Development of Self Support Device and Control for Operating the

Wheelchair for Upper Limb Disabled Persons

Taku Itami

, Toshihito Yabunaka

, Ken’ichi Yano

, Yasuyuki Kobayashi

Takaaki Aoki

and Yutaka Nishimoto

Dept. of Mechanical Engineering, Mie University, 1577 Kurimamachiya-cho, Tsu City, 514-8507, Japan

LUMINOUS JAPAN CO., LTD., 68 Hagae, Murakami, 959-3134, Japan

Dept. of Rehabilitation and Orthopedics, Gifu University Hospital, 1-1 Yanagido, Gifu, 501-1193, Japan

Dept. of Surgical Nursing, Gifu University School of Medicine, 1-1 Yanagido, Gifu, 501-1193, Japan

Keywords:

Exoskeleton, Rehabilitation, Operating Analysis, Electromyographic(EMG) Signals.

Abstract:

Nowadays, it has been actively expanded to develop assist robots attached directly. In this study, we focused

on developing an exoskeletal robot, speciﬁcally, a force transmission robot with a rotary drive type ratchet

mechanism, to enable users with upper limb disability to make use of their residual function to achieve better

function of their upper limbs. A lock/unlock mechanism on the elbow joint is effectively used to transmit the

user’s residual function around the shoulder to the hand. We conducted veriﬁcation experiments on whether the

developed mechanism enables the user to transfer the remaining force in the shoulder joint to operation force

in the hand. Three subjects with C5 and C6 spinal cord injury with disabilities affecting their hands, lower

limbs, and trunk muscles performed the veriﬁcation experiment with the developed device. We conﬁrmed that

they could operate a wheelchair on a slope and on grass when using the developed device, and they could use

their residual function around the shoulder more strongly. It can be expected to rehabilitation effect.

1 INTRODUCTION

The number of disabled people who have lost body

function due to accident or illness has recently been

increasing. Injuries sustained in a road or sports acci-

dent where damage to the cervical cord means the per-

son cannot feel pain or temperature, and cannot move

their upper limbs and lower limbs in many cases.

These injuries are one of the main factors why peo-

ple need to use a wheelchair in daily life. There are

over 100,000 people with spinal cord injury, and it is

estimated that more than 5,000 people newly suffer

such injury every year.

In recent years, there has been consider-

able research on exoskeletal robots that substi-

tute for users’ the loss of function or amplify

their power (Tsukahara et al., 2015),(M. H. Rah-

man and Archambault, 2015),(Juanjuan Zhang,

2015),(Mao et al., 2015),(Hsieh et al., 2015),

and we have also developed upper limb mo-

tion support robots(N.Mizutani and Y.Kobayashi,

2013),(T.Watanabe et al., 2011),(T.Yabunaka et al.,

2014).

However, the power assist exoskeletal robots de-

Figure 1: Motion assist robot for the upper limbs.

veloped in previous studies have required a high-

power drive system because the robots support move-

ment directly by using motors and other mechanisms.

This has meant that people with cervical cord in-

jury are not able to make use of their residual func-

tion because their limbs are moved by external power

sources. This also introduces many device problems

such as increased weight due to using a battery and

466

Itami, T., Yabunaka, T., Yano, K., Kobayashi, Y., Aoki, T. and Nishimoto, Y.

Development of Self Support Device and Control for Operating the Wheelchair for Upper Limb Disabled Persons.

DOI: 10.5220/0005990404660471

In Proceedings of the 13th Inter national Conference on Informatics in Control, Automation and Robotics (ICINCO 2016) - Volume 2, pages 466-471

ISBN: 978-989-758-198-4

large motor, reduced portability, and limits on the du-

ration of continuous use. These problems reduce the

utility, and there are also concerns over the loss of

residual function resulting from support by an exter-

nal power source.

We are therefore focusing our present work on the

residual function of the upper limbs in people who

use manual wheelchairs. We aim to develop a robot

orthosis that can make use of their residual function.

In cases of paralysis due to stroke or cervical cord

injury, the paralysis has a great effect on terminal

nerves. However, muscles which are nearer to the

trunk often experience milder symptoms. In partic-

ular, although people might not be able to exert force

because they have strong paralysis of the hands and

forearms. However, the area within the shoulder cir-

cumference on the trunk is often not paralyzed and

has muscles to exert force. There are many relatively

large muscles within the shoulder circumference, and

large movements can be achieved by using them. If

people can utilize these muscles as an assist source,

the robot orthosis is able to support the movement

based on their intention. The robot does not require a

high-power drive system as used in previous studies.

The robot orthosis is able to provide support by using

only a drive system with the minimum requirements

for transmitting residual function to help operate the

hand.

2 UPPER LIMB DYSFUNCTION

The Modiﬁed Zancolli classiﬁcation is usually used

to evaluate upper limb function in people with cervi-

cal spinal cord injury. According to this system, the

C5 classiﬁcation indicates the person can operate a

manual wheelchair, C4 or lower injury indicates the

need to use an electric wheelchair as they have little

residual function, and C5 or C6 indicates the bound-

ary between using a manual wheelchair and an elec-

tric wheelchair.

First, we measured the remaining power in the

shoulder joint circumference and the elbow joint cir-

cumference of a person with cervical cord injury. The

subject was a man with modiﬁed Zancolli classiﬁca-

tion of C6B1 in his right arm and C5B in his left arm.

In this experiment, we measured the extension and

ﬂexion power of his shoulder and elbow joints ﬁve

times each using a push-pull gauge, and calculated

the joint torque by taking the average. Table 1 shows

the experimental results.

In the experimental results, the extension torque

of the elbow joint could not be measured. In other

words, the former power of the elbow to extend was

fully lost.

As the next step, we measured the tangential op-

erating force on the wheel rim of the wheelchair com-

paring 1 link of the elbow joint (Subject’s elbow is

locked at 90[deg].) and 2 link of the elbow joint (Sub-

ject can operate the wheelchair freely.) by using the

Push-pull gauge. We experimented that the subject’s

hand position started from 140[deg] of the wheelchair

backward to 30[deg] of the wheelchair forward by

10[deg]. The experiment was performed twice for

each hand position, and calculated the average value

as the operating force. Fig 2 shows the results of

the experiment. As the result, the subject whose el-

bow joint is locked can put power to the wheel rim of

the wheel chair to operate in the entire operation area

compared to 2 link of the elbow joint. In particular,

a meaningful difference was seen in initial operation

interval. This result is thought that the power of the

shoulder can translate to the hand directly by locking

the elbow joint. In other words, the subject can add

torques to operate on such steps with high resistance.

We also conﬁrmed that the power loss in the elbow

joint had a large inﬂuence during wheelchair opera-

tion.

Table 1: Residual torque at the shoulder and elbow joints

for modiﬁed Zancolli classiﬁcation of C6B1.

Subject Extension torque Flexion torque

Shoulder joint 8.6[Nm] 11.3[Nm]

Elbow joint 0.0[Nm] 20.3[Nm]

2 link model

1 link model

020406080100120140160

Contact angle of hand and wheel rim [deg]

Tangential direction force [N]

Figure 2: Measured results of operation force between 2

link and 1link.

3 DEVELOPED ROBOT

ORTHOSIS

The lock mechanism in the elbow joint is created by

Development of Self Support Device and Control for Operating the Wheelchair for Upper Limb Disabled Persons

467

using a static element such as an orthosis and a free

mechanism depending on the situation by using a

drive system. The developed robot orthosis is in-

tended to support the independence of disabled people

in everyday life. Therefore it is desirable for the or-

thosis to be light, compact, and inexpensive. It is also

desirable for it to offer long-term use and durability.

We use a ratchet mechanism that is able to withstand

the load of the lock mechanism and that has ﬂexibility

in the rotational direction. We use the small-size DC

motor to drive the component, and we control these

parts to lock or unlock in a timely fashion. Figure

1 shows the robot orthosis that transfers the residual

function from the shoulder joint circumference to the

hands.

Our device consists of a drive mechanism, bat-

tery, control board in one part, and this device has

the special orthosis, and the glove sensor attached

to the hand that can detect contact to the rim of the

wheelchair. Our robot orthosis can be programmed

to drive independently. The sensors used are a small

rotary encoder for detecting the elbow joint, a glove

sensor for detecting contact with the wheel rim of

the wheelchair, a photo interrupter for controlling the

motor. When people with a cervical cord injury op-

erate a wheelchair, they operate it by pushing their

hand on the rim because they are unable to grasp the

rim. We therefore use a force sensor to detect contact

with the rim. In terms of the orthosis, we developed

a special orthosis by using two types of materials to

transfer power from the muscles within the shoulder

to their hands. This orthosis can provide legged robot-

enhancing capability, and also provides pronation and

supination to the forearm.

3.1 Rotary Drive Type Ratchet

Mechanism And Control

We developed a component that employs a ratchet

mechanism which has an externally toothed gear and

an internally toothed gear in one part. Two claws that

correspond with teeth on each are attached to this part.

This mechanism is therefore able to lock both direc-

tion linking the ﬂexion and extension of the elbow

joint by using only one part. This makes it possi-

ble to reduce the size of the robot orthosis. We con-

sider the physiological excursion of the elbow, which

is ﬂexural, to be about 140[deg] from the greatest ex-

tension and we removed the area that is not needed.

We created a tilt of 8.0[deg] at the point where the as-

pect of the gear comes into contact with the aspect of

the forearm frame. This is because the robot ortho-

sis might block the movement of ﬂexural extension of

the elbow joint. The claw is linked to the DC mo-

Figure 3: Upper left; Pattern of the meshing of ratchet

mechanism which elbow joint moves freely. Upper right;

Pattern of the meshing of ratchet mechanism which elbow

joint is completely locked. Lower left; Pattern of the mesh-

ing of ratchet mechanism which elbow joint is completely

locked in the direction of the ﬂexion movement and moves

freely in the direction of the extension movement. Lower

right; Pattern of the meshing of ratchet mechanism which

elbow joint is completely locked in the direction of the ex-

tension movement and moves freely in the direction of the

ﬂexion movement.

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

0.1

0.2

0.3

Glove sensor value [V]

Current value of the motor [A]

Time[s]

Glove sensor

Motor current

Figure 4: Monitoring motor current.

tor through two leaf springs. The system is such that

each claw is inserted and removed from the gear. We

use the small DC motor for driving the ratchet mech-

anism. There are two leaf springs of size 0.5[mm]

thickness between the claw of the ratchet and the DC

motor, and the two leaf springs are branched to hold

the eccentric cam that is attached to the output shaft

edge of the DC motor. The eccentric cam has two ﬂat

surfaces on the bottom, and each surface can touch

the upper arm frame. Because the DC motor occurs

resistance when the ﬂat surface of the cam comes into

contact with the upper arm frame by rotating, the elec-

tric current to the motor increases. This system sends

a stop instructiomjn by monitoring the increase in the

electric current levels to allow the system to stop the

motor at the desired position. Figure 4 shows the re-

sults for the motor electric current levels. In this ex-

periment, we lock or unlock by pushing the glove sen-

sor. The electric current levels increase after pushing

ICINCO 2016 - 13th International Conference on Informatics in Control, Automation and Robotics

468

Figure 5: Block diagram of basic control.

the glove sensor. After that, the electric current lev-

els decrease, and then increase again at around 0.3[s].

This is because the ﬂat surface of the cam comes into

contact with the upperframe. In this system, sending

the stop instruction to the motor by monitoring the in-

crease the electric current levels at the second salient.

For this reason, We can reduce power consumption

and the total lock/unlock frequency of this system are

41,000 in one charge.

3.2 Basic Control And Timing Chart

Figure 5 shows a block diagram of the basic control.

We use the FSR406 of the glove sensor, the rotary po-

sition sensor attached to the device and the manual

switch for changing mode. The FSR406 is the force

sensor that can measure the force over a large area,

and the system judges the contact with the wheel rim

or other objects. We use the encoder for sensing the

elbow angle. V

is the control voltage, θ

is the an-

gle of the device in the block diagram. In the basic

movement, if the output voltage from the signal re-

ceived from the glove sensor is more than 0.847[V],

a voltage instruction is sent to lock the motors. Af-

ter this, the system changes into PWM drive mode

from the signals from the photo interrupters and is

driven at 50% duty ratio. We give consideration for

the cam to make softly make contact with the upper

arm frame by changing the PWM drive mode. This

makes it possible to reduce electricity consumption

by using PWM drive mode to increase the longevity

of the battery. Figure 6 shows the timing chart for this

as a series of events.

If output voltage from the signal received from the

Glove sensor

Control voltage

Photo interrupter

Motor current

(CCW)

(CW)

PWM PWM

Figure 6: Timing chart.

glove sensor is less than 1.8[V], a voltage instruction

is sent to unlock the motors. The system changes the

PWM drive mode from the signals from the photo

interrupters, and sets the drive duty ratio to 50% to

unlock the motor. In other words, the system judges

the state of contact between the wheel rim and the

FSR406 of the glove sensor. If the objects make con-

tact with each other, the motors rotate direction to

lock, and when the FSR406 of the glove sensor de-

taches from the wheel rim, the motors rotate in the

unlock direction. We can change the output voltage

of the glove sensor and the duty ratio. For this reason,

it is possible to set the lock/unlock of the elbow joint

at the timing of the operator.

4 WHEELCHAIR OPERATION

EXPERIMENTS INVOLVING

SUBJECTS WITH CERVICAL

CORD INJURY

The subjects were three people with cervical cord in-

jury. The modiﬁed Zancolli classiﬁcations were as

follows: Subject 1, C6B1 for the right arm and C5B

for the left arm; Subject 2, C5A for the right arm

and C5B for the left arm; Subject 3, C6 for the right

and left arm. We veriﬁed differences between using

and not using the developed device, by using motion

analysis and electromyography for a road with high

resistance and on a slope. The subject who has the

most serious symptoms among the three is not able to

Figure 7: Left; Situation of the experiments on a slope

which set the inclination to 3.4[deg] in accordance with the

barrier-free law of Japan. Right; Situation of the experi-

ments on a lawn way simulated the irregular ground.

Development of Self Support Device and Control for Operating the Wheelchair for Upper Limb Disabled Persons

469

0 1 2 3 4 5 6 7 8 9 10

-1

Trapezius(Upper)

EMG[V]

0 1 2 3 4 5 6 7 8 9 10

0.5

1.5

iEMG[V]

0 1 2 3 4 5 6 7 8 9 10

-5

Biceps brachii

EMG[V]

0 1 2 3 4 5 6 7 8 9 10

Time [s]

iEMG[V]

0 1 2 3 4 5 6 7 8 9 10

-1

Trapezius(Upper)

EMG[V]

0 1 2 3 4 5 6 7 8 9 10

0.5

1.5

iEMG[V]

0 1 2 3 4 5 6 7 8 9 10

-5

Biceps brachii

EMG[V]

0 1 2 3 4 5 6 7 8 9 10

Time [s]

iEMG[V]

Figure 8: Left; activity of the upper trapezius and biceps brachii muscles without the proposed device by a subject with C5A

for the right arm and C5B for the left arm of cervical cord injury. Right; activity of the upper trapezius and biceps brachii

muscles with the proposed device by a subject with C5A for the right arm and C5B for the left arm of cervical cord injury.

perform extension movement of the elbow joint and

has very weak ﬂexion movement and we describe his

results. This study was approved by an institutional

review board and was performed in accordance with

the Declaration of Helsinki governing human studies.

Figure 7 shows the situation of the experiments.

In the experiment on a slope, we measured the

muscle power of the shoulder circumference by using

an EMG sensor. We used a treadmill to simulate the

slope and set the inclination to 3.4[deg] in accordance

with the barrier-free law of Japan. Figure 8 shows the

muscle power of the shoulder circumference.

0 2 4 6 8 10 12

With device

Without device

With device

Without device

Moving distance[m]

Time[s]

Velocity[m/s]

0.5

1.5

2.5

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Figure 9: Experimental results of overlapped waveforms of

operation in grass with/without the proposed device by C5A

and C5B of cervical cord injury.

From the motion analysis results, in the case of

not using the developed robot orthosis, the subject

reached the end of the slope after six strokes and about

13[s]. In the case of using the developed robot ortho-

sis, he reached it after four strokes and about 8[s].

The movement distance per stroke is thus increased.

In the muscle power measurement results, the mus-

cle power of the upper trapezius became more active

than without the device and the muscle power of the

biceps brachii became active in the last part of the op-

eration. This is because the movement distance per

stroke is higher, and the graph shows that continuous

operating power is added in the last part of the op-

eration when the operation occurs at the wheel rim

to the front of the wheelchair. Simulation of irregular

ground includes a rough way, a lawn way, and a gravel

path. We selected using a lawn way. The largest static

coefﬁcient of friction of the artiﬁcial lawn which we

set placed in this experiment was approximately 0.07.

We measured the behavior of the wheelchair by using

a motion capture system. Figure 9 shows the behavior

of the wheelchair. The upper graphs show the move-

ment distance, and the lower graphs show the veloc-

ity of the wheelchair. In this experiment, the subject

when using the developed robot orthosis moved a dis-

tance of 2[m] about 2[s] faster than when not using

the developed robot orthosis, and he was able to oper-

ate the wheelchair strongly during the ﬁrst stroke op-

eration when using the developed robot orthosis. In

the motion analysis results, when not using the de-

veloped robot orthosis, the subject reached a distance

of 2[m] after about ﬁve strokes, compared with about

four strokes when using it.

In other words, the movement distance per

ICINCO 2016 - 13th International Conference on Informatics in Control, Automation and Robotics

470

stroke was also increased, and he could operate the

wheelchair strongly.

5 DISCUSSION AND

CONCLUSION

We tested the robot orthosis on three subjects with

cervical cord injury and veriﬁed the difference be-

tween using and not using the developed device based

on motion analysis and electromyography at a road

with high resistance and on a slope where it is difﬁcult

to operate a wheelchair. We performed experiments

on a slope and lawn path. With the robot orthosis, the

movement distance per stroke was found to increase

and the muscle power of the upper trapezius became

more active compared to without it, and the muscle

power of the biceps brachii became active in the last

part of operation.

Biceps brachii muscles support ﬂexion of the el-

bow joint, and trapezius muscles support putting up

the shoulder. This set of experiment results show

that the subject can put up his shoulder by using bi-

ceps brachii muscles and also hold the rim of the

wheelchair by using his hand. Therefore, it is indi-

cated that the residual power of his shoulder transmit-

ted to his hand strongly by using the force transmis-

sion orthosis which can lock elbow joint.

In the future, we need to investigate protection

against dust, waterprooﬁng, safety for the device and

make further improvements. It is also necessary to ac-

cumulate more data by increasing the number of test

subjects. Further development is needed to employ

this mechanism and control system for industrial ma-

chinery and other products.

ACKNOWLEDGEMENTS

This work was supported by Projects for Developing

Assistive Products in Ministry of Health Labour and

Welfare of Japan.

REFERENCES

Hsieh, H.-C., Chien, L., and Lan, C.-C. (2015). Mechan-

ical design of a gravity-balancing wearable exoskele-

ton for the motion enhancement of human upper limb.

In 2015 IEEE International Conference on Robotics

and Automation (ICRA), pages 4992–4997.

Juanjuan Zhang, C. C. C. (2015). Passivity and stability

of human-robot interaction control for upper-limb re-

habilitation robots. IEEE Transactions on Robotics,

31:233–245.

M. H. Rahman, M. J. Rahman, O. L. C. M. S. J. P. K.

and Archambault, P. S. (2015). Development of a

whole arm wearable robotic exoskeleton for rehabil-

itation and to assist upper limb movements. Robotica,

33:19–39.

Mao, Y., Jin, X., Dutta, G. G., Scholz, J. P., and Agrawal,

S. K. (2015). Human movement training with a ca-

ble driven arm exoskeleton(carex). IEEE Transactions

on Neural Systems and Rehabilitation Engineering,

23:84–92.

N. Mizutani, T. Watanabe, K. T. Y. and Y. Kobayashi

(2013). A wheelchair operation assistance control for

a wearable robot using with user’s residual function.

In IEEE Int Conf Rehabil Robot.

Tsukahara, A., Hasegawa, Y., Eguchi, K., and Sankai, Y.

(2015). Restoration of gait for spinal cord injury pa-

tients using hal with intention estimator for preferable

swing speed. IEEE Transactions on Neural Systems

and Rehabilitation Engineering, 23:308–318.

T. Watanabe, K. Yano, T. Aoki, and Y. Nishimoto (2011).

Extension motion assistance for upper limb using

proxy-based sliding mode control. In Systems, Man,

and Cybernetics (SMC), 2011 IEEE International

Conference on.

T. Yabunaka, K. Yonezawa, N. Kato, K. Yano, Y.

Kobayashi, T. Aoki, and Y. Nishimoto (2014). A

wheelchair operation with an exoskeletal robot using

user’s residual function. In Micro-NanoMechatronics

and Human Science (MHS), 2014 International Sym-

posium on.

Development of Self Support Device and Control for Operating the Wheelchair for Upper Limb Disabled Persons

471