthesis to actuate more naturally by
adding conditions and classifications to the Electromyogram sensor.

Keywords: prosthetic hand; MyWare sensor; force sensing resistors; human hand anatomy







Citation: Dunai, L.; Novak, M.; 1. Introduction

García Espert, C. Human Hand More than 3 million people suffer from hand amputations or loss due to health disor-
Anatomy-Based Prosthetic Hand. ders caused by infections, congenital absence, diabetes, cancer or others [1,2]. Over 75%
Sensors 2021, 21, 137. https:// of the amputations are partial [3]. Hand loss has an important impact on the person’s
dx.doi.org/10.3390/s21010137
functional aspect. Many of the people with a loss of the hand have the possibility of using
a prosthetic hand. The development of prosthetic hands has been less based on their
Received: 27 October 2020
functionality, relying more on human hand aesthetic aspects [4–8]. With the technological
Accepted: 22 December 2020
advances in biotechnology, the innovation reached the area of robotics and prosthetic hand
Published: 28 December 2020
development. Consequently, current commercial prosthetic hands have become more
sophisticated. They are fitted with sensors and actuators, so that the fingers are motorized
Publisher’s Note: MDPI stays neu-
tral with regard to jurisdictional claims
and can perform grasping movements. Nevertheless, automatized prosthetic hands are
in published maps and institutional
expensive and not accessible to all social strata. Usually, the most common prosthetic
affiliations. hands are passive, and their goal is to substitute the human hand more esthetically than
functionally. Powered prosthetic hands are classified in body powered and external pow-
ered prosthetic hands [9]. Body powered prosthetic hand mechanisms are actuated by
human body movement through wires or cables. Usually, these types of devices are simple
Copyright: © 2020 by the authors. Li- devices with grasping movement and are relatively lightweight. Moreover, body powered
censee MDPI, Basel, Switzerland. This prosthetic hands require harnessing. External powered prosthetic hands are based on
article is an open access article distributed external power and actuators. Some of these types of prosthetic hands are controlled
under the terms and conditions of the
by Electromyograms (EMGs) [10,11] for grasping. The most common EMG-controlled
Creative Commons Attribution (CC BY)
prosthetic hands use surface EMG [12] while a few others use intramuscular EMG [13,14].
license (https://creativecommons.org/
EMG prosthetic hands are amplitude-based measurement devices and, usually, the control
licenses/by/4.0/).

Sensors 2021, 21, 137. https://dx.doi.org/10.3390/s21010137 https://www.mdpi.com/journal/sensors

Sensors 2021, 21, 137 2 of 15

is slow. Because most of the prosthetic hands are controlled by a single input, the control
of individual fingers or joins is not allowed. Usually, the prosthetic hands based on EMG
use electrical signals of two antagonist muscle contractions. They allow two directions
of movement: flexion and extension; one is for start grasping while the other is to start
extension. As the EMG based prosthetic hands do not have external cables, these de-
vices are more esthetical. To obtain more than two movements for the prosthetic hand,
it is required to introduce more conditions, such as triggering or artificial intelligence
(pattern recognition and classification). Prosthetic hands that are aimed to perform move-
ments for all fingers operate in a sequential order with time delay. In some prosthetic
hands, the movements of the different fingers are performed by using several contractions
of the same muscle (quick contractions of the same muscle) or by alternating both muscle
contractions to control different joint movements. Another control system is based on
force-sensing resistors, pull or push switches or Inertial Measurement Unit (IMU) [15].
Prosthetic hands also include hybrid prostheses. Hybrid prostheses are body pow-
ered and externally powered devices. Often, these devices are used in cases of upper
limb amputations, including transhumeral and shoulder. Regarding external powered
prosthesis devices, these can be classified as those with one degree of freedom and those
with multiple degrees of freedom. Devices with one degree of freedom perform only the
extension and flexion movements. Usually these devices are robust [16,17]. The ones with
multiple degrees of freedom, also known as multiarticulated prostheses, are fitted with
several actuators for different fingers and/or interphalangeal joins [17,18]. They use small
actuators that perform the required movement. Despite the high accuracy of the EMG
signals, the researchers are still looking for the best methods of prosthetic hand control by
combining EMG with artificial vision [19], microphone [20], tongue control system [21], etc.
Table 1 presents the characteristics of the most advanced bionic prosthetic hands.
The present study introduces novel prosthetic hand development, based on a com-
bined control system that uses EMG, buttons, and force-sensing resistors. The device design
is totally based on the human hand anatomy. All phalanges are human hand-scanned
phalanges. The ligaments and joins are strictly developed as real ones. The device has
15 DOF (Degree of Freedom) and the joins have different speeds and forces. The soft mate-
rial joins provide the prosthetic hand with a high level of adaptation to the object surface.
They increase the DOF of each join, enabling the small abduction/adduction. The use of
force sensitive resistors allows the prosthetic hand to simulate the touch pressure sensing
that stops grasping movement.
The paper is structured as follows: Section 2 describes the materials and methods of
the prosthetic hand development. Section 3 presents the experimental results and, finally,
in Section 4, the conclusions are provided.
 multiple
multiple degrees
multiple
multiple
degrees of freedom.
degrees
degrees
of of of
freedom. Devices
freedom.
freedom.
Devices with
Devices one
Devices
with one degree
with
withone
degreeoneof freedom
degree
ofdegree of of
freedom perform
freedom
freedom
perform only
onlythe
perform
perform exten-
only
the only thethe
exten- exten-
exte
sion
sionand
and flexion
sion
sionandand
flexion movements.
flexion
flexion
movements. Usually
movements.
movements.
Usually these
Usually
Usually
these devices
these
these
devices are
devicesrobust
devices
are robust [16,17].
areare
robust
robust
[16,17]. The ones
[16,17]. The
[16,17].
The ones with
Theones
with mul-
oneswith
mul- mul-
with mu
tiple degrees
tiple
tiple of
degreesfreedom,
degrees of of also
freedom,
freedom, known
also
also as
knownmultiarticulated
known as multiarticulated
as prostheses,
multiarticulated
tiple degrees of freedom, also known as multiarticulated prostheses, are fitted with sev- are
prostheses, fitted
prostheses, are with
fitted
are sev-
fitted with sev-
with se
eral actuators
eral
eral
eral actuators for
actuators different
actuators forfor
for different fingers
different
different and/or
fingersfingers
fingers
and/or interphalangeal
and/or
and/or joins
interphalangeal
interphalangeal
interphalangeal [17,18].
joins
joins
joins [17,18]. They
[17,18]. use
They
[17,18].
They use small
They ac-
useuse
small small
ac-smallac-a
tuators that
tuatorstuators perform
tuators
that thatthat
perform the
the required
perform
perform thethe
required movement.
required
required
movement. Despite
movement.
movement.
Despite the
Despite high
Despite
the high accuracy
thethehigh
high
accuracy of
of the
accuracy theEMG
accuracy of
EMG sig-
thethe
of EMGEMG
sig- sig-
si
Sensors 2021, 21, 137 3 of 15
nals,
nals, the
the researchers
nals,
nals,thethe
researchers are
are still
researchers
researchers are
still looking
still
are still
looking for
lookingthe
looking
for best
thefor the
for
best methods
thebest
best
methods of
of prosthetic
methods
methods of of hand
hand control
prosthetic
prosthetic prosthetic hand
hand
control by
control
control
by byb
combining
combining EMG
EMG with
combining
combining EMGEMG
with artificial
with
with
artificial vision
artificial
artificial[19],
visionvision microphone
vision
[19], [19],
[19],
microphone [20],
microphone[20], tongue
microphone [20],
tongue[20],control
tongue
tongue
control system
control
control
system [21],
system
system
[21], [21],
[21
etc. Table
etc. 1
etc.presents
Table
Table 1 1 the
presentscharacteristics
presents the the of
characteristics the
characteristics most
of the
of advanced
most
the most bionic
advanced
advanced
etc. Table 1 presents the characteristics of the most advanced bionic prosthetic hands. prosthetic
bionic
bionic hands.
prosthetic
prosthetic hands.
hands.
Table 1. Bionic prosthetic hands characteristics.
Table
Table1.1.Bionic
Table 1.prosthetic
Table
Bionic Bionic
1. Bionic
prosthetic hands
handscharacteristics.
prosthetic hands
prosthetic characteristics.
hands characteristics.
characteristics.
Miquelangelo Vincent Hand
Model
Miquelangelo [22]
Miquelangelo i-Limbi-Limb
i-Limb [23] [23][23] BeBebionic
Bionic [24]
[24] Sensor
Sensor Hand
Hand [25]
[25] Vincent Hand [26]
Miquelangelo [22] [22]
Miquelangelo
[22] [22] i-Limb
i-Limb
[23] [23] Be BeBe
bionic bionic
[24] [24]
bionic [24] Sensor
Sensor Hand Hand
Sensor [25]
Hand
[25] [25]Vincent
Vincent Vincent
Hand
[26] Hand [26]
Hand
[26] [26

Model
ModelModel
Model
Characteristic
Characteristic
Characteristic
Characteristic

Characteristic
Developer
Developer Developer Otto
Developer OttoBockOtto
BockOttoBockBock Touch
TouchBionics
Touch
Touch
Bionics Bionics Otto
Bionics OttoBock Otto
Bock Bock
Otto Bock Otto-Bock Otto-Bock
Otto-Bock Vincent
Otto-Bock
Vincent Systems
Vincent
Vincent Vincent
Systems Systems
Systems
Developer
Weight, gr
Weight,
Weight, gr gr 510 Otto
510 Bock
510 Touch599
599 Bionics
599 Otto
500 Bock500500 Otto-Bock
500 500500 410 410 410
Weight, gr 510 599 500 500 410
Systems
Operating
Operating voltage,
Operating
Operating
voltage, VVvoltage,
voltage, V V11,1
11,1 11,111,1 7,4
7,4 7,4 7,4 7,4
7,4 7,4 7,4 6–8
6–8 6–86–8
Weight, gr 510 599 500 500 410
Battery type
BatteryBattery type
Battery
type type li-ion li-ion Lithium
li-ion li-ion Lithium polymer
Lithium
Lithium
polymer polymer li-ion
polymer li-ion li-ion
li-ion li-Pol
li-Pol li-Pol li-Pol
Operating
Battery, mAhvoltage, V 11.1 7.4 7.4 6–8
Battery,Battery, mAh
Battery,
mAh mAh 1500 1500 1500 1500 1300–24001300–2400
1300–2400 1300–2400 1300–2200
1300–22001300–2200
1300–2200 1300–2600
1300–2600 1300–2600
1300–2600
N°
N°Actuators
N°N°
Actuators
Battery Actuators
Actuators
type 22 2li-ion
2 66 Lithium 6 6 55
li-ion 5 5 11 1 1 66
li-Pol 6 6
DC motor
DCDCwithpolymer
motorworm
motor withwithworm
worm
Type DC motor with worm
Typeof actuators
ofType of of
Type actuators
actuators actuators
gear
DC
DCmotor DCDC
motor head
motor
head screw
head
motor
screwhead DC
DCmotor
screw
screw DCDC
motor DC
DCmotor
motor
motor DCDC
motor worm
motor
motor
worm gear
worm
worm
gear gear
ge
Battery, mAh 1500 gear
gear1300–2400gear 1300–2200 1300–2600
Active
N◦fingers
Active Active fingers
Active
fingers
Actuators fingers 33 3 32 55 65 5 555 5 5 221 2 2 565 5 5
Force,
Force, NNForce, NN
Force, 70
70 70 70 100
100 100100 140
140 140140 100
100 100100 6060 60 60
DC motor with DC motor head DC motor
Type of actuators SequentialSequential
Sequential Sequential DC motor Single
Single trigger
Single
Single
trigger trigger
trigger
Movement
Movement control
Movement
Movement EMG,
EMG,4channels
controlcontrol
control EMG,
EMG,4channels
4channels Mobile app,
MobileMobile
4channels worm
EMG
Mobile
app, app,
EMG gearEMG
app, EMG screw EMG
EMG EMG EMG worm gear
EMG
EMG EMG EMG EMG
EMG EMG EMG
Active fingers 3 5 5 2 Switch 5
Double and triple im- Switchsignal
Switch
Switch
signal signal
signal
Double Double
Double
and and
triple andtriple
im- im-im-
triple
Movement
Movement command
Force,
Movement
Movement Ncommand
command Switching
command Switching
Switching 70
Switching 100 Co-contractions
140Co-contractions 100
Co-contractions
Co-contractions Co-contractions
60
Co-contractions
Co-contractions
Co-contractions
pulse
pulse pulse pulse
Mobile app, Sequential Double
Double impulse
SingleDouble Double
impulse
trigger impulse
impulse
Movement control EMG, 4channels Audible bip
Audible bipbip EMG
EMG Audible
EMG Audible
bip EMG
Feedback
Feedback Feedback
Feedback NO
NO NONO NO
NO NONO NO
NO NONO Vbrations
Vbrations Vbrations
Vbrations
vibrationsvibrations
vibrations vibrations
Switch signal
Double and
Movement command Switching
The Co-contractions Co-contractions
The present
TheThe
present study
present
present
study introduces
triplestudy
study
introduces
impulse novel
novel prosthetic
introduces
introduces novel
novel
prosthetic hand
hand development,
prosthetic
prosthetic hand
hand based
based on
development,
development,
development, aa com-
based
onbased onon
com- a com-
a com
Double impulse
bined
binedcontrol
bined
bined
control system
control
control
system that
system uses
system
that that
uses EMG,
uses
that
EMG, buttons,
usesEMG,
EMG,
buttons, and
andforce-sensing
buttons, and
buttons, and resistors.
force-sensing
force-sensing
force-sensing resistors. The
Thedevice
resistors. The
resistors. The
device de-
device
device
de- de-d
sign is totally
sign issign
sign
totally based
is totally
isbased on
totally the
based
onbased
thehuman
on on
human hand
thethe
human Audible
human
hand anatomy.
hand
handbip
anatomy. All
anatomy.phalanges
anatomy. AllAll are
phalanges
All phalanges human
phalanges areare
are human hand-scanned
human
humanhand-scanned
hand-scanne
hand-scanned
Feedback NO NO NO Vbrations
phalanges. The vibrations
phalanges. Theligaments
phalanges.
phalanges. The
The
ligaments and
andjoins
ligaments
ligaments and
joins are
and
are strictly
joins areare
joins
strictly developed
strictly
strictly
developed as
asreal
developed
developed ones.
realas real
as
ones. The
ones.
real
The device
The
ones. The
device has 15
device15hashas
device
has 15

2. Materials and Methods

2.1. Prototype Design
The prototype was built using a human hand anatomy-based design. All the elements
of the prosthetic hand were based on real human hand measurements that included
the dimensions, proportions, and human hand functionality. The idea of the proposed
prosthesis relied on the reproduction of the human finger motions. For the prosthetic
phalanges design, the real human hand phalanges were 3D scanned and then designed
by using a 3D drawing technology—the Autodesk Inventor Professional 2019. The whole
prosthetic hand structure for actuators and processing supports was modelled with the
same 3D drawing tool. Before proceeding to the prosthetic hand design and assembly,
the main design specifications based on the human hand behavior as joins and movement
capabilities were analyzed. All hard elements were constructed by using 3D printing
technology with Polylactic Acid (PLA) filament that has good functional and structural
characteristics and that are suitable for 3D printing. One of the main novelties of this
prototype relies on the employed materials, which are ideal to reproduce human tendons,
ligaments, fibrous sheaths, joins, etc.
The human hand consists of Carpal bones, Metacarpal bones, Proximal,
Middle and Distal Phalanges. All fingers are based on four bones: Metacarpal bone,
Proximal, Middle and Distal Phalanges (see Figure 1). The thumb finger is different and has
one phalange less than the rest of the fingers: Metacarpal bone and the Proximal and Distal
phalanges. The joins are located between phalanges. There are fourteen joins for the whole
hand. The join between Carpal and Metacarpal bones does not have any Degree of Free-
dom (DOF). The Thumb is the only one with a Metacarpal join with abduction/adduction
Sensors 2021, 21, 137 4 of 15

movement
Sensors 2021, 21, x FOR PEER REVIEW with respect to the sagittal plane. The rest of joins have one DOF, flexion4 of
and15
extension movement with respect to the frontal plane.

(b)

(a) (c)
Figure
Figure 1.
1. Human
Humanhand
handanatomy.
anatomy.(a)
(a)Human
Humanhand
handbones
bonesand
andjoins;
joins;(b)(b)
volar plate
volar and
plate collateral
and ligaments;
collateral (c) (c)
ligaments; extensor
exten-
hooks.
sor hooks.

The
Takingabduction/adduction
into consideration theas number
well as the flexion/extension
of ligaments and their of the thumb are
characteristics, the inde-
artifi-
pendently controlled
cial ligaments by the
are chosen fromcontrol
rubbersystem. Thewith
materials kinematics
differentofhardness
the Indexand
finger are repre-
elasticity char-
sented in Figure 2. The five fingers are driven by six actuators; each finger is controlled by
acteristics.
one actuator,
It is knownexcept
that on
thethe
boneThumb, whichare
dimensions is important
actuated by fortwo. The purpose
the prosthetic handofdesign
this archi-
and
tecture, relyingThe
development. on using at least
phalanges one
and actuator per
Metacarpal finger,
bone is toconsidered
lengths allow the prosthetic hand to
for the prosthetic
perform
hand arefinger
those movements
corresponding independently.
to an adult female.
The average lengths of the human hand are presented in Table 2. The lengths of
the phalanges significantly affect the object manipulation and hand movement. As the
prosthetic phalanges are based on human hand anatomy, the length of the fingers is 99% of
the real hand; the 1% remaining depends on the joins. As the joins are reproductions of the
human hand joins, the abduction/adduction and rotation for each join is possible.

Table 2. Anatomic human hand dimensions.

Bone Thumb Index Middle Ring Little

Metacarpal Bone 1.3567 2.049 1.906 1.719 1.578
Proximal Phalange 1.134 1.489 1.683 1.563 1.254
Intermediate phalanges 0.864 1 0.994 0.719
Distal Phalange 0.74 0.757 0.798 0.778 0.698
Figure 2. Three-dimensional model of the prosthetic hand assembly and its kinematics.
As can be seen in Figure 2, the prosthetic hand kinematics is based on real-human
The abduction/adduction as well as the flexion/extension of the thumb are indepen-
hand anatomy. The solution for the joins is to design the volar plate, collateral ligaments,
dently controlled by the control system. The kinematics of the Index finger are represented
and extensor ligaments, as shown in Figure 1b,c using rubber materials with different
in Figure 2. The five fingers are driven by six actuators; each finger is controlled by one
hardness. The developed join elements are presented in Figure 3. The joins can perform 2
actuator, except on the Thumb, which is actuated by two. The purpose of this architecture,
DOF at each join that allows them to increase their functionality. Nevertheless, the abduc-
tion/adduction movement of the phalanges is so small that it does not have the sense to
 (a) (c)
Figure 1. Human hand anatomy. (a) Human hand bones and joins; (b) volar plate and collateral ligaments; (c) extensor
hooks.

Sensors 2021, 21, 137

The abduction/adduction as well as the flexion/extension of the thumb are inde- 5 of 15
pendently controlled by the control system. The kinematics of the Index finger are repre-
sented in Figure 2. The five fingers are driven by six actuators; each finger is controlled by
one actuator, except on the Thumb, which is actuated by two. The purpose of this archi-
relyingrelying
tecture, on using
onatusing
leastat
one actuator
least per finger,
one actuator per is to allow
finger, is tothe prosthetic
allow hand tohand
the prosthetic perform
to
finger movements
perform independently.
finger movements independently.

Figure 2. Three-dimensional model of the prosthetic hand assembly and its kinematics.
Figure 2. Three-dimensional model of the prosthetic hand assembly and its kinematics.
As can be seen in Figure 2, the prosthetic hand kinematics is based on real-human
handAsanatomy.
can be seen
Thein Figure for
solution 2, the
the prosthetic hand kinematics
joins is to design is based
the volar plate, on real-human
collateral ligaments,
hand anatomy. The solution for the joins is to design the volar plate, collateral
and extensor ligaments, as shown in Figure 1b,c using rubber materials with different ligaments,
2021, 21, x FOR PEER REVIEW
and extensorThe
hardness. ligaments, as shown
developed in Figure
join elements are1b,c using rubber
presented materials
in Figure 3. 5 The with different
of 15 joins can per-
hardness.
form 2 DOF at each join that allows them to increase their functionality.can
The developed join elements are presented in Figure 3. The joins perform 2
Nevertheless,
DOF at each join that allows them to increase their functionality. Nevertheless,
the abduction/adduction movement of the phalanges is so small that it does not have the abduc-
the
tion/adduction movement of the phalanges is so small that it does not have
sense to introduce it in the prosthetic hand. The only existing abduction/adduction and the sense to
introduce it in the prosthetic hand. The only existing abduction/adduction and flexion/ex-
flexion/extension movement in the four fingers is between the Metacarpal bone and the
tension movement in the four fingers is between the Metacarpal bone and the Proximal
Proximal Phalange, in the MCP join.
Phalange, in the MCP join.

(a)

(b)

Figure 3. (a)Figure
Finger3.flexor
(a) Finger flexor
tendon routetendon route
and joins and
(two joins
side (two side
collateral collateral
ligaments andligaments and(b)
volar plate). volar plate).
Finger extensor tendons.
(b) Finger extensor tendons. Three-dimensional
Three-dimensional model of the prosthetic index finger assembly. model of the prosthetic index finger assembly.

To reduce the movement

To reduce the range, the role range,
movement of the stopper
the role ofis played by the
the stopper is tendon
played rail.
by the tendon rail.
All fingers areAll
actuated through wires (tendons), which substitute
fingers are actuated through wires (tendons), which substitute flexor and extensor
flexor and extensor
tendons connected to the
tendons actuatorstopulley.
connected The mechanism
the actuators pulley. The formechanism
finger movement is based
for finger movement is based
on the endless onrouting tendons,
the endless in which
routing the flexor
tendons, and the
in which extensor
flexortendons are connected
and extensor tendons are connected
to the same actuator
to the pulley (see Figure
same actuator 4). This
pulley (see architecture
Figure 4). This enables driving the
architecture pulley
enables in
driving the pulley
both directionsinatboth
the same time. The
directions assembly
at the same time.of the
Theprosthetic
assemblyhand is presented
of the prosthetic in Fig-is presented in
hand
Figureversus
ure 5. The difference 5. Theanother
difference versus
devices another
is that, devices
in this is that,
prosthetic in thisthe
model, prosthetic
inter pha- model, the inter
langes-driven pulley is not used. The tendon passes through tendons rails of each pha- rails of each
phalanges-driven pulley is not used. The tendon passes through tendons
lange and endsphalange and ends
on the distal on the To
phalange. distal phalange.
avoid tendons Totearing,
avoid tendons tearing,15%
an additional an additional
of 15% of
tendon is added tendon is finger.
to each added to each finger.
 on the endless routing tendons, in which the flexor and extensor tendons are connected
to the same actuator pulley (see Figure 4). This architecture enables driving the pulley in
both directions at the same time. The assembly of the prosthetic hand is presented in Fig-
ure 5. The difference versus another devices is that, in this prosthetic model, the inter pha-
Sensors 2021, 21, 137langes-driven pulley is not used. The tendon passes through tendons rails of each pha- 6 of 15
lange and ends on the distal phalange. To avoid tendons tearing, an additional 15% of
tendon is added to each finger.

Sensors 2021, 21, x FOR PEER REVIEW 6 of 15

Figure 4. Mechanical architecture
Figure of the architecture
4. Mechanical tendon transmission.
of the tendon transmission.

For the hand control, five force sensing resistors are used. The sensors are placed on
the Distal Phalange muscle and are built using soft flex material. The Artificial Abductor
Muscle is fabricated with rubber and prevents the movement more than it is necessary. It
also enables to complete the palm of the prosthesis. As the thumb join with the trapezoidal
carpal bone is made by tendons and collateral ligaments, it allows the join to perform
3DOF (flexion/extension, abduction/adduction and turn).
The maximum motion angle of the thumb abduction is 80 degrees. For the other four
fingers (the Index, Middle, Ring and the Little fingers), the motion flexion/extension angle
is from 0 degrees to 90 degrees. Each join of the phalanges is dotted with artificial carti-
lages to avoid phalanges friction. The tendon sheath covers and holds the flexor and ex-
tensor tendons rails.

Figure 5. Mechanical assembly of the prosthetic hand.

Figure 5. Mechanical assembly of the prosthetic hand.
For the hand control, five force sensing resistors are used. The sensors are placed on
The actuators
the Distal Phalangeand EMGand
muscle are are
placed
builtinusing
the prosthesis forearm,The
soft flex material. while the servomotor
Artificial Abductor
for
Muscle is fabricated with rubber and prevents the movement more than itbones
the abduction/adduction movement of the thumb is located in the carpal of the
is necessary.
prosthesis.
It also enables to complete the palm of the prosthesis. As the thumb join with the trape-
zoidal carpal bone is made by tendons and collateral ligaments, it allows the join to perform
2.2.
3DOFControl System
(flexion/extension, abduction/adduction and turn).
Theelectronic
The maximumsystem motionofangle of the thumb
the prosthetic handabduction
is based onis 80 degrees. EMG
a Myware For the other five
sensor, four
fingers (the Index, Middle, Ring and the Little fingers), the motion flexion/extension
step motors, five drivers, a servo motor, an Arduino AtMega 2560, two pushbuttons, a angle
is from
shield 0 degrees
and to 90
five force degrees.
sensing Each join
resistors. Theofcharacteristics
the phalangesofisthedotted with artificial
electronic cartilages
components are
to avoid phalanges
presented in the Table friction.
3. The tendon sheath covers and holds the flexor and extensor
tendons rails.
Table The actuators
3. Electronic and EMG are placed in the prosthesis forearm, while the servomotor
components.
for the abduction/adduction movement of the thumb is located in the carpal bones of the
Components Specifications
prosthesis.
Input Voltage 7–12 V
2.2. Control System Analog Input Pins 16
The electronic system of the prosthetic hand is I/O
DC Current per basedPinon40 amA DC Current
Myware EMG for 3.3
sensor,
five step motors, five drivers, a servo motor, an ArduinoVAtMega Pin 50 mA2560, two pushbuttons,
a shield and five force sensing resistors. The characteristicsClock speed
of the16 MHz components
electronic
Arduino AtMega 2560
are presented in the Table 3. EEPROM 4 KB
SRAM 8 KB
Flash memory 256 KB
Analog inputs Pins 16
Digital Inputs 54
Sensors 2021, 21, 137 7 of 15

Table 3. Electronic components.

Components Specifications
Input Voltage 7–12 V
Analog Input Pins 16
DC Current per I/O Pin 40 mA DC Current for
3.3 V Pin 50 mA
Clock speed 16 MHz
Arduino AtMega 2560
EEPROM 4 KB
SRAM 8 KB
Flash memory 256 KB
Analog inputs Pins 16
Digital Inputs 54
Operating voltage 2.9 V–5.7 V
Myware EMG Operating Current 9 mA–14 mA
Output RAW and filtered signal
Measuring range 0–2 kg
Thickness <0.25 mm
Precission ±2.5%
Force sensing resistors
Initial resistance >10 Mohm
Voltage DC 3.3 V
Response time 1 ms
Operating Voltage 4.8 V
Operating current 50 mA
Servo motor Speed 0.12 at 4.8 V
Torque 1.8 kg/cm
Degree 180◦
Operating Voltage 5 V–12 V
Operating current 2.5 A
Step motor
Speed 0.1
Torque 0.34 kg/cm
Operating Voltage 8.2 V–45 V
DRV8825 driver
Operating current 2.5 A
Pushbutton Switch 12 mm button square

The main control unit of the prosthetic hand is the Arduino AtMega 2560.
Prosthetic hand movement activation is controlled by a surface Myoware muscle sen-
sor. The sensor electrodes are placed on the human hand skin above the flexor muscle and
the reference electrode is installed in a neutral place (over the join bones), as shown in
Figure 6.
The block diagram of the electronic circuit is shown in Figure 7. It provides the control
signals STEP and DIR to all 5 stepper motors. A DRV8825 stepper motor driver is used.
Each motor has its individual driver. The driver sets the motor current to the desired
value to control the torque. To assure a more precise positioning, 1/32 microstepping is
enabled. The used stepper motor type is 28BYJ-48 with an integrated reduction gearbox.
The stepper motors operate the pulleys, as shown in Figure 4. The motion of the thumb is
done with a servo, type Microservo SG90, connected directly to the Arduino. The force
sensors, button and EMG sensor are connected to the Arduino as well. The FSR relates to
a voltage divider to an analog input of the Arduino.
 all stepper motors and the servo are stopped. In this state, the start command is awaited.
It is assumed that the initial position of the hand is kept the same through the flexibility
of the rubber tendons. The start signal can come from the button or the EMG sensor. When
the start command is issued, the movement starts. The stepper motors and the servos are
controlled independently to allow for the grasping of objects with different shapes. The
Sensors 2021, 21, 137 8 of 15
grasping is completed when the FSR in the thumb signal is above a given threshold. Posi-
tion (number of steps) for all steppers is stored in memory and used for the release phase.

2021, 21, x FOR PEER REVIEW 8 of 15

Once the extension is required, the user presses the button for the extension and the
Figure Figure
6. Positioning ofpasses
the Positioning
6. EMG ofofthe
sensors EMG
the sensors
human hand.of the human hand.
hand automatically to the initial position and waits for the new command.

Figure 7. Motion Figure Motion

sensing7.block sensing
diagram of block diagram hand.
the prosthetic of the prosthetic hand.

The control
The signals received by EMG algorithm
sensorsoperates as a state
are amplified andmachine.
rectified.ItFigure
starts8a
in illustrates
an IDLE state, where all
stepper motors and the servo are stopped. In this state, the start command is awaited.
the EMG signal and in Figure 8b the filtered signal. The input amplitude signal of the
It is assumed that the initial position of the hand is kept the same through the flexibility
EMG is in millivolts. Once the EMG signal is received, it should be filtered with a band-
of the rubber tendons. The start signal can come from the button or the EMG sensor.
pass filter or by using a low-pass filter and a high-pass filter. Firstly, the signal is filtered
When the start command is issued, the movement starts. The stepper motors and the
with a digital Low-Pass Filter based on the Kirchoff’s Law in order to reduce signal noises.
servos are controlled independently to allow for the grasping of objects with different
The amplification is calculated by:
shapes. The grasping is completed when the FSR in the thumb signal is above a given
𝑥 𝑛 (number
threshold. Position = α ∗ y nof steps)
1 αfor∗ally nsteppers
1 is stored in memory(1) and used for the
release phase.
where α is the smoothing factor that varies from 0 to 1, x[n] is the resulting filtered discrete
Once the extension is required, the user presses the button for the extension and the
signal, y[n] is the discrete signal received by EMG. An example of an EMG filtered signal
hand automatically passes to the initial position and waits for the new command.
is presented in Figure 6b. In that figure, the EMG signal was filtered at α = 0.05. The EMG
The signals received by EMG sensors are amplified and rectified. Figure 8a illustrates
signal was tested for
the EMG different α values
signal and from
in Figure 8b0.05 to 1. Forsignal.
the filtered the α The
= 0.05 value,
input the filter
amplitude is of the EMG
signal
slower and clear.
is inInmillivolts.
Figure 8b Once
the filtered signal
the EMG is represented
signal is received,initred color.be filtered with a band-pass
should
filter or by using a low-pass filter and a high-pass filter. Firstly, the signal is filtered with
a digital Low-Pass Filter based on the Kirchoff’s Law in order to reduce signal noises.
The amplification is calculated by:

x[n] = α ∗ y[n] + (1 − α)∗y[n − 1] (1)

where α is the smoothing factor that varies from 0 to 1, x[n] is the resulting filtered discrete
signal, y[n] is the discrete signal received by EMG. An example of an EMG filtered signal is
presented in Figure 6b. In that figure, the EMG signal was filtered at α = 0.05. The EMG
signal was tested for different α values from 0.05 to 1. For the α = 0.05 value, the filter is
slower and clear. In Figure 8b the filtered signal is represented in red color.

(a) EMG signal

 𝑥 𝑛 =α∗y n 1 α ∗y n 1 (1)
where α is the smoothing factor that varies from 0 to 1, x[n] is the resulting filtered discrete
signal, y[n] is the discrete signal received by EMG. An example of an EMG filtered signal
Sensors 2021, 21, 137
is presented in Figure 6b. In that figure, the EMG signal was filtered at α = 0.05. The9 EMG
of 15
signal was tested for different α values from 0.05 to 1. For the α = 0.05 value, the filter is
slower and clear. In Figure 8b the filtered signal is represented in red color.

Sensors 2021, 21, x FOR PEER REVIEW 9 of 15

(a) EMG signal

Flexion movement

Normal level

(b) Low Pass Filtered EMG signal for α = 0.05

Figure
Figure 8.
8. Example
Example of
of EMG
EMG signal.
signal.

The
The filtered
filtered signal
signal is
is then
then classified
classified and
and analyzed
analyzed for
for finger
finger flexion
flexion and
and extension.
extension.
The
The features
features are are extracted
extracted andand discriminated
discriminated fromfrom the
the EMG signal.
signal. The
The power
power for
for grasp
grasp
movement
movement is is considered
considered in in the
thepresent
presentstudy.
study. The
The posture
posture of
ofthe
thefingers
fingersisisnot
notconsidered.
considered.
In order
In order to to quickly
quickly calculate
calculate thethe EMG
EMG signal,
signal, the
the features
features areare extracted
extracted in the time
domain. Furthermore,
domain. Furthermore,the the grasping
grasping function
function starts
starts with
with the the condition
condition of EMG
of EMG signal
signal am-
amplitude.
plitude. It starts
It starts grasping
grasping whenwhenthe EMGthe signal
EMG signal amplitude
amplitude exceedsexceeds the predefined
the predefined thresh-
threshold.
old. 
1 i f x > threshold
f (x) = 1 𝑖𝑓0 𝑥 > 𝑡ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑 (2)
𝑓 𝑥 = otherwise (2)
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
where f (x) is the EMG input signal. The threshold is defined in accordance with the EMG
where f(x) is the EMG input signal. The threshold is defined in accordance with the EMG
signal voltage.
signal voltage.
Figure 8 presents, in the red rectangle, the grasping movement. The signal below the
Figure 8 presents, in the red rectangle, the grasping movement. The signal below the
amplitude threshold is not considered.
amplitude threshold is not considered.
The grasping end depends on many factors as contact points, force closure, grasp
control, external force, friction, etc. it means that it is necessary to realize object surface
exploration.
Let us consider that external force is defined as f and depends on the external wrench
Sensors 2021, 21, 137 10 of 15
Sensors 2021, 21, x FOR PEER REVIEW 10 of 15

The grasping end depends on many factors as contact points, force closure, grasp con-
3. Results
trol, external force, friction, etc. it means that it is necessary to realize object surface
The experiments carried out with the prosthetic hand aimed to verify the correct
exploration.
functionality of the device.
Let us consider Moreover,
that external force in these experiments,
is defined the evaluation
as f and depends of thewrench
on the external device
structure was also performed. To this end,
w, at a moment m, the contact force p and torque τ. the joins equilibrium was studied. As the joins
are made of rubber,
The torque can it
beisdefined
important
as: to analyze the correct fingers flexion/extension trajec-
tory, as well as the limits of the possible τdeviations
= p ∗ J T of the phalanges under pressure. Dur- (3)
ing the experiments, the prosthetic hand was placed in a vertical position fix-mounted to
where T
J is theexperimental
Jacobian matrix for manipulation.
the table. The study was based on the same manipulation motion as de-
scribedTheinforce
[27]. balance
The main −G *was
f =target w can be calculated
to establish by taking
the grasp into consideration
movement. that the
In the experiments,
applied contact force must balance the external force applied to the object,
the same force sensing resistors threshold is used for all objects. That means that the where G pros-
is the
grasp matrix.
thetic hand applied the same force to all objects. The grasp experiments were conducted
with The
threegrasping
differentforce
sizeisobjects;
also proportional
the objectstoemployed
the actuators current
in the f = k * I. Ifare:
experiments considering
a ball, a
that the force-sensitive resistance sensor voltage v, then the contact
pencil, and a note block, as shown in Figure 9. The ball diameter was 64 mm, force p = k * v.the note
blockThewidthgrasping
was 4 mmstop and
function then isdiameter
the pencil calculated
was as:10 mm. During the experiments, the
movement of the fingers was observed.  This movement allows us to better perceive the
1 i f p > threshold
flexion, extension, abduction/adduction v= as well as the rotation for adaptation to the objects (4)
0 otherwise
surface. The fingers trajectory of the hand for a simple grasping was also studied to ana-
lyze the small abductions/adductions on the fingers joins represented in Figure 9a. Gen-
3. Results
erally, the trajectory defined by the prosthetic hand is cylindric pointing to the carpal
The experiments carried out with the prosthetic hand aimed to verify the correct
bones. In the grasping of the tennis ball, shown in the Figure 9b, the ball is held into the
functionality of the device. Moreover, in these experiments, the evaluation of the device
hand. The grasping stops when the force sensing resistor detects the threshold. Then, it
structure was also performed. To this end, the joins equilibrium was studied. As the joins
stops.
are made of rubber, it is important to analyze the correct fingers flexion/extension trajectory,
The hand can grasp the ball without any glove, but the use of a glove is also tested.
as well as the limits of the possible deviations of the phalanges under pressure. During the
The glove
experiments, used theinprosthetic
the experiments
hand was wasplaced
a standard glove made
in a vertical of latex.
position Becausetothe
fix-mounted theglove
table.
has
Theaexperimental
dry surface, itstudy
was impossible
was based on to grasp the tennis
the same ball since
manipulation it slipped
motion from thein
as described hand.
[27].
Additionally,
The main target thewas
glove design influences
to establish the graspthe prostheticInhand
movement. grasping experiments.
the experiments, the same forceAf-
terwards, the ball grasping is tested without the latex glove, and the
sensing resistors threshold is used for all objects. That means that the prosthetic hand prosthetic hand can
grasp
applied thethe
ball,
sameas observed inobjects.
force to all Figure 9. TheFigure
grasp9b shows that were
experiments the prosthetic
conductedhand withfinger
three
trajectory has been
different size slightly
objects; displaced,
the objects adapting
employed to the
in the object surface,
experiments are: aexcept
ball, aon the thumb
pencil, and a
that
noteperforms
block, as ashown
greaterindisplacement.
Figure 9. TheFor ballthe pen grasping,
diameter was 64 amm, clamp themovement
note blockiswidthper-
ceived. The thumb applies the force to the pencil and the index
was 4 mm and the pencil diameter was 10 mm. During the experiments, the movementDistal Phalange. When the
force threshold is detected, the prosthetic hand stops grasping.
of the fingers was observed. This movement allows us to better perceive the flexion,For the notebook grasp,
only the index
extension, and thumb fingers
abduction/adduction as are
welloperative, as shown
as the rotation in Figureto
for adaptation 9d.
theInobjects
this case, the
surface.
experiment
The fingers consists
trajectory in of
controlling
the hand thefor thickness and correct
a simple grasping was manipulation
also studiedof tothin objects.
analyze the
Figure 9a,b presents the abduction/adduction
small abductions/adductions on the fingersofjoins the index and little
represented in finger.
Figure The9a. maximum
Generally,
abduction of the
the trajectory Index by
defined finger
the is over 30° hand
prosthetic and 20° for the little
is cylindric finger.toThis
pointing the deviation
carpal bones.de-
pends on the object size. The soft joins allow the fingers to adapt to the
In the grasping of the tennis ball, shown in the Figure 9b, the ball is held into the hand. objects surface in
order to obtain a better grasp.
The grasping stops when the force sensing resistor detects the threshold. Then, it stops.

(a) (b) (c) (d)

Figure
Figure 9.
9. Prosthetic
Prosthetic hand
hand manipulation.
manipulation. (a)
(a)Grasping
Graspingtrajectory
trajectoryrepresented
representedwith
withgreen
greencolor,
color, (b)
(b) grasp
grasp of
of aa tennis
tennis ball
ball
without
without glove,
glove, (c)
(c) pencil
pencil grasp
grasp and
and (d)
(d) notebook grasp.
notebook grasp.
Sensors 2021, 21, 137 11 of 15

The hand can grasp the ball without any glove, but the use of a glove is also tested.
The glove used in the experiments was a standard glove made of latex. Because the glove
has a dry surface, it was impossible to grasp the tennis ball since it slipped from the
hand. Additionally, the glove design influences the prosthetic hand grasping experiments.
Afterwards, the ball grasping is tested without the latex glove, and the prosthetic hand can
grasp the ball, as observed in Figure 9. Figure 9b shows that the prosthetic hand finger
trajectory has been slightly displaced, adapting to the object surface, except on the thumb
that performs a greater displacement. For the pen grasping, a clamp movement is perceived.
The thumb applies the force to the pencil and the index Distal Phalange. When the force
threshold is detected, the prosthetic hand stops grasping. For the notebook grasp, only the
index and thumb fingers are operative, as shown in Figure 9d. In this case, the experiment
consists in controlling the thickness and correct manipulation of thin objects. Figure 9a,b
presents the abduction/adduction of the index and little finger. The maximum abduction
of the Index finger is over 30◦ and 20◦ for the little finger. This deviation depends on the
object size. The soft joins allow the fingers to adapt to the objects surface in order to obtain
a better grasp.
The grasping time is different and depends on the object’s thickness or diameter.
The basic grasping time was 1.3 s from the open hand position.
For a better object manipulation, independent phalange tendons that can be controlled
by the same actuator incrementing joins kinematics and mechanics are necessary.
With regard to the design of the prosthetic hand, the rubber materials resistance was
studied. The joins and the rubber-made extensor tendon, as well as the artificial adductor
policis (muscle), are analyzed. The hand was evaluated and tested during 6 months in
laboratory conditions (without patients) to test its functionalities and materials. During this
period, multiple grasping movement experiments were performed. Some wear in the
extensor tendon and artificial adductor muscle was observed after this period. The rubber
started cracking and changed color in the joins zone, as well as where other rubber material
was, as can be seen in the Figure 10. In addition, after 6 months, the artificial adductor
muscle, made by the same material, broke when big objects were tested, such as the
tennis ball. The rubber extensor tendon generated resistance when the grasping was
performed, and this resistance required high-powered actuators. To solve this disadvantage,
an additional extensor tendon was attached to the first, by using wires passing through
the tendon rail. Small servomotors with the torque of 1.8 kg/cm were not enough. Due to
this problem, the usage of the stepper motors was adopted in the current prosthetic hand.
The weight, size and power supply of the prosthetic hand depended on the electronics
and the prosthetic hand functionality. Finger load is correlated with the rate torque of
the actuators.
One of the objectives of the presented prototype was to improve prosthetic hand
functions with its design. The conventional prosthetic hands with robotic joins can achieve
just 1 DOF per join [28,29]. Nevertheless, the proposed design has flexible joins, so that the
small abduction/adduction on the joins will allow the fingers to easily adapt to the object
forms. Basically, this improvement in the joins can provide multiple additional features to
the prosthetic hand, such as hook, spherical grasp, cylindrical grasp, tip, etc. The proposed
design also avoids for abduction/adduction limitations and permits a better prosthesis
function, increasing the range of motion. The use of the force sensitive resistor placed on
the distal phalange of the thumb allows one to define a better pressing over objects and
stop the grasping of the prosthetic. Technical characteristics of the proposed prosthetic
hand are presented in Table 4.
 and this resistance required high-powered actuators. To solve this disadvantage, an addi-
tional extensor tendon was attached to the first, by using wires passing through the ten-
don rail. Small servomotors with the torque of 1.8 kg/cm were not enough. Due to this
problem, the usage of the stepper motors was adopted in the current prosthetic hand. The
Sensors 2021, 21, 137
weight, size and power supply of the prosthetic hand depended on the electronics and the
12 of 15
prosthetic hand functionality. Finger load is correlated with the rate torque of the actua-
tors.

(a) (b) (c)

Figure
Figure 10.
10. Rubber
Rubber materials defects after
materials defects after66months
monthsofofusing.
using.(a)
(a)The
Therubber
rubber glove
glove form
form prevents
prevents thethe correct
correct abduction
abduction of
of the
the thumb. The 90 degrees of the thumb placement does not allow the prosthetic hand to correctly manipulate the
thumb. The 90 degrees of the thumb placement does not allow the prosthetic hand to correctly manipulate the objects. (b)objects.
(b) Extensor tendon wear due to friction with the tendon sheath. (c) Artificial Adductor policis (muscle) fracture.
Extensor tendon wear due to friction with the tendon sheath. (c) Artificial Adductor policis (muscle) fracture.

One of the objectives of the presented prototype was to improve prosthetic hand
functions with its hand
Table 4. Prosthetic design. The conventional prosthetic hands with robotic joins can achieve
characteristics.
just 1 DOF per join [28,29]. Nevertheless, the proposed design has flexible joins, so that
Weight, gr 480
the small abduction/adduction on the joins will allow the fingers to easily adapt to the
Operating voltage, V 12
object forms. Basically, this improvement in the joins can provide multiple additional fea-
tures to the prosthetic hand,
Battery type such as hook, spherical grasp, cylindrical
Li-ion grasp, tip, etc. The
proposed designBattery
also avoids
mAh for abduction/adduction limitations 3000and permits a better
prosthesis function, increasing the range of motion. The use of the force sensitive resistor
N actuators 6
placed on the distal phalange of the thumb allows one to define a better pressing over
Type of actuators DC motor
Active fingers 5
Force, N 17
Movement control EMG, FSR
Movement command Buttons, co-contractions
Max flexion degree per join, ◦ 90
Max abduction per join, ◦ 30
Max Thumb abduction, ◦ 90
Joins soft
Feedback NO

In Table 5, we presented the advantages and disadvantages of our prosthetic hand

with respect to the existing ones.
Sensors 2021, 21, 137 13 of 15

Table 5. Prosthetic hand advantages and disadvantages.

Advantages Disadvantages
Miquelangelo - Difficulties on co-contractions too much effort
- Better
- External device requires as mobile phone, buttons, etc.
i-Limb grasping wit
- No need of gyroscope
multifunc-
Be - Pattern recognition requires much more training
tional
bionic - Expensive
fingers
Sensor - Expensive repair
- Wrist rotation
Hand - Heavy objects the fingers open unintentionally
- Multiple grip
- Insufficient force
Vincent types
- Grasp of static objects
Hand - Mechanical robustness

- Soft joins
- Adaptation
to the object
surface - Difficulties on co-contractions and or bottom (needs the
Our - Multiple grip other hand)
Hand types - Insufficient force
- Combination - Mechanical robustness
of 2 control
systems EMG
and FSR

The technique applied in the actual prosthetic hands are based on the fusion of the
EMG and force sensitive technology. The combination of both techniques compensates the
weakness of each of the techniques when operating independently.
Actual EMG prosthetic hands require some different movements done by the user
to choose a specific movement manipulation as well as the pressure applied to the object
surface [30]. This specific movement as multiple muscle contractions in a short time are
sometimes not well accepted by the users [31]. To correctly control the force applied to the
objects and the adequate fingers movement, the prosthetic hand requires the use of force
sensitive sensors applied to all fingers individually. This method allows the prosthetic
hand to correctly achieve the manipulation movement. In the proposed prosthetic hand,
the EMG signals are used only to start the grasping. Nevertheless, the stop end of the grasp
is controlled by the force-sensing resistors.
The sensory controlled prosthetic hand allows the user to actively enable the desired
task. The sensor feedback gives to the user a more adequate control over the prosthetic hand
and the external objects. The experimental results with the force sensitive resistors show
good results on the grasping end, as well as imply minimum computational costs. When the
pressure threshold is detected, then the grasping is stopped. Therefore, the grasping stop
generates a new grasping order (flexion or extension), waiting for the new EMG signal and
force sensitive resistor. When the new flexion (grasping) order is generated, the cycle is
repeated again.
The integration of 5 force-sensing resistors will allow the actuators to work au-
tonomously, i.e., with independent movement of the fingers. In this case, it will not be
necessary to introduce muscle movement classifications for all types of grasping. This will
decrease the EMG classification processing time and will reduce the memory space required
for the signal processing. All these improvements will facilitate the real time functionality
of the prosthetic hand, avoiding time delays. The sensor feedback allows the user to control
the prosthetic hand and take decisions on the prosthetic hand activity. Vision also helps the
user to define the prosthetic hand activity.
Sensors 2021, 21, 137 14 of 15

4. Conclusions
In the present work, the design and development of a prosthetic hand that emulates the
human hand motion is presented. The proposed mechanical architecture of the prosthetic is
based on the human hand anatomy and offers a broad range of movements. The articulation
joins increase the degrees of freedom of the fingers and improve the hand flexibility,
comparing with the existing bionic prosthetic hands mentioned in Table 5. The proposed
prosthetic hand can perform fine movements and grasp different sized objects. The soft
joins allow the prosthetic hand to achieve abduction and gyro on the fingers that adapt
to the object surface. In the experiments with the tennis ball, the index and little fingers
perform an abduction/adduction of 20–30 degrees. The weight of the device is 480 gr,
which is below 500 gr and meets the requirements for bionic prosthetic hands. The system
is externally powered and has robust and simple finger kinematics. The work shows that
the employed materials must have enough flexibility and hardness to enable a correct
use of the hand. The wire-driven tendons methodology that is employed in the proposed
prosthetic hand solution shows a good grasping performance. The friction of the tendons
is very low, which is another advantage of the prototype. The friction generated by the
extensor tendons with respect to the tendon sheath, makes it that actuators with higher
power are required.
On the other hand, the use of the EMG to control the prosthetic hand enables an
increase in its ability when using human muscle actuation. The use of advanced signal
processing (signal acquisition, filtering, classification, and training) is also remarkable to
enable a correct simulation of the different hand movements. Finally, the usage of force-
resistive sensors to end the grasping movement allows the prosthetic hand to simulate the
touch pressure of the real hand.
In conclusion, the proposed solution shows interesting advantages versus available
alternatives, enhancing the functionality and ergonomic nature of the device and not only
relying on esthetic aspects.

Author Contributions: L.D. produced the prosthetic hand design, structural analysis of the compo-
nents, prototyping, designed the experiments and wrote the paper; M.N. developed the software of
the prosthetic hand and performed the experiments and revised the paper; C.G.E. contributed with
the human hand anatomy content and revised the paper. All authors have read and agreed to the
published version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: Data available on request due to restrictions. The data presented in this
study are available on request from the corresponding author. The data are not publicly available.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Raichle, K.A.; Hanley, M.A.; Molton, I.; Kadel, N.J.; Campbell, K.; Phelps, E.; Ehde, D.; Smith, D.G. Prosthesis use in persons with
lower- and upper-limb amputation. J. Rehabil. Res. Dev. 2008, 45, 961–972. [CrossRef] [PubMed]
2. Uellendahl, J.E.; Uellendahl, E.N. Experience Fitting Partial Hand Prostheses with Externally Powered Fingers; Bentham Science:
Sharjah, UAE, 2012; pp. 15–27.
3. O&P Almanac. Amputation data from community hospitals. O&P Almanac 2016, 65, 8.
4. Bethge, M.; Von Groote, P.; Giustini, A.; Gutenbrunner, C. The world report on disability: A challenge for rehabilitation medicine.
Am. J. Phys. Med. Rehabil. 2014, 93, S4–S11. [CrossRef] [PubMed]
5. Sahu, A.; Sagar, R.; Sarkar, S.; Sagar, S. Psychological effects of amputation: A review of studies from India. Ind. Psychiatry J. 2016,
25, 4–10. [CrossRef] [PubMed]
6. Solgajová, A.; Sollár, T.; Vörösová, G. Gender, age and proactive coping as predictors of coping in patients with limb amputation.
Kontakt 2015, 17, e67–e72. [CrossRef]
7. Cavanagh, S.R.; Shin, L.M.; Karamouz, N.; Rauch, S.L. Psychiatric and emotional sequelae of surgical amputation. Psychosomatics
2006, 47, 459–464. [CrossRef]
8. Abeyasinghe, N.L.; de Zoysa, P.; Bandara, K.M.; Bartholameuz, N.A.; Bandara, J.M. The prevalence of symptoms of post-traumatic
stress disorder among soldiers with amputation of a limb or spinal injury: A report from a rehabilitation centre in Sri Lanka.
Psychol. Health Med. 2012, 17, 376–381. [CrossRef]
Sensors 2021, 21, 137 15 of 15

9. Childress, D.S. Historical aspects of powered limb prosthesis. Clin. Prosthet. Orthot. 1985, 9, 2–13.
10. Parker, P.A.; Scott, R.N. Myoelectric control of prostheses. Crit. Rev. Biomed. Eng. 1986, 13, 283–310.
11. Shenoy, P.; Miller, K.J.; Crawford, B.; Rao, R.P.N. Online electromyographic control of a robotic prosthesis. IEEE Trans. Biomed. Eng.
2008, 55, 1128–1135. [CrossRef]
12. Khushaba, R.N.; Kodagoda, S.; Takruri, M.; Dissanayake, G. Toward improved control of prosthetic fingers using surface
electromyogram (EMG) signals. Expert Syst. Appl. 2012, 39, 10731–10738. [CrossRef]
13. Weir, R.F.; Troyk, P.R.; Schorsch, J.F.; Maas, H. Implantable myoelectric sensor (IMESs) for intramuscular electromyogram
recording. IEEE Trans. Biomed. Eng. 2009, 56, 159–171. [CrossRef] [PubMed]
14. Malesevic, N.; Björkman, A.; Andersson, G.S. A database of multi-channel intramuscular electromyogram signals during isometric
hand muscles contractions. Sci. Data 2020, 7, 10. [CrossRef] [PubMed]
15. Resnik, L.; Klinger, S.L.; Etter, K. The DEKA Arm: Its features, functionality, and evolution during the Veterans Affairs study to
optimize the DEKA arm. Prosthet. Orthot. Int. 2014, 38, 492–504. [CrossRef]
16. Belter, J.T.; Segil, J.L.; Dollar, A.M.; Weir, R.F. Mechanical design and performance specifications of anthropomorphic prosthetic
hands: A review. J. Rehabil. Res. Dev. 2013, 50, 599–618. [CrossRef]
17. Be Bionic, Technical Manual. Available online: https://shop.ottobock.us/media/pdf/bebionicTechManualSmall.pdf
(accessed on 25 August 2020).
18. i-Limb, Ossur. Available online: https://www.ortosur.es/catalogo-de-productos/protesis/miembro-superior/mano-
mioelectrica/i-limb/ (accessed on 27 October 2020).
19. Dosen, S.; Cipriani, C.; Kostic, M.; Controzzi, M.; Carrozza, M.C.; Popovic, D.B. Cognitive vision system for control of dexterous
prosthetic hands: Experimental evaluation. J. Neuroeng. Rehabil. 2010, 7, 42. [CrossRef]
20. Mainardi, E.; Davalli, A. Controlling a prosthetic arm with a throat microphone. In Proceedings of the 29th Annual International
Conference of the IEEE Engineering in Medicine and Biology Society, Lyon, France, 22–26 August 2007; pp. 3035–3039.
21. Johansen, D.; Cipriani, C.; Popovic, D.B.; Struijk, L.N. Control of a Robotic Hand Using a Tongue Control System-A Prosthesis
Application. IEEE Trans. Biomed. Eng. 2016, 63, 1368–1376. [CrossRef]
22. Feix, T.; Romero, J.; Ek, C.H.; Schmiedmayer, H.B.; Kragic, D.A. Metric for Comparing the Anthropomorphic Motion Capability
of Artificial Hands. Robotics. IEEE Trans. Robot. 2013, 29, 82–93. [CrossRef]
23. Connolly, C. Prosthetic hands from Touch Bionics. Ind. Robot. Int. J. 2018, 35, 290–293. [CrossRef]
24. Medynski, C.; Rattray, B. Bebionic prosthetic design. In Proceedings of the MyoElectric Controls/Powered Prosthetics Symposium
(MEC), Fredericton, NB, Canada, 14–19 August 2001; pp. 1–4.
25. Available online: https://www.ottobockus.com/prosthetics/upper-limb-prosthetics/solution-overview/myoelectric-devices-
speedhands/ (accessed on 30 November 2020).
26. VINCENT Hand. Weingarten (Germany): Vin-Cent Systems. 2013. Available online: http://handprothese.de/vincent-hand/
(accessed on 30 November 2020).
27. Ma, R.R.; Dollar, A.M. On dexterity and dexterous manipulation. In Proceedings of the 15th International Conference on
Advanced Robotics (ICAR), Tallinn, Estonia, 20–23 June 2011; pp. 1–7.
28. Mnyusiwalla, H.; Vulliez, P.; Gazeau, J.P.; Zeghloul, S. A new desxteros had based on bio-inspired finger design for inside-hand
manipulation. IEEE Trans. Syst. 2016, 46, 809–817.
29. Abdul Wahit, A.A.; Ahmad, S.A.; Marhaban, M.H.; Wada, C.; Izhar, L.I. 3D printed robot hand structure using four-bar linkage
mechanism for prosthetic applications. Sensors 2020, 20, 4174. [CrossRef] [PubMed]
30. Favieiro, G.W.; Balbinot, A.; Barreto, M.M. Decoding arm movements by myoeletric signals and artificial neural networks.
In Proceedings of the ISSNIP Biosignals and Biorobotics Conference, Vitoria, Brazil, 6–8 January 2011; pp. 1–6.
31. Franzke, A.W.; Kristoffersen, M.B.; Bongers, R.M.; Murgia, A.; Pobatschnig, B.; Unglaube, F.; Van Der Sluis, C.K. Users’ and
therapists’ perceptions of myoelectric multi-function upper limb prostheses with conventional and pattern recognition control.
PLoS ONE 2019, 14, e0220899. [CrossRef] [PubMed]