Development of Wrist Bending Rehabilitation Robot
Hyeon-Min Kim
1
, Tae-Kyung Hong
1
, Hyung-Je Cho
2
and Gab-Soon Kim
1
1
Department of Control & Instrumentation Engineering (ERI), Gyeongsang National University, Jinju, South Korea
2
YESON Rehabilitation Hospital, Seoul, South Korea
Keywords: Four-axis Force/Moment Sensor, Interference Error, Wrist Rehabilitation Robot, Rehabilitation Exercise.
Abstract: This paper describes the wrist bending rehabilitation robot using a four-axis force/moment sensor. The robot
can be used to exercise the wrist bending rehabilitation for severe stroke patients lying in bed wards or at
home. The manufactured four-axis force/moment sensor which can detect two directional force Fx, Fy and
two directional moment Mx, My, was attached to the developed rehabilitation robot and allows the
rehabilitation robot to measure a bending force (Fx) exerted on a wrist, the signal force Fy and moments
Mx, My which in turn allows the device to be used safely. The results of a characteristics test for the
developed rehabilitation robot showed that it was safely operated while the wrist bending flexibility
rehabilitation exercise was performed. Therefore, it is thought that the developed rehabilitation robot can be
safely used with severe stroke patients.
1 INTRODUCTION
The numbers of patients with severe strokes are
increasing rapidly in the world, and their wrists must
receive rehabilitation exercises so that they can be
used in everyday life. The wrist bending flexibility
rehabilitation exercise bends the wrist in a counter-
clockwise direction (right hand) or clock wise
direction (left hand) based on the wrist, like with
rehabilitation therapist, it conducts rehabilitation
exercises until the patient starts feel pain, and then
the robot pauses the exercise for a while (about 4s).
It then bends back the wrist again in the opposite
direction, and then pauses for a while. Rehabilitation
therapists, nurses and physicians for rehabilitation
find it very difficult to deal with the number of
severe stroke patients and because of this some of
the severe stroke patients don’t receive adequate
rehabilitation exercise. Therefore, the wrist bending
rehabilitation robot has been developed for safe and
adequate wrist bending rehabilitation and to ease the
burden on rehabilitation therapists, nurses and
physicians.
Pan developed three degrees of freedom
rehabilitation robot for mild stroke patients sitting in
a chair attached to a robot, that provided elbow
rehabilitation exercise. Both Pan and Culmer did
research on how to control the robot using the fuzzy
theory. Kim developed the six degrees of freedom
robot for upper limb rehabilitation of the minor
stroke patients, and this robot was able to perform
the rehabilitation exercises for mild stroke patients’
upper limbs. Li developed an upper limb
rehabilitation robot that could be used to exercise
bending and straightening of a patients arm. The
robot can measure the moment needed to the bend of
the patient's arm and the control of the robot is
performed using measured moment. Culmer
developed a six degrees of freedom rehabilitation
robot to rehabilitate patient's upper limb using two
robots, and he researched the control of the robot for
mild stroke patient's upper limb rehabilitation
exercise.
These robots can perform the rehabilitation
exercise for mild patients’ wrists, but can't perform
that of severe stroke patients’ wrists. Thus, a wrist
bending rehabilitation robot, such as rehabilitation
therapist, that helps rehabilitate severe stroke
patients lying in bed wards or at home by providing
wrist rehabilitation exercises must be developed.
The Rehabilitation robots must measure the
moments of wrist rotated to counter-clockwise and
clockwise, and be controlled using the measured
moments. A four-axis force/moment sensor
measures x, y directional force Fx, Fy and moment
Mx, My, and it is composed of a body in order to
reduce the size of the sensor in general. The four-
axis force/moment sensor built into a wrist bending
rehabilitation robot should be suitable in the size,
with a rated capacity and price, and it should be easy
272
Kim H., Hong T., Cho H. and Kim G..
Development of Wrist Bending Rehabilitation Robot.
DOI: 10.5220/0005007402720279
In Proceedings of the 11th International Conference on Informatics in Control, Automation and Robotics (ICINCO-2014), pages 272-279
ISBN: 978-989-758-040-6
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
to attach to the robot. Current developments of a
multi-axis force sensor [Nagai, Kim and Kim] are
not suitable for a wrist bending rehabilitation robot
in terms of size, price and additional conditions.
Therefore, the wrist bending rehabilitation robot
needs to be developed for severe stroke
rehabilitation lying in bed wards or at home.
In this paper, the wrist bending rehabilitation
robot using a four-axis force/moment sensor was
developed. The four-axis force/moment sensor was
designed and fabricated for measuring the applied
force (Fx) of patient's wrist and force Fy and
moment Mx, My for a safe control of movement.
And then a robot body was designed and
manufactured. The characteristic test of a wrist
bending flexibility rehabilitation exercise of a
normal people and a stroke patient was carried out
using the developed the robot.
2 DESIGN
AND MANUFACTURING
OF WRIST BENDING
REHABILITATION ROBOT
2.1 Design and Manufacturing of Wrist
Bending Rehabilitation Robot
Figure 1 shows the manufactured wrist bending
rehabilitation robot, it was designed and
manufactured to perform the rehabilitation exercise
of a wrist for a stroke patient lying in a hospital
ward. The rehabilitation robot was composed of a
robot body, a hand fixing block, a hand fixture
Velcro, an arm fixing block, an arm fixture Velcro,
an up-down supporter, a wrist bending motor, a
four-axis force/moment sensor, a high-speed
controller and so on. The robot body was
manufactured using aluminum square rods (size:
40mm40mm), and its size is 800mm 600mm
400mm, and most of the robot’s parts were fixed to
the robot body. The hand fixing block and the hand
fixture Velcro were designed to be fixed to the
patient's hand safely, and they were also fixed to the
four-axis force/moment sensor that transmits the
bending force (Fx) of the wrist bending motor to
patient's wrist. The arm fixing block and arm fixture
Velcro were designed to be fixed patient's arm
safely, and they were also fixed to the robot body
and the up-down supporter. The up-down supporter
was designed to be adjusted the height of the
patient's bed and to the height of the robot body with
one side being fixed to the wrist bending motor and
other side being fixed to the four-axis force/moment
sensor. The wrist bending motor applies the bending
force to patient's wrist, and it was fixed to the up-
down supporter.
The four-axis force/moment sensor was designed
and manufactured, it measures the bending force
(Fx) applied to a patient's wrist during the
rehabilitation exercise and also measures force Fy
and moments Mx, My in case of emergency. The
manufactured four-axis force/moment sensor is
described in detail in the next section. The high-
speed control system was manufactured to measure
the forces and moments from the four-axis
force/moment sensor, and to control the wrist
bending motor with the measured force Fx, and also
returns the robot to the initial position of the wrist
bending motor using the measured force Fy and
moments in case of an emergency situation.
Figure 1: Manufactured rehabilitation robot for wrist
bending exercise.
The wrist rehabilitation exercise using the wrist
bending rehabilitation robot is conducted by, first,
fixing patient's arm with the arm fixing block and
the arm fixture Velcro. The patient's hand is then
fixed to the hand fixing block and hand fixture
Velcro, after which the stroke patient lies in bed.
The robot bends the wrist counter-clockwise and
clockwise by the rotating of the bending motor
repeatedly, and the bending force exerted on the
wrist (force Fx) is measured by the four-axis
force/moment sensor. The robot is controlled with
the reference of the measured bending force (Fx). In
an emergency situation, the patient applies any force
to the robot, the four-axis force/moment sensor
measures forces Fx, Fy and moments Mx, My at the
same time, and the high-speed control system stops
the robot from using the forces and moments. The
robot is then returned to the initial position of the
motor.
DevelopmentofWristBendingRehabilitationRobot
273
2.2 Design and Manufacture of the
Four-Axis Force/Moment Sensor
Figure 2 shows the structure of the four-axis
force/moment sensor, and the sensor detects forces
Fx, Fy, and moments Mx, My simultaneously. The
sensor is composed of a force/moment transmitting
block, F/M, two
fixing block, F1 and F2, two
moving block, M1 and M4, four parallel plate beam,
PPB1~PPB4. PPB1 and PPB2 are for perceiving
force Fx and moment My, and PPB3 and PPB4 are
for force Fy and moment Mx. The sizes of the plate-
beams composed of PPB1 and PPB2 are width
1
b ,
thickness
1
t , length
1
l , those for PPB3 and PPB4 are
width
2
b
, thickness
2
t
, length
2
l
. The size of the
sensor is 104mm20mm20mm. Theoretical
equations should be derived to determine the size of
the sensor.
Figure 2: Structure of four-axis force/moment sensor.
Figure 3 shows the free body diagram of plate beams
for a four-axis force/moment sensor under the forces
Fx(or Fy). When the force Fx is applied to the point
O of the force/moment transmitting block, force
Fxx
F
and moment
Fxy
M are generated on the end of
the plate-beam. The force
Fxx
F and moment
Fxy
M
can be derived from the equations of force and
moment equilibrium-condition
0Fx ,
0Fy
and
0
o
M
of the force/moment transmitting
block and the moving block M1. And the equations
UFx
and
LFx
for analyzing the strains on the
upper and lower surfaces of the plate-beams are
derived by substituting the derived force
Fxx
F and
moment
Fxy
M into the equation of bending moment
Pz
EZM
1
/
, and they can be written as
)
2
(z
2
3
1
3
11
l
hEb
F
x
UFx
(1)
where,
E
is the modulus of longitudinal elasticity.
Figure 3: Free body diagram of plate beam for a four-axis
force/moment sensor under the forces Fx(or Fy).
Figure 4 shows the free body diagram of plate beams
for a four-axis force/moment sensor under the
moment My(or Mx). When the moment My is
applied to the point O of the force/moment
transmitting block, forces
Myx
F ,
Myz
F and moment
Myy
M are generated on the end of the plate-beam.
The forces
Myx
F ,
Myz
F and moment
Myy
M can be
derived from the equations of force and moment
equilibrium-condition
0Fx ,
0Fy and
0
o
M of the force/moment transmitting block
and the moving block M1. And the equation of the
rotational angle of force/moment transmitting block
can be derived using forces
Myx
F ,
Myz
F and
moment
Myy
M .
)
32
(
484
)
2
(
48
11
2
1
1
1
2
211
1
3
1
11
ld
l
EI
l
EdAl
d
l
dEI
M
y
(2)
And the equations
UMy
and
LMy
for analyzing the
strains on the upper and lower surfaces of the plate-
beams are derived by substituting the derived forces
Myx
F ,
Myz
F and moment
Myy
M into the equation of
bending moment
Pz
EZM
1
/
and the equation of
tension and compression
12
/ld
, and they can be
written as
1
211
2
1
11
1
3
1
1
))
32
(
6
))
2
(
6
l
dld
l
tl
d
l
zt
UMy
(3)
The design variables of the modeled the four-axis
force/moment sensor are the size of body, the rated
output of each sensor, the rated forces and moments
ICINCO2014-11thInternationalConferenceonInformaticsinControl,AutomationandRobotics
274
Figure 4: Free body diagram of plate beam for a four-axis
force/moment sensor under the forces My(or Mx).
of each sensor, the size (width
1
b , thickness
1
t ,
length
1
l
) of plate-beams of PPB1 and PPB2, the
size (width
2
b , thickness
2
t , length
2
l ) of plate-
beams of PPB3 and PPB4. And,
1
d is the distance
between the vertical center-line of the force/moment
transmitting block and the right end point of PPB1,
2
d
is the distance between the horizontal center-line
of the force/moment transmitting block and the right
end point of PPB1. The processing of the sensor
design is as follow, first the size of sensor's body,
the rated output of each sensor and the rated load of
each sensor are determined. And then the width
1
b
,
thickness
1
t , length
1
l , width
2
b , thickness
2
t ,
length
2
l ) of plate-beams are calculated by
substituting the determined them into the equations
(1) and (3).
Wheastone bridges are made for each sensor in
the four-axis force/moment sensor. The rated strain
of each sensor under each rated force or moment is
determined from the strain values of the attached
strain-gages to each sensor. Total strain of
Wheastone bridge can be calculated as following
equation.
112 2TCT C


(4)
where,
is total strain from Wheastone bridge,
1T
is strain of a tension strain-gage
1
T ,
1C
is strain of
a compression strain-gage
1
C ,
2T
is strain of a
tension strain-gage
2
T ,
2C
is strain of a
compression strain-gage
2
C .
And, the rated output can be calculated as below
equation.
1
4
o
i
E
K
E
(5)
where,
i
E is the input voltage of Wheastone bridge,
o
E is the output voltage of Wheastone bridge,
K
is
the factor of strain-gage (about 2.0),
is total strain
gotten from equation (4).
The rated outputs of the variables for designing
the four-axis force/moment sensor is determined by
about 0.5
VmV / respectively, the rated force and
moment of each sensor of the four-axis
force/moment sensor are Fx=Fy=100N, Mx=6Nm
and My=7Nm, the size of the sensor is
104mm
20mm20mm, the attaching locations of
strain-gages in length-direction is 1.5mm from the
end of the plate-beam, the attaching locations of
strain-gages in width-direction is on the center line
of the plate-beams, and the rated strain at the
attaching location of strain-gage is about 250
mm /
.
The sizes of sensing elements of the four-axis
force/moment sensor are determined by substituting
the determined variables into equations (1) and (3).
The lengths
1
l and
2
l are 10mm respectively, the
widths
1
b
and
2
b
are 20mm and 12mm, the
thicknesses
1
t and
2
t are 1.14mm and 1.36mm, the
distance
1
d and
2
d are 12mm and 5.34mm
respectively. The material of the sensor id Al 2024-
T351.
Figure 5 shows the attachment locations of strain
gages on each sensing element of four-axis
force/moment sensor. The attaching locations for
each sensor in the four-axis force/moment sensor
were determined in consideration of the results of
theoretical analysis. The attaching locations of
strain-gages for Fx sensor are S1~S4, those for Fy
sensor are S5~S8, those for Mx sensor are S13~S16,
those for My sensor are S9~S12. The attaching
location of each strain-gage is 1.5mm from the end
of plate-beam in the length-direction, and that in the
width-direction is the center line of plate-beam.
These locations were determined in consideration of
the interference error 0% calculated by equation (4).
The rated strains of each sensor in the four-axis
force/moment sensor at each attachment location of
strain-gages were calculated by theoretical analysis.
The rated strains for Fx sensor and Fy snsor were
992
mm /
, and those for Mx sensor and My sensor
were 992
mm /
and 968 mm /
respectively. The
maximum error of each sensor was less than 3.2%,
reflecting the fact that the sensing element is
0.01mm thick, taking the manufacturing processing
into consideration.
DevelopmentofWristBendingRehabilitationRobot
275
Figure 5: Attachment locations of strain gages on each
sensing element of four-axis force/moment sensor.
The four-axis force/moment sensor was fabricated
by using the bond (M-200) and the strain gauges
(N2A-13-S1452-350, made in Micro-Measurement
Company, gauge constant 2.03, size 3
5.2mm), and
constructed Wheatstone bridge. Figure 6 shows a
photograph of the manufactured four-axis
force/moment sensor, the rated forces of Fx sensor
and Fy sensor are each 100N, and the rated moments
of Mx sensor and My sensor are each 6.00Nm and
7.00Nm. The size of the four-axis force/moment
sensor is 104mm
20mm20mm. The characteristic test
of the manufactured four-axis force/moment sensor
must be carried out to get the rated output, the
repeatability error, the non-linearity error and the
interference error of each sensor. The characteristic
test was performed with a multi-axis force/moment
sensor calibration system [Kim] and a measuring
device (DMP40).
Figure 6: Manufactured four-axis force/torque sensor.
In order to detect the rated output of each sensor and
the interference error of the four-axis force/moment
sensor, the moment and the force from each sensor
was measured under the rated force and the moment
of each sensor and it was repeated three times, and
the rated output of each sensor was derived by
averaging the values. Table 1 shows the rated output
from theoretical analysis and characteristic test of
each sensor of four-axis force/moment sensor. The
rated outputs from theoretical analysis were by
substituting the gage factor 2.03 of strain-gage and
the calculated rated strain from equation (4) into the
equation (5). The rated outputs of Fx sensor, Fy
sensor, Mx sensor and My sensor were 0.4927mV/V,
0.5145mV/V, 0.5084mV/V and 0.4893mV/V
respectively. The maximum error of the rated output
from theoretical analysis and characteristic test was
less than 2.21%. Table 2 shows the interference error
of each sensor of four-axis force /moment sensor,
The maximum interference error of the four-axis
force/moment sensor was less than 0.96%.
In order to get the repeatability error and the
non-linearity error of each sensor, the force and the
moment from each sensor was measured under the
force and the moment 10%~100% of the rated force
and the moment with 10% step of each sensor and it
was repeated three times. The maximum
repeatability error and the maximum non-linearity
error of each sensor were less than 0.03%
respectively. Therefore, the four-axis force/moment
sensor manufactured in this paper is similar to the
multi-force sensors previously developed [Kim and
ATI INDUSTRIAL AUTOMATION] in level of the
errors. Thus, it is thought that the manufactured
four-axis force/moment sensor can be used for the
wrist bending rehabilitation robot. The developed
four-axis force / moment sensor is suitable for use in
rehabilitation robotics, because it is inexpensive, and
its size is 104
2020mm.
Table 1: Rated output from theoretical analysis and
characteristic test of each sensor of four-axis force
/moment sensor.
Sensor
Rated output )/( VmV
Theo. Exp. Error(%)
Fx 0.5034 0.4927 2.13
Fy 0.5034 0.5145 -2.21
Mx 0.5034 0.5084 0.99
My 0.4913 0.4893 0.41
Table 2: Interference error of each sensor of four-axis
force /moment sensor.
Sensor
Force
Interference error(%)
Fx Fy Mx My
Fx=100N - 0.23 0.52 0.96
Fy=100N 0.31 - 0.27 0.38
Mx=6Nm 0.48 0.18 - 0.62
My=7Nm 0.59 0.31 0.24 -
The four-axis force/moment sensor with the high-speed
ICINCO2014-11thInternationalConferenceonInformaticsinControl,AutomationandRobotics
276
control system was calibrated using the multi-axis
force/moment sensor calibration machine [Kim] to
measure the forces and moments. The calibration was
performed as follows: first, in order to calibrate Fx
sensor of the four-axis force/moment sensor, the
calibration machine applied the rated force of
Fx=100.0N to x-direction, and the indicated value of
Fx sensor was adjusted to 100.0N (resolution: 0.2N).
The Fy sensor, Mx sensor and My sensor were
calibrated as that of the Fx sensor, and they were
adjusted to 100.0N (resolution: 0.1N), 3.00Nm
(resolution: 0.01Nm), 5.00Nm (resolution: 0.01Nm)
respectively.
3 THE CHARACTERISTIC TEST
OF A WRIST BENDING
REHABILITATION ROBOT
AND ITS FINDINGS
3.1 Characteristic Test of Wrist
Bending Flexibility Rehabilitation
Exercise
Figure 7 shows the characteristic test for the wrist
bending flexibility rehabilitation exercise using the
wrist bending rehabilitation robot. The wrist bending
flexibility rehabilitation exercise is performed as
follows; first, a person's wrist and arm are safely
fixed using the fixtures and Velcro. The reference
bending force and the reference bending angle (the
measured bending angle add 5
) in counter-
clockwise and clockwise are then measured and are
inputted into the high-speed control system. The
initial position (the reference angle (0
)) of the hand
fixture and person's right hand is that the palm is
straight in the axial direction of the arm and vertical
with the ground, when person was lying in bed. The
method of wrist bending flexibility rehabilitation
exercise is as follows; first, the right hand is bended
to counter-clockwise from the bending force of 0.0N
(the initial position) to the reference bending force
using the motor attached to the robot, the robot is
controlled with the reference bending force for about
4s. Then it is rotated to clockwise like counter-
clockwise. Such a process carried out continuously.
Figure 8 shows the results of the characteristic
test for the wrist bending flexibility rehabilitation
exercise using the wrist bending rehabilitation robot.
The two forces and two moments are outputs from
the four-axis force/moment sensor during the wrist
bending flexibility rehabilitation exercise. The force
Fx is the generated force that the robot pushes the
Figure 7: Photographs of characteristic test for wrist
bending flexibility rehabilitation exercise using the wrist
bending rehabilitation robot (right hand).
Figure 8: Graphs of characteristic test for wrist bending
flexibility rehabilitation exercise using the wrist bending
rehabilitation robot (right hand).
palm or the back of hand, The force Fy is the
generated force that the hand presses the hand fixing
block in upward and downward, the moment Mx is
the generated moment that the end of fingers or the
part of hand presses the hand fixing block in upward
and downward (y-direction), and the moment My is
the generated moment that the end of fingers or the
part of hand presses the hand fixing block left and
right direction (x-direction). As the results of the
measurement tests of subject, the reference bending
force (Fx) and the reference angle in counter-
clockwise were -14.1N and 90.5
respectively. And
in clockwise, they were each 18.8 and 89.3
.
Therefore, the reference bending force and the
reference bending angle (the measured bending
angle add 5
) in counter-clockwise and clockwise
are then measured and are inputted into the high-
speed control system.
As shown in Figure 8, the reference wrist
bending forces (Fx) were applied to subject's right
hand in counter-clockwise and clockwise during the
DevelopmentofWristBendingRehabilitationRobot
277
wrist bending flexibility rehabilitation exercise, and
then the PI control at the reference bending force
was performed for about four seconds. The force Fy
at reference bending force in counter-clockwise was
2.5N, and 6.2N in clockwise. The moment Mx at
reference bending force in counter-clockwise was -
0.07Nm(-7Ncm), and -0.35Nm(-35Ncm) in
clockwise. The moment My at reference bending
force in counter-clockwise was 0.02Nm(2Ncm), and
-0.15Nm(-15Ncm) in clockwise. The generated
forces and moments in clockwise were greater than
those in counter-clockwise, because the wrist
withstand greater force in counter-clockwise
(bending to backward). And the generated force Fy
and the moments Mx, My were that the fixing block
pushes the palm to the end of finger direction.
Thus, the Fx sensor of the four-axis
force/moment sensor attached to the robot can be
used to control for performing the wrist bending
rehabilitation exercise by measuring the bending
force (Fx) of wrist, and the Fy sensor and Mx sensor
can be used to sense the fixing situation of patient's
arm during the wrist bending rehabilitation exercise
by measuring forces Fx, Fy and moments Mx, My.
3.2 Characteristics Test of Wrist
Bending Flexibility Rehabilitation
Exercise of Severe Stroke Patient
The characteristics test of the wrist bending
flexibility rehabilitation exercise was carried out to
confirm applying the wrist bending rehabilitation
robot for severe stroke patient. The characteristic
test was carried out at YESON Rehabilitation
Hospital, and a severe stroke patient who was almost
unable to move the right hand participated in the
test. She received her wrist bending flexibility
rehabilitation exercise using the wrist bending
rehabilitation robot for thirty minutes each day.
Figure 9 shows the stroke patient's characteristic
test for the wrist bending flexibility rehabilitation
exercise using the wrist bending rehabilitation robot.
Figure 10 shows the outputs of the two forces and
the two moments from the four-axis force/moment
sensor in a severe stroke patient's wrist bending
flexibility rehabilitation exercise using the wrist
bending rehabilitation robot(for three days). At the
results of rehabilitation exercise for eight days, the
reference bending forces in counter-clockwise and
clockwise were -5.4N and 9.4N, and the reference
bending angles were 48.23
and 47.97 respectively.
The bending force after eight days was 3.1N bigger
than after three days in counter-clockwise, and that
in clockwise was 4.5N bigger. And the bending
Figure 9: Photographs of stroke's characteristic test for
wrist bending rehabilitation exercise using the wrist
bending rehabilitation robot (right hand).
Figure 10: Graphs of stroke's wrist bending flexibility
rehabilitation exercise using the wrist bending
rehabilitation robot (right hand).
angle after eight days was 0.72 bigger than after
three days in counter-clockwise, and that in
clockwise was 3.33
bigger. The reference bending
forces from the sensor and the reference bending
angles were determined when the robot rotated the
patient's wrist until she felts pain in counter-
clockwise and clockwise in the everyday. To reveal
bigger moment Mx=-0.09Nm(9.00Ncm) was fixed
to the patient's wrist axis in discord to the central
axis of the hand fixture, and the part of wrist pushed
the wrist fixing block as shown in Figure 9. It is that
the four-axis force/moment sensor was composed of
a body.
4 CONCLUSIONS
In this paper, the wrist bending rehabilitation robot
using the four-axis force/moment sensor was
ICINCO2014-11thInternationalConferenceonInformaticsinControl,AutomationandRobotics
278
developed. The four-axis force/moment sensor of the
wrist bending rehabilitation robot was designed and
manufactured, and the interference error of the four-
axis force/moment sensor and the maximum
repeatability error and the maximum non-linearity
error of each sensor are similar to that of the
developed sensor [Kim and ATI INDUSTRIAL
AUTOMATION]. Thus, the four-axis force/moment
sensor can be used for the wrist bending
rehabilitation robot. In the characteristic test of the
wrist bending flexibility rehabilitation exercise, the
robot was accurately operated with the reference
bending force in counter-clockwise and clockwise
motions. The robot was safely operated when severe
stroke patient received the wrist bending flexibility
rehabilitation exercise. Therefore, it is thought that
the developed wrist bending rehabilitation robot can
be applied to severe stroke patient for the wrist
bending flexibility rehabilitation exercise.
ACKNOWLEDGEMENTS
This research was supported by Basic Science
Research Program through the National Research
Foundation of Korea(NRF) funded by the Ministry
of Education (2012R1A1A2A10041417)
REFERENCES
Pan, L., Song, A., Xu, G., Li, H., Zeng, H. and Xu, B.,
2013. Safety Supervisory Strategy for an Upper-Limb
Rehabilitation Robot Based on Impedance Control.
International Journal of Advanced Robotic Systems,
vol. 10 pp. 1-12. Ars-Journal.
Kim, H., Miller, L. M., Fedulow, I. M. S., Abrams, G. M.
N. B., and Rosen, J., 2013. Kinematic Data Analysis
for Post-Stroke Patients Following Bilateral Versus
Unilateral Rehabilitation With an Upper Limb
Wearable Robotic System. IEEE Transactions on
Neural Systems and Rehabilitation Engineering, vol.
21 no. 2 pp. 153-164. IEEE.
Li, C., Inoue, Y., Liu, T., Shibata, K. and Oka, K., 2010.
A Self-controlled Robot for Upper Limb
Rehabilitation, Proceedings of the 2010 IEEE
International Conference on Information and
Automation, pp. 115~118, June 20 - 23, Harbin, China.
IEEE.
Culmer, P. R., Jackson, A. E., Makower, S., Richardson,
R., Cozens, J. A., Levesley, M. C. and Bhakta, Bipin
B., 2010. A Control Strategy for Upper Limb Robotic
Rehabilitation With a Dual Robot System.
IEEE/ASME Transactions on Mechatronics, vol. 15,
no. 4, PP. 575-585. IEEE/ASME.
Nagai, K., Ito, Y., Yazaki, M., Higuchi, K. and Abe, S.,
2004. Development of a small Six-component
force/torque sensor based on the double-cross structure.
Journal of the Robotics Society of Japan, vol.22 no.3
pp. 361-369. RSJ.
Kim, G. S. and Park, J. J., 2005. Development of the 6-
axis force/moment sensor for an intelligent robot's
gripper. Sensors and Actuators A, Vol. 118 pp.127-
134. Sensors and Actuators A.
Kim, H. M., Yoon, J. W., Shin, H. S. and Kim, G. S.,
2010. Development of finger-force measuring system
with six-axis force/moment sensor for measuring a
spherical-object grasping force. Korean Society of
Precision Engineering, vol. 27 no. 11 pp.37-45.
KSPE.
Kim, G. S., and Yoon, J. W., 2007. Development of
Calibration System for Multi-Axis Force/Moment
Sensor and its Uncertainty Evaluation. Korean Society
of Precision Engineering, vol. 24 no. 10 pp. 91-98.
KSPE.
Kim, G. S., 2010. Development of force sensors of toes
and heel for humanoid robot’s intelligent foot. Korean
Society of Precision Engineering, vol.27 no. 10 pp.61-
68. KSPE.
ATI Industrial Automation, 2005. Multi-Axis
Forcre/Torque Sensor, ATI Industrial Automation. pp.
4-45.
DevelopmentofWristBendingRehabilitationRobot
279