ArticlePDF Available

Legless soft robots capable of rapid, continuous, and steered jumping

Springer Nature
Nature Communications
Authors:

Abstract and Figures

Jumping is an important locomotion function to extend navigation range, overcome obstacles, and adapt to unstructured environments. In that sense, continuous jumping and direction adjustability can be essential properties for terrestrial robots with multimodal locomotion. However, only few soft jumping robots can achieve rapid continuous jumping and controlled turning locomotion for obstacle crossing. Here, we present an electrohydrostatically driven tethered legless soft jumping robot capable of rapid, continuous, and steered jumping based on a soft electrohydrostatic bending actuator. This 1.1 g and 6.5 cm tethered soft jumping robot is able to achieve a jumping height of 7.68 body heights and a continuous forward jumping speed of 6.01 body lengths per second. Combining two actuator units, it can achieve rapid turning with a speed of 138.4° per second. The robots are also demonstrated to be capable of skipping across a multitude of obstacles. This work provides a foundation for the application of electrohydrostatic actuation in soft robots for agile and fast multimodal locomotion.
Content may be subject to copyright.
ARTICLE
Legless soft robots capable of rapid, continuous,
and steered jumping
Rui Chen 1,9, Zean Yuan 1,9, Jianglong Guo 2,9, Long Bai 1,9, Xinyu Zhu1, Fuqiang Liu 3,
Huayan Pu 4, Liming Xin 5, Yan Peng 6, Jun Luo1,4, Li Wen 7& Yu Sun 8
Jumping is an important locomotion function to extend navigation range, overcome obstacles,
and adapt to unstructured environments. In that sense, continuous jumping and direction
adjustability can be essential properties for terrestrial robots with multimodal locomotion.
However, only few soft jumping robots can achieve rapid continuous jumping and controlled
turning locomotion for obstacle crossing. Here, we present an electrohydrostatically driven
tethered legless soft jumping robot capable of rapid, continuous, and steered jumping based
on a soft electrohydrostatic bending actuator. This 1.1 g and 6.5 cm tethered soft jumping
robot is able to achieve a jumping height of 7.68 body heights and a continuous forward
jumping speed of 6.01 body lengths per second. Combining two actuator units, it can achieve
rapid turning with a speed of 138.4° per second. The robots are also demonstrated to be
capable of skipping across a multitude of obstacles. This work provides a foundation for the
application of electrohydrostatic actuation in soft robots for agile and fast multimodal
locomotion.
https://doi.org/10.1038/s41467-021-27265-w OPEN
1State Key Laboratory of Mechanical Transmissions, Chongqing University, Chongqing 400044, China. 2School of Science, Harbin Institute of Technology
(Shenzhen), Shenzhen 518055, China. 3College of Mechanical and Vehicle Engineering, Chongqing University, Chongqing 400044, China. 4School of
Mechatronics Engineering and Automation, Shanghai University, Shanghai 200444, China. 5School of Computer Engineering and Science, Shanghai
University, Shanghai 200444, China. 6Research Institute of Unmanned Surface Vessel Engineering, Shanghai University, Shanghai 200444, China. 7School
of Mechanical Engineering and Automation, Beihang University, Beijing 100191, China. 8Department of Mechanical and Industrial Engineering, University of
Toronto, Toronto, Canada.
9
These authors contributed equally: Rui Chen, Zean Yuan, Jianglong Guo, Long Bai. email: cr@cqu.edu.cn;
phygood_2001@shu.edu.cn
NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications 1
1234567890():,;
Content courtesy of Springer Nature, terms of use apply. Rights reserved
As an important locomotion function of terrestrial robots,
jumping or leaping is useful for them to effectively extend
their navigation range, overcome challenging obstacles,
and enhance their adaptability in unstructured environments13.
However, enhancing single-jump performance (jumping height
(JH) and jumping distance (JD)) of soft jumping robots to
improve their obstacle-crossing ability and accelerating jumping
frequency to increase their navigation efciency at the same time
are two grand engineering challenges. Researchers have developed
soft or partially soft jumping robots capable of forward naviga-
tions and driven by integrated springs417, shape memory alloys
(SMAs)1820, magnetic actuators21, light-powered actuators22,23,
dielectric elastomer actuators (DEAs)2426, pneumatic
actuators2729, chemical actuators3033, motors3436, and poly-
vinylidene diuoride (PVDF) actuators37. Some of them, which
are energy-storing jumping robots425, have excellent single-jump
performance, but always at the cost of navigation efciency due to
the necessity of the additional elastic energy-storage process. The
prolongation of the energy-storage process increases the JH but
decreases the landing stability and reduces the jumping fre-
quency. On the other hand, soft jumping robots actuated by
pneumatic actuators2729, chemical actuators3033, and
motors3436 have been demonstrated but have complicated
navigation strategies and structures. Lightweight soft hopping
robots based on DEAs26 and PVDF actuators37 can simply jump
by bending their body parts without additional energy-storing,
which can lead to rapid jumping frequencies, but their JHs and
JDs are not enough (<0.25 body height) to meet the requirements
of crossing obstacles.
Hydraulically amplied self-healing electrostatic (HASEL)3841
actuators, which can achieve linear motion by electro-
hydrostatically changing the distribution of internal liquids, have
been demonstrated to achieve remarkable continuous actuation
performance with actuation strains up to 118%, strain rates of
about 7500% s1, and a peak specic power of 156 W/kg. The
stacked quadrant donut HASELs presented by Mitchell et al.40
can achieve continuous vertical jumping by rapidly changing
internal liquid distribution with a JH of about 1.67 body heights.
This electrohydraulic actuation method, which can generate the
energy required for jumping in a very short time without the need
for a complicated energy-storing process, is a potential solution
for rapid obstacle-crossing robots. However, the diffusion-like
isotropic liquid ow of a donut HASEL actuator cannot generate
the energy of the forward jumping since the kinetic energy of the
liquid in all directions was canceled out, which caused the loss
and waste of energy. The jumping caused by the partial expansion
of the liquid pouch can only keep the partial actuator off the
ground and the backow of the dielectric liquid relies solely on
gravity, making it slow to return to the original state after landing
(Supplementary Fig. 1a). Therefore, it is still challenging for
HASEL jumpers to (1) achieve enhanced single-jump perfor-
mance without stacking, (2) achieve rapid restoration, and (3)
generate forward jumping and steered jumping.
Based on the exible electrical-driven liquid redistribution
method of HASEL series actuators3841, the actuator structure
was redesigned to make liquid ow anisotropically, utilize the
kinetic energy generated by liquid redistributions, and achieve
forward jumping. In addition, using saddle-shaped bending based
on an elastically deformable frame-membrane structure to
achieve locomotion is common in DEAs42, which inspired us to
use rapid bending and rebound based on electrohydrostatic
principle and frame to enhance the jumping performance of
actuators. In this work, we propose an electrohydrostatically
driven tethered legless soft jumping robot (LSJR) with rapid,
continuous, steered jumping and obstacle-crossing capabilities
based on a soft electrohydrostatic bending actuator (sEHBA). We
show that the characteristic of sEHBAs rapid response leads to a
short actuation time (~10 ms). The LSJR can be used to achieve a
JH of 7.68 body heights, a JD of 1.46 body lengths in a single
jump, and a continuous forward jumping speed of 390.5 mm/s
(6.01 body lengths per second) with a frequency of 4 Hz. We also
demonstrate that the integration of two LSJRs can readily achieve
rapid steered jumping. The turning speed of the dual-body LSJR
was able to reach 138.4°/s1, which is the fastest among existing
soft jumping robots. Furthermore, we show that LSJRs rapid
continuous jumping locomotion can cross various obstacles,
including slopes, wires, single steps, continuous steps, ring
obstacles, gravel mounds, and cubes of different shapes, some of
which are larger than the robot.
Results
Design concept and movement principle of the LSJR. In order
to use the anisotropic liquid ow to achieve forward jumping
caused by unbalanced energy, we heat-sealed a HASEL like
actuator into a semicircular separated HASEL (SCS-HASEL)
actuator composed of two semicircular liquid pouches based on
the zipping mechanism43 and it showed better jumping perfor-
mance (Supplementary Fig. 1b). Then, the dielectric liquid in the
rear semicircular pouch of the SCS-HASEL actuator was replaced
with an equal volume of air and the covered electrodes of the rear
semicircular pouch were removed so that the dielectric liquid can
ow anisotropically relative to the entire actuator. As expected, it
can be found that the special liquidair layout can make the
liquidair actuator jump forward even though the air pouch
dragged on the ground (Supplementary Fig. 1c). This was because
the electrodes squeezed the liquid dielectric to make it ow for-
ward quickly, thus giving it the initial kinetic energy that can be
used to provide forward kinetic energy to the liquidair actuator.
The detailed experimental results can be seen in Supplementary
Movie 1.
By observing the three types of actuators above (Supplemen-
tary Fig. 1), it can be found that their jumps were all caused by the
partial expansion of the liquid pouch. This kind of jump was
unstable and unreliable, and the JH was insufcient. Furthermore,
the edges of the SCS-HASEL actuator and liquidair actuator
were unrestricted so that every jump had randomness and the
initial state cannot be completely restored. Fixing a predeformed
frame on the edge is a well-known restriction method in DEAs42
and has not been used for HASEL series actuators. Combining the
frame to transform the linear motion of the electrohydraulic
actuators into a saddle-shaped bending motion provided the
possibility of an efcient type of jumping method. Therefore, the
non-prebending ring frame and the prebending ring frame were
combined to the SCS-HASEL actuator and the liquidair
actuator, and then their single-jump performance was tested,
respectively (Supplementary Fig. 2). The left half of Supplemen-
tary Fig. 2 shows the moment (11.76 ms after applying voltage)
the actuators took off from the ground. The prebending frames
xed on the edges of these actuators can guide the direction of
deformation, making the deformation of the actuators with
prebending frames more regular and closer to the saddle shape
than the actuators with plane frames and achieve better jumping
performance. The liquidair actuator with a prebending frame,
which was also the sEHBA we needed, can jump higher and
further, and it can return to its original state immediately after
landing. Clearly, using the frame with more regular deformation
to touch the ground and generate jumping energy is a more stable
and reliable jumping mechanism than relying on the partial
expansion of the liquid pouch (Supplementary Movie 1).
In this paper, we propose an electrohydrostatically driven LSJR
(Fig. 1) that has rapid, continuous, steered jumping and obstacle-
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w
2NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
crossing capabilities based on the sEHBA. The special liquidair
layout and the semicircular zipping structure can be used to make
the internal liquid ow anisotropically and rapidly to generate a
great amount of forward kinetic energy, which is offset and
wasted in HASEL actuators. Meanwhile, the prebending frame
xed on the edge of the sEHBA can be used to guide the
deformation direction to achieve fast bending locomotion which
can be used to generate vertical and forward kinetic energy. The
special liquidair layout and prebending frame structure greatly
enhance the jumping performance of electrohydraulic actuators
and enable rapid, continuous, and steered jumping without
stacking. The LSJR can jump simply by quickly liquid owing and
body bending, thereby greatly shortening the propulsive interval
(~10 ms). The stored elastic energy associated with body bending
Fig. 1 LSJR detailed design and motion principle. a The LSJR consists of two plastic semicircular pouches printed with exible electrodes. The front pouch
is lled with a dielectric liquid, and the rear is lled with air with the same volume. A exible plastic ring frame is xed on the edge and is prestrained. Note
that the rear air pouch functions to ensure that the pre-curved frame is consistent and maintains structural balance during the ight. bThe LSJR prototype
(1.1 g). Scale bar, 1 cm. cSchematic diagram of the LSJR jumping process. By the application of a high voltage to the two electrodes, the LSJR is energized to
bend itself to generate forces and energy for forward jumping. During the voltage application, Maxwell stress squeezes the dielectric liquid and makes it
ow laterally into the portion of the front pouch that is not covered by the electrodes (from the liquid outow area to the liquid inow area). dCross-
sectional views (ee and ff) of the LSJR: ee denotes the deformation of the front pouch, whereas ff shows the ee deformation-driven whole-body
bending and jumping. eSnapshots of the LSJR jumping, where 10 kV is applied to the actuator. Scale bar, 2 cm.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w ARTICLE
NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved
can help the robot quickly restore its original shape to avoid
affecting the next jump. The detailed design iterations are
demonstrated in Supplementary Fig. 3.
The LSJR consists of two exible plastic semicircular pouches
printed with exible electrodes that were connected with two
conductive tapes for potential electric wire connections (Fig. 1a).
The two pouches were made of biaxially oriented polypropylene
(BOPP) lms. The front pouch was lled with a dielectric liquid
and the rear was lled with the same volume of air. A exible
plastic polyvinyl chloride (PVC) ring frame was xed on the edge
and prestrained. Note that the rear air pouch, which is similar to
the tail of animals and is used to maintain the balance of the
jumping and landing posture, played an important role in the
whole structure of the LSJR. To further enhance the jumping
performance of LSJR, the air in this pouch can be replaced with
helium or other less dense and non-explosive gas. The prototype
of an LSJR (1.1 g) is shown in Fig. 1b. The detailed fabrication
process can be seen in Methods.
The LSJR is energized to bend itself to generate forces and
energy for forward jumping. After the application of a high
voltage to the electrodes, Maxwell forces attract the electrodes,
squeezing the dielectric liquid from the outow area to the inow
area with no electrode coverage (Fig. 1c). The rapid and
anisotropic ow can generate a horizontal initial kinetic energy.
The increased electrostatic force between the electrodes of the
front pouch causes a rapid liquid ow, which increases the
thickness of the cross-section F-F and decreases its lateral length
from X
1
to X
2
, as the pouch is inextensible. This deformation,
plus the pre-strain of the ring frame to keep the bottom electrode
from touching the substrate and facilitate the further bending of
the frame during actuation (cross-section G-G), pulls the front
and rear ends closer to each other (Fig. 1d). The instantaneous
partial deformation of the frame results in instant bending of the
overall frame, which propels the robot body into the air.
Snapshots of a rapid (~10 ms) take-off process is presented in
Fig. 1e. After take-off, the initial horizontal velocity of LSJR is
determined by the horizontal ground reaction forces at the frame
ends. The forces are caused by the moving dielectric liquid ow
and frame bending. The initial vertical velocity of the LSJR is
determined by the vertical ground reaction forces at the frame
ends, which are caused by the frame bending. During the leaping
state, the ring frame quickly releases its elastic energy, and the
dielectric liquid ows back, restoring the robot to the original
state in preparation for the next jump after landing. Note that the
low-prole robot design makes both the jumping and landing
stable with no capsizing. The detailed theoretical analysis of the
locomotion mechanism is demonstrated in Supplementary
Note 1.
Single-jump characterization. JD and JH are two important
performance measures that can be used to characterize the
jumping performance of the LSJR. If the same robot material and
size are maintained, the electrode area/nonelectrode area ratio,
magnitude of the voltage application, the mass of the load, and
the ring frame prebending level are important parameters inu-
encing the jumping performance of the LSJR. Figure 2a shows the
untethered single-jump process and related parameters. The two
aluminum electrodes connected to the tail of the LSJR were freely
placed on two copper electrodes (top left inset in Fig. 2a), aiming
to eliminate the inuence of electric wires on the jumping per-
formance. Note that (1) three LSJRs with the same parameters
were fabricated, and the differences in the JD and JH results were
all within 10%, and (2) for each jumping experiment, 10 repeated
tests were performed in the same laboratory environment, and
average and one standard deviation values of each 10 results were
reported. Furthermore, applying a voltage of the same polarity
would cause the charge to be retained and to accumulate inside
the actuators, which would prevent the actuators from fully
returning to their initial position, thus affecting the results of the
next experiment. Therefore, the polarity would be reversed and a
waiting time of 60 s would be used to alleviate charge retention
after each experiment (top right inset in Fig. 2a).
We dene r=electrode area:nonelectrode area (of the front
pouch) and fabricated three robots with r=2:1, 1:1, and 1:2.
Figure 2b, c shows that the robot with r=1:1 produced a larger
JD and JH. At a low voltage (0~3 kV), the robot deformed slowly,
and the deformation force was not large enough for jumping.
Average values of JD =95.0 mm (1.46 body lengths) and
JH =30.7 mm (7.68 body heights) were achieved when the
applied voltage was 10 kV and r=1:1. Carrying a 1 g load (0.91
body weight) decreased the JD (56.0 mm) and JH (20.0 mm) to
59% and 65% of the no-load condition, respectively, bringing a
jumping performance reduction of 38%. Carrying a 2 g load (1.82
body weight) decreased the JD (33.7 mm) and JH (8.1 mm) to 35
and 26% of the no-load condition, resulting in a jumping
performance reduction of 70%, which was nearly twice that of a
1 g load condition. We used different body heights (BH) to
represent different prebending levels and fabricated three robots
with BH =2, 4, and 6 mm. Figure 2d, e shows that the robot with
BH =4 mm produced a larger JD and JH. In addition, the LSJR
with 2 mm BH had a greater average relative jumping height
(RJH) of 9.4 when 10 kV was applied, whereas the average RJHs
of the LSJR (BH =4 mm) and the LSJR (BH =6 mm) were 7.7
and 4.2, respectively. Comparing the experiment results (Fig. 2b,
c) of the LSJR (r=2:1, load =0 g), the LSJR (r=1:1, load =0 g),
and the LSJR (r=1:2, load =0 g), an apparent difference in their
JHs but a very small difference in JDs can be found. According to
Supplementary Note 1, the amount of dielectric liquid ow had a
much greater impact on the initial horizontal velocity v
x
than on
the initial vertical velocity v
y
. The bigger ratio r, which can affect
the size of the electrode coverage area, led to a bigger volume
ΔV
oil
of dielectric liquid ow, a faster velocity v
oil
of dielectric
liquid ow, and the faster bending rate of the ring frame. It
caused a bigger horizontal initial kinetic energy of the moving
dielectric liquid ow and greater vertical ground reaction forces,
leading to a bigger JD and JH. However, the excessive r(e.g.
r=2:1) not only affected the exibility of BOPP lms and
hindered the normal bending of the frame, reducing the vertical
ground reaction forces, but also led to a smaller ΔV
oil
and a lower
v
oil
, reducing the horizontal initial kinetic energy of the moving
dielectric liquid ow. Therefore, the ratio (r=1:1) was the most
appropriate in this experiment.
Continuous jumping on different substrates. Rapid continuous
forward jumping is a useful locomotion capability. Continuous
forward jumping speed (CFJS) is an important performance
feature characterizing a continuous forward jumping robot. If the
same robot is used, the substrate surface roughness, actuation
frequency, and magnitude of the applied voltage are important
parameters inuencing the jumping performance of the LSJR.
Sufcient substrate surface roughness prevents the robot from
slipping during continuous motions, and an appropriate actua-
tion frequency enables a quicker continuous movement. Note
that (1) for each jumping experiment, 10 repeated tests were
performed in the same laboratory environment, and the obtained
results were reported as the average and one standard deviation,
and (2) the polarity was reversed to alleviate charge retention.
It is shown in Fig. 1e that applying 10 kV to the robot for 10 ms
or more was a necessary condition for a successful jump. This
resulted in a jumping time of almost 250 ms (Supplementary
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w
4NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Movie 2) and a typical actuation frequency of 4 Hz. Therefore, we
set the power-on time to 10 ms and selected a series of test
frequencies, i.e., 8, 4, 2, 1, 0.5, and 0.25 Hz, to test the inuence of
actuation frequency. In the experiments, the LSJR moved for four
cycles at each frequency, and CFJS was calculated. In Supple-
mentary Movie 3, we demonstrated continuous forward jumping
locomotion of the soft robot on four substrates, i.e. a glass plate, a
paper plate, a PVC plate, and a wood plate, with different surface
roughness (Supplementary Fig. 10). Figure 3ad shows that at
4 Hz and 10 kV, the robot achieved a greater average CFJS =
390.5 mm/s (6.01 body lengths per second) on the wood
substrate, whereas an average CFJS =95.6 mm/s (1.47 body
lengths per second) was achieved on the glass plate because it was
difcult for the smooth glass plate to provide sufcient friction.
Fig. 2 Single-jump characterization results. See also Supplementary Movie 2. aUntethered single-jump process and parameters. The top left inset shows
the electrical connections. The top right inset is the voltage application strategy in the experiment. bThe relationship between JD and applied voltage under
different loads (0, 1, and 2 g) and different electrode area/nonelectrode area ratios (2:1, 1:1, and 1:2). cThe relationship between JH and applied voltage
under different loads (0, 1, and 2 g) and different electrode area/nonelectrode area ratios (2:1, 1:1, and 1:2). dThe relationship between JD and applied
voltage at different body heights (2, 4, and 6 mm). eThe relationship between JH, RJH, and applied voltage at different body heights (2, 4, and 6 mm).
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w ARTICLE
NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Fig. 3 Continuous jumping on different substrates. See also Supplementary Movie 3. aThe relationship between CFJS and actuation frequency on the
glass plate, where the slowest average CFJS =95.6 mm/s (1.47 body lengths per second) was obtained at 4 Hz and 10 kV. bThe relationship between CFJS
and actuation frequency on the paper plate. cThe relationship between CFJS and actuation frequency on the PVC plate. dThe relationship between CFJS
and actuation frequency on the wood plate, where the fastest average CFJS =390.5 mm/s (6.01 body lengths per second) was obtained at 4 Hz and 10 kV.
eThe voltage application strategy and the corresponding robot motion states. fComposite image of the initial position and four landing points in
continuous jumping on the PVC plate when CFJS =250.1 mm/s at 4 Hz and 10 kV. The angle deviation of each jump was less than 8°. Scale bar, 5 cm.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w
6NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
As the actuation frequency increased, the CFJS increased. At a
frequency higher than 8 Hz, after the LSJR took off and left the
substrate, another actuation cycle started while the robot was still
in the air. Each actuation cycle should only occur after the LSJR
had landed smoothly; otherwise, the next jumping performance is
impaired. Figure 3e shows the constant voltage application
waveform and the corresponding robot states. Applying a voltage
of the same polarity would cause the charge to be retained and to
accumulate inside the actuators, which would prevent the
actuators from fully returning to their initial position. Reversing
the polarity mitigated charge retention within the actuator during
continued cycling. The voltage application period was required to
be long enough to prevent the LSJR from entering the next
actuation cycle before landing. Therefore, changing the voltage
waveform in real time through visual recognition and other
means and making the LSJR enter the next actuation cycle
immediately after landing may provide solutions to the random-
ness of ight time and further increase the CFJS. Figure 3f
demonstrates the initial position and four landing points during
continuous forward jumping on the PVC plate with a speed of
250.1 mm/s (3.85 body lengths per second). The angle deviation
of each jump was less than 8°, which means that the robot can
achieve a reasonably good straight-line movement. More details
on the LSJRs motion precision can be seen in the Supplementary
Note 2 and Supplementary Movie 8.
Steered jumping of a dual-body LSJR. Steered jumping is a
useful function for animal locomotion in unstructured and
complex terrains. It is desirable for soft jumping robots to
replicate this ability. Two LSJRs were connected abreast, resulting
in a dual-body LSJR (Fig. 4a, b) with the ability to adjust its
locomotion direction42. Applying a voltage (V
1
) to one unit of the
LSJRs causes the unit to jump: it deforms, generates forward
kinetic energy, and bears ground friction forces. The center of
gravity of the dual-body LSJR is not in the direction of the initial
speed and friction forces, resulting in turning behavior (each cycle
achieves a turning angle of α, as shown in Fig. 4a). Selectively
controlling V
1
on the left and right LSJRs results in steered
jumping. To illustrate the turning performance of the dual-body
LSJR, it was made to turn at least 60° on each of the four sub-
strates. Figure 4c shows that the robot on the wood plate achieved
a greater turning speed (TS). An average TS =138.4°/s was
achieved at 10 kV and 4 Hz, which, to the best of the authors
knowledge, is the fastest among existing soft jumping robots. An
average TS of only 27.9°/s was achieved on the glass plate at 10 kV
and 4 Hz because the surface was smooth. Figure 4d demonstrates
Fig. 4 Turning results of the dual-body LSJR. See also Supplementary Movie 4. aSchematic diagram of the dual-body LSJR turning process, which
consists of the rest state, the turning state, and the landing state. Over each voltage cycle, the robot turns by an angle of α.bCenter of gravity of the dual-
body LSJR. cThe relationship between TS and applied voltage on the four substrates. dComposite image of the initial and nal positions during a
continuous turning procedure on the PVC plate with a speed of 65.0°/s. The robot took 1.23 s to turn 80° at 10 kV and 4 Hz. Scale bar, 5 cm.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w ARTICLE
NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved
the initial and nal positions during a continuous turning pro-
cedure on the PVC plate with a speed of 65.0°/s at 10 kV and
4 Hz. Sufcient substrate surface roughness not only prevents the
robot from slipping during continuous motions but also hinders
the movement of the unpowered LSJR and thus the turning
behavior. Therefore, a potential solution is to apply a large V
1
to
one LSJR and a small V
0
to the other LSJR. This will lead to a
greater jump by one unit, thus resulting in turning behavior, and
will be considered in our future work. More details on the LSJRs
motion precision can be seen in Supplementary Note 2 and
Supplementary Movie 8.
Obstacle-crossing ability of the LSJR. A main function of a
jumping robot is obstacle avoidance so that it can conduct
explorations, inspections, and reconnaissance tasks in complex
and unstructured environments. Both the single-body and dual-
body LSJRs can be used to achieve decent obstacle-crossing
capability, as shown in Figs. 5and 6(Supplementary Movies 5
and 6). Under an applied voltage of 10 kV and an actuation
frequency of 4 Hz, the single-body LSJR was found to (1) climb
on the glass plate (tilt angle of 3°) with a CFJS of 16.3 mm/s (0.25
body lengths per second), (2) jump cross an electric wire (dia-
meter of 6.3 mm), (3) jump across a square step (8 mm high), and
(4) jump across continuous steps (consisting of the square step
and a 5 mm high round step). In crossing tests with an obstacle
height interval of 4 mm, the maximum height that the LSJR can
cross was 14 mm for cuboids, and 18 mm for triangular prisms
and cylinders, as shown in Supplementary Fig. 11 (Supplementary
Movie 7). Affected by the leaping posture and wires, the
Fig. 5 Single-unit LSJR obstacle crossing at 10 kV and 4 Hz. See also Supplementary Movie 5. aClimbing on the glass plate (tilt angle of 3°) with a CFJS
of 16.3 mm/s. bCrossing an electric wire (diameter of 6.3 mm). cJumping across a square step (height of 8 mm). dJumping across continuous steps
(heights of 8 and 5 mm). Scale bar, 5 cm. eComposite image of the LSJRs locomotion on a gravel mound (gravel size: 36 mm). Scale bar, 2 cm.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w
8NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
maximum obstacle height that the LSJR can cross was less than
the max JH. In order to further prove the robots ability to
overcome obstacles in realistic unstructured environments, we
made a gravel mound with many gravels (size: 3 to 6 mm), and
tested LSJRs locomotion on the gravel mound (Fig. 5e). The
steered jumping (simultaneous jumping and turning) capability
of the dual-body LSJR made it (1) jump across a 5 mm high
round step and (2) jump across a ring object with a height of
8 mm, an inner diameter of 77 mm, and an outer diameter of
83 mm. It was challenging for the single-body LSJR to avoid the
ring obstacle. The collaborative movement of two LSJR units
results in much greater exibility and improved obstacle-crossing
ability for soft jumping robots, thus making the robot more
adaptable to complex terrains.
Discussion
We proposed an electrohydrostatically driven, low-prole (0.85-
mm-thick), lightweight (1.1 g), modular, and cost-effective teth-
ered LSJR based on a sEHBA. The robot features capabilities of
rapid, continuous, and steered jumping, load-carrying, and
obstacle-crossing through a simple control strategy. The
inspiration for the design of the sEHBA stemmed from the
electrohydrostatic jumping of HASEL actuators40 and the peri-
odic saddle-shaped bending caused by the predeformed frame of
DEAs42. In the design process, a mechanical analysis model and a
dielectric liquids center of gravity moving equivalent model were
built to guide the optimization of size parameters.
Most existing soft jumping robots (Table 1)havealargeJHbut
require a long actuation/energy-storage time and righting time, which
leads to a slow CFJS and lack of exibility. A few soft jumping
robots26,37 have a high actuation frequency and high speed mobility,
but they can only achieve a small JH (<0.25 body height), which is
not suitable for overcoming challenging obstacles. Our LSJR used the
special liquidair layout and the edge-xing prebending frame to
achieve rapid continuous forward and steered jumping locomotion
caused by periodic saddle-shaped bending and anisotropic liquid
ow, which made up for the limitations of HASEL actuators40,
including (1) unachievable forward and steered jumping, (2) weak
single-jump performance without stacking, and (3) incapability for
rapid restoration. As a result, the LSJR had a short actuation time
(~10 ms) and was able to achieve a JH of 7.68 body heights, a JD of
1.46 body lengths in a single jump (Fig. 2and Supplementary
Movie 2), and a CFJS of 390.5 mm/s (6.01 body lengths per second)
with a frequency of 4 Hz (Fig. 3and Supplementary Movie 3). The
angle deviation of each jump can be controlled within 8° in con-
tinuous forward jumping locomotion. The integration of two LSJRs
was able to readily achieve rapid steered jumping (Fig. 4and Sup-
plementaryMovie4).TheTSofthedual-bodyLSJRwasableto
reach 138.4°/s, which is the fastest among existing soft jumping
robots. Experiments also veried that LSJRs rapid continuous
jumping locomotion could be applied to cross many obstacles,
Fig. 6 Dual-body LSJR obstacle crossing. See also Supplementary Movie 6. aStraight jumping across a round step (height of 5 mm). bSteered jumping
across the round step (height of 5 mm). cJumping across a ring obstacle (height of 8 mm, inner diameter of 77 mm, and outer diameter of 83 mm). Scale
bars, 10 cm.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w ARTICLE
NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Table 1 Comparison between this work and some soft forward jumping robots.
Soft
jumping robots
Energy-storing
jumping
Actuation
methods
Weight (g) Jumping distance Jumping height Propulsive
interval time (s)
Landing
stability
Straight line
jumping
capability
Direction-
adjusting
capability
Unit cost
Kovač(2013)7Yes Spring 14.00 2.55 BD 3.44 BD 5.00 Medium Low Directional
jumping (
directions)
High
Zhakypov
(2019)18
Yes SMA 9.70 3.97 BL 2.50 BH >24.00 High High No High
Huang (2018)19 Yes SMA 3.00 2.00 BL 1.00 BH 3.00 High High No High
Hu (2018)21 Yes Magnetic N/A 1.63 BL 2.44 BL >10.00 Low Medium Steered jumping
(15.0°/s)
N/A
Ahn (2019)22 Yes Light N/A 8.00 BL 5.00 BH 100.00 High High No N/A
Hu (2017)23 Yes Light N/A N/A 5.00 BH 10.56 High High No N/A
Duduta (2020)25 Yes DEA 0.90 1.34 BL 1.16 BL 6.00 High High No Low
Zhao (2019)26 No DEA 6.50 0.29 BL <0.25 BH 0.03 High High No Low
Ni (2015)27 No Pneumatic N/A 0.64 BL 0.64 BH 0.75 High High No
High
Liu (2020)29 No Pneumatic 0.45 0.90 BL 0.80 BL 0.28 High High No
Low
Tolley (2014)30 No Chemical 510.00 7.50 BH 7.50 BH 0.03 Low Low Directional
jumping (3
directions)
High
Loepfe (2015)31 No Chemical 2100.00 2.78 BD 1.11 BD 4.50 Medium Low No High
Bartlett (2015)32 No Chemical Tethered: 478.60 Tethered: N/A Tethered: 2.35 m 2.45 High Low Directional
jumping (3
directions)
High
Untethered: 964.60 Untethered:
0.50 BL
Untethered:
6.00 BH
Churaman
(2011)33
No Chemical 0.314 21.78 BL 80 mm Low Low No Low
Li (2017)34 No Motor 250.00 2.57 BL 1.00 BH 10.00 Medium Low Steered jumping
(0.6°/s)
High
Mintchev
(2018)36
No Motor 37.00 3.35 BD 2.86 BD 3.00 Medium Low No High
Wu (2019)37 No PVDF 0.06 0.11 BL <0.25 BH <0.01 High High Steered jumping
(0.8°/s)
Low
This work No sEHBA 1.10 1.46 BL 7.68 BH 0.01 High High Single-body: No Low
Dual-body: steered
jumping (138.4°/s)
Notes: BL body length, BH body height, BD body diameter, N/A not available. Landing stability, straight line jumping, and unit cost were evaluated in three (high, medium, and low) levels. The following are the detailed level judgment criteria:
Landing stability:
(1) High. After the robot lands, it does not roll and does not need articial/self-righting before the next jumping.
(2) Medium. The robot is capable of self-righting. After landing, it rolls a distance and sometimes needs self-righting based on its posture.
(3) Low. The robot is not capable of self-righting. After landing, it rolls a distance and sometimes needs articial-righting based on its posture.
Straight line jumping capability:
(1) High. There is no need for articial direction adjustment in continuous forward jumping process. Connecting several continuous landing points as a line, it is basically a straight line.
(2) Medium. The robot needs articial direction adjustment in continuous forward jumping process. Connecting several continuous landing points as a line, it is basically a straight line.
(3) Low. The landing points have some randomness in continuous jumping process. Connecting several continuous landing points as a line, it is a curve.
Unit cost:
We estimate the unit price based on the main material and component price (from a Chinese e-commerce website) of the robot.
(1) High. The unit price exceeds 10 RMB.
(2) Low. The unit price does bot exceeds 10 RMB.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w
10 NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
including slopes, wires, single steps, continuous steps, ring obstacles,
gravel mounds, and cubes of different shapes. The maximum height
of the robot can reach up to 18 mm (Figs. 5and 6, Supplementary
Movie 5, Supplementary Fig. 11 and Supplementary Movies 6 and 7).
The jumping performances of the LSJR rely on not only the
applied voltages but also the surface textures of various substrates
(Supplementary Fig. 10). Under the same applied voltage (10 kV,
4 Hz), the glass substrate with the smoothest surface provided the
lowest friction among all substrates, leading to a lower CFJS of
95.6 mm/s (1.47 body lengths per second) and a lower TS of
27.9°/s. This currently limits the application of the robot to
jumping on relatively smooth surfaces but can be mitigated if
electroadhesion is applied by adding additional passive electrodes
to the rear pouch.
The LSJR can be applied to detect and record environmental
changes such as temperature and ultra-violet light by attaching a
light and soft temperature sensor/paste and photochromic dyes
(Supplementary Fig. 15 and Supplementary Movie 9). Through
integrating other sensors, it is expected to detect more environ-
mental factors, such as pollutants in industrial environments and
civil buildings. In addition, future work may also include (1)
study of the scalability and parametric optimization of the sEHBA
to achieve better jumping performance, (2) development of an
untethered LSJR and its applications, and (3) investigation of
sEHBAs other soft robotics applications such as wall-climbing
robots, swimming robots, and apping wing robots.
Methods
Materials of the LSJR. The inextensible pouch shell was made by heat-sealing two
16-μm-thick BOPP lms (Jiazhixing Co., China) and injected with air and a
dielectric liquid made of 25# mineral transformer oil (Aokelai Lubricants Co.,
China). BOPP has reasonably good dielectric breakdown strength (~700 V µm1)
and tensile strength (~300 N mm2)44. The dielectric liquid has favorable dielectric
properties and low viscosity45. The electrodes were screen printed on the BOPP
lms using LN-GCI-3 graphene conductive inks (Jining Leadernano Tech., China).
The plastic ring frame was fabricated by laser cutting a 0.5-mm-thick PVC sheet
(Lizhiyuan Plastic Industry, China), which has good mechanical properties and low
density46. All materials are low in cost and easy to procure.
Fabrication procedure of the LSJR. The fabrication procedure of an LSJR involves
nine major steps (Supplementary Fig. 4). First, we screen printed conductive inks
onto BOPP lms using a screen-printing machine (Mingming Screen-Printing
Equipment Co., China) (Supplementary Fig. 5). The electrodes were cured at room
temperature for 12 h, resulting in a exible graphene electrode with a thickness of
20 µm. Second, we stacked two BOPP-electrode composite lms with the electrodes
facing outwards and put them on a Teon high-temperature cloth as a load-
dispersing layer. Third, the soldering iron was set to 200 °C to heat-seal the BOPP
lms, creating two semicircular pouches (both 55 mm in diameter) and leaving two
ll ports in the seal of each pouch. Fourth, we lled the front pouch with the
dielectric liquid of 1 mL and the rear pouch with the same volume of air by using
two syringes. Fifth, we squeezed the air bubbles out of the front pouch and heat-
sealed the two ll ports to prevent uid compression. Sixth, we lifted the four
corners of the top BOPP lm in sequence and placed the laser-cut exible PVC
ring frame (thickness of 0.5 mm, inner diameter of 58 mm, and outer diameter of
62 mm) between the two BOPP lms. Seventh, we xed the edge position of the
three-layer composite membrane and moved the soldering iron on the lms to
deform the ring frame and heat-seal it following the rebound trajectory. In this
step, the prebending levels of the nal LSJR can be controlled by moving different
distances (d). Eighth, we heat-sealed along the inside and outside circles of the
predeformed frame and removed the pushpins to rebound the frame to generate
prebending. Finally, excess BOPP lm was cut away, and two aluminum tapes were
attached to the ends of electrodes to create a completed LSJR with a total body
weight of 1.1 g and a full body length of 65 mm.
Experimental control strategy of the LSJR. After electrical connections were
made to the LSJR, applying a high voltage to the LSJR for 10 ms enabled it to
complete a jump. In the single-jump tests, after it took-off, the aluminum pieces at
the tail detached from the copper electrodes xed on the ground, thereby cutting
off the voltage. The voltage waveform was a square wave and the duty cycle was
1:6001, as shown in the top right inset of Fig. 2a. Voltage reversal and a wait time of
60 s were conducted to alleviate the residual charge issue. Different voltages caused
different jumping distances and JHs. In the continuous locomotion tests, the wires
were used for electrical connection instead of using a separable aluminumcopper
contact structure. The voltage waveform was a square wave and the duty cycle was
adjustable, as shown in Fig. 3e. The power-on time of the robot in each cycle was
10 ms. Different voltages and frequencies caused different continuous forward
jumping speeds. In the direction-adjusting tests, two LSJRs were abreast connected
and the voltage was selectively applied to one of them. The electrical connections
were made the same as in the continuous locomotion tests. The voltage waveform
was a square wave and the frequency was 4 Hz. The power-on time of the robot in
each cycle was 10 ms. Different voltages caused different turning speeds. In the
obstacle-crossing tests, the electrical connections were the same as those in the
continuous locomotion tests and direction-adjusting tests. The voltage waveform
was a square wave and the frequency was 4 Hz. The voltage was 10 kV and the
power-on time of the robot in each cycle is 10 ms.
Experimental set up of the LSJR. For single-jump tests, a high-speed camera
(2F01, Revealer, China) was used to capture the motions. For other low-speed tests,
a Nikon D3400 DSLR camera and an Apple iPhone XR were used to record the
movements. For substrate surface morphology characterization, a 3D Laser Scan-
ning Confocal Microscope (OLS4000, Olympus Corp., Japan) was used to measure
the areal surface texture. Adobe Premiere (version 7.0.0) and Adobe Photoshop
(version 19.1.4) were used to process the videos and to obtain the experimental
data. An EMCO high-voltage amplier (E101CT, EMCO High Voltage Co., USA)
was used with two high-voltage relays (CRSTHV-14KV-A, CRST, China), con-
trolled by a DMAVR-128 board (Ningbo Xinchuang Electronic Technology Co.,
China), to energize the LSJR.
Data availability
The authors declare that data supporting the ndings of this study are available within
the paper and its Supplementary Information les.
Received: 23 June 2021; Accepted: 10 November 2021;
References
1. Armour, R., Paskins, K., Bowyer, A., Vincent, J. & Megill, W. Jumping robots:
a biomimetic solution to locomotion across rough terrain. Bioinspir. Biomim.
2, S65 (2007).
2. Zhang, Z., Zhao, J., Chen, H. & Chen, D. A survey of bioinspired jumping
robot: takeoff, air posture adjustment, and landing buffer. Appl. Bionics
Biomech. 2017,122 (2017).
3. Zhang, C., Zou, W., Ma, L. & Wang, Z. Biologically inspired jumping robots: a
comprehensive review. Robot. Auton. Syst. 124, 103362 (2020).
4. Kovač, M., Fuchs, M., Guignard, A., Zufferey, J. & Floreano, D. A miniature 7
g jumping robot. In IEEE International Conference on Robotics and
Automation 373378 (IEEE, 2008).
5. Kovač, M. The bioinspiration design paradigm: a perspective for soft robotics.
Soft Robotics 1,2837 (2014).
6. Kovač, M., Schlegel, M., Zufferey, J. & Floreano, D. A miniature jumping robot
with self-recovery capabilities. In IEEE International Conference on Intelligent
Robots and Systems. 583588 (IEEE, 2009).
7. Kovač, M., Schlegel, M., Zufferey, J. & Floreano, D. Steerable miniature
jumping robot. Auton. Robot. 28, 295306 (2010).
8. Stoeter, S. A., Rybski, P. E., Gini, M. & Papanikolopoulos, N. Autonomous
stair-hopping with scout robots. In IEEE International Conference on
Intelligent Robots and Systems 721726 (IEEE, 2002).
9. Beck, A. et al. Jump stabilization and landing control by wing-spreading of a
locust-inspired jumper. Bioinspir. Biomim. 12, 066006 (2017).
10. Zaitsev, V. et al. A locust-inspired miniature jumping robot. Bioinspir.
Biomim. 10, 066012 (2015).
11. Ho, T. & Lee, S. A novel design of a robot that can jump and roll with a single
actuator. In IEEE International Conference on Intelligent Robots and Systems
908913 (IEEE, 2012).
12. Lambrecht, B. G. A., Horchler, A. D. & Quinn, R. D. A small, insect-inspired
robot that runs and jumps. In IEEE International Conference on Robotics and
Automation 12401245 (IEEE, 2005).
13. Haldane, D. W., Plecnik, M. M., Yim, J. K. & Fearing, R. S. Robotic vertical
jumping agility via series-elastic power modulation. Sci. Robot. 1, eaag2048
(2016).
14. Zhao, J. et al. MSU jumper: a single-motor-actuated miniature steerable
jumping robot. IEEE Trans. Robotics 29, 602614 (2013).
15. Yang, K. et al. A water-walking robot mimicking the jumping abilities of water
striders. Bioinspir. Biomim. 11, 066002 (2016).
16. Jung, G. & Cho, K. Froghopper-inspired direction-changing concept for
miniature jumping robots. Bioinspir. Biomim. 11, 056015 (2016).
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w ARTICLE
NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications 11
Content courtesy of Springer Nature, terms of use apply. Rights reserved
17. Jung, G., Choi, H. & Cho, K. The effect of leg compliance in multi-directional
jumping of a ea-inspired mechanism. Bioinspir. Biomim. 12, 026006 (2017).
18. Zhakypov, Z., Mori, K., Hosoda, K. & Paik, J. Designing minimal and scalable
insect-inspired multi-locomotion millirobots. Nature 571, 381386 (2019).
19. Huang, X. et al. Chasing biomimetic locomotion speeds: creating untethered
soft robots with shape memory alloy actuators. Sci. Robot. 3, eaau7557 (2018).
20. Huang, X. et al. Highly dynamic shape memory alloy actuator for fast moving
soft robots. Adv. Mater. Technol. 4, 1800540 (2019).
21. Hu, W., Lum, G. Z., Mastrangeli, M. & Sitti, M. Small-scale soft-bodied robot
with multimodal locomotion. Nature 554,8185 (2018).
22. Ahn, C., Liang, X. & Cai, S. Bioinspired design of light-powered crawling,
squeezing, and jumping untethered soft robot. Adv. Mater. Technol. 4,
1900185 (2019).
23. Hu, Y. et al. Electrically and sunlightdriven actuator with versatile biomimetic
motions based on rolled carbon nanotube bilayer composite. Adv. Funct.
Mater. 27, 1704388 (2017).
24. Duduta, M. et al. Electrically-latched compliant jumping mechanism based on
a dielectric elastomer actuator. Smart Mater. Struct. 28, 09LT01 (2019).
25. Duduta, M. et al. Tunable multi-modal locomotion in soft dielectric elastomer
robots. IEEE Robot. Automat. Lett.5, 38683875 (2020).
26. Zhao, J. et al. Soft and fast hopping-running robot with speed of six times its
body length per second. Soft Robot. 6, 713721 (2019).
27. Ni, F., Rojas, D., Tang, K., Cai, L. & Asfour, T. A jumping robot using soft
pneumatic actuator. In IEEE International Conference on Robotics and
Automation 31543159 (IEEE, 2015).
28. Niiyama, R., Nagakubo, A. & Kuniyoshi, Y. Mowgli: a bipedal jumping and
landing robot with an articial musculoskeletal system. In IEEE International
Conference on Robotics and Automation 25462551 (IEEE, 2007).
29. Liu, Z. et al. A 1 mm-thick miniatured mobile soft robot with
mechanosensation and multimodal locomotion. IEEE Robot. Automat. Lett. 5,
32913298 (2020).
30. Tolley, M. T. et al. An untethered jumping soft robot. In IEEE International
Conference on Intelligent Robots and Systems 561566 (IEEE, 2014).
31. Loepfe, M., Schumacher, C. M., Lustenberger, U. B. & Stark, W. J. An
untethered, jumping roly-poly soft robot driven by combustion. Soft Robot. 2,
3341 (2015).
32. Bartlett, N. W. et al. A 3D-printed, functionally graded soft robot powered by
combustion. Science 349, 161165 (2015).
33. Churaman, W. A., Gerratt, A. P. & Bergbreiter, S. First leaps toward jumping
microrobots. In IEEE International Conference on Intelligent Robots and
Systems 16801686 (IEEE, 2011).
34. Li, S. & Rus, D. JelloCube: a continuously jumping robot with soft body. IEEE-
ASME Trans. Mech. 24, 447458 (2019).
35. Li, S., Katzschmann, R. & Rus, D. A soft cube capable of controllable
continuous jumping. In IEEE International Conference on Intellignet Robots
and Systems 17121717 (IEEE, 2015).
36. Mintchev, S., Zappetti, D., Willemin, J. & Floreano, D. A soft robot for
random exploration of terrestrial environments. In IEEE International
Conference on Robotics and Automation 74927497 (IEEE, 2018).
37. Wu, Y. et al. Insect-scale fast moving and ultrarobust soft robot. Sci. Robot. 4,
eaax1594 (2019).
38. Acome, E. et al. Hydraulically amplied self-healing electrostatic actuators
with muscle-like performance. Science 359,6165 (2018).
39. Kellaris, N., Venkata, V. G., Smith, G. M., Mitchell, S. K. & Keplinger, C.
Peano-HASEL actuators: Muscle-mimetic, electrohydraulic transducers that
linearly contract on activation. Sci. Robot. 3, eaar3276 (2018).
40. Mitchell, S. K. et al. An easy-to-implement toolkit to create versatile and high-
performance HASEL actuators for untethered soft robots. Adv. Sci. 6, 1900178
(2019).
41. Wang, X., Mitchell, S. K., Rumley, E. H., Rothemund, P. & Keplinger, C. High-
strain Peano-HASEL actuators. Adv. Funct. Mater. 30, 1908821 (2019).
42. Gu, G., Zou, J., Zhao, R., Zhao, X. & Zhu, X. Soft wall-climbing robots. Sci.
Robot. 3, eaat2874 (2018).
43. Rothemund, P., Kellaris, N. & Keplinger, C. How inhomogeneous zipping
increases the force output of Peano-HASEL actuators. Extreme. Mech. Lett. 31,
100542 (2019).
44. Ho, J. & Jow, R. Characterization of High Temperature Polymer Thin Films for
Power Conditioning Capacitors (No. ARL-TR-4880, Army Research
Laboratories, 2009).
45. Meshkatoddini, M. R. & Abbospour, S. Aging study and lifetime estimation of
transformer mineral oil. Am. J. Eng. Appl. Sci. 1, 384388 (2008).
46. Alfrey, T., Wiederhorn, N., Stein, R. & Tobolsky, A. Plasticized polyvinyl
chloride: structure and mechanical behavior. Ind. Eng. Chem. Res. 41, 701703
(1949).
Acknowledgements
This work was supported by the National Key Research and Development Project of
China (grant no. 2020YFB1313000) and the National Natural Science Foundation of
China (grant nos. 52075051 and 51975070). J.G. wants to thank Prof. Jonathan Rossiter
for supporting this work whilst working at SoftLab Bristol, University of Bristol, under
grant EP/M020460/1. In addition, he wants to thank Prof. Jinsong Leng for the support
while working at HITSZ. We also thank Yi Sun and Thomas Bamber for the initial review
of the paper, and Wenbo Liu for the assistance in experiments.
Author contributions
R.C. conceived the research. R.C. and Z.Y. jointly designed and implemented the soft
electrohydrostatic bending actuator. J.G. rened the legless soft jumping robot concept,
and proposed the robot design rationale and rened paper novelty clarications. Z.Y. and
X.Z. conducted the structural design, device manufacturing and experiments. Z.Y., R.C.,
L.B. and X.Z. analyzed and interpreted the results. Z.Y. drafted the manuscript. Z.Y., X.Z.
and J.G. designed and optimized the gures, tables, and videos. Z.Y., J.G., L.W. and Y.S.
wrote the abstract, introduction, and discussion part, and fully revised the whole
manuscript. L.B., F.L., H.P., L.X., Y.P. and J.L. reviewed and commented on the paper.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information The online version contains supplementary material
available at https://doi.org/10.1038/s41467-021-27265-w.
Correspondence and requests for materials should be addressed to Rui Chen or Huayan
Pu.
Peer review information Nature Communications thanks Manolo Garabini and Cecilia
Laschi for their contribution to the peer review of this work. Peer reviewer reports are
available.
Reprints and permission information is available at http://www.nature.com/reprints
Publishers note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional afliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the articles Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
articles Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2021
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27265-w
12 NATURE COMMUNICATIONS | (2021) 12:7028 | https://doi.org/10.1038/s41467-021-27265-w | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
... The lack of flexibility in rigid robots limits their application scenarios and adaptions to complex environments. Soft robots, made of flexible materials with soft drives (e.g., stimuli-active materials such as dielectric elastomers [17][18][19][20][21] , shape memory polymers [22][23][24][25] , liquid crystal elastomers [26][27][28][29] , etc.), offer advantages of light weight, strong deformability and good environmental adaptability, and provide a convincing solution to resolve the flexibility challenge of rigid robots as demonstrated in developing flying soft robots [30,31] , swimming soft robots [32][33][34] , jumping soft robots [35,36] , crawling soft robots [37][38][39] , etc. Despite these notable advances in soft robots, soft drives based on stimuli-active materials still have certain disadvantages such as high driving voltage, high driving temperature, low power density and poor biocompatibility, which greatly hinder their practical utilities, particularly in the fields of biomedicine and organ-on-a-chip system. ...
Article
Full-text available
Biohybrid robots (bio-bots), made of biocompatible skeletons with living drives (e.g., biological living tissues or cells), represent a new direction of robotics technology due to their attractive advantages of softness, flexibility, adaptability and biocompatibility, accompanied by the remarkable capabilities of self-assembly, self-healing, and self-replication. This paper provides a brief review of recent advances of bio-bots from a functional view, including walking, swimming and non-locomotion bio-bots, by exploring their structure designs along with their operational principles. The performances of these bio-bots are summarized and compared followed by the discussions of challenges and perspectives, which provide valuable insight and guidance for future developments of bio-bots.
... Bioinspired solutions also include multi-DOF locomotion, which demonstrates even higher manoeuvrability for complex environments. Example designs include soft drones such as those developed by Park and Cha [293], Chen et al. [294], and Shah et al. [295], demonstrating adaptable motion using caterpillar motion at a small scale. Other examples include earthworm and snake-like locomotion [296][297][298]. ...
Article
Full-text available
With the ever-increasing demand for harvesting wind energy, the inspection of its associated infrastructures, particularly turbines, has become essential to ensure continued and sustainable operations. With these inspections being hazardous to human operators, time-consuming and expensive, the door was opened for drone solutions to offer a more effective alternative. However, drones also come with their own issues, such as communication, maintenance and the personnel needed to operate them. A multimodal approach to this problem thus has the potential to provide a combined solution where a single platform can perform all inspection operations required for wind turbine structures. This paper reviews the current approaches and technologies used in wind turbine inspections together with a multitude of multimodal designs that are surveyed to assess their potential for this application. Rotor-based designs demonstrate simpler and more efficient means to conduct such missions, whereas bio-inspired designs allow greater flexibility and more accurate locomotion. Whilst each of these design categories comes with different trade-offs, both should be considered for an effective hybrid design to create a more optimal system. Finally, the use of sensor fusion within techniques such as GPS and LiDAR SLAM enables high navigation performances while simultaneously utilising these sensors to conduct the inspection tasks.
Article
Simulating soft robots offers a cost-effective approach to exploring their design and control strategies. While current models, such as finite element analysis, are effective in capturing soft robotic dynamics, the field still requires a broadly applicable and efficient numerical simulation method. In this paper, we introduce a discrete differential geometry-based framework for the model-based inverse design of a novel snap-actuated jumping robot. Our findings reveal that the snapping beam actuator exhibits both symmetric and asymmetric dynamic modes, enabling tunable robot trajectories (e.g., horizontal or vertical jumps). Leveraging this bistable beam as a robotic actuator, we propose a physics-data hybrid inverse design strategy to endow the snap-jump robot with a diverse range of jumping capabilities. By utilizing a physical engine to examine the effects of design parameters on jump dynamics, we then use extensive simulation data to establish a data-driven inverse design solution. This approach allows rapid exploration of parameter spaces to achieve targeted jump trajectories, providing a robust foundation for the robot's fabrication. Our methodology offers a powerful framework for advancing the design and control of soft robots through integrated simulation and data-driven techniques.
Article
Full-text available
Multimodal microrobots are of growing interest due to their capabilities to navigate diverse terrains, with promising applications in inspection, exploration, and biomedicine. Despite remarkable progress, it remains challenging to combine the attributes of excellent maneuverability, low power consumption, and high robustness in a single multimodal microrobot. We propose an architected design of a passively morphing wheel that can be stabilized at distinct geometric configurations, relying on asymmetric bending stiffness of bioinspired tentacle structures. By integrating such wheels with electromagnetic motors and a flexible body, we develop a highly compact, lightweight, multimodal microrobot (length ~32 mm and mass ~4.74 g) with three locomotion gaits. It has high motion speed (~21.2 BL/s), excellent agility (relative centripetal acceleration, ~206.9 BL/s ² ), low power consumption (cost of transport, ~89), high robustness, and strong terrain adaptabilities. Integration of batteries and a wireless control module enables developments of an untethered microrobot that maintains high motion speed and excellent agility, with capabilities of traveling in hybrid terrains.
Article
Endowing variable-stiffness soft robots (VSSRs) with sensing capabilities will help to improve their safe interaction capabilities and intelligence. Therefore, this work adopts direct ink writing (DIW) to directly fabricate functional soft components of the VSSR, including heating and sensing. To improve the printing accuracy of DIW, the liquid metal (LM)/sodium alginate (SA) composite ink was prepared by sonicating LM in a SA aqueous solution and centrifuging. Functional soft components were printed on the bottom and top of the soft robot with LM/SA ink by DIW and were used to achieve low-voltage (as low as 1.5 V) Joule heating and to sense the bending deformation of the soft robot with a low response time (about 80 ms), respectively. Then, the photoabsorber Sudan I (0.02 wt %) was added to the commercial photosensitive resin to improve the precision of digital light processing for manufacturing the soft robot body. Through control of the flexible heating circuit to realize the solid–liquid transition of the low-melting-point alloy (47 °C), the stiffness of the soft robot can be changed. This work has great prospects for the design and manufacture of a VSSR with integrated sensing capability.
Article
With the robust and self-trapped properties, recent advances about soliton dynamics in multi-stable mechanical metamaterials have led to many innovative techniques from signal processing to robotics. This work proposes a multi-stable mechanical metamaterial driven by nonlinear dissipative solitons, in which the coupling and decoupling of multiple locomotion modes can be achieved. Based on a cylinder network with asymmetric energy landscape, the uniform field model of Landau theory is developed. During the theoretical calculation, the analytical solutions of several dissipative solitons are derived, which allow multiple special behaviors of solitary waves, such as wave velocity gaps, directional propagation and spiral phase transition. By incorporating such effects into robotic designs, a variety of complex movements can be achieved by a single structure, including hopping, rolling, rotating, swinging, bending and translational components. In particular, as excitation positions change, the mechanical metamaterial can flexibly switch multiple locomotion modes without changing configurations, e.g., spinning and spin-less, straight and oblique as well as coupled multimode movements. This work wishes to provide some new inspirations for the applications of nonlinear elastic wave metamaterials and phase transition theory in robotics.
Article
Full-text available
Soft robots are intrinsically safe for use near humans and adaptable when operated in unstructured environments, thereby offering capabilities beyond traditional robots based on rigid components. Soft actuators are key components of soft robots; recently developed hydraulically amplified self‐healing electrostatic (HASEL) actuators provide a versatile framework to create high‐speed actuators with excellent all‐around performance. Peano‐HASEL actuators linearly contract upon application of voltage, closely mimicking the behavior of muscle. Peano‐HASEL actuators, however, produce a maximum strain of ≈15%, while skeletal muscles achieve ≈20% on average. Here, a new type of HASEL is introduced, termed high‐strain Peano‐HASEL (HS‐Peano‐HASEL) actuator, that achieves linear contraction up to ≈24%. A wide range of performance metrics are investigated, and the maximum strain of multiunit HS‐Peano‐HASEL actuators is optimized by varying materials and geometry. Furthermore, an artificial circular muscle (ACM) based on the HS‐Peano‐HASEL acts as a tubular pump, resembling the primordial heart of an ascidian. Additionally, a strain‐amplifying pulley system is introduced to increase the maximum strain of an HS‐Peano‐HASEL to 42%. The muscle‐like maximum actuation strain and excellent demonstrated all‐around performance of HS‐Peano‐HASEL actuators make them promising candidates for use in artificial organs, life‐like robotic faces, and a variety of other robotic systems. High‐strain Peano‐HASEL (hydraulically amplified self‐healing electrostatic) actuators are electrohydraulically driven artificial muscles that feature high‐speed linear contraction of ≈24%, matching average values for skeletal muscles. Materials and geometry are optimized, and application as a soft pump is demonstrated; other potential uses include bioinspired robots, robotic faces, and artificial organs.
Article
Full-text available
Existing robots capable of hopping or running mostly rely on rigid actuators, but soft hopping or running robots based on muscle-like actuators have not yet been achieved. In this article, we report a tethered soft robot capable of hopping-running on smooth and rough surfaces with high speed. The hopping-running robot is composed of two soft joints that simulate the foreleg and hind leg driven by dielectric elastomer. The mass, length, width, and height of the robot are 6.5 g, 8.5 cm, 4.8 cm, and 50 cm, respectively. The robot can run at a speed of 51.83 cm/s (6.10 body lengths/s), which is much faster than previously reported locomotion robots driven by soft responsive materials. The robot also shows good adaptability to different terrains, such as marble, wood, rubber, sandpaper, and slopes. The robot can carry a load equal to its weight, can maintain a high locomotion speed, and demonstrates the potential ability to carry its power supply and control circuitry.
Article
Full-text available
Jumping mechanisms are useful in robotics for locomotion in unstructured environments, or for self-righting abilities. However, most rigid robots rely on impact with the ground to jump, thereby requiring a relatively rigid, and flat environment. Moreover, they need to be able to absorb high impact forces during landing in order to maintain structural integrity. In this paper we investigate soft systems, capable of jumping repeatedly in unstructured environments with no need for precise landings. Our impulsive approach is based on a soft electro-mechanical transducer, a dielectric elastomer actuator (DEA). The design is inspired by click-beetles and simple bio-mechanical models, which convert the flexing around a hinge into jumping. Our actuator is power amplified by the addition of a stiffer strip, allowing for rapid shape transitions (22 ms) between flat and curved states. The transition is controlled by an electric latch: the DEA is discharged faster than the actuator can deform. The mechanical energy stored in the composite beam is released rapidly, leading to impulsive motions (jumps of a full body length: i.e. 5 cm). This demonstration of an electrically-latched power amplification mechanism shows that relatively simple electro-mechanical systems can exhibit impulsive behavior which may enable new types of locomotion in compliant machines.
Article
Full-text available
A ten-gram insect-inspired robot capable of crawling, walking, jumping, somersaulting and performing collective tasks is built from low-cost, assembly-free components, demonstrating its scalability for collective applications with expanded mobility.
Article
Full-text available
Light has been recently intensively explored to power robots. However, most existing light‐driven robots have limited locomotion modalities, with constrained locomotion capabilities. A light‐powered soft robot with a bioinspired design is demonstrated, which can crawl on ground, squeeze its way through a small channel, and jump over a barrier. The arch‐shaped robot is made up of liquid crystal elastomer–carbon nanotube composite. When a light source with a power intensity of around 1.57 W cm⁻² is scanned over the surface of the robot, it deforms and crawls forward. With an increase in the light scanning speed, it can deform sufficiently to pass through a channel 25% lower than its body height. Subjected to light irradiation, it can also deform to a closed loop, gradually store elastic energy, and suddenly release it to jump over a wall or onto a step quickly, with a jumping distance around eight times its body length and jumping height around five times its body height. Mathematical models for quantitatively understanding the multimodal locomotion of this light‐powered soft robot are also presented.
Article
Soft robots require strong, yet flexible actuators for locomotion and manipulation tasks in unstructured environ- ments. Dielectric elastomer actuators (DEAs) are well suited for these challenges in soft robotics because they operate as compliant capacitors and directly convert electrical energy into mechanical work, thereby allowing for simple design integration at a minimal footprint. In most demonstrations, DEA-based robots are limited to a single mode of locomotion, for example crawling, swimming, or jumping. In this work, we explored a range of actuation patterns in combination with a novel actuator design to enable multi-modal locomotion, whereby an actuation pattern is defined by an actuation voltage (proportional to the applied electric field) and frequency (the actuation rate). We present a DEA robot capable of three different gaits including crawling, hopping, and jumping. In addition, our robot can set itself upright by performing a roll, for example to prepare for the next jump after landing on its side. These results demonstrate that DEAs can be used as versatile experimental devices to validate locomotion models, in both natural and engineered systems.
Article
The miniature soft robots have many promising applications, including micro-manipulations, endoscopy, and microsurgery, etc. Nevertheless, it remains challenging to fabricate a miniatured robot device that is thin, flexible, and can perform multimodal locomotor mobility with sensory capacity. In this study, we propose a miniatured, multi-layer (two shape memory polymer layers, a flexible copper heater, a silk particle enhanced actuator layer, and a sensory layer) four-limb soft robot (0.45-gram, 35mm-long, 12mm-wide) with a total thickness of 1mm. A precise flip-assembling technique is utilized to integrate multiple functional layers (fabricated by soft lithography, laser micromachining technologies). The actuator layer's elastic modulus increased ~100% by mixing with 20% silk particles by weight, which enhanced the mechanical properties of the miniature soft robot. We demonstrate that the soft robot can perform underwater crawling and jumping-gliding locomotion. The sensing data depicts the robot's multiple bending configurations after the sensory data been processed by the microprocessor mounted on the robot torso. The miniatured soft robot can also be reshaped to a soft miniatured gripper. The proposed miniatured soft robots can be helpful for studying soft organisms‘ body locomotion as well as medical applications in the future.
Article
Applying concepts and methods of bionics to endow autonomous robots with elegant and agile mobility just like natural living beings is gradually becoming a hot research topic in intelligent robot field. Compared with walking, crawling, rolling and other motion modes, jumping performs considerable advantages that can leap across obstacles and move to different heights in agility and flexibility. In this paper, we specifically review the developments of biologically inspired jumping robots in the past decades, and give comprehensive analysis on some key technologies for implementing a practical jumping robot effectively. First, the jumping mechanism of frog (amphibian, quadruped), locust (arthropod, hexapod), kangaroo (mammality, bipedalism) as examples of typical animals good at jumping is introduced and analyzed, from which it is concluded that power sources, limbs coordination and control are key elements for excellent jumping performances, which should be synthetically improved by combination with structure design and model establishment. Then, spring loaded inverted pendulum (SLIP), bio-inspired open-chain and closed-chain multi-linkage as representative jumping mechanical structures, their characteristics are explored accompanied with dynamic analysis. After a detailed analysis to actuators and energy storage devices and a comprehensive summarization to functional and soft materials commonly applied in jumping robots, different control methods and strategies adopted to achieve better jumping performance are reviewed and analyzed, from self-righting, driving control to path planning. Especially, how to analyze the stability of a jumping control system and how to stabilize it are explained theoretically by taking a vertical monopedal jumping robot as an example and via limit cycle analysis. Finally, some feasible and potential future developments in bio-inspired jumping robots are also presented after detailed discussions on current status and existing deficiencies.
Article
Research in soft robotics has yielded numerous types of soft actuators with widely differing mechanisms of operation that enable functionality that is difficult or impossible to reproduce with hard actuators such as electromagnetic motors. The Peano-HASEL (hydraulically amplified self-healing electrostatic) actuator is a new type of electrostatic, linearly contracting, soft actuator that features large strains, fast actuation, and high energy densities. Peano-HASEL actuators are comprised of pouches, which are made of flexible dielectric polymer films, filled with a liquid dielectric, and covered with flexible electrodes. When a voltage is applied to the electrodes, they “zip” together due to the Maxwell stress, which displaces the liquid inside the pouch, and causes the actuator to contract. Zipping can occur homogeneously or inhomogeneously. In this letter we analyze inhomogeneous zipping and its influence on the performance of Peano-HASEL actuators. We develop a theoretical model that describes inhomogeneous as well as homogeneous zipping of the electrodes and characterize the behavior of actuators experimentally. Inhomogeneous zipping occurs (depending on the size of the electrodes) predominantly at large loads, because it allows for larger areas of the electrodes to be zipped. Inhomogeneous zipping increases the blocking force of the actuators and leads to larger actuation strains near the blocking force. Exploiting inhomogeneous zipping by increasing the electrode size enables an increase in the blocking force of the actuators by up to 47%.
Article
Mobility and robustness are two important features for practical applications of robots. Soft robots made of polymeric materials have the potential to achieve both attributes simultaneously. Inspired by nature, this research presents soft robots based on a curved unimorph piezoelectric structure whose relative speed of 20 body lengths per second is the fastest measured among published artificial insect-scale robots. The soft robot uses several principles of animal locomotion, can carry loads, climb slopes, and has the sturdiness of cockroaches. After withstanding the weight of an adult footstep, which is about 1 million times heavier than that of the robot, the system survived and continued to move afterward. The relatively fast locomotion and robustness are attributed to the curved unimorph piezoelectric structure with large amplitude vibration, which advances beyond other methods. The design principle, driving mechanism, and operating characteristics can be further optimized and extended for improved performances, as well as used for other flexible devices.