ArticlePDF Available

Design and Validation of a Lightweight Hip Exoskeleton Driven by Series Elastic Actuator With Two-Motor Variable Speed Transmission

Authors:

Abstract and Figures

To overcome the different requirements of torque-velocity characteristics for walking, running, stand-to-sit, sit-to-stand, and climbing stairs, we propose a novel concept for actuator design, namely, a series elastic actuator with two-motor variable speed transmission. The two-motor variable speed transmission can be adjusted in real-time to realize variable torque-velocity characteristics. A novel lightweight wearable hip exoskeleton driven by a series elastic actuator with two-motor variable speed transmission, named SoochowExo, has been developed in this paper for use in the elderly population. The weight of the whole hip exoskeleton is 2.85 kg (excluding batteries), including two actuators and the frame. The proposed hip exoskeleton can match the weight of the state-of-the-art hip exoskeleton while offering suitable torque and velocity for sitting-to-standing, walking, running on level ground, and climbing stairs. The benchtop tests and the preliminary human subject tests further confirm the design.
Content may be subject to copyright.
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
1
AbstractTo overcome the different requirements of torque-
velocity characteristics for walking, running, stand-to-sit, sit-to-
stand, and climbing stairs, we propose a novel concept for actuator
design, namely, a series elastic actuator with two-motor variable
speed transmission. The two-motor variable speed transmission
can be adjusted in real-time to realize variable torque-velocity
characteristics. A novel lightweight wearable hip exoskeleton
driven by a series elastic actuator with two-motor variable speed
transmission, named SoochowExo, has been developed in this
paper for use in the elderly population. The weight of the whole
hip exoskeleton is 2.85 kg (excluding batteries), including two
actuators and the frame. The proposed hip exoskeleton can match
the weight of the state-of-the-art hip exoskeleton while offering
suitable torque and velocity for sitting-to-standing, walking,
running on level ground, and climbing stairs. The benchtop tests
and the preliminary human subject tests further confirm the
design.
Index TermsHip Exoskeleton, Variable Speed Transmission,
Series Elastic Actuator, Locomotion modes, Multiple Tasks.
I. INTRODUCTION
GING is accompanied by physical performance and sense
reduction due to muscle weakness and motor unit loss [1].
The major concerns for senior citizens are handicaps related to
sitting-to-standing, climbing stairs, and walking. The hip
exoskeleton [2-9] has become a rising research topic in recent
years due to the small inertia added to the leg and the device’s
ability to assist elderly people in walking, sitting-to-standing,
and climbing stairs. However, the current hip exoskeletons still
have limitations, such as wearability, weight, and assistive
performance.
Human hip joints are very important for many activities that
are classified as activities of daily living (ADL), including
walking, running, sitting-to-standing, climbing stairs, and the
real-time adjustment of the step width to avoid falling [10, 11].
Furthermore, the hip joint has different characteristics for the
different types of motion. In particular, the velocity and torque
requirements for walking and sitting-to-standing are opposite.
The joint needs the high torque and slow velocity of the hip
flexion/extension (HFE) joints during sitting-to-standing,
whereas it needs a high velocity with small torque for walking
and running [12, 13]. To satisfy the requirement of high torque
This work was supported in part by the National Key R&D Program of China
(2020YFC2007804), the Natural Science Foundation of the Jiangsu Higher
Education Institutions of China (19KJA180009), the Natural Science
Foundation of Jiangsu Province (BK20191424), the Jiangsu Frontier Leading
assistance for sitting-to-standing, most exoskeletons have been
designed using high-reduce radio transmission or high-power
motors [14-16]. This design will inevitably increase the inertia
and weight. To improve the wearability, in addition to
improving the actuator’s ability, the hip exoskeleton should be
as light as possible, and the structure should be slimly designed.
To reduce weight, most hip exoskeletons adopt low-power
motors or low reduce radio gear to meet the high-velocity
requirement for walking and running. However, this will
compromise the loss of the ability of high-torque assistance. Su
proposed a quasi-direct drive actuation using a custom high
torque density motor and low ratio gear transmission for a hip
exoskeleton that demonstrates mechanical versatility for being
lightweight (3.4kg overall mass), highly-backdrivable (0.4 Nm
back drive torque) with high nominal torque (17.5 Nm) and
high control bandwidth (62.4 Hz) [17]. However, the nominal
torque of 17.5Nm is far away from the human biological hip
moment requirement of the sitting-to-standing.
An exciting concept, namely, a dual reduction actuator for
the knee joint of the flexible exoskeleton GEMS L-Type, has
been recently introduced [18]. The two reduction radios were
fixed and switched through a clutch. And the exoskeleton has
to stop when the reduction radios switch. Yang, et al. [19] and
Rouse, et al. [20] designed hip and knee exoskeletons so that
the gear ratio can be changed during movement. Vanderborght,
et al. [21] designed a variable stiffness actuator that used
another motor to change the actuator stiffness. Horst, et al. [22]
designed a lead screw and belt-based gear variable actuator for
Orthotics. Jang designed an active-type twisted string actuator
(TSA)-based on continuously variable transmission (CVT) for
the robot application [23]. Lee designed a new actuator with
continuously variable transmission using parallel dual-motors
and a planetary gear for mobile robots [24]. Tran developed a
powered knee prosthesis with actively variable transmission for
multiple tasks [25, 26]. However, there has been comparatively
little work on developing a hip exoskeleton with variable speed
transmission to meet the different torque-velocity for the
different gait tasks.
In this paper, a lightweight, compact, compliant robotic hip
exoskeleton is developed for the elderly population and the
population with lower-limb impairments. To overcome the
different requirements of torque-velocity characteristics for
walking, running, stand-to-sit, sit-to-stand, and climbing stairs,
Technology Fundamental Research Project (BK20192004D), and the
Distinguished Professor of Jiangsu province.
T. Zhang, C. Ning, L. Yang and M. Wang are with the Robotics and
Microsystems Center, College of Mechanical and Electrical Engineering,
Soochow University, Suzhou 215000, China. (e-mail: zhangt.hit@gmail.com).
Design and Validation of a Lightweight Hip
Exoskeleton Driven by Series Elastic Actuator
with Two-Motor Variable Speed Transmission
Ting Zhang, Member IEEE, Chuanxin Ning, Yang Li, Meng Wang
A
This article has been accepted for publication in IEEE Transactions on Neural Systems and Rehabilitation Engineering. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TNSRE.2022.3201383
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
2
we propose a novel concept for actuator design, namely, a series
elastic actuator with two-motor variable speed transmission.
The speed transmission ratio can be adjusted in real-time to
realize variable torque-velocity characteristics. Furthermore,
two-motor variable speed transmission can adapt to elderly
people’s personalized motion characteristics due to age and
physical condition.
The main contributions of this paper are as follows:
1) introduce the concept of the series elastic actuator with
two-motor variable speed transmission to the hip exoskeleton
design to meet the different requirements of torque-velocity
characteristics for walking, running, stand-to-sit, sit-to-stand,
and climbing stairs;
2) develop a novel lightweight, wearable hip exoskeleton
driven by series elastic actuator with two-motor variable speed
transmission, named SoochowExo, to promote the independent
living of the elderly and the population with lower-limb
impairments.
The paper is structured as follows. The hardware and
controller design are described in Sections II and III. Section III
and V present the bench-top testing and the preliminary
human subject tests, which show the performance of the
SoochowExo. The discussion and conclusion are described in
Sections VI and VII, respectively.
II. HARDWARE DESIGN
A. Design Objectives
The main objective of the hip exoskeleton is to assist the
elderly and the population with lower-limb impairments in
sitting-to-standing, walking, running, and climbing stairs. The
actuator requirements of torque and speed are different for
different locomotion modes, such as walking, running, stand-
to-sit, sit-to-stand, and climbing stairs [13, 27, 28]. The
actuators of hip exoskeleton are required to working in
distinctively different torque-speed load conditions. To achieve
this, a series elastic actuator with two-motor variable speed
transmission is designed to meet the different requirements for
the assistive torque and velocity. The peak torque and velocity
are designed to meet the velocity requirement, gain higher
torque output, and reduce the weight. To ensure a wide range of
speeds, the actuator’s requirement of the hip exoskeleton firstly
is to have high velocity, and then can offer large assistive torque
for the sitting-to-standing and slow walking. The design
requirements of the hip exoskeleton are summarized in Table I.
The hip exoskeleton was designed to offer assistive torque of
60 Nm with the velocity of 2.3rad/s at the highest transmission
ratio, while output torque of 12Nm with the velocity of 5.6rad/s
at the lowest transmission ratio.
Table I summarizes the hip exoskeleton design goals.
target
actual
Total active DOFs
Hip flex. /ext.
Hip flex. /ext.
Speed transmission ratio
range
6:1 ~ 32:1
6:1 ~ 32:1
Peak torque range (Nm)
12 ~ 60
12 ~ 62
Peak velocity range (rad/s)
5.6 ~ 2.3
5.9 ~ 2.8
Actuator
SEA
SEA
Weight (excluding the power
supply system)
As light as possible
2.85 kg
B. Design Overview
Fig. 1 shows the overall system of the hip exoskeleton. This
hip exoskeleton is designed to deliver assistive torque to both
HFE joints. The hip exoskeleton includes the flexible suit and
two module actuators that can connect to the flexible suit. The
hip exoskeleton consists of a pair of series elastic actuators with
two-motor variable speed transmission for the HFE joints, and
a pair of 3D-print customized frames. The layout of the hip
exoskeleton is designed to minimize the extra inertia on the
wearer’s limb and maximize wearing comfort. The interface
between the hip exoskeleton and the wearer consists of four
parts: a belt to conform to the wearer’s waist shape, two braces
for both thighs, and a suspender. The two actuators, the
embedded control system, and the battery are located close to
the back, which is in the proximity of the center of mass of the
wearer. The overall hip exoskeleton system weight without a
battery is 2.85 kg. The material and weight for different parts of
the hip exoskeleton are shown in Table II.
(a) Frame of the hip exoskeleton
(b) System overview of the hip exoskeleton
Fig. 1 System overview of the proposed hip exoskeleton: SoochowExo
The hip exoskeleton includes the following core elements:
1) Two-motor variable speed transmission: To meet the
assistance of various tasks, such as level-walking and sitting-
to-standing, the transmission ratio of the HFE actuators of the
hip exoskeleton could continue to change adaptively. The series
elastic actuator with two-motor variable speed transmission
uses a novel transmission system that adapts the actuator torque
and speed to the demand of different ambulation tasks. For
example, the speed transmission ratio adjusts to high for tasks
that require high torque and low speed, such as sitting-to-
standing tasks, and adjusts to low for tasks that require high
speed with low output torque, such as walking and running.
2) Series elastic actuator (SEA): The actuator is designed
based on the SEA concept presented in our previous works [7].
The SEA configuration can realize accurate interaction force
control and complaint interaction with the wearer, as well as the
This article has been accepted for publication in IEEE Transactions on Neural Systems and Rehabilitation Engineering. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TNSRE.2022.3201383
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
3
torque sensor, which measures the displacement under the load.
Table II The material and weight for different parts of the hip exoskeleton
Quantity
Material
weight
2
aluminum alloy
0.875
2
Plastic, textile
0.15
1
textile
0.1
2
plastic
0.2
1
fabric
0.2
1
-
0.1
2.85kg
C. The principle of the two-motor variable speed transmission
The principle of the two-motor variable speed transmission
for the hip exoskeleton is based on the planetary gear
mechanism. A planetary gear includes four parts: sun gear, ring
gear, planet gears, and carrier. There have two motion control
inputs, one is the sun gear and the other is the ring gear rotation,
to adjust the rotation speed of the carrier and the planet gears,
as shown in Fig. 2.
Fig. 2 Basic concept of two-motor variable speed transmission
When the ring gear is fixed (󰇗 ), the velocity relation
between the input 1 and output is:
󰇗 󰇡
󰇢󰇗 (1)
When the sun gear is connected with input 1 and the ring gear
is connected with input 2, the output velocity 󰇗 can be
expressed as 󰇗 󰇡
󰇢󰇗󰇡
󰇢󰇗 (2)
where ,
, and is the number of sun gear teeth, the
planetaries teeth, and the ring gear teeth, respectively.
Fig. 3 shows the concept of the planetary gear-based two-
motor variable speed transmission in our design. Two planetary
gear subsystems are series-connected together in the two-motor
variable speed transmission. The basic principle is that two
inputs contribute to one output. The first planetary gear’s
reduction ratio is constant, while the second planetary gear’s
reduction ratio is determined by two inputs and . The main
input comes from the output of the first planetary gear and is
connected to the sun gear of the second planetary stage. The
second input is connected to the ring gear. A worm and gear
driven by the small-power motor were applied at the input
to change the reduction ratio of actuators and can also provide
a self-locking function.
Fig.3 Schematic diagram of the transmission system
When the small-power motor is powered off (󰇗 ), the
two-motor variable speed transmission worked in the highest-
transmission ratio mode. The speed transmission ratio is:
󰇗
󰇗 (3)
󰇗
󰇗 (4)

󰇛󰇜󰇛󰇜 (5)
When the 󰇗 rotated in the opposite direction to the main
drive motor, then Two motor variable speed transmission
worked in the low-transmission ratio mode. The speed
transmission ratio is computed as follows:
󰇗
󰇗
󰇗 (6)
󰇗
󰇗 (7)
󰇗
󰇗 (8)
󰇗 󰇗 (9)
󰇛󰇜
󰇛󰇜󰇛󰇜 (10)
where , are the number of sun gear teeth and ring gear
teeth of the planetary gear 1, respectively, , are the number
of sun gear teeth and ring gear teeth of the planetary gear 2,
respectively, , is the number of worms and worm gear
teeth, respectively, is proportion factor of the velocity
between the main motor and secondary motor.
D. Fully-integrated SEA-with-two-motor-variable-speed-
transmission
Fig. 4 shows the CAD of the SEA-with-two-motor-variable-
speed-transmission. The system consists of the main motor,
secondary motor, two planetary gear modules, two motor
variable speed transmission, and strings. The main motor is
directed connected with the sun gear of the first planetary gear
module using a coupler. The carrier of the first planetary gear
module was mounted next to the two-motor variable speed
transmission. The two-motor variable speed transmission
includes a sun gear, ring gear, three planet gears, a carrier, a
worm-and-gear, and a secondary motor. The output of the first
planetary gear module is connected with the sun gear of the
two-motor variable speed transmission. The ring gear of the
two-motor variable speed transmission connected with the
second motor through a worm-and-gear.

Input 1
Planetary Gear Carrier
Output

Input 2
Input 1
Input 2
Output


Input 2
Output
󰇗
󰇗
󰇗
This article has been accepted for publication in IEEE Transactions on Neural Systems and Rehabilitation Engineering. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TNSRE.2022.3201383
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
4
Fig. 4 CAD of the SEA-with-two-motor-variable-speed-transmission
Fig. 5 shows the fabricated fully-integrated SEA-with-two-
motor-variable-speed-transmission with dimensions of Ø90
mm × 51mm. The overall actuator weight is 680 g (actuator
with link weight is 875 g), including the main motor and second
motor, two planetary gear-based two-motor variable speed
transmission systems, and two high-resolution encoders.
Furthermore, the motor driver and STM32 microcontroller-
based embedded control system are integrated into the actuator
as a module.
To meet the small dimension, lightweight, high power
density ratio requirements of hip exoskeleton design, a custom
high torque density motor (Our motor was manufactured by
Unitree Robotics) was adopted into our design to work as the
main motor of the SEA-with-two-motor-variable-speed-
transmission. The customized high torque density motor (280
g), which can output nominal torque of 2 Nm with a normal
speed of 2100 RPM works as the main input  of the two-
motor variable speed transmission.
The second input of the two-motor variable speed
transmission is driven through a small power DC motor (EC10,
Maxon, Sachseln, Switzerland), a harmonic gear drive (HDUC-
5-80-1U-CC, Harmonic Drive, Limburg, Germany), and a
worm-and-gear with a reduction ratio of 75:1 (MÄDLER,
modulus of 0.5 mm). This motor and transmission ratio
configure can ensure the maximum velocity of the second input
make the output 󰇗 in the (5) equal to zero, then the SEA-
with-two-motor-variable-speed-transmission work at the
lowest-transmission mode.
Considering assistive torque and velocity requirements
during level running and sitting-to-standing, we chose the gear
parameters , , , , , and . The relative ratio between
the maximum low- transmission and high- transmission modes
for the hip exoskeleton was set to 4. The transmission ratio
reached 32:1 in the highest-transmission mode, and the
resulting output torque reached 64 Nm with a nominal velocity
of 5.6 rad/s. In the low-transmission mode, we set the small-
power motor motion at maximum speed, resulting in the
minimum reduction ratio for the HEF joint being 6:1. The
resulting nominal velocity was up to 25 rad/s with an output
torque of 16 Nm. The lowest-transmission mode allows the
wearer to run up to 2.5 m/s with 20% biological torque
assistance.
Fig. 5 Fully-integrated SEA-with- two-motor-variable-speed-transmission
This article has been accepted for publication in IEEE Transactions on Neural Systems and Rehabilitation Engineering. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TNSRE.2022.3201383
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
5
Furthermore, the series elastic actuator concept was also used
in the design. There is a customized torsion spring serving as an
elastic element sandwich between the output of the
transmission and the torque output. This SEA configuration can
realize accurate torque control and complaint interaction, as
well as the torque sensor [8, 29]. The designed stiffness of the
monolithic disc-shaped torsion spring is 400 Nm/rad. The
maximum of the deformation and torque is 10° and 60Nm,
respectively. The torsion spring, with a high-resolution
quadrature kit encoder (Avago AEDA-3300 Series, Broadcom
Limited, South Carolina) is needed to measure the deflection; it
also operates as a torque sensor with a resolution of 0.03 Nm.
III. CONTROLLER DESIGN
A. Embedded System
The hip exoskeleton includes the onboard and off-board
custom electronic boards. The onboard custom electronic
circuit is devoted to controlling and driving the motors (two
motors for the HFE joint, and one motor for the step-width
adjustment joint), collecting data on the encoders, and CAN bus
communication. The chip of the onboard embedded system is
STM32 32-bit Arm Cortex MCUs. The onboard embedded
system is integrated into the modular joint, as shown in Fig. 5.
The off-board and battery are placed at the back of the
exoskeleton. One DSP chip (TMS320F2812)-based embedded
system operates controllers that read all sensor data to compute
high-level control algorithms. The high-level controller
communicates with a motor controller that is integrated into the
actuators through a CAN bus. There are two IMUs on both
thighs to estimate the acceleration of hip flexion/extension and
adduction/abduction. There is one more IMU placed at the back
to measure the angles, velocity, and acceleration of the upper
body.
B. Low-Level Torque Control of the SEA-with-two-motor-
variable-speed-transmission
For the low-level controller, the controller performs closed-
looped torque tracking control of the SEA-with-two-motor-
variable-speed-transmission. The torque sensor data measured
by the deformation of the torsion spring are fed back to the low-
level controller. The torque control is transformed into the
torsion spring’s deflection control:
󰇛󰇜  (11)
where is the torsion spring’s deflection, and is the stiffness
of the torsion spring.
The torsion spring’s deflection control is based on the
proportional-integral differential (PID) controller. Reference
torque signal is the input and is converted into a command
spring deflection angle for the feedback term. The output of the
PID torque controller is the desired velocity 󰇗. Then we can
get the control command of the main motor and secondary
motor through:
󰇗

󰇗

󰇗 (12)
C. High-Level Assistive Control
The hip exoskeleton works in various modes, such as level
walking (LW), stair ascent (SA), stair descent (SD), sitting-to-
standing, and standing-to-sitting. The relationship between the
assistive timing and the gait cycle is different for the various
locomotion modes inspired by the human hip joint biological
torque [30, 31]. The control goal is to offer effective assistive
torque to adapt to the different locomotion modes. Starting from
the heel strike (), the controller generates the desired
assistive torque. The assistive torque profile is a mixture of four
halves of minimum jerk curves joined together at their tops, as
shown in Fig. 6.
Fig. 6 Parameterization of the hip torque profile
The assistive torque profile is:
when  ,
󰇛󰇜 
 
 (13)
when  ,
󰇛󰇜  
 
 (14)
when  ,
󰇛󰇜   
  
 (15)
when  ,
󰇛󰇜  
 
 (16)
otherwise 󰇛󰇜 (17)
where 󰇛󰇜 is the assistive torque;
, , and
 are the hip angles corresponding to gait events of the
heel strike, heel off, and toe-off, respectively;  and  are
the flexion and extension peak torques, respectively;  is the
hip flexion angle corresponding to the assistive torque at the
peak timing; and is the hip joint angle. The gait cycle begins
and ends at the heel-strike event.  is maximum flexion
angle, which is an average value based on previous strides
measured through IMUs [32]. The proposed assistive torque
profile mimics the human biological hip moment profile. Over
a gait cycle, the assistive torque profile can generate an assistive
torque for both hip flexion and extension. Three parameters,
 ,  and  , are adjusted in real-time to adapt to the
different locomotion modes (LW, SD, SA, running) and the
This article has been accepted for publication in IEEE Transactions on Neural Systems and Rehabilitation Engineering. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TNSRE.2022.3201383
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
6
different speeds.  is defined as follows:
  (18)
where is a constant, which is to determine the hip flexion
assistive torque onset timing (increasing from toe-off and end
at the maximum flexion angle) in the gait cycle. And will be
set as different values for the LW, running, SD, and SA. The
onset timing is determined by the gait events.  and  are
scaled for the different locomotion modes. Additionally, 
and  are also scaled for different walking speeds and
benefits from the actuator being able to actively vary the
reduction ratio. The three key parameters (,  and ,) can
dictate the assistance profile that is offset timing, assistance
duration, and assistance magnitude.
IV. BENCHTOP EXPERIMENTS
A. Backdrivability and zero-torque control performance of the
SEA-with-two-motor-variable-speed-transmission
We conducted benchtop testing and quantified the
performance in terms of the torque and velocity capabilities and
backdrivability. The fully integrated SEA-with-two-motor-
variable-speed-transmission was fixed during the benchtop
experiments. The output shaft of the actuator was manually
pushed. The torque and angle data are measured through a
spring-based toque sensor and encoders.
Fig. 7 Backdrivability Experiment. (a) The actuator worked at the highest-
reduction ratio mode (the stiffness of the SEA), (b) closed-looped zero torque
control
In the first test, the actuator was working in the highest-
reduction ratio mode, both motor M1 and M2 are power off. Fig.
7 (a) shows the relationship between the back-drive torque and
the angle. The proposed hip exoskeleton’s passive compliance
is due to the elasticity introduced by the customized torsion
spring. In the second test, SEA-with-two-motor-variable-
speed-transmission worked at close-looped zero torque control
to testing the active compliance. The back-drive torque of the
fully integrated SEA-with-two-motor-variable-speed-
transmission within 0.1 Nm for the different motion frequencies,
as shown in Fig. 7 (b).
B. Closed-loop Torque Control Bandwidth with Different
Reduction Ratio
To test the torque and velocity characteristics of the fully
integrated series elastic actuator with two-motor variable speed
transmission, the actuator was velocity-controlled motion with
different loadings. The peak torque can reach 12 Nm when the
actuator operates in the lowest-transmission ratio mode (6:1),
and the normal velocity is more than 5.9 rad/s. The actuator is
operated with a peak torque of 62 Nm and a normal velocity of
2.8 rad/s when operating in the highest-transmission ratio mode
(32:1).
Fig. 8 Step response of the torque control
Table III The performance of the actuator with different transmission ratios
6: 1
12: 1
24: 1
32: 1
Peak Torque (Nm)
12
23
46
62
Peak velocity (rad/s)
5.9
5.18
3.75
2.8
Cutoff frequency (Hz)
43.75
29.17
19.44
12.52
Transmission ratio adjusting
time (increasing) (ms)
13±4
28±5
21±3
Transmission ratio adjusting
time (decreasing) (ms)
11±4
24±3
17±5
Step response tests were performed for the SEA-with- two-
motor-variable-speed-transmission with the output of the
device rigidly fixed. The output of the SEA-with-two-motor-
variable-speed-transmission was blocked while step responses
were tested (i.e., 10 Nm, 20 Nm, 40 Nm, and 60 Nm), as shown
in Fig. 8. Table III summarized the performance of the actuator
with different speed transmission ratios. For the SEA-with-
two-motor-variable-speed-transmission, the cutoff frequency
of the closed-loop torque control is 43.75 Hz, 29.17 Hz, 19.44
Hz, and 12.52 Hz at 10 Nm, 20 Nm, 40 Nm, and 60Nm,
respectively. We also test the adjustment time of the speed
transmission ratio increasing from 6:1 to 12:1, 24:1, and 32:1,
and then decreasing from 32:1 to 24:1, 12:1, and 6:1 five times.
The adjustment time is 13±4, 28±5, and 21±3s for increasing
from 6:1 to 12:1, from 12:1 to 24:1, and from 24:1 to 32:1,
respectively. And the adjustment time is 17±5, 24±3, and 11
±4s for decreasing from 32:1 to 24:1, from 24:1 to 12:1, and
from 12:1 to 6:1, respectively.
C. Continual Torque control of the SEA-with-two-motor-
variable-speed-transmission
In this experiment, the proposed SEA-with-two-motor-
variable-speed-transmission torque and velocity characteristics
This article has been accepted for publication in IEEE Transactions on Neural Systems and Rehabilitation Engineering. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TNSRE.2022.3201383
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
7
are examined in the condition that the speed transmission ratio
continually adjusting, while outside of SEA-with-two-motor-
variable-speed-transmission is interacted with an external
motion. A high-power motor (300W) with a reduction gear of
200:1 was used as external motion. A commercial torque sensor
is connected between the SEA-with-two-motor-variable-speed-
transmission and external motion to measure the load torque.
Test 1: The SEA-with-two-motor-variable-speed-
transmission is controlled to track a torque increased from 0Nm
to 60 Nm within the 60 s, while the external motor is braked.
Test 2: The SEA-with-two-motor-variable-speed-
transmission is controlled to track a torque increased from 0Nm
to 60 Nm within the 60 s, while the external motor with a fixed
speed of 3 rad/s.
Test 3: The SEA-with-two-motor-variable-speed-
transmission is controlled to track a fixed torque of 10 Nm,
while the external motion is increased from 0 rad/s to 10 rad/s
within the 60 s.
Fig. 9 Continual torque control. Note: The M2 velocity in the plots is the
harmonic gear drive output side’s velocity.
The SEA-with-two-motor-variable-speed-transmission
output torque, the velocity and current of the main and
secondary motor, and external motion velocity were recorded.
Fig. 11 shows the test results, from top to the bottom (Fig. 9 (a)
- Fig. 9 (c)) is the results of Test 1, Test 2, and Test 3. For Test
1, as shown in Fig. 9 (a), we can see that only the main motor
M1 motion. The SEA-with-two-motor-variable-speed-
transmission worked at the highest-transmission ratio mode.
The motor current increased with torque increasing. For Test 2,
as shown in Fig. 9 (b), the secondary motor M2 began motion
at T1 and stopped at T2 to adjust the reduction ratio to meet
both the torque and velocity requirements. For Test 3, as shown
in Fig. 9 (c), the secondary motor M2 began motion at T1 to
adjust the reduction ratio.
V. PRELIMINARY HUMAN SUBJECT TESTS
Five healthy subjects (age 28.56 y.o., weight 78.59.9 kg,
height 1.760.26 m) participated in the preliminary human
subject tests. All subjects signed an informed consent form. The
experiments were approved by the Ethical Committee of
Soochow University. The hip exoskeleton was powered by a
lithium-ion battery (400 g), which can support the device in
operation for 1.5 hours. The actuator, electric circuit, and
battery were fixed in the bag, which was fixed at the wearer’s
back.
The hip joint angle and assistive torque of the hip
exoskeleton were recorded to analyze the performance of the
assistance. The hip joint angles and assistive torque were
measured by an encoded spring-based torque sensor integrated
into a fully integrated series elastic actuator with actively
variable reduction ratio transmission. The hip angles during the
free motion were performed without wearing the hip
exoskeleton (no exo) and were measured through an inertial
motion capture system (Perception Neuron 2.0, Noitom
Technology Ltd.). The heel-strike event, which was detected
through two loading cells placed at each shoe, was used to
normalize the gait periods.
A. Sit-to-stand test
The subjects wore the hip exoskeleton and performed sit-to-
stand trials with different peak torque assistances. The subjects
were asked to keep the same speed to sitting-to-standing at each
trial. For the sitting-to-standing, the assistive torque profile is:
󰇛󰇜  󰇡
󰇢 󰇡
󰇢 (19)
where  is the peak assistive torque,  is the hip
flexion angle when the human stands, and  is the hip flexion
angle when the human sits. We set  to 5 degrees; the hip
exoskeleton will work in zero torque mode when the hip flexes
and is smaller than 5 degrees, which allows the subject to freely
move. The peak assistive torque was set as 50 Nm, 30 Nm, 10
Nm, and 0 Nm (power off) for the four test groups.
Fig. 10 shows the HFE joint angle, assistive torque under the
peak assistive torque was set as 50 Nm, 30 Nm, 10 Nm, and 0
Nm (power off). There were no significant changes in the hip
HFE joint angle on the peak assistive torque was set as 50 Nm,
30 Nm, 10 Nm, and 0 Nm (power off). We can see from Fig. 8
that the assistive torques start from the beginning sitting on the
chair and are gradually reduced to zero with standing. The
assistive torques show that the hip exoskeleton can output
assistive torque up to 50 Nm.
This article has been accepted for publication in IEEE Transactions on Neural Systems and Rehabilitation Engineering. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TNSRE.2022.3201383
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
8
Fig. 10 Sit-to-Stand transition experiment
B. Different locomotion modes assistance tests
The subjects walked on a treadmill for 10 min at a fixed speed
of 1.2 m/s, ran on a treadmill for 10 min at a fixed speed of 2.2
m/s, and ascended and descended a stair (four floors, 230 stairs)
under three conditions: wearing the hip exoskeleton with
assistance (assist on), wearing the hip exoskeleton without
assistance (assist off), and without wearing the hip exoskeleton
(no exo) [35]. Participants were asked to keep the selected
cadences constant during the ascending stair and descending
stairs. We set as different values for LW, SA, SD, and
running based on the human biological hip moment [12, 13].
Table IV Assistance parameters for the different locomotion modes


Speed
LW
0.2Nm/kg
-0.2Nm/kg
65%
1.2 m/s
SA
0.3Nm/kg
-0.5Nm/kg
20%
-
SD
0.4Nm/kg
-0.4Nm/kg
30%
-
Running
0.1Nm/kg
-0.1Nm/kg
20%
2.2m/s
 and  are the flexion and extension peak torque, respectively, is a
constant, which is to determine the hip flexion assistive torque onset timing
(increasing from toe-off and end at the maximum flexion angle) in the gait cycle.
Fig. 11 shows the average kinematic and assistive force
delivered by the hip exoskeleton under three conditions: assist
on, assist off, and no exo. Fig. 11 shows that the assistive torque
begins at the heel-strike event, smoothly increases to the
negative peak and ends at heel-off, then increases to the positive
peak and ends at maximum hip flexion for all locomotion
modes. Table V summarized the performance of hip
exoskeleton assistance for the different locomotion modes. The
peak torque reached 0.22 Nm/kg, 0.12 Nm/kg, 0.53 Nm/kg,
0.42 Nm/kg and 0.21 Nm/kg and 0.12 Nm/kg, 0.32 Nm/kg, and
0.42 Nm/kg at extension and flexion for walking, running, SA,
and SD, respectively. The assistive torque up to peak flexion
torque reached 20%, 20%, 30%, and 40% for the walking,
running, SA, and SD, respectively, while up to extension torque
at 20%, 20%, 30%, 40%. In summary, the hip exoskeleton only
introduced minimum changes to gait kinematics. The assistive
force tracked the designed profiles for different locomotion
modes. Furthermore, the assistive torque could also adapt to
different subjects’ independent walking speeds.
Table V Performance of assistance for the different locomotion modes
LW
Running
SA
SD
Angle (deg)
(-14, 23)
(-12, 18)
(4, 39)
(-14, 37)
Peak torque
(Nm/kg)
0.22
0.12
0.53
0.42
Peak velocity
(rad/s)
2.43
5.48
2.98
3.21
Torque
tracking error
(RMSE)
0.189
0.234
0.201
0.209
Fig.11 The results of the different locomotion modes assistance tests
VI. DISCUSSION
The design of hip exoskeletons for elderly individuals is
challenging in terms of the weight of the exoskeleton, the
compromise between force and velocity, the comfortable
This article has been accepted for publication in IEEE Transactions on Neural Systems and Rehabilitation Engineering. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TNSRE.2022.3201383
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
9
wearerexoskeleton interface, and assistance performance. The
hip joints work together on many activities, such as sitting-to-
standing, walking, running, and climbing stairs. There are
diverse characteristics of the different activities. The torque and
velocity requirements for walking/running and sitting-to-
standing are opposite. The motivation of this paper is to address
this opposite and to avoid using a high-power motor to reduce
the weight of the hip exoskeleton.
Benefitting the planetary gear’s multiport characteristics, a
series elastic actuator with two-motor variable speed
transmission is designed for our hip exoskeleton, which can
support different speed transmissions. In the lowest-
transmission ratio mode, the transmission ratio for the HFE
joints is 6:1. Based on our benchtop tests, the peak velocity is
more than 5.9 rad/s with an output torque of 16 Nm, which
allows the wearer to run at 2.5 m/s, as shown in Fig. 11. In
contrast, in the highest-transmission ratio mode, the
transmission ratio for the HFE joints becomes 32:1. The
resulting peak torque can reach 60 Nm, which can assist sitting-
to-standing with 25% of the biological hip flexion torque, as
shown in Fig. 10. Based on the two-motor variable speed
transmission and custom high torque density motor, the fully
integrated SEA weighing 870 g can generate assistive torque up
to 60 Nm in the highest- transmission ratio mode and can allow
the hip exoskeleton to run at a high speed of 2.5 m/s in the
lowest-reduction mode. These performances are achieved by
adopting a high torque density motor and two-motor variable
speed transmission rather than using a high transmission ratio
or high-power motor due to their inherent heavy and bulky
characteristics.
The proposed SEA-with-two-motor-variable-speed-
transmission can continuously adjust the speed transmission
ratio similar to [26], and not only discretely adjust between
tasks like [25, 26], and [18]. Different from the two same motor
parallel arranged input to the planetary gear in [24], there has a
high-power motor work as the main input to output the drive
torque, and a low-power motor with the worm-and-gear as
secondary input work to adjust the speed transmission ratio.
Different from [24], the reduction ratio of the proposed SEA-
with-two-motor-variable-speed-transmission is reduced with
the increasing velocity of the secondary motor. Furthermore,
there have two planetary gears series connections as the
variable speed transmission. These configure reduced the axial
dimension and the weight of the actuator. Besides, our two-
motor variable speed transmission connected with a torsion
spring assemble as a series elastic actuator to realize accurate
interaction force control and complaint interaction with the
wearer, as well as the torque sensor.
The proposed wearable hip exoskeleton with the two-motor
variable speed transmission series elastic actuator is to promote
the independent living of the elderly and the population with
lower-limb impairments. The SEA-with-two-motor-variable-
speed-transmission can meet the individual needs of the elderly
because different elderly people and people with lower-limb
impairments have different assistance level requirements.
Furthermore, our exoskeleton also has the potential to reduce
joint loadings of able-bodied populations for regular activities.
The motivation of this paper is to propose an actuator design
concept for the hip exoskeleton. The speed transmission ratio
adjusts rang can choose based on the user’s requirement.
Whatever how our effort, the hip exoskeleton still introduces
nonnegligible mass. Currently, the hip exoskeleton focuses on
reducing the weight to assist energy-efficient walking, which
hasn’t a frame to support the weight. A biomechanics study [34]
showed that the energy expenditure increased the smallest when
additional loading was applied to the waist compared with the
thigh, shank, or foot. The energy expenditure will not
significantly increase when the 4 kg mass loading is at the waist.
The proposed hip exoskeleton is 2.85 kg (excluding batteries).
And this weight is light than the most of hip exoskeletons [4, 6-
9, 15, 16]. Our hip exoskeleton with the rigid frame can connect
with the weight support frame, such as Kim designed bio-
inspired knee joint [35] or Lenzi designed self-aligning
mechanism [36], to balance the weight of the exoskeleton, it
will be beneficial for the elderly to wear the hip exoskeleton
long time.
Several limitations related to the device’s design and controls
still limit the possibility of adoption without extra work. There
is a tradeoff between the actuator’s backdrivability and the two-
motor variable speed transmission. The backdrivability of the
SEA-with-two-motor-variable-speed-transmission depends on
the stiffness of the customized torsion spring when both motor
M1 and M2 power off. The back-drive torque is big than [17,
37]. A simplistic control scheme was used in this paper to
evaluate the performance of the hip exoskeleton during steady-
state walking, running, SA, SD, RA, and RD. The peak timing
and peak force were fixed. Previous studies [38] have shown
that online peak force and timing optimization can improve gait
performance and reduce energy expenditure. The load cells at
the shoes were used in the controller to detect the gait event and
introduce cables at the shoes and legs. In the future, detecting
gait events based on IMUs placed on the thighs will eliminate
cable issues. With the design of the hip exoskeleton now
validated, additional elderly people and people with lower-limb
impairments clinical validation can conduct investigate clinical
outcomes in the further [39].
VII. CONCLUSIONS
In this paper, a novel concept for actuator design, namely, a
series elastic actuator with two-motor variable speed
transmission is introduced in the hip exoskeleton design. The
reduction ratio can be adjusted in real-time to meet the different
requirements of torque-velocity characteristics for walking,
running, stand-to-sit, sit-to-stand, and climbing stairs. The
design, control, and experimental validation of a novel
lightweight hip exoskeleton for elderly individuals and people
with lower-limb impairments are presented in this study. The
proposed lightweight hip exoskeleton driven by SEA-with-two-
motor-variable-speed-transmission can application in elderly
people and people with lower-limb impairments, as well as the
able-bodied populations for regular activities.
REFERENCES
[1] P. T. Chinimilli, S. M. Rezayat Sorkhabadi and W. Zhang, "Assessment
of Human Dynamic Gait Stability With a Lower Extremity Assistive
Device," in IEEE Transactions on Neural Systems and Rehabilitation
Engineering, vol. 28, no. 3, pp. 669-678, 2020
[2] F. Giovacchini, F.Vannetti, M. Fantozzi, M. Cempini, M. Cortese, A.
Parri, T. Yan, D. Lefeber and N. Vitiello, A Light-weight Active
Orthosis for Hip Movement Assistance, Robotics and Autonomous
Systems, vol. 73, pp. 123-134, 2015.
This article has been accepted for publication in IEEE Transactions on Neural Systems and Rehabilitation Engineering. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TNSRE.2022.3201383
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
10
[3] K. Yasuhara, K. Shimada, T. Koyama, T. Ido, K. Kikuchi, Y. Endo,
Walking Assist Device with Stride Management System, Honda R&D
Technical Review, vol. 21, no.2, pp. 5766, 2009.
[4] Q.Wu, X. Wang, F. Du and X. Zhang, “Design and Control of a Powered
Hip Exoskeleton for Walking Assistance”, International Journal of
Advanced Robotic Systems, vol.12, Issue 3, pp. 1-11, 2015.
[5] A. T. Asbeck, K. Schmidt and C. J. Walsh. Soft Exosuit for Hip
Assistance, Robotics and Autonomous Systems, vol. 73, pp. 102-110,
2015.
[6] T. Zhang, M. Tran, H. Huang, “Admittance Shaping-based Assistive
Control of SEA-driven Robotic Hip Exoskeleton”, IEEE/ASME
Transactions on Mechatronics, vol. 24, no. 4, pp. 1508-1519, 2019.
[7] T. Zhang, H. Huang, “Design and Control of a Series Elastic Actuator
with Clutch for Hip Exoskeleton for Precise Assistive Magnitude and
Timing Control and Improved Mechanical Safety”, IEEE/ASME
Transactions on Mechatronics, vol. 24, no. 5, pp. 2215-2226, 2019.
[8] I. Kang, D. D. Molinaro, S. Duggal, Y. Chen, P. Kunapuli and A. J. Young,
"Real-Time Gait Phase Estimation for Robotic Hip Exoskeleton Control
During Multimodal Locomotion," IEEE Robotics and Automation Letters,
vol. 6, no. 2, pp. 3491-3497, 2021.
[9] M. K. Ishmael, D. Archangeli and T. Lenzi, “Powered hip exoskeleton
improves walking economy in individuals with above-knee amputation”
Nature medicine, vol. 27, no. 10, pp. 1783-1788, 2021.
[10] M. K. Ishmael, M. Tran and T. Lenzi, "ExoProsthetics: Assisting Above-
Knee Amputees with a Lightweight Powered Hip Exoskeleton," 2019
IEEE 16th International Conference on Rehabilitation Robotics (ICORR),
2019, pp. 925-930.
[11] Y. Lee et al., "Biomechanical Design of a Novel Flexible Exoskeleton for
Lower Extremities," IEEE/ASME Transactions on Mechatronics, vol. 22,
no. 5, pp. 2058-2069, 2017
[12] M. Grimmer and A. Seyfarth, “Mimicking human-like leg function in
prosthetic limbs,” Neuro-Robotics, Berlin, Germany: Springer, 2014, pp.
105155.
[13] D. A. Winter, “The biomechanics and motor control of human gait:
Normal, elderly and pathological, Journal of Biomechanics, vol. 25, no.
8, pp. 949-949, 1992.
[14] A. M. Dollar and H. Herr, Lower Extremity Exoskeletons and Active
Orthoses: Challenges and State of the Art, IEEE Transactions on
Robotics, vol. 24, no. 1, pp. 144-158, 2008.
[15] T. Yana, M. Cempinia, C. M. Oddoa, N. Vitielloa, Review of Assistive
Strategies in Powered Lower-Limb Orthoses and Exoskeletons, Journal
Robotics and Autonomous Systems, vol. 64, Issue C, pp. 120-136, 2015.
[16] A. J. Young and D. P. Ferris, "State of the Art and Future Directions for
Lower Limb Robotic Exoskeletons," IEEE Transactions on Neural
Systems and Rehabilitation Engineering, vol. 25, no. 2, pp. 171-182, 2017.
[17] S. Yu et al., "Quasi-Direct Drive Actuation for a Lightweight Hip
Exoskeleton with High Backdrivability and High Bandwidth,"
IEEE/ASME Transactions on Mechatronics, vol. 25, no. 4, pp. 1794-1802,
2020.
[18] Y. Lee et al., "Flexible Gait Enhancing Mechatronics System for Lower
Limb Assistance (GEMS L-Type)," IEEE/ASME Transactions on
Mechatronics, vol. 24, no. 4, pp. 1520-1531, 2019.
[19] I. Kang, H. Hsu and A. Young, "The Effect of Hip Assistance Levels on
Human Energetic Cost Using Robotic Hip Exoskeletons," IEEE Robotics
and Automation Letters, vol. 4, no. 2, pp. 430-437, 2019.
[20] M. K. Shepherd and E. J. Rouse, "Design and Validation of a Torque-
Controllable Knee Exoskeleton for Sit-to-Stand Assistance," IEEE/ASME
Transactions on Mechatronics, vol. 22, no. 4, pp. 1695-1704, 2017.
[21] T. Bacek, M. Moltedo, C. Rodriguez-Guerrero, J. Geeroms, B.
Vanderborght and D. Lefeber, Design and evaluation of a torque-
controllable knee joint actuator with adjustable series compliance and
parallel elasticity,” Mechanism and Machine Theory, vol. 130, pp. 71-85,
2018.
[22] R. W. Horst and R. R. Marcus, "FlexCVA: A Continuously Variable
Actuator for Active Orthotics," 2006 International Conference of the
IEEE Engineering in Medicine and Biology Society, 2006, pp. 2425-2428.
[23] J. Jang, Y. -U. Song and J. -H. Ryu, "Active-Type Continuously Variable
Transmission System Based on a Twisted String Actuator," IEEE
Robotics and Automation Letters, vol. 7, no. 2, pp. 2605-2612, 2022.
[24] H. Lee and Y. Choi, "A New Actuator System Using Dual-Motors and a
Planetary Gear," IEEE/ASME Transactions on Mechatronics, vol. 17, no.
1, pp. 192-197, 2012.
[25] M. Tran, L. Gabert, M. Cempini and T. Lenzi, "A Lightweight, Efficient
Fully Powered Knee Prosthesis With Actively Variable Transmission,"
IEEE Robotics and Automation Letters, vol. 4, no. 2, pp. 1186-1193, 2019.
[26] Lenzi, Tommaso, et al. "Design, development, and testing of a lightweight
hybrid robotic knee prosthesis." The International Journal of Robotics
Research, vol. 37, no. 8, pp. 953-976, 2018.
[27] D.A. Winter, "Biomechanical Motor Patterns in Normal Walking,"
Journal of Motor Behavior, vol. 15, pp. 302-330, 1983.
[28] S.R. Hamner and S.L. Delp, "Muscle contributions to fore-aft and vertical
body mass center accelerations over a range of running speeds," Journal
of Biomechanics, vol. 46, pp. 780-787, 2012.
[29] T. Zhang, M. Tran, H. Huang, “Design and Experimental Verification of
Hip Exoskeleton with Balance Capacities for Walking Assistance”,
IEEE/ASME Transactions on Mechatronics, vol. 23, no. 1, pp. 274-285,
2018.
[30] X. Liu and Q. Wang, "Real-Time Locomotion Mode Recognition and
Assistive Torque Control for Unilateral Knee Exoskeleton on Different
Terrains," IEEE/ASME Transactions on Mechatronics, vol. 25, no. 6, pp.
2722-2732, 2020.
[31] W. Huo, S. Mohammed, Y. Amirat and K. Kong, "Fast Gait Mode
Detection and Assistive Torque Control of an Exoskeletal Robotic
Orthosis for Walking Assistance," IEEE Transactions on Robotics, vol.
34, no. 4, pp. 1035-1052, 2018.
[32] Y. Ding, I. Galiana, C. Siviy, F. A. Panizzolo and C. Walsh, "IMU-based
iterative control for hip extension assistance with a soft exosuit," 2016
IEEE International Conference on Robotics and Automation (ICRA),
Stockholm, 2016, pp. 3501-3508.
[33] J. Kim, et al., “Reducing the metabolic rate of walking and running with
a versatile, portable exosuit,” Science, vol. 365, no. 6454, pp. 668-672,
2019.
[34] R. C. Browning, J. R. Modica, R. Kram, and A. Goswami, “The effects
of adding mass to the legs on the energetics and biomechanics of walking,”
Medicine and Science in Sports and Exercise, vol. 39, no. 3, pp. 515525,
2007.
[35] T. Kim, M. Jeong and K. Kong, "Bioinspired Knee Joint of a Lower-Limb
Exoskeleton for Misalignment Reduction," IEEE/ASME Transactions on
Mechatronics, vol. 27, no. 3, pp. 1223-1232, 2022
[36] S. V. Sarkisian, M. K. Ishmael and T. Lenzi, "Self-Aligning Mechanism
Improves Comfort and Performance with a Powered Knee Exoskeleton,"
IEEE Transactions on Neural Systems and Rehabilitation Engineering,
vol. 29, pp. 629-640, 2021.
[37] H. Zhu, C. Nesler, N. Divekar, V. Peddinti and R. D. Gregg, "Design
Principles for Compact, Backdrivable Actuation in Partial-Assist
Powered Knee Orthoses," IEEE/ASME Transactions on Mechatronics,
vol. 26, no. 6, pp. 3104-3115, 2021.
[38] Y. Ding, M. Kim, S. Kuindersma, C. J. Walsh, “Human-in-the-loop
optimization of hip assistance with a soft exosuit during walking,” Science
Robotics, vol. 3, issue. 15, eaar5438, 2018.
[39] T. Elery, S. Rezazadeh, C. Nesler and R. D. Gregg, "Design and
Validation of a Powered KneeAnkle Prosthesis with High-Torque, Low-
Impedance Actuators," IEEE Transactions on Robotics, vol. 36, no. 6, pp.
1649-1668, 2020.
This article has been accepted for publication in IEEE Transactions on Neural Systems and Rehabilitation Engineering. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TNSRE.2022.3201383
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
... DMAs are also used in prosthetic hands where limitations in size and weight make it more difficult to include a compact and lightweight actuator which can provide a sufficient grip force and an appropriate grasp speed [41,42]. Recently, a lightweight hip exoskeleton was presented in [43] combining the DMA technology with series elastic actuation. ...
... A multitude of SEA designs for powered lower-limb exoskeletons has been proposed. A common implementation of SEAs for powered exoskeletons consists of using a torsional spring in combination with planetary or harmonic-drive gearboxes (Chen et al., 2017;Kim and Bae, 2017;Aguirre-Ollinger and Yu, 2021;Zhang et al., 2022). These solutions tend to be quite compact but have some drawbacks. ...
Article
Full-text available
Powered exoskeletons need actuators that are lightweight, compact, and efficient while allowing for accurate torque control. To satisfy these requirements, researchers have proposed using series elastic actuators (SEAs). SEAs use a spring in series with rotary or linear actuators. The spring compliance, in conjunction with an appropriate control scheme, improves torque control, efficiency, output impedance, and disturbance rejection. However, springs add weight to the actuator and complexity to the control, which may have negative effects on the performance of the powered exoskeleton. Therefore, there is an unmet need for new SEA designs that are lighter and more efficient than available systems, as well as for control strategies that push the performance of SEA-based exoskeletons without requiring complex modeling and tuning. This article presents the design, development, and testing of a novel SEA with high force density for powered exoskeletons, as well as the use of a two degree-of-freedom (2DOF) PID system to improve output impedance and disturbance rejection. Benchtop testing results show reduced output impedance and damping values when using a 2DOF PID controller as compared to a 1DOF PID controller. Human experiments with three able-bodied subjects ( N = 3) show improved torque tracking with reduced root-mean-square error by 45.2% and reduced peak error by 49.8% when using a 2DOF PID controller. Furthermore, EMG data shows a reduction in peak EMG value when using the exoskeleton in assistive mode compared to the exoskeleton operating in transparent mode.
... So, to compensate for the limitations of these rigid systems, SEA was developed and widely used in areas where humans and robots collaborate (Paine et al. 2013;Zhang et al. 2022). However, in the case of SEA, there is a limitation that it can be used only for specific tasks because it has only one stiffness, so it has high stability but low accuracy. ...
Article
Full-text available
The series elastic actuators (SEAs) have been popularly applied to various robotic applications. The rigid actuator is vulnerable to applications where back drivability is required despite many automated assembly tasks requiring the function for enhanced robot operation. Since the SEA comprises only one stiffness, there is a limit to the usable application. Therefore, the variable stiffness actuator (VSA) with multiple stiffnesses is to be developed. This study aims to develop a linear multi-degree of freedom VSA. Instead of the conventional rotary-type VSA with one degree of freedom, we propose an actuator that can control stiffness with more than two degrees of freedom. The proposed VSA is advantageous for tasks such as screwing and deburring that require a change in stiffness depending on the circumstances. The mechanism is configured using a linear spring, and the stiffness is controlled by changing the tension spring to a nonlinear form. We also develop the new VSA through this mechanism and perform simulation analysis. In addition, the validity of the proposed VSA is verified through experiments by manufacturing a VSA based on the design.
... What's more, it is worth mentioning that current walking assistive devices and cooperative robots usually utilize elastic components in series with stiff actuators (named as series elastic actuator, SEA) to guarantee safety in physical human-robot interactions [42][43][44]. Ning et al. [45] and Zhang et al. [46] have developed a SEA with an active-type continuously variable transmission for the exoskeleton design, which can meet the different requirements of daily life. Lee et al. [47] have proposed a tendon type of SEA for a knee assistive exosuit, which can offer power in the knee joint when walking up and down the stairs. ...
Article
Full-text available
In order to improve the elderly people's quality of life, supporting their walking behaviors is a promising technology. Therefore, based on one ultrasonic motor, a wire-driven series elastic mechanism for walking assistive system is proposed and investigated in this research. In contrast to tradition, it innovatively utilizes an ultrasonic motor and a wire-driven series elastic mechanism to achieve superior system performances in aspects of simple structure, high torque/ weight ratio, quiet operation, quick response, favorable electromagnetic compatibility, strong shock resistance, better safety, and accurately stable force control. The proposed device is mainly composed of an ultrasonic motor, a linear spring, a steel wire, four pulleys and one rotating part. To overcome the ultrasonic motor's insufficient output torque, a steel wire and pulleys are smartly combined to directly magnify the torque instead of using a conventional gear reducer. Among the pulleys, there is one tailored pulley playing an important role to keep the reduction ratio as 4.5 constantly. Meanwhile, the prototype is manufactured and its actual performance is verified by experimental results. In a one-second operating cycle, it only takes 86 ms for this mechanism to output an assistive torque of 1.6 N·m. At this torque, the ultrasonic motor's speed is around 4.1 rad/s. Moreover, experiments with different operation periods have been conducted for different application scenarios. This study provides a useful idea for the application of ultra-sonic motor in walking assistance system.
Article
Full-text available
Above-knee amputation severely reduces the mobility and quality of life of millions of individuals. Walking with available leg prostheses is highly inefficient, and poor walking economy is a major problem limiting mobility. Here we show that an autonomous powered hip exoskeleton assisting the residual limb significantly improves metabolic walking economy by 15.6 ± 2.9% (mean ± s.e.m.; two-tailed paired t-test, P = 0.002) in six individuals with above-knee amputation walking on a treadmill. The observed metabolic cost improvement is equivalent to removing a 12-kg backpack from a nonamputee individual. All participants were able to walk overground with the exoskeleton, including starting and stopping, without notable changes in gait balance or stability. This study shows that assistance of the user’s residual limb with a powered hip exoskeleton is a viable solution for improving amputee walking economy. By significantly reducing the metabolic cost of walking, the proposed hip exoskeleton may have a considerable positive impact on mobility, improving the quality of life of individuals with above-knee amputations. In individuals with above-knee amputation, a powered hip exoskeleton, which reduces the metabolic cost of walking by adding energy at the hip joint, greatly improves walking economy and has the potential to improve mobility.
Article
Full-text available
Misalignments between powered exoskeleton joints and the user’s anatomical joints are inevitable due to difficulty locating the anatomical joint axis, non-constant location of the anatomical joint axis, and soft-tissue deformations. Self-aligning mechanisms have been proposed to prevent spurious forces and torques on the user’s limb due to misalignments. Several exoskeletons have been developed with self-aligning mechanisms based on theoretical models. However, there is no experimental evidence demonstrating the efficacy of self-aligning mechanisms in lower-limb exoskeletons. Here we show that a lightweight and compact self-aligning mechanism improves the user’s comfort and performance while using a powered knee exoskeleton. Experiments were conducted with 14 able-bodied subjects with the self-aligning mechanism locked and unlocked. Our results demonstrate up to 15.3% increased comfort and 38% improved performance when the self-aligning mechanism was unlocked. Not surprisingly, the spurious forces and torques were reduced by up to 97% when the self-aligning mechanism was unlocked. This study demonstrates the efficacy of self-aligning mechanisms in improving comfort and performance for sit-to-stand and position tracking tasks with a powered knee exoskeleton.
Article
Misalignment between an exoskeletal robot and the wearer’s body must be minimal to prevent injury and discomfort during active assistance. It is particularly difficult for the robot to imitate the knee joint motion of a human because the human knee joint has a complex structure, which cannot be realized with a single revolute joint. In this article, a knee joint mechanism that closely realizes the human knee joint motion using the curved guide rail and bearings is proposed. For the optimal design of the proposed mechanism, the motions of the tibia and the femur are captured, and the guide rail of the proposed mechanism is designed to realize the captured human knee joint motion. A simulation study is performed based on kinematic calculations, and the shape of the guide rail is optimized for the proposed device to precisely imitate the human motion. This article also introduces an experimental method for the quantitative evaluation of the misalignment; the pressure inside the brace is measured using an air-pressure sensor pad, and the pressure measurements are utilized for objective comparison of the misalignment with respect to the joint mechanism. An additional experiment is performed to verify the noninterfering of the human walking motion.
Article
This letter presents an active-type twisted string actuator (TSA)-based continuously variable transmission (CVT). The proposed mechanism improves the advantage of the intrinsic CVT nature of TSA by compensating its nonlinear transmission ratio (TR) with the active and continuous conversion of the TR during contraction at a constant load as well as variable load. To change the TR, the proposed CVT adjusts the ratio of rotation between individual twists and overlap twists, which are two different types of string-twisting patterns. The performance of the TSA-based active CVT was experimentally assessed with the fabricated prototype and compared with a mathematical model. This adjustment in the rotation ratio between individual and overlap twists is realized using a planetary gear. By adjusting the rotating speed ratio between the sun and ring gear, the ratio of rotation between individual and overlap twists can be adjusted. Unlike the conventional TSA-based CVT, the proposed CVT can actively and continuously change the TR independent of the external load, even during operation. Consequently, it allows an increase in the operating range of the TSA while maintaining a high contraction speed.
Article
We developed and validated a gait phase estimator for real-time control of a robotic hip exoskeleton during multimodal locomotion. Gait phase describes the fraction of time passed since the previous gait event such as heel strike, and is a promising framework for appropriately applying exoskeleton assistance during cyclic tasks. A conventional method utilizes a mechanical sensor to detect a gait event and uses the time since the last gait event to linearly interpolate the current gait phase. While this approach may work well for constant treadmill walking, it shows poor performance when translated to over ground situations where the user may change walking speed and locomotion modes dynamically. To tackle these challenges, we utilized a convolutional neural network-based gait phase estimator that can adapt to different locomotion mode settings to modulate the exoskeleton assistance. Our resulting model accurately predicted the gait phase during multimodal locomotion without any additional information about the user's locomotion mode}, with a gait phase estimation RMSE of 5.04 0.79%, significantly outperforming the literature standard (p < 0.05). Our study highlights the promise of translating exoskeleton technology to more realistic settings where the user can naturally and seamlessly navigate through different terrain settings.
Article
This paper presents the design and validation of a backdrivable powered knee orthosis for partial assistance of lower-limb musculature, which aims to facilitate daily activities in individuals with musculoskeletal disorders. The actuator design is guided by design principles that prioritize backdrivability, output torque, and compactness. First, we show that increasing the motor diameter while reducing the gear ratio for a fixed output torque ultimately reduces the reflected inertia (and thus backdrive torque). We also identify a tradeoff with actuator torque density that can be addressed by improving the motor's thermal environment, motivating our design of a custom Brushless DC motor with encapsulated windings. Finally, by designing a 7:1 planetary gearset directly into the stator, the actuator has a high package factor that reduces size and weight. Benchtop tests verify that the custom actuator can produce at least 23.9 Nm peak torque and 12.78 Nm continuous torque, yet has less than 2.68 Nm backdrive torque during walking conditions. Able-bodied human subjects experiments (N=3) demonstrate reduced quadriceps activation with bilateral orthosis assistance during lifting-lowering, sit-to-stand, and stair climbing. The minimal transmission also produces negligible acoustic noise.
Article
In this article, we present the design of a powered knee–ankle prosthetic leg, which implements high-torque actuators with low-reduction transmissions. The transmission coupled with a high-torque and low-speed motor creates an actuator with low mechanical impedance and high backdrivability. This style of actuation presents several possible benefits over modern actuation styles in emerging robotic prosthetic legs, which include free-swinging knee motion, compliance with the ground, negligible unmodeled actuator dynamics, less acoustic noise, and power regeneration. Benchtop tests establish that both joints can be backdriven by small torques ( $\sim$ 1–3 N $\cdot$ m) and confirm the small reflected inertia. Impedance control tests prove that the intrinsic impedance and unmodeled dynamics of the actuator are sufficiently small to control joint impedance without torque feedback or lengthy tuning trials. Walking experiments validate performance under the designed loading conditions with minimal tuning. Finally, the regenerative abilities, low friction, and small reflected inertia of the presented actuators reduced power consumption and acoustic noise compared to state-of-the-art powered legs.
Article
Stroke may lead to considerable physical impairment and functional disability which affects walking ability. As a potential way to assist gaits, lower-limb exoskeletons have been developed. To supply appropriate assistive torque, real-time accurate recognition of current gait mode is important. In this paper, a hierarchical control system was proposed to recognize locomotion modes in real time for a unilateral knee exoskeleton on different terrains with specific assistive control strategies. Support vector machine classifier was used to recognize the locomotion mode based on two inertia measurement units. The corresponding assistive control strategy was designed according to the recognition result. Real-time recognition experiments under assistive torque control were conducted on five able-bodied subjects and one stroke patient, respectively. For the able-bodied subjects: (1) No significance was found on the total recognition accuracies whichever the leading leg was for the five subjects (p=0.057), which indicated the proposed method in this paper was suitable whichever the leading leg is as far as the overall classification accuracy was concerned. (2) Transitions occurred in swing phase when the leading leg was the paretic leg and transitions occurred in stance phase when the leading leg was the sound leg. No significance was found on mean delay time for the five subjects (p=0.785) whichever the leading leg was, which indicated the proposed method in this paper was suitable for these two leading legs as far as the mean delay time were concerned. (3) The method of generating the assistance based on the previous gait cycle time was demonstrated to be reasonable and the tracking performance of the torque could meet the requirement. For the stroke patient, similar experimental results were obtained.
Article
This paper focuses on assessing gait stability by metrics derived from dynamical systems theory to understand the influence of unilateral robot assistance on the human walking pattern. A motorized assistive robot is applied to the right knee joint to provide stance support. The metrics related to global stability (the maximum Floquet multiplier, max FM), local stability (short-term and long-term divergence exponents, $\lambda _{\text {s}}$ and $\lambda _{\text {l}}$ ), and variability (median absolute deviation, MAD) are considered. These metrics are derived for bilateral hip, knee, and ankle joint angles. Additionally, a biomechanical metric, the minimum margin of stability is assessed. Experiments are conducted on 11 healthy participants with different robot controllers. The max FM and $\lambda _{\text {s}}$ yield statistically significant results, showing that the unassisted (left) leg is more stable in right knee assistance conditions when compared to the normal walking condition due to inter-limb coordination. Moreover, MAD and $\lambda _{\text {l}}$ show that the variability and chaotic order of walking pattern during assistance are lower than those of normal walking. The proposed control strategy (automatic impedance tuning, AIT) improves local and orbital gait stability compared to existing controllers such as finite-state machine (FSM). The assessment of dynamic gait stability presented in this paper provides insights for further improving control strategies of assistive robots to help a user reach improved gait stability while maintaining appropriate variability.