Active Lower Limb Orthosis

with One Degree of Freedom for Paraplegia

Takuhiro Sunada

, Goro Obinata

and Yanling Pei

Department of Robotic Science and Technology, Chubu University, 1200 Matsumoto-cho, Kasugai 487-8501, Japan

Department of Electronic Control and Robot Engineering, Aichi University of Technology,

50-2 Nishihasama-cho, Gamagori 443-0047, Japan

Keywords: Assistive Robot Device, Lower Limb Orthosis, Active Control, Mechanism of Linkage, All in One.

Abstract: This paper describes a new design of active lower limb orthosis which is called as oneDHALO (one-actuator

Drive Hip and Ankle Linked Orthosis). The oneDHALO has a linking mechanism which connects both ankle

joints with a medial hip joint and an actuator which drives the rotation angle. The joints linkage mechanism

keeps feet always in parallel with the floor to avoid stumbling, and assists swinging of the leg. One servo

motor has been introduced to assist and control the movement constrained by the mechanism. To match the

active movement to walking phase, optical sensors have been introduced at the soles for detecting the distance

between the feet and floor. The control device which consists of internal communication system, sensor

interfaces and a single board computer (Raspberry Pi) is designed for all in one with the mechanical part of

the orthosis. The system has achieved continuous walking based on the feedback signals from the sensors.

This paper reveals the preliminary experimental results of the system to show the good points of the design.

1 INTRODUCTION

Several hip-knee-ankle-foot orthotic systems have

been developed

for

the bipedal locomotion of

paraplegics (Rose, K. G., et al., 1979). However,

most

of existing orthotic systems have problems in use;

a large energy consumption for bipedal walking on

flat floor

(Stallard, J. and Major, E. R., 1998), 2)

bulky, 3) difficult to don/doff (Merati at al., 2000). To

solve those disadvantages, recent studies for lower

limb orthoses have aimed

at the usability with simple

and lightweight design (Kirtley, et al., 1996).

Problems of those systems are that the strides

were

short because the characteristic of the

horizontal rotation of

the pelvis in the orthosis,

otherwise the patient feel pain

(Saito et al., 1996,

1997). Genda et al. (2004) proposed an

orthosis,

HALO (Hip and Ankle Linked Orthosis), which

has

a link mechanism connecting ankle joints with a

medial

single hip joint. HALO partly solved the

problems of short strides

or large rotation of the

pelvis (Genda et al., 2004). The orthosis allows the

users to keep their

both feet always parallel with the

floor to avoid stumbling, and it

assists the swinging

of the leg when the contralateral ankle

is fixed

dorsally by loading. The energy consumption of the

user is a problem of HALO remained unsolved.

The

consumption energy of the users was

about

five

times larger than normal walking (Genda et al.,

2007). To reduce the

energy consumption of

walking with HALO, the authors proposed

extension of HALO w i t h one-actuator drive, which

is called as oneDHALO in our previous study

(Michal et al., 2017). In this paper, we describe a new

software development system and an all-in-one

design for oneDHALO.

The next chapter shows the mechanical

configuration and the motion behaviour of HALO. In

chapter 3, the extension with one actuator has been

proposed for the all in one system including the power

source. The active control system and the software

development system are proposed in chapter 3. The

comparison with another active assistive devices is

also given from the aspect of the usability,

performance and cost in the chapter. Chapter 4 shows

the concept of motion planning and the preliminary

experimental results. The role of each subsystem and

the system integration seeking suitable active motions

are explained. The preliminary experimental results

of the system are also shown to understand the good

points of this new design in chapter 5. The concluding

remarks are given in the final chapter.

504

Sunada, T., Obinata, G. and Pei, Y.

Active Lower Limb Orthosis with One Degree of Freedom for Paraplegia.

DOI: 10.5220/0007959905040509

In Proceedings of the 16th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2019), pages 504-509

ISBN: 978-989-758-380-3

2 MECHANICAL

CONFIGURATION OF HALO

The features of HALO come from the link

mechanism and the connections between the medial

hip joint and the ankle joints by Boden wires. The

hip

joint has two pulleys which rotate

independently around the same hip joint axis. Each

pulley is connected with the contralateral

knee-ankle-

foot orthosis. One Boden wire is set for coupling the

ankle

joint with a certain moment arm to the same

side

pulley at the hip joint. It is noted that the pulley

combined with the other side link, which is shown

in Figure 1. The wire of the other pulley is set in similar

manner for the other couple.

When dorsal flexion at

one ankle occurs with loading, the

wire connected

to the heel of the same side pulls and rotates

the

pulley of the hip joint connected with the

contralateral

knee-ankle-foot part of orthosis, and

then the generated torque at hip joint assists the

other

side leg in swinging forward.

Figure 1: Medial hip joint and the connections.

The Figure 2 explains the operation of Boden wire

in sagittal plane. The lower bar B

is connected to

the foot link and rotates around the axis O; the

upper bar B

is connected to the contralateral link

system and rotates around O’. The Boden wire W

connects the two Bars to transfer the rotation of B

to the upper bar B

(See (a) in Figure 2). The tilt

occurs, then the loading around O by the foot link

keeps B

in the horizontal orientation. This causes

that the bar B

pushes the bar B

up r

θ, where θ

is the rotation angle of B

. When r

= r

, the bar B

rotates the same angle θ. This means that the bar

is always parallel to B

(See (b)). If 2r

= r

, the

bar B

rotates double (See(c)). In this case, the

contralateral link system connected to the bar B

shown in Figure 2 is in a symmetric position with

respect to the line A-A’. The Boden wire of the

contralateral link makes the lower bar in parallel to

Figure 2: Operation of Boden wire.

the bar B

. When r

= r

and the link works as

supporting leg, the clearance to the floor bigger

than lsinθwill be required for the swinging leg

(the contralateral link). In the case of 2r

= r

, the

swinging leg (the contralateral link) declines

oppositely the same angle as the supporting leg; the

lower bar is parallel to the floor. Moreover, the

height of the swinging leg is equal to the supporting

leg. This causes that zero clearance is required for

the swinging leg.

Figure 3: Schematic illustration of mechanism operation.

The illustration in Figure 3 explains the way in which

the link mechanism works as an assistive device for

walking with paraplegia. When the upper body

inclines forward, a certain amount of dorsiflexion

torque is generated at the ankle joint because the body

weight is loaded on the supporting leg. First, the

torque is converted by the lower bar to the force,

which pulls the wire to rotate the upper bar which is

just the pulley at hip joint. Second, the torque

generated at the hip joint rotates the contralateral

knee-ankle-foot orthosis (the swinging leg). Third, it

results in the forward swing of the contralateral leg.

Active Lower Limb Orthosis with One Degree of Freedom for Paraplegia

505

Moreover, the loop connection of the Boden wire

keeps the lower link at the ankle joint parallel to

the floor as explained above with Figure 2. This

means that the user can enjoy his/her foot clearance

of the swinging leg without being anxious while

walking with HALO.

3 CONFIGURATION OF ACTIVE

CONTROL SYSTEM

In previous study, Genda et al. (2007) revealed the

pelvic rotation with Loftstrand crutches helped

enough

for the physiologically normal level by the

gait analysis with HALO in

the experiments.

However, the consumption energy of the user was

five times larger than normal walking. In this

paper, one servo motor with a communication port

has been introduced to assist and control the

movement constrained by the mechanism of HALO.

All in one orthosis including power source is the

target for our new design. There are several ways

when active control is applied to HALO mechanism.

An active extension of HALO with two electric

motors was proposed (Lee et al., 2015), which is

called as powered HALO (pHALO). In the extension,

the problem is solved how two motors are set into the

one degree-of-freedom mechanism. The direct

introduction of two motors in the two Boden wires of

HALO’s both legs provides a function for changing

the relative positions between the three links at the

two joints, ankle and hip. The function was used for

Table 1: Comparison of HALO and powered HALO

(Obinata, G., et al., ,2015).

Stride

length[m]

Gait speed

[m/s]

Displacement

of CoG [m]

Power

[Nm/s]

HALO

617.9

×10

−3

225.5

×10

−3

m/s

51.6

×10

−3

12.9

Nm/s

pHALO

637.9

×10

−3

283.5

×10

−3

m/s

25.7

×10

−3

7.84

Nm/s

cf.

[%]

+3.23 % +5.76 % -50.2 %

-39.2 %

pushing up the supporting leg to decline the upper

body forward in pHALO. Experimental results of

such a function showed the advantageous effect on

strides, gait speed, fluctuation of body COG,

mechanical power as shown in Table 1. The results

looks good; on the other hand, the total weight of the

device including the orthosis was more than 12Kg.

Taking out the link mechanism with Boden wires and

setting four motors at hip joint and at knee joints make

a multi-actuated assistive device. Device called as

Rewalk (ReWalk Robotics, 2011) is a typical and

successful one with four actuators. If the control

scheme is good for the user, Rewalk may work well

because it has enough degree-of-freedom to achieve

an ideal motion for the user’s severity of impairment.

However, the weight is 22Kg and it is expensive.

Lighter weight of assistive device is desired for users’

handling and make it possible bring the device

everywhere.

To reduce the weight, only one motor

has been introduced into HALO in this pa per. With

one motor, the movement of HALO which is

described in Chapter 2 will be compensated by

Figure 4: oneDHALO, the actuator and the sensors.

appropriate motor control scheme. The preliminary

results of one-drive HALO (oneDHALO) showed the

similar effects as powered HALO with two motors

(Michal et al., 2017). However, the device can work

only in the case that the control signal to the motor

and the electric power are provided from outside

through wires. In other words, the subsystems: the

controller, the sensor interface and the power source

were located outside of the HALO mechanism. On

the other hand, all in one system has been achieved

including the mechanism, the sensors, the actuator,

the controller, and the power source in a new design

of oneDHALO. To match the active movement to

walking phase, optical distance sensors (Ambient

light sensor, VCNL4010) have been introduced at the

soles of feet. The mechanical part, the actuator, the

power source, controller and the sensors are shown in

Figure 4. The control system which consists of

internal communication system (Ethernet), interfaces

for sensors and a single board computer (Raspberry

Pi) is designed for all in one with the mechanical part

of the orthosis, which is shown in Figure 5. The

ICINCO 2019 - 16th International Conference on Informatics in Control, Automation and Robotics

506

sensors have infrared emitter to analyze the

proximity, and transmit the measured distance to the

board computer (main controller) by I2C communica-

tion interface. The signals from the sensor are used

for the main controller to decide which leg is

supporting one. The total weight including the

mechanical part and the power source is 7.1kg. The

control system includes a router of TCP/IP, which

Figure 5: Software development system for oneDHALO.

communicates with a high specification personal

computer through Wi-Fi. The PC enjoys several kinds

of high-level programming languages for developing

the control algorithms of the active assistive orthosis.

In this design, the PC is used only for the software

development. The developed program for controlling

the active orthosis in real time will be sent to the

board computer, and be executed locally. The

actuator (HEBI X8-16, HEBI Robotics) is an

Ethernet-enabled device that integrates a brushless

motor, gear-train, a rotary encoder and control

electronics into a compact package. It runs on

standard DC voltages and communicates using

standard 10/100Mbps Ethernet. The module run a real

time operating system in the modules itself that

process commands and feedback at 1kHz. The

actuator accepts commands and responds to requests

to feedback through the network.

4 MOTION PLANNING AND

EXPERIMENT

4.1 Motion Planning

Assume that right leg is in supporting phase with

oneDHALO. When the motor gives torque of a

positive magnitude to swing the left leg forward, this

action assists the user’s walking forward. In this

situation, the torque of a negative magnitude

interferes walking forward. Hence, the controller

must detect which leg is in supporting phase. For this

purpose, proximity and ambient light sensing

modules (

VCNL4010) have been introduced for

measuring the distance between the feet and the floor.

The distance below a certain value means that the foot

is in supporting phase. To avoid malfunction of the

sensors due to unevenness of the floor surface, four

sensing modules are set for one foot. To plan the

motion of motor, the following parameters have to be

defined: the stride, the cycle time, and the assist level

of walking with the actuator. These parameters define

the pattern of hip joint angel while walking as in

Figure 7, or our simulation technology can generate a

suitable pattern for the hip joint angle (Obinata et al.,

2015). Assume that h*(t) is a cyclic function desired

for the hip joint. The following control system is

considered to achieve the desired motion of the hip

joint (See Figure 6). The controller is a simple

position controller with P, I, D elements for hip joint

angle h(t). The generated torque of hip joint is the

summation of the motor torque and the passive torque

generated by loading on the supporting leg. The

assistive level of active control can be adjusted by

tuning the controller parameters. The level of active

control L

is defined as follows:











()













(



)







 (1)

The passive torque can be considered as a disturbance

for the controller. If the passive torque t

(t) is

satisfactory and enough to generate exactly the joint

motion h*(t), the active torque t

(t) generated by the

Figure 6: Proposed control system.

Active Lower Limb Orthosis with One Degree of Freedom for Paraplegia

507

motor takes zero. If the passive torque t

(t) is short,

then the active torque t

(t) compensates to achieve the

joint motion h*(t). Therefore, if we can define

available reference joint angle h*(t) as the

reference,the control system automatically achieves

the required torque by compensating for the passive

torque. From the energy efficiency viewpoint, the

timing for applying h*(t) to the control system has to

be matched to the generated passive torque. For this

purpose, the phase matching between the passive

torque and the reference joint angle is important. The

sensors at feet for detecting walking phase can

provide the signals for the phase matching of the

reference joint angle h*(t) to the passive torque.

4.2 Experiment

Operation tests with the control system described

in the former chapter were conducted. The tests

were on the tracking performance of angle position

with a sinusoidal reference input and on the timing

problem with signals of the proximity sensors. The

tests was conducted without any load on the

actuator. The tracking performance of joint angle

to the reference was near perfect. The result is

shown in Fig. 7 with the signals of the proximity

sensor. It seems sufficient that the device achieves

a certain tracking performance and provides

enough assistive swing force to the user. The

control program made the actuator start when the

measured distance from the floor of the swinging

leg exceeded 2mm. This is the typical usage of the

proximity sensor. If the actuator starts before the

swing leg leaves the floor, it may cause turning in

situ. Signals from eight sensors (four sensors per

one foot) can be used to estimate upper body

movements of the user. Such estimation may be

useful to match the active assistive torque to the

passively generated torque specially for achieving

the coincidence between the two torques.

5 CONCLUDING REMARKS

This paper proposes a new design of active control

orthosis for paraplegic walking. After explaining

the operation of the link mechanism, the

configurations of the active control system and the

software development system are given. The total

weight including power source is 7.1kg, which is

the lightest device of active lower limb assistive

device for paraplegia in the world. This is achieved

because the device has only one electric motor as

the actuator and it works in concert with the

passively generated torque induced by the shift of

the upper body’s CoG (center of gravity). Finally,

the concept of motion planning and the preliminary

experimental results are given in this paper. The

tracking performance of active control system

indicates enough potential for the purpose of

assisting paraplegic walking in the preliminary

experiment. Moreover, the proximity sensors

introduced here will be useful for various control

schemes, which may provide the adaptability to

several types of the user. Defining and taking the

Figure 7: Reference, measured value of angle, and signals from the proximity sensors.

ICINCO 2019 - 16th International Conference on Informatics in Control, Automation and Robotics

508

timing of the reference input for the device

movement in concert with the user’s action or

intension is the main issue that should be solved in

further research.

ACKNOWLEDGEMENTS

This works is supported by JSPS KAKENHI Grant

Number JP16H03214. The authors thank to JSPS,

for the grant and to Dr. Genda for his original idea

of HALO.

REFERENCES

Genda, E.,Oota, K., Suzuki, Y., Koyama K., Kasahara, T.

(2004) “A

new walking orthosis for paraplegics: hip

and ankle linkage system,”

Prosthet Orthot Int, Vol.28,

pp.69-74.

Genda, E., Koyama, K., Oota, K., Suzuki, Y., Hayashi, M.

(2007) "Energy efficiency of walking with Hip and

Ankle Linked Orthosis: HALO," Gishisougu-

Gakkaishi (Journal of the Japanese Society of

Prosthetics and Orthotics), vol. 23, pp. 65-70, (in

Japanese).

HEBI X-SERIES ACTUATOR Ⓡ at: https://www.

hebirobotics.com/x-series-smart-actuators (Accessed:

20 April 2019).

Kirtley, C., McKey, K. S. (1992) “A medially-mounted

orthotic hip joint

for paraplegic walking systems

preliminary report on the polymedic

walker Device,”

Polymedic Technical Note, Australia.

Lee, J., Mizumoto, R., Obinata, G., Genda, E., Stefanov, D.,

Aoki, H., et al. (2015) "Gait generation for powered

Hip-Ankle-Linkage-Orthosis," Engineering in

Medicine and Biology Society 37th Annual

International Conference of the IEEE, pp. 5732-5735.

Merati, G., Sarchi, P., Ferrarin, M., Pedotti, A., Veicsteinas,

A. (2000)

“Paraplegic adaptation to assisted-walking;

energy expenditure during

wheelchair versus orthosis

use,” Spinal Cord, vol.38, pp.37-44.

Michal, G., Obinata, G., Genda, E., Babjak, J., Pei, T.

(2017) “Active lower limb orthosis with one degree of

freedom for people withparaplegia,” International

Conference of Rehabilitation Robotics of the IEEE,

pp.258-263.

Moore, P., Stallard, J. (1991) “A clinical review of adult

paraplegic patients

with complete lesions using the

ORLAU Para Walker,” Paraplegia, vol.

29, pp.191-196.

Obinata, G., Mizumoto, R., Lee, J., Genda, E., Hase, K.

(2015) "Simulation of knee-locked walking for

designing powered orthosis," Engineering in Medicine

and Biology Society 37th Annual International

Conference of the IEEE.

Rewalk, 2011. at: https://rewalk.com/ (Accessed: 20 April

2019)

Rose, K. G. (1979) “The principles and practice of hip

guidance articulations,”

Prosthet Orthot Int, vol.3,

pp.37-43.

Saitoh, E., Suzuki, T., Sonoda, S., Fujitani, J., Tomita Y.,

Chino, N. (1996)

“Clinical experience with a new hip-

knee-ankle-foot orthotic system

using a medial hip

joint for paraplegic standing and walking,” Am J

Phys

Med Rehahil, vol.75, pp.198-203.

Saitoh, E., Baba, M., Sonoda, S., Tomita, Y., Suzuki M.,

Hayashi, M. (1977)

“A new medial single hip joint for

paraplegic walkers,” In Proc. The

world congress of the

international rehabilitation medicine association,

Kyoto Japan, August 31, pp.1299-1305.

Stallard, J., Major, E. R. (1998) “A review of reciprocal

walking systems

for paraplegic patients: factors

affecting choice and economic justification,” Prosthet

Orthot Int, vol. 22, pp.240-247.

Active Lower Limb Orthosis with One Degree of Freedom for Paraplegia

509