Development of Exoskeletons and Applications on Rehabilitation

Abstract

For over a century, the development of exoskeletons experienced five periods including sprout, exploration, dormancy, accumulation and climax period from a concept in 19th century to applications in distinctive fields in 21th century. Recently, exoskeletons are applied in military, civilian and rehabilitation to augment the travel and loading abilities of soldiers, increase an operator’s load-handling capabilities, reduce the occurrence of musculoskeletal disorders, and improve the lost functions and quality of life of patients, respectively. Aiming at lessening the strain on physical therapists to train patients with severe or degenerative disabilities, motor cognitive limitation and improving their quality of life, exoskeletons are applied on the field of rehabilitation, mainly on patient training and locomotion. Although great progress has been made in the century long effort to design and implement exoskeletons, many design challenges still remain including powered devices, the comfort of human-machine interface and how to effectively understand the wearer’s intensions.

Full text

Development of Exoskeletons and Applications on Rehabilitation
Xinyu Guan, Linhong Ji and Rencheng Wang
Division of Intelligent and Bio-mimetic Machinery, The State Key Laboratory of Tribology, Tsinghua University, China
Abstract. For over a century, the development of exoskeletons experienced five periods including sprout, exploration,
dormancy, accumulation and climax period from a concept in 19th century to applications in distinctive fields in 21th
century. Recently, exoskeletons are applied in military, civilian and rehabilitation to augment the travel and loading
abilities of soldiers, increase an operator’s load-handling capabilities, reduce the occurrence of musculoskeletal
disorders, and improve the lost functions and quality of life of patients, respectively. Aiming at lessening the strain on
physical therapists to train patients with severe or degenerative disabilities, motor cognitive limitation and improving
their quality of life, exoskeletons are applied on the field of rehabilitation, mainly on patient training and locomotion.
Although great progress has been made in the century long effort to design and implement exoskeletons, many design
challenges still remain including powered devices, the comfort of human-machine interface and how to effectively
understand the wearer’s intensions.
1 Introduction
The term ‘exoskeleton’ was used in biology referring to
the chitinous or calcified external skeleton used by
numerous animal taxa for structural support and defense
against predators [1]. Now, the exoskeletons are
generally regarded as a technology that extends,
complements, substitutes or enhances human function
and capability or empowers the human limb where it is
worn. Different from other robots, the operator of an
exoskeletons is human who need to make decisions [2]
and perform tasks with exoskeletons. Through combining
human intelligence and machine power exoskeletons
enhance the abilities of both human power and machine
intelligence [3].
Since the concept of exoskeleton was produced in the
19th century, the development of exoskeletons have
undergone five phases, i.e. sprout period, exploration
period, dormancy period, accumulation period and climax
period [4, 5]. Exoskeletons apply and merge manifold
techniques involving mechanical and electronic
engineering, automation technology, biological, medical,
and material science [3]. Recently, exoskeletons are
applied in military, civilian and rehabilitation [4]. For
military purposes, the exoskeletons are designed to
augment the travel and loading abilities of soldiers [6];
for civil applications, the exoskeletons are used to
increase an operator’s load-handling capabilities and
reduce the occurrence of musculoskeletal disorders, or for
rescue; for rehabilitation, exoskeletons are aiming at
improving the lost functions [6, 7] and the quality of life
of patients with severe or degenerative disabilities, motor
cognitive limitation [8].
2 Development of exoskeletons
2.1 Sprout period
The sprout period lasted more than one century from
1830 to 1960. During this period, (1) A British inventor
Robert Seymour proposed the concept to help people
walk by a wearable device which was propelled by steam
in 1830; (2) An American inventor Ira C. C. Rinehart
conceptually designed a walking machine which enabled
an individual to step seven feet and four inches at an
ordinary stride in 1889; (3) From 1889 to 1890, Nicholas
Yagn, of St.Petersburg, Russia, designed a walking,
jumping, and running assisted device using a giant leaf
spring; (4) In 1890, another inventor Yagn designed an
exoskeleton with long leaf springs in parallel to the legs
to help people run faster and jump higher. In stance phase,
the weight of body can be transferred to the ground
directly by the spring to reduce the forces on the standing
leg. Most exoskeletons were conceptual design in sprout
period due to the limitations of the technology at that
time.
DOI: 10.1051/
C
Owned by the authors, published by EDP Sciences, 201
/
00 (
201 )
201conf
Web of Conferences
MATEC
atecm
0
,
00
0
6
6
6
40
40
2
2
4
4
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits
distribution, and reproduction in any medium, provided the original work is properly cited.
XQUHVWULFWHGXVH
Article available at http://www.matec-conferences.org or http://dx.doi.org/10.1051/matecconf/20164002004

MATEC Web of Conferences
(a) (b)
(c) (d)
Figure 1. Exoskeletons in sprout period: (a) exoskeleton
designed by Robert Seymour; (b) exoskeleton designed by Ira C.
C. Rinehart; (c) exoskeleton designed by Nicholas Yagn; (d)
exoskeleton designed by Yagn
2.2 Exploration period
The exoskeleton HARDIMAN developed by the US
Department and General Electric in 1965 marked the
exoskeleton development entering the exploration period.
HARDIMAN aimed at augmentation that the individual
who worn it could lift 1,500 lbs (682kg) [4]. In fact, only
one arm of HARDIMAN was developed and achieved to
lift 750lbs (341kg) until the 1970s. The failure of
HARDIMAN was mainly caused by which the energy
supplies were too huge to be portable, and the speed of
data processing and function control was slow [4]. In the
late 1960s and 1970s, an active anthropomorphic
exoskeletons with pneumatic power and partly
kinematical program for paraplegics was developed at the
Mihailo Pupin Institute under the Prof. Vukobratovic’s
guidance. At the same time, the theory of legged
locomotion systems was first put forward by Prof.
Vukobratovic, which established foundation for present
modern high-performance exoskeletons[9]. The
researchers of University of Wisconsin started to develop
a full lower limb exoskeleton in 1968 [10]. This
exoskeleton was designed to help those paraplegics with
complete upper limb capabilities to walk again. The
wearer can implement the sit-to-stand, stand-to-sit
translation and walk at 50% of normal speed. The hip and
ankle joint both had three rotational degrees of freedom
(DOF) and the knee joint had one rotational DOF. The
joints at hip and knee for flexion/extension were actuated
by hydraulic power, and the other joints were passive
[11]. Although this exoskeleton was developed for
paraplegics, there was not any report about the relevant
tests [6].
(a) (b) (c)
Figure 2. Exoskeletons in exploration period: (a) HARDIMAN
by General Electric; (b) active anthropomorphic exoskeleton by
Mihailo Pupin Institute; (c) lower limb exoskeleton by
University of Wisconsin
2.3 Dormancy period
The development of exoskeleton entered the dormancy
period in the 1980s. In the middle 1980s, the exoskeleton
concept “Pitman” was put forward by Jeffrey Moore at
the Los Alamos National Laboratory (Los Alamos, NM)
to apply in military to augment the soldiers’ capabilities.
However, this exoskeleton program was not funded by
the U.S. Defense Advanced Research Projects Agency
(DARPA). In 1988, Prof. Jichuan Zhang started to
research the electric walking machine for high leg
paraplegia patients at Tsinghua University. Using bar
linkage mechanism, the ipsilateral hip joint and knee joint
of the exoskeleton were actuated by only one motor. This
structure decreased the weight of the exoskeleton and
became more compact and portable. In 1990, G. John
Dick and Eric A. Edwards developed SpringWalker
according to the mechanism that a device in series with
the human leg can reduce the metabolic cost of running
by lowering impact losses and by providing energy return
[12]. However, SpringWalker can only enhance jumping
height. For running, it even increased metabolic cost by
20% compared to locomotion without it [5].
(a) (b)
Figure 3. Exoskeletons in dormancy period: (a) Electric
walking machine for high leg paraplegia patients of Tsinghua
University; (b) SpringWalker
2.4 Accumulation period
02004-p.2

ICMES 2015
From 1990 to 2000, the research of exoskeleton went into
the accumulation period. In 1992, Prof. Yoshiyuki Sankai
of University of Tsukuba started to develop a wearable-
type robot ‘Robot Suit HAL’ (Hybrid Assistive Limb),
which was intended to physically support a wearer’s daily
activities and heavy work [13]. The first prototype named
HAL-1 adopted DC motors and ball screws to augment
the wearer’s joint torque [14]. In 1994, researchers of
Kanagawa Institute of Technology developed a wearable
power assisting suit for nurses to enhance their muscle
strength to lift patients and avoid back injuries [15]. The
movement of the joints at arms, waist and legs of the suit
were sensed by strain sensors to detect the muscle force
and actuated by pneumatic rotary actuators with
concentric round boxes sliding each other [16, 17].
Compared to the over-ground exoskeletons, Hocoma AG
developed an immobile exoskeleton Lokomat consisting
of an over-ground exoskeleton, an advanced body weight
support system (BWS) and a treadmill in 2000 at
Switzerland [18, 19]. The Lokomat with repetitive
walking on one hand help to improve circulation,
strengthen bones and muscles and gain a natural walking
pattern [20], on the other hand decrease the physical
effort and constraint of the therapists [19].
(a) (b)
(c)
Figure 4. Exoskeletons in accumulation period: (a) HAL-1 of
University of Tsukuba; (b) suit sliding boxes actuator of
Kanagawa Institute of Technology; (c) Lokomat of Hocoma AG
2.5 Climax period
Exoskeletons attracted much more attention of
researchers from different countries including US, Japan,
Israel, France, Switzerland, South Korea, China, etc. and
the development of exoskeletons went into the climax
period since 2000.
One representative of exoskeletons applied in military,
Berkeley Lower Extremity Exoskeleton (BLEEX) was
developed to increase soldier’s load capacity, lessen the
risk of leg and back injury, decrease the metabolic
consumption and reduce the perceived level of difficulty
[5]. BLEEX adopted the hybrid hydraulic-electric
portable power supply [21] in order to achieve carrying
its own power source [22]. The hip and ankle joint of
BLEEX had three DOFs, respectively, among which hip
flexion/extension, abduction/adduction and ankle
flexion/extension were actuated by linear hydraulic
actuators. Its knee joint had one DOF actuated for
flexion/extension [23]. The control system of BLEEX
mainly collected sensory information from exoskeletons
to determine the kinematic and dynamic parameters [23].
It was reported that the soldier who wore BLEEX can
walk at 0.9 m/s with load up to 75 kg and 1.3m/s without
load [6].
The representative civil application of exoskeletons
was the Robot Suit Hybrid Assistive Limb (HAL)-5
developed by Professor Yoshiyuki Sankai at University
of Tsukuba for both power augmentation and walk
assistance [24, 25]. The hip and ankle joint of HAL-5
were actuated by a DC motor with harmonic drive for
flexion/extension, respectively, and the ankle joint for
flexion/extension DOF was passive with springs to return
a normal angle [26]. HAL-5 adopted joint torque
augmentation at the hip, knee and ankle joint, which is
different from BLEEX transferring the load to ground.
HAL-5 had two types of control systems: ĀCybernic
Voluntary Control System” and “Cybernic Autonomous
Control System” [26]. Cybernic Voluntary Control
System understood the wearer’s voluntary intention
according to the surface electromyographic (sEMG)
signals through placing the sEMG electrodes below the
hip and above the knee [5]. Then the power units of
HAL-5 generated power assist torque by amplifying the
wearer’s joint torque estimated from sEMG signals [26].
Cybernic Autonomous Control System was developed to
provide effective physical supports for the handicaps by
the potentiometers, ground reaction force sensors, a
gyroscope and accelerometer on the backpack to estimate
the posture since the signals of handicaps could cause a
broken walking pattern [1, 5, 26].
ReWalk from Argo Medical Technologies has been
commercialized for fundamentally changing the health
and life experiences of individuals with spinal cord
injuries (SCI). It consisted of a wearable brace support
suit with DC motors at hip and knee joint, respectively,
rechargeable batteries, a computer-based controller
contained in a backpack, a wireless mode selector, and an
array of sensors that measure upper-body tilt angle, joint
angles, and ground contact [27]. ReWalk utilized a
closed-loop algorithm software control and triggered and
maintained the walking pattern by detecting the wearer’s
upper-body movements. Additionally, ReWalk can also
help the wearer climb stairs, transform from sitting to
standing and vice versa. The crutches were necessary to
keep balance.
02004-p.3

MATEC Web of Conferences
(a) (b) (c)
Figure 5. Exoskeletons in climax period: (a) BLEEX; (b) HAL-
5; (c) ReWalk
3 Rehabilitation applications
Rehabilitation exoskeletons are usually applied on two
aspects. One can provide greater repeatability and lessen
the strain on physical therapists to train stroke survivors.
The other is to help the patients with spinal cord injury or
muscle atrophy walk and improve their activities of daily
living (ADL).
Lower-limb exoskeletons for rehabiliation were
grouped as treadmill gait trainers, foot-plate-based gait
trains, overground gait trainers and stationary gait trainers
according to rehabilitation principle (Fig. 6) [28].
Treadmill gait trainer system combines BWS and
exoskeleton type robots. Foot- plate-based gait trains
system achieves to simulate different gait patterns
through controlling the movement trajectory of the
separate foot plates on which patients’ feet position on.
Overground gait trainers system allow patients move
under their own control rather than moving them through
predetermined movement patterns. Stationary gait
trainers system aims at obtaining effcient strengthening
of the muscles, endurance development, joint mobility
and movement coordination.
(a) (b)
(c) (d)
Figure 6. Lower-limb rehabilitation exoskeletons for training:
(a) treadmill gait trainers; (b) foot-plate-based gait trains; (c)
overground gait trainers; (d) stationary gait trainers
Exoskeletons for locomotion can be divided into open-
chain structure and coupling structure based on
mechanical structure. The open-chain structure is that the
exoskeleton joints are separately actuated while the
coupling structure refers to the structure that one actuator
simultaneously actuates more than one joint coupled by
links or cables. The exoskeleton developed by Tsinghua
University (Fig. 3a) used coupling structure which has
advantages of light weight and less power. However,
most exoskeletons (e.g. HAL-5, ReWalk, etc.) adopted
the open-chain structure [18, 23, 26, 29] as this structure
is designed easily and controlled simply compared to the
coupling structure. Additionally, this structure can be
redesigned or extended conveniently based on the
original level. The movements of exoskeleton joints with
open-chain structure are coordinated through the control
system.
4 Challenges
Although great progress has been made in the century
long effort to design and implement exoskeletons, many
design challenges still remain.
The first challenge is the power of exoskeleton. Heavy
powered devices limit torque and power. Therefore,
developing light power system and small actuators are
necessary. The second challenge is the human-machine
interface. The interface designs often cause discomfort
when wearing the exoskeletons, which limit the length of
time that a device can be worn. It is certainly an
achievable goal to combine the exoskeletons and the
human body harmoniously and effectively to guarantee
comfort. The third challenge is how to understand the
intension of wearers, especially the patients with SCI
who lost neurotransmission below injured spinal canal.
Although rehabilitation exoskeletons like HAL-5 collect
and analyze surface electromyography (sEMG) signal to
detect the wearer’s intension, there are less direct
information exchange between current exoskeletons and
wearers’ nervous system. Thus it is critically important to
develop neural technology on human-machine direct
information exchange.
02004-p.4

ICMES 2015
5 Conclusions
In this paper, an insight review has been conducted on the
development of exoskeletons, their applications on
rehabilitation and current challenges. The development of
exoskeletons experienced five periods including sprout,
exploration, dormancy, accumulation period and climax
period. Especially in the climax period, more and more
rehabilitation exoskeletons were developed to lessen the
strain on physical therapists to train or assist patients.
Although great progress has been made in the century
long effort to develop exoskeletons, many design
challenges still remain.
References
1. E. Guizzo and H. Goldstein, Spectrum, IEEE 42,
50-56 (2005).
2. J. M. P. Gunasekara, R. A. R. C. Gopura, T. S. S.
Jayawardane, and S. W. H. M. T. Lalitharathne, in
System Integration (SII), 19-24 (2012).
3. C. J. Yang, J. F. Zhang, Y. Chen, Y. M. Dong, and Y.
Zhang, Proceedings of The Institution of Mechanical
Engineers Part C-journal of Mechanical
Engineering Science 222, 1599-1612 (2008).
4. R. Bogue, Industrial Robot: An International Journal
36, 421-427 (2009).
5. H. Herr, J NeuroEngineering Rehabil 6, 1-9 (2009).
6. A. M. Dollar and H. Herr, Robotics, IEEE
Transactions on 24, 144-158 (2008).
7. F. Yakub, A. Z. Md Khudzari, and Y. Mori,
International journal of rehabilitation research.
Internationale Zeitschrift fur
Rehabilitationsforschung. Revue internationale de
recherches de readaptation 37, 9-21 (2014).
8. T. Saito, T. Shibata, K. Wada, and K. Tanie, in
Computational Intelligence in Robotics and
Automation, 2003.1, 119-124 (2003).
9. M. K. Vukobratovic, International Journal of
Humanoid Robotics 4, 459-486 (2007).
10. J. Grundmann and A. Seireg, Theory Pract. Robots
Manipulators,233-240 (1977).
11. A. Seireg and J. G. Grundmann, Biomechanics of
Medical Devices,569-644 (1981).
12. T. A. McMahon and P. R. Greene, Journal of
biomechanics 12, 893-904 (1979).
13. Y. Sankai, in Robotics Research, 25-34 (2011).
14. F. Ichihashi, Y. Sankai, and S. Kuno, International
Journal of Sport and Health Science 4, 617-627
(2006).
15. K. Yamamoto, M. Ishii, H. Noborisaka, and K.
Hyodo, in Robot and Human Interactive
Communication, 2004. 661-666, (2004).
16. K. Yamamoto, K. Hyodo, and T. Matsuo, in
Proceedings of the JFPS International Symposium
on Fluid Power, 415-420, (1996).
17. K. Yamamoto, K. Hyodo, M. Ishii, and T. Matsuo,
JSME International Journal Series C Mechanical
Systems, Machine Elements and Manufacturing 45,
703-711 (2002).
18. G. Colombo, M. Joerg, R. Schreier, and V. Dietz, J
Rehabil Res Dev 37, 693-700 (2000).
19. R. Riener, L. Lunenburger, S. Jezernik, M.
Anderschitz, G. Colombo, and V. Dietz, IEEE
transactions on neural systems and rehabilitation
engineering 13, 380-94 (2005).
20. M. Wellner, T. Thuring, E. Smajic, J. von Zitzewitz,
A. Duschau-Wicke, and R. Riener, Studies in health
technology and informatics 125, 497-9 (2007).
21. K. Amundson, J. Raade, N. Harding, and H.
Kazerooni, in Intelligent Robots and Systems, 2005.,
3453-3458 (2005).
22. H. Kazerooni and R. Steger, Journal of dynamic
systems, measurement, and control 128, 14-25
(2006).
23. A. B. Zoss, H. Kazerooni, and A. Chu, Mechatronics
11, 128-138 (2006).
24. H. Kawamoto and Y. Sankai, in Computers Helping
People with Special Needs. 2398, 196-203(2002).
25. H. Kawamoto, L. Suwoong, S. Kanbe, and Y. Sankai,
in Systems, Man and Cybernetics, 2003. 2, 1648-
1653 (2003).
26. Y. Sankai, in Robotics Research. 66, 25-34 (2011).
27. G. Zeilig, H. Weingarden, M. Zwecker, I.
Dudkiewicz, A. Bloch, and A. Esquenazi, J Spinal
Cord Med 35, 96-101 (2012).
28. I. Díaz, J. J. Gil, and E. Sánchez, Journal of Robotics,
2011 (2011).
29. A. Esquenazi, M. Talaty, A. Packel, and M. Saulino,
American Journal of Physical Medicine &
Rehabilitation,91, 911-921 (2012).
02004-p.5