Soft robot review

Abstract

Soft robots are often inspired from biological systems which consist of soft materials or are actuated by electrically activated materials. There are several advantages of soft robots compared to the conventional robots; safe human-machine interaction, adaptability to wearable devices, simple gripping system, and so on. Due to the unique features and advantages, soft robots have a considerable range of applications. This article reviews state-of-the-art researches on soft robots and application areas. Actuation systems for soft robots can be categorized and analyzed into three types: variable length tendon, fluidic actuation, and electro-active polymer (EAP). The deformable property of soft robots restricts the use of many conventional rigid sensors such as encoders, strain gauges, or inertial measurement units. Thus, contactless approaches for sensing and/or sensors with low modulus are preferable for soft robots. Sensors include low modulus (< 1 MPa) elastomers with liquid-phase material filled channels and are appropriate for proprioception which is determined by the degree of curvature. In control perspective, novel control idea should be developed because the conventional control techniques may be inadequate to handle soft robots. Several innovative techniques and diverse materials & fabrication methods are described in this review article. In addition, a wide range of soft robots are characterized and analyzed based on the following sub-categories; actuation, sensing, structure, control and electronics, materials, fabrication and system, and applications.

Full text

International Journal of Control, Automation and Systems 15(1) (2017) 3-15
http://dx.doi.org/10.1007/s12555-016-0462-3
ISSN:1598-6446 eISSN:2005-4092
http://www.springer.com/12555
Soft Robot Review
Chiwon Lee, Myungjoon Kim, Yoon Jae Kim, Nhayoung Hong, Seungwan Ryu, H. Jin Kim, and Sungwan Kim*
Abstract: Soft robots are often inspired from biological systems which consist of soft materials or are actuated by
electrically activated materials. There are several advantages of soft robots compared to the conventional robots;
safe human-machine interaction, adaptability to wearable devices, simple gripping system, and so on. Due to the
unique features and advantages, soft robots have a considerable range of applications. This article reviews state-
of-the-art researches on soft robots and application areas. Actuation systems for soft robots can be categorized
and analyzed into three types: variable length tendon, fluidic actuation, and electro-active polymer (EAP). The de-
formable property of soft robots restricts the use of many conventional rigid sensors such as encoders, strain gauges,
or inertial measurement units. Thus, contactless approaches for sensing and/or sensors with low modulus are prefer-
able for soft robots. Sensors include low modulus (<1 MPa) elastomers with liquid-phase material filled channels
and are appropriate for proprioception which is determined by the degree of curvature. In control perspective, novel
control idea should be developed because the conventional control techniques may be inadequate to handle soft
robots. Several innovative techniques and diverse materials & fabrication methods are described in this review arti-
cle. In addition, a wide range of soft robots are characterized and analyzed based on the following sub-categories;
actuation, sensing, structure, control and electronics, materials, fabrication and system, and applications.
Keywords: Biological systems, flexible materials, smart structure, soft robotics, soft structure.
1. INTRODUCTION
Biological mechanism and locomotion systems have in-
spired many robot engineers and scientists to study mul-
tifunctional systems [1]. Innovative and creative results
from such research have organized a new field of robotics
called soft robotics. Soft robots have distinguishable fea-
tures compared to the conventional robots. Conventional
robot’s structures are made with high stiffness materi-
als such as steel, aluminum, titanium, stainless steel, etc.
These parts can be manufactured by mechanical machin-
ing tools including milling, lathe, computerized numeri-
cal control (CNC) machine, and are mechanically assem-
bled. On the other hand, soft robots adopt hyper elastic
materials for main body and moving parts such as poly-
mer, rubber, silicone, or other flexible materials. These
are manufactured using a three dimensional (3D) printer
or 3D mold. Material stiffness of soft robots are often in
Manuscript received July 29, 2016; accepted October 29, 2016. Recommended by Editor-in-Chief Young Hoon Joo. This work was
supported by the Bio and Medical Technology Development Program of the NRF funded by the Korean Government, MSIP (NRF-
2014M3A9E3064623), the Interdisciplinary Research Initiatives Program from College of Engineering and College of Medicine, Seoul
National University (Grant No. 800-20160095), and the Research Program 2016 funded by Seoul National University College of Medicine
Research Foundation (Grant No. 800-20160072). Chiwon Lee, Myungjoon Kim, Yoon Jae Kim, and Nhayoung Hong contributed equally to
this work (co-first author).
Chiwon Lee is with the Institute of Medical and Biological Engineering, Medical Research Center, Seoul National University, Seoul 03080,
Korea (e-mail: lcwkf16@snu.ac.kr). Myungjoon Kim, Yoon Jae Kim, and Nhayoung Hong are with the Interdisciplinary Program for
Bioengineering, Graduate School, Seoul National University, Seoul 08826, Korea (e-mails: jinmj08@gmail.com, kyj182731@naver.com,
jhong407@naver.com). Seungwan Ryu and H. Jin Kim are with the School of Mechanical and Aerospace Engineering, Seoul National
University, Seoul 151-742, Korea (e-mails: rsy7942@gmail.com, hjinkim@snu.ac.kr). Sungwan Kim is with the Department of Biomedical
Engineering, Seoul National University College of Medicine, Seoul 03080, Korea (e-mail: sungwan@snu.ac.kr).
* Corresponding author.
the order of 104−109Pa corresponding with biological
skin or muscle tissue [2]. Due to the differences in ma-
terial and manufacturing method utilized, protection and
stability strategies are contrasted with that of rigid-body
robots. Conventional robots require complex protection
or stability control algorithm whereas soft robots do not
require such methods due to shock absorbing property of
the materials used. For this reason, conventional robots
are designed for specific tasks in controlled environments,
while soft robots are made to perform in unstructured en-
vironments. Conventional robots are usually actuated by
electrical motor, hydraulic pump or pneumatic compres-
sor which is capable of producing force ranging from a
few millinewton (mN) to meganewton (MN). However,
one of the main disadvantages of soft robot is being un-
able to produce a large force owing to its elastic structure
and usage of pneumatic compressor, shape-memory alloy
(SMA), electro-active polymer (EAP), etc.
c
⃝ICROS, KIEE and Springer 2016

4Chiwon Lee, Myungjoon Kim, Yoon Jae Kim, Nhayoung Hong, Seungwan Ryu, H. Jin Kim, and Sungwan Kim
Fig. 1. Soft Robots. (a) A resilient, untethered robot [3]. (b) Multigait soft robot [4]. (c) Untethered jumping robot [5].
(d) Jamming skin enabled locomotion (JSEL) [6]. (e) OctArm [7]. (f) Compliant and underactuated robotic hand
[8]. (g) Manta swimming robot [9]. (h) Hydraulic autonomous soft robotic fish [10]. (i) Soft robotic gloves [11].
(j) Octopus robot [12]. (k) Meshworm robot [13]. (l) GoQbot [14]. (m) Universal granular jamming [15]. (n)
Jamming granular robot [16].
Despite this disadvantage, there are several advantages
of soft robot using soft materials. It can interact with hu-
man and environment more safely than conventional rigid
robot system by reducing the risk of injury to itself and
surroundings from collision [2]. This allowed its appli-
cation to be extended to a wearable devices and medical
usage. In addition, soft robots have inherent benefits over
conventional rigid robot system in gripping motion having
simpler structure and control algorithm. Due to the unique
features and advantages of soft robot, it has various appli-
cation areas.
In the following sections, state-of-the-art researches on
soft robot are categorized and analyzed based on the fol-
lowing sub-categories; actuation, sensing, structure, con-
trol and electronics, materials, fabrication and system, and
applications. Then, specific issues regarding soft robot are
discussed followed by suggestion of its future directions.
2. SOFT ROBOTS
2.1. Actuation
Conventional robots are actuated with rigid motors and
various complicated mechanisms are realized by rota-
tional and linear actuation systems. In contrast to the
rigid-body robots with limited degrees of freedom (DOFs)
of motion, soft robots have continuum deformation of the
flexible body resulting in high DOFs. According to the
previous studies [2,17], most soft robots are actuated in
one of three ways: variable length tendons, fluidic actua-
tion, or EAP.
Variable-length tendons in the form of tension cables or
SMA actuators may be embedded in soft segments deliv-
ering a controlled force to deform the segment in a desired
way. An octopus inspired robot used silicone arm with
embedded cables to replicate the functionality of muscular
structure of the octopus arm [18]. When tension is applied
to cables using spooler motors, the granular media such as
grained coffee grabs the target object by granular jamming
[19]. A meshworm robot, which is inspired by peristaltic
locomotion of Oligochaetes, uses nickel titanium (NiTi)
SMA as its mode of actuation [20]. GoQBot is a cater-
pillar inspired soft bodied rolling robot and also utilizes
SMA coils which yields larger strain than the straight form
[21]. Relatively small strain produced by SMA has been
pointed out as a disadvantage, but it can be overcame by
using spiral-shaped SMA. Another soft robot inspired by
mechanical characteristic of octopus uses both tension ca-
ble (longitudinally) and spiral SMA (transversally) [22].
Fluidic actuator inflates channels within the soft body
to deform the structure in a controlled manner [2]. As one
of earlier versions of fluidic actuator, pneumatic artificial
muscle (PAM) was suggested. PAM, also called McK-
ibben actuator, is a flexible linear soft actuator consisting
of deformable elastomer tubes encased by fiber sleeves
[23,24]. Progressively, more soft robots using fluidic elas-
tomer actuator (FEA) are reported. FEA is a novel type of
highly deformable and adaptable soft actuator. It has syn-
thetic elastomer layers and deforms by pressurized fluid
expanding the embedded channels. This type of architec-
ture is often called Pneu-Nets (PN). Little or no additional
energy is required to hold its shape once the actuator is
deformed by pressure. Pressure used in FEA can be either
pneumatic or hydraulic. Due to the omnipresence, less
inviscid and lightweight nature of air in the environment,
pneumatic systems are often preferred over hydraulic sys-
tems. A recent study [25] reported resilient quadrupedal
soft robot, it can adapt to various environmental condi-
tions. It is an advanced version of its previous study [26].

Soft Robot Review 5
The new model employs a modified PN architecture for
more rapid and stable actuation differentiating from its
previous design [27]. A different quadrupedal soft robot
model can camouflage itself in an environment through
altering its color, contrast, pattern, apparent shape, lumi-
nescence, and surface temperature [28]. It also uses PN
design for pneumatic pressurization in an independent net-
work of micro-channels embedded in highly extensible
elastomers. A snake inspired soft robot consists of four
bidirectional fluidic elastomers [29], and it provides ser-
pentine locomotion despite some rigid components such
as passive wheels. Another pneumatic-based locomotion
is seen in a robot with jamming skin [30]. The locomotion
is provided by deformation of unjammed cells caused by
pneumatic expansion of a volume variable actuator inside
the soft robot.
In some soft robots, not only omnipresent air but also
other chemicals were used to power the actuators. The
chemical decomposition of hydrogen peroxide into oxy-
gen gas in a closed container was used to self-regulate the
actuation pressure [31]. In other cases, explosive actua-
tors using butane gas are utilized for directed jumping mo-
tion [32]. The FEAs have been used not only in locomo-
tive robots but also in manipulators. One of early models
of FEA, actuated with pneumatic pressure and channels
made from rubber, was suggested in the early 1990s [33].
It had three DOFs, pitch, yaw, and stretch, which make it
suitable for robotic mechanisms such as fingers, arms, or
legs. OctArm continuum manipulator was pneumatically
actuated and tested in both open-air and in-water environ-
ments [34]. A gripper implemented with lithography tech-
niques and PN architecture have been reported previously
[35], and a hand robot for dexterous grasping with modi-
fied version of PN called PneuFlex was also reported [36].
Another research focused on wrist torsional motion by he-
lically arranged tubes rather than finger grasping [37]. As
for another case, the previously mentioned quadrupedal
soft robot [26] was modified to conduct both locomotion
and grasping [38]. Another type of pneumatically actu-
ated soft robot, a manta ray inspired swimming robot was
reported [39]. It tested various cross sections of bend-
ing pneumatic actuators and developed manta swimming
robot based on the results.
Hydraulic actuation is not as common as pneumatic but
can produce a larger force. A soft robotic fish capable
of 3D swimming, such as forward swimming, diving, and
turning, was reported [40]. For medical purpose, a soft
robotic glove for at-home rehabilitation was introduced
[41]. It contains hydraulic pumps and a water-reservoir
for actuation.
EAPs are biocompatible polymers that exhibit a change
in size or shape when stimulated by an electric field. First
proposed by Wihelm Rontgen in 1880s, EAPs are classi-
fied into two categories: electronic EAP and ionic EAP
[42]. The electronic EAPs generally require high activa-
tion fields (>150 V/
µ
m) and hold the induced displace-
ment when a DC voltage is applied. The electronic EAPs
have high energy density as well as rapid response time in
the range of milliseconds. In contrast, ionic EAP such as
ionic polymer-metal composite (IPMC) requires low ac-
tuating voltages (<5 V) [42]. Additionally, it has a nat-
ural bi-directional actuation property deforming in differ-
ent direction based on the voltage polarity [43]. IPMCs
are advantageous due to its ability to achieve large defor-
mation with relatively low electric field (<10 kV/m) and
therefore are more efficient and safer [44]. Besides, they
have low power consumption, and demonstrate rapid ac-
tuating response (>10 Hz in water). Also simple actu-
ator structure enables miniaturization of soft robots. The
ionic EAPs have limitation of having to operate in wet
conditions in solid electrolyte, and therefore require en-
capsulation or protective layer in open air condition [43].
They produce relatively lower bending force than elec-
tronic EAPs.
Since IPMC was first discovered in 1990s, its use in
soft robots has been investigated [17,42]. When driv-
ing voltage is applied, the IPMC membrane bends to-
ward the anode because the hydrophilic positive ions or
cations move toward the cathode. Considering the fact that
IPMC works in the cantilevered form, the geometric prop-
erties of the beam, such as thickness, width, and length,
should be tuned to achieve desired level of force and de-
flection [17]. Previous studies investigated about IPMC
based bioinspired soft robots such as snake-like swim-
ming robot [45,46] and multi-DOF micro-robot manip-
ulators [47,48]. Whereas early studies used simple strip-
shaped IPMC, which only enables simple bending motion
[49,50], patterned IPMC can provide multi-DOF motions.
It enables individual control of each segment of IPMC and
is more efficient in performing complex motions, such as
snake-like wavy motion.
As reviewed in this section, actuation methods used
in soft robots can be classified into three types, variable
length tendon, fluidic actuation, and EAP. Actuation ap-
proaches using variable length tendon can be subcatego-
rized into tension cable based and SMA based methods.
Fluidic pressure actuation system can be further subcate-
gorized into pneumatic and hydraulic actuation. In case
of EAP, it can be classified into electronic and ionic types.
Every approach has unique advantages compared to the
others, and have demonstrated notable performances in in-
dividual researches. Especially, pneumatic pressure based
soft robots has wide application areas and reported fre-
quently. The major advantages of each actuation method
are summarized in below Table 1.
2.2. Sensing
Sensors allow proprioception of robotic hardware but
the deformable characteristic of soft robots prevents the
use of many conventional sensors including encoders,

6Chiwon Lee, Myungjoon Kim, Yoon Jae Kim, Nhayoung Hong, Seungwan Ryu, H. Jin Kim, and Sungwan Kim
Table 1. Advantages of three actuation approaches.
Actuation approach Advantages
Length
variable
tendon
Tension
cable
- Use of conventional motor in
external platform for actuation
Shape-
memory
alloy
(SMA)
- Easy manufacturing and
programming
- A spiral shaped SMA capable of
300% deformation
(Beam shape: 5-8%)
Fluidic
actuation
Pneumatic
- Omnipresent air
- Environmentally benign
- Light weight
- Less inviscid (less time delay)
Hydraulic
- Larger force compare to
pneumatic actuators
- Ability to be implement in
untethered in-water soft robots
Electro-
active
polymer
(EAP)
Electronic
- High energy density
- Larger force compare to ionic
class
- Rapid response time in the
range of milliseconds
- Long operation time
Ionic
- Low actuation voltage
- Large bending displacement
- Bi-directional actuation is
possible
strain gauges, or inertial measurement units [2]. Thus,
alternative sensing methods are preferred over the con-
ventional sensors. Contactless or low-modulus sensors
can be the alternatives. Sensors with low modulus (<1
MPa) elastomers along with liquid-phase materials are ap-
propriate for proprioception of curved structure and this
type of sensors apply a small impedance change to soft
robot hardware [2]. These types of soft sensors are con-
ventionally manufactured by soft lithography where thin
elastomer layers are patterned with microfluidic channels.
These microfluidic channels are filled with liquid conduc-
tors, such as eutectic gallium-indium [51–54] and con-
ductive carbon grease [55]. Deformation of the structure
causes a change in geometry of the channels and their
electrical resistance and through measuring the change in
resistance, strain can be calculated. This type of sensors
can be tuned by modulation of channel geometries and
can measure various types of strains [51–53]. For the fab-
rication of the sensors, various approaches such as mask
deposition of the conductor [56], and direct 3D printing
of conductive material [55], have been suggested recently.
As for the sensing methods, not only resistive changes ac-
cording to deformation of sensor’s hardware but also ca-
pacitive [54] and inductive changes [57,58] can be used.
The capacitive approach is applied to elastomer layers
patterned with microfluidic channels, and the inductive
change is used to measure displacement of the McKibben
actuators. The geometry of conductive fibers surrounding
McKibben actuator changes as it contracts or stretches,
and the inductance of the structure changes with it.
IPMC, which was described in the actuation section,
is a smart composite material exhibiting characteristics
of both actuators and sensors (energy harvester) and this
property makes it unique [42]. IPMC based sensors are
also employed in soft robots [59]. IPMC demonstrates
a linear voltage output when quasi-static displacement is
placed at the tip of IPMC. Besides, IPMC also responds to
dynamic stimulus such as impact or shock loading, and a
damped electrical variation is observed [59,60]. Although
both IPMC and piezoelectric sensor can detect mechanical
stimulus, IPMC has higher sensitivity. Moreover, IPMC is
suited for sensing any modes of deformation such as bend-
ing, tension, compression, torsion and shear.
Although, the application of soft sensors in soft robots
is not very prevalent, they are an important component for
attaining accurate control of soft robots. Moreover, the
applicability of the soft sensors extends beyond the soft
robot. Recently in eHealthcare areas, unobtrusive sens-
ing and wearable devices are receiving much attention
[61], and flexible and stretchable soft sensors with low
impedance are suitable for this purpose.
2.3. Structure
The structures of soft robots are unlike those of rigid-
body robots. This is because soft robots are usually manu-
factured using soft materials and therefore are not suitable
for exploiting conventional electric motors. In this sense,
researchers have developed several special structures for
soft robotic systems such as those made up of soft materi-
als, mainly silicone rubber and SMA, soft materials with
medium enclosed in them.
For example, GoQBot is a soft robotic system with the
structure based on silicone and SMA [21]. This system
consists of a flexible silicone body and two tensile actu-
ators. To generate forward displacement, this system is
divided into two sections: an anterior part and a posterior
part. Meshworm is one another case which uses this type
of structure [20]. To achieve sequential antagonistic mo-
tion, this system mimics circular and longitudinal muscle
structure and arrangement of oligochaetes, and uses mesh
materials and NiTi SMA. Lastly, an octopus soft robot
also uses the structure based on silicone and SMA and
engages in biomimetic [22]. The structure of this system
directly mimics the muscle system of octopus. It consists
of four longitudinal actuators and one radially allocated
actuator, which represent four longitudinal muscle bun-
dles and transverse muscles of octopus respectively and
resemble the tissues.
Researchers have also developed soft robotic systems
using some medium encased within the silicone material.
Manta swimming robot is one of the kind [39]. To achieve
two degrees of freedom of bending motion, it used sil-
icone rubber which has two internal chambers. A soft
snake robot also employs similar structure [29]. In this
research, they designed and developed a snake-like robot

Soft Robot Review 7
system based on fluidic system. The proposed system con-
sists of four fluidic actuator parts and passive wheels be-
tween every two actuator parts. A multigait soft robot is
also developed based on this kind of structure [26]. One
different feature is that it is composed of elastomeric poly-
mers inspired by animal tissue. The proposed system con-
sists of five parts which can be independently actuated
by pneumatic valves. Based on similar concept, a soft
robot gripper has been also developed [35,62]. This grip-
per can be actuated based on pneumatic system. Since it
used soft materials, it offers enhanced safety compared to
a rigid system. Jamming skin enabled locomotion is a soft
robotic system inspired by a different actuating system to
the previous works described but shares similar structure
[30]. It consists of several silicone cells and an expand-
able actuator, and is filled with fluid. The silicone cells
can be divided into two types, jammed and unjammed
cells. Similarly, a jammable manipulator using granular
media was developed [19]. This robot is composed of
five serial jammable parts. Unlike other granular systems,
ground coffee has been used as granular media. Similarly,
a universal robotic gripper, which can be integrated with a
6-axis robot arm as an end-effector, has also been devel-
oped [63,64]. The structure of this system is quite simple.
This system do not have many parts like other silicone and
medium based system, it only has a single compartment
filled with granular material.
As for other soft robotic systems, there are some sys-
tems which only use soft material for their actuating parts,
not in overall systems. Soft robotic fish are examples of
this kind of structure which mimicked fish’s body [65,66].
However, the system can be seen as a rigid robot system
as a whole except for the part from dorsal section to anal
section. This is due to the purpose of actuating its tail in
order to provide force to navigate the entire body, like real
fish does. SmartBird is another system which utilizes sim-
ilar structure [67]. As it can be inferred from the name of
the system, it mimics birds for developing a system capa-
ble of flying. Similar with soft robotic fish, most of its
system can be considered as rigid-body system. However,
it utilizes soft material on wings of the system mimicking
material property of a real bird wing in order to achieve
the better performance.
2.4. Control and electronics
Control methods used for soft robot can be classified
into two categories, open-loop control and closed-loop
control. The most representative philosophy of soft robots
is to exploit material properties, in particular, softness.
Because of this feature, it is important to consider compli-
ance enabling the robots’ body and functions to be suited
to their environments. This aspect makes it difficult to
implement suitable sensors or acquire a reliable dynamic
model, based on which the controller can be designed.
Accordingly, there have been many attempts to develop
soft robots without using specific sensors or feedback con-
trol, which is referred to open-loop control. Whereas
many problems are encountered for open-loop control us-
ing compensated input to determine output, closed-loop
control enables more accurate and robust actuation of im-
plemented systems. Through utilizing proprioceptive sen-
sors, closed-loop control reflects the outputs of a system to
make control inputs. With the dynamic model, it is possi-
ble to design an efficient controller which takes the char-
acteristics of the system into account. In soft robot ap-
plication, linear model based adaptive controllers [68,69]
were introduced. Additionally, model predictive control
[70] for pneumatic actuation and linear quadratic regula-
tor for SMA actuator were reported. Even though sev-
eral cases of soft robot control have been reported, con-
trol still remains as the main challenge in soft robotics
because most of the control approaches are not yet gen-
erally defined compared to conventional robots. Further-
more, there should be more progress on soft robot control
for autonomous behavior, so that the robot can be used
in multiple cases: high level tasks, cognition, and inter-
actions with their environment. Online learning may help
configure models or conduct tasks in unstructured envi-
ronments where there are many uncertainties.
Soft robot electronics have employed rigid electronics
to hold control algorithms and interconnect the systems’
actuators, sensors, and power sources [2]. This rigid elec-
tronics is a mismatch with biological materials which are
soft, elastic, and curved [71]. There has been a progress
in electronics area that allows flexible and stretchable na-
ture. This is also known as “stretchable electronics”. Soft
robots have barely incorporated a soft electronic system
for all the soft robotic parts as of now but for enhanced
performance and better suitability to the environment es-
pecially biological application, use of stretchable or flexi-
ble electronics is crucial.
Rigid electronics are either incorporated in soft robot
itself becoming a part of the soft robot [25,40,66,72,73]
or housed externally in a physically different location to
the soft robot body [22,29,34,63,71,74,75]. Usually,
manipulation system has a rigid body consisting of power
source, control system, and other necessary electronics,
and a soft robotic arm/manipulator. The speed of pro-
cess, commercial availability, cost and level of integration
are the major benefits of using rigid electronics. How-
ever, rigid electronics have limitations to miniaturization,
biointegrity, withstanding extreme mechanisms including
stretching, twisting, and compressing.
2.5. Materials
Materials of soft robotic systems were mostly found
to be the silicone or rubber because of the fact that
the soft robotic system should perform flexible motion
[19–22,29,30,39,62,65]. However, soft robotic systems
such as a multigait soft robot utilized elastomeric poly-

8Chiwon Lee, Myungjoon Kim, Yoon Jae Kim, Nhayoung Hong, Seungwan Ryu, H. Jin Kim, and Sungwan Kim
mers in order to attain more flexibility [26]. SMA was
also widely used materials for soft robotic systems. Since
a conventional electrical motor cannot be directly used in
soft robots, SMA is chosen to be actuators [20–22].
2.6. Fabrication and system
The enhancement of fabrication technologies of soft
material, such as shape deposition manufacturing (SDM),
smart composite microstructure (SCM), and 3D printing
technology, has a close relationship with the growth of re-
search activities related with soft robots. This is because
the fabrication method for manufacturing rigid robot sys-
tem is not appropriate for the fabrication of soft material
and therefore the fabrication technology for soft material
is highly demanded.
SDM manufacturing technology, which was developed
for rapid-prototyping of rigid material [76], was first used
for developing force sensing robotic finger [77]. The ad-
vantage of this technology is that it could print both rigid
and soft materials to fabricate an integrated part. Thus,
SDM has been widely used for developing soft robotic
system [78–80]. Similarly, some researchers used a cus-
tom designed mold to fabricate their systems [81]. The
SCM manufacturing process is also a well-used type of
soft robot fabrication technique. SCM enables the fab-
rication of mesoscale gears and links to be directly used
in mesoscale system whereas traditional parts cannot due
to large friction. Based on SCM, polymer films and re-
inforced prepregs can be integrated layer by layer. Thus,
soft gear and link structure can be fabricated in mesoscale.
Several systems have been manufactured based on this
process [82–84]. 3D printing technology has also brought
many advantages to fabricating soft robot system since
any 3D model can be printed. Therefore, it has become
relatively easy to fabricate soft-robot-related parts. Ac-
cordingly, production costs and time have been reduced
compared with traditional fabrication methods. GoQBot
is an example of 3D printed soft robot system [21]. Based
on 3D printing technology, ABS plastic molds can be con-
structed and using this plastic mold, silicone rubber can
be made into the desired shape. There is also a soft robot
system in which the body was entirely fabricated using 3D
printing technology [85].
Traditional robotic systems involve complex calcula-
tions and constant analysis of the system, and the environ-
ment limits tasks to repetitive and non-versatile functions
reliable only in pre-defined conditions. There is a grow-
ing need for more versatile and reliable robotic systems
which can adapt to uncertain and/or versatile conditions
and perform required tasks. In this section, how soft robot
systems differ from the rigid robot system and analysis of
current soft robot systems are discussed.
There are essentially two types of systems for soft
robots in control mechanism: continuum robots with in-
verse kinetic algorithms and with neural network algo-
rithms [86]. Caterpillar-like soft robot is controlled by in-
ternal compressive force through inverse kinematics and
dynamics [87]. Many of the soft robots developed are
continuum robots which include modeling of the system
by piecewise-constant-curvature approximation (PCCA)
[21,88–90]. Another continuum manipulator control sys-
tems include steady-state model inspired by the octopus
arm and its speed was faster than conventional PCCA [91].
Soft robotic systems overcome the limitations of rigid
robots, such as limited degree of freedom and need for
complex calculations, through embodied intelligence and
morphological computation [92,93]. Embodied intelli-
gence highlights the interaction of morphological struc-
ture and environment to shape adaptive behavior and these
results in the adaptability and robustness seen in living
organisms. This control concept, which usually involves
neural network, could be quite different from the current
control mechanisms for conventional robots, estimating
unknown model dynamics as it does not involve com-
plex analytical calculations. Neural network is typically
trained using the backpropagation algorithm [94].
Soft robotic octopus arm mimicked the adaptability of
octopus’ morphology, soft and compliant skin with lon-
gitudinal and transverse muscular structures, to the en-
vironment achieving variable stiffness and good dexter-
ity [92]. A feed-forward neural network learning of soft
cable-driven manipulator to perform grasping of various
objects proved its feasibility of conventional inverse dy-
namic calculation methods [86,95,96]. PhysX utilizes
spiking neural networks to perform gait motion [97].
2.7. Applications
Soft robots have a wide variety of applications rang-
ing from industrial to medical usages. In this section, we
will introduce soft robots developed upto date according
to their potential application areas; human-machine inter-
face and interaction, locomotion and exploration, manip-
ulation, medical and surgical applications, rehabilitation,
and wearable robots.
2.7.1 Human-machine interface and interaction
Soft robots are essentially more compatible for human
interactions as their soft and easily deformable bodies en-
sure a minimal damage and load given to the human and
environment [98]. Soft robot’s ability to adapt to curved
and irregular surfaces allows overcoming the shortcom-
ings of rigid robots. A robotic system much like the bi-
ological system in terms of material and mechanics al-
lows better human-machine interaction in almost all ar-
eas including but not limited to medical, healthcare, pack-
aging, etc. Human-machine interface is important espe-
cially for the medical robots since communication of ac-
curate and real-time data is essential for safe and robust re-
sponse. Data can be communicated in largely three ways;
by wire [30,36,37,39,63,64], wireless [66,74] and in-

Soft Robot Review 9
dependent controller with determined action algorithm or
self-decision making algorithm [20,25,31,32,99]. Most
studies require human interventions to perform complex
tasks and few simple tasks are self-determined.
2.7.2 Locomotion and exploration
Soft robots are capable of performing locomotions that
cannot be seen in rigid robots and thus are able to navi-
gate and explore unknown terrains. They are also able to
more efficiently and more robustly move in environments
where rigid robots have difficulties such as underwater.
Studies of caterpillar-like and underwater creatures have
inspired many soft robotic systems in control and actua-
tion methods. They perform rolling, jumping, and crawl-
ing motion. GoQBot simulates caterpillar movement and
rolls for locomotion [21]. A highly deformable 3D printed
soft robot can inch and crawl like a catepillar and can gen-
erate dynamically-shape-dependent frictional force differ-
entiated from conventional rigid robots [85]. Inchworm-
inspired soft robot composed of smart soft composite per-
form locotion via abdominal contraction [100]. Mesh-
worm involves peristaltic locomotion similar to that of
an earthworm and the use of flexible mesh materials al-
lows external shock absorption [20]. Self-contained ser-
pentine soft robot actuated by FEAs can synthesize snake-
like motion with wheels at the current stage [29]. Another
animal-inspired soft robot is the multigait soft robot which
produces complex gait motion with relatively simple ac-
tuation [26]. A more advanced version of this robot is
untethered quadrupedal soft robot and it is capable of ex-
ploring indoor and harsh conditions outdoor environments
with the goal of surveilance [25].
Many underwater animals have also inspired many soft
robots to move robustly and in a controlled manner in wa-
ter. The soft robotic fish has a rigid head with conven-
tional electronics housed within and a soft-bodied tail to
perform diving, turning and swimming motion like fish
[40]. The octopus soft robot mimics the movement of oc-
topus through utilizing eight legs and has the ability to
travel through small apertures and unstructured surfaces
as well as perform grasping using its leg [101,102]. The
shell-like soft robot inspired by cephalopds propells itself
underwater by dynamic activity of its shell [103].
There are non-animal inspired soft robots. A circu-
lar deformable soft robot can crawl and jump to move in
a rough terrain [104]. Untethered robots perform jump-
ing by pneumatic and explosive actuators [32,105]. 3D
printed soft robot powered by combustion with on board
electronics [106].
2.7.3 Manipulation
Rigid robots have a natural limitation to the types of
manipulation they can perform due to low DOF, complex-
ity, and difficulty in calculation of grasping action. Soft
robotic manipulators are more compliant and can manipu-
late fragile and unknown objects by a simple control algo-
rithm having advantage over rigid robots [2]. Soft robot
manipulators engage in either granular or octopus-like
manipulation methods. A tentacle-like manipulator with
embedded pneumatic networks allows grasping of variety
of materials, and other functionalities can be added such as
needle and visual system [107]. OctArm can grasp objects
with varying size, shape, and payload under dynamic dis-
turbances [34]. Many other manipulators are inspired by
octopus or other cephalopod limbs [18,108]. A multiseg-
ment soft manipulator with real time positional feedback
system have autonomous grasp-and-place ability [109].
2.7.4 Medical and surgical applications
Soft robots inherently have advantage of being compli-
ant with the natural tissues of human and living organ-
isms. Minimally invasive surgery (MIS) is one of the re-
search areas with big potential of adopting soft robotics.
This is because it overcomes limitation of traditional MIS
methods such as low DOFs [110]. A controllable-stiffness
in combination with granular jamming system can be uti-
lized in laparoscopic surgery and endoscopy [111]. Stiff
flop surgical manipulator capable of modifying its me-
chanical properties as needed and stiffness-controllable
end-effectors for minimizing damage to surrounding soft
tissues and increases accessibility in vivo [75,112]. A cu-
cumber tendril-inspired tactile sensor sleeve for soft ma-
nipulators for use in MIS can be used to overcome limita-
tion in lack of haptic feedback in MIS [113]. A soft en-
doscopic system inspired by elephant’s trunk is developed
for cardiac ablation in MIS allowing access via confined
space and manipulation [114].
2.7.5 Rehabilitation and wearable robots
Biocompatibility and biointegrity are vital criteria for
wearable and human assistant applications and these are
the main advantages of soft robot over conventional rigid
robot systems. Since soft robots utilize elastic and soft
material, they exhibit mechanical properties similar to that
of living organisms and therefore, are compliant to be used
in wearable devices supporting functions of humans and
perhaps other animals [98]. Use of soft and elastic mate-
rials absorbs mechanical stress and minimizes the chance
of injury to both the robot itself and the user.
Soft robotic gloves includes soft actuators for use in
hand rehabilitation for patients with grasp pathologies
[41]. Similarly, there exists a soft robotic glove for thumb
rehabilitation [115]. Another soft robotic glove for hand
rehabilitation with human-machine interaction (with clin-
ician) performs specific tasks for training [116]. There are
many other works on hand rehabilitation by soft exoskele-
ton systems [117,118].
Other wearable soft robotic devices exist for rehabilita-
tion and other applications. The gait rehabilitation soft
robot for spinalized rodents uses soft pneumatic actua-

10 Chiwon Lee, Myungjoon Kim, Yoon Jae Kim, Nhayoung Hong, Seungwan Ryu, H. Jin Kim, and Sungwan Kim
tors in conjunction with rigid frames and their inherent
softness is suitable for interacting with soft tissues [119].
Another gait assisting soft robotic exosuit exists for hu-
mans, and it can be worn like normal clothing, is light
and minimizes unintentional interference with the wearer
[120]. The ankle-foot rehabilitation wearable robotic
device utilizes pneumatic artificial muscle actuators and
mimic muscle-tendon ligament morphology and function
[121]. There are many stretchable sensors; a tremor neu-
robot is a form of sleeve which assesses and attenuates
pathological tremors and can satisfy aesthetic appearance
preferred by the patients [122]. There is a soft oral reha-
bilitation robot for people with mandibular mobility dis-
orders [123].
3. DISCUSSIONS
In terms of actuation, three major types, variable length
tendon, fluidic actuation, and EAP, have been used. Es-
pecially, pneumatic pressure based fluidic actuation has
been frequently applied to wide application including lo-
comotion, gripping, and rehabilitation. Major advantages
of each actuation approach have been discussed. Sensing
is an important function for proprioception of soft robots.
For the least effect on the impedance of robot hardware,
sensors with low modulus are recommended. Multiple-
layered elastomer filled with liquid conductor and IPMC
have been frequently reported by the previous researchers.
As for the structure of the soft robotic systems, they have
been mostly inspired by biomimetics, which is actually the
former stage of soft robotics. However, researchers started
to use some entirely new approaches and new materials to
develop soft robotic systems since the biomimetics also
suffers from several limitations. Soft robotic systems re-
quire less complex calculations and therefore can perform
more versatile tasks than rigid robots such as packaging
various goods in industries. Due to the absence or lack
of hard components and their flexible nature, soft robots
are distinguishably more compatible with biological ma-
terials and humans. These appealing properties of soft
robots create new opportunity areas where conventional
robots have come to limit like exploration, manipulation,
medical, rehabilitation, and wearable fields. A hybrid sys-
tem much like many living creatures would be the most
promising system comprising benefits of both.
Soft robots will offer unique properties and therefore
opportunities which are previously unexplored. Recently,
there is a growing need for devices which can interact
with human such as smart watches and unmanned vehi-
cles. Soft robots are suited for interaction with humans
due to their soft and compliant nature and their ability to
perform works with uncertainties. They minimize damage
to the user and the environment via shock-absorbability
and are more compatible and reliable with uses regard-
ing biological materials owing to their flexibility. User-
friendly mechanical properties of soft robots are the re-
sults of advances in new flexible materials and electronics.
4. CONCLUDING REMARKS
Soft robot is a rapidly growing area in robotics as draw-
backs of conventional robots such as human-machine in-
teraction and adaptability are alleviated. Also, robots
have been evolved from performing labor-intensive, repet-
itive, and simple tasks to more interactive, dexterous, and
high-level tasks. The demands for robots which interact
with humans in the areas of military, medicine, rehabilita-
tion, assistive technology, etc. will continue to grow and
soft robots could possibly answer these demands. Future
soft robots with embedded artificial intelligence could be
developed into self-contained navigating (surveillance of
difficult to reach areas, rescue mission in disaster zones,
etc.) and manipulating (packaging and organization of
goods) systems which bring industrial, medical and mil-
itary benefits. Soft robots are usually inspired by the liv-
ing creatures and mimic their muscular (actuation), skin
(barrier) and, observatory organs (sensors) which are ac-
knowledged more efficient and robust movements. This
gives soft robots many benefits compared to conventional
robots such as ability to compensate uncertainties in the
environments and safe interaction with humans and living
organisms. Many animals are vertebrates or insects hav-
ing rigid structural frameworks along with soft materials,
so a mixture of soft robots with conventional robots will
be essential to maximize its benefits to achieve the best
solution.
This review paper has covered a wide range of soft
robots being developed over the past few decades and
would benefit soft robot research community by provid-
ing overview with some details and up-to-date informa-
tion. Also, detailed information on the research trends as
well as advantages and drawbacks of each soft robot can
be found in this review paper.
REFERENCES
[1] M. Hülse, S. Wischmann, and F. Pasemann, The Role of
Non-linearity for Evolved Multifunctional Robot Behavior,
Springer Berlin Heidelberg, Berlin, Heidelberg, 2005.
[2] D. Rus and M. T. Tolley, “Design, fabrication and control
of soft robots,” Nature, vol. 521, pp. 467-475, 2015. [click]
[3] Untethered soft robot. https://www.youtube.com/watch?v=
_OJrwCP24cI
[4] Soft Robot Walking and Crawling. https://www.youtube.
com/watch?v=2DsbS9cMOAE
[5] Soft Robot Uses Explosions to Jump. https://www.
youtube.com/watch?v=dkUtNPwm2wc
[6] iRobot’s Soft Morphing Blob ’Bot Takes Its First Steps.
https://www.youtube.com/watch?v=SbqHERKdlK8

Soft Robot Review 11
[7] Field experiments with the OctArm continuum manipula-
tor. https://www.youtube.com/watch?v=6fhpt6MWDDE
[8] A novel type of compliant, underactuated robotic hand
for dexterous grasping. https://www.youtube.com/watch?
v=Tbg-vo5Tx34
[9] Soft Robotics 03: Manta Robot. https://www.youtube.com/
watch?v=mTjhXKIXc90
[10] Autonomous, self-contained soft robotic fish at MIT. https:
//www.youtube.com/watch?v=BSA_zb1ajes
[11] Soft Robotic Glove. https://www.youtube.com/watch?v=
Ef6Ebc8RtLQ
[12] Octopus-Inspired Robots Can Grasp, Crawl, and Swim.
https://www.youtube.com/watch?v=L7FEJJsvHRQ
[13] Soft autonomous earthworm robot at MIT. https://www.
youtube.com/watch?v=EXkf62qGFII
[14] GoQBot. https://www.youtube.com/watch?v=a-
1AiExU3Vk
[15] Universal robotic gripper based on the jamming of
granular material.https://www.youtube.com/watch?v=
bFW7VQpY-Ik
[16] MIT’s Jammable Robot Manipulator. https://www.
youtube.com/watch?v=0DNJwM8lyBo
[17] Y. Bahramzadeh and M. Shahinpoor, “A review of ionic
polymeric soft actuators and sensors,” Soft Robotics, vol.
1, pp. 38-52, 2014. [click]
[18] M. Calisti, M. Giorelli, G. Levy, B. Mazzolai, B. Hochner,
C. Laschi, and P. Dario, “An octopus-bioinspired solution
to movement and manipulation for soft robots,” Bioinspi-
ration & Biomimetics, vol. 6, p 036002, 2011. [click]
[19] N. G. Cheng, M. B. Lobovsky, S. J. Keating, A. M. Se-
tapen, K. I. Gero, A. E. Hosoi, and K. D. Iagnemma, “De-
sign and analysis of a robust, low-cost, highly articulated
manipulator enabled by jamming of granular media,” 2012
IEEE International Conference on Robotics and Automa-
tion (ICRA), pp. 4328-4333, 2012.
[20] S. Seok, C. D. Onal, K. J. Cho, R. J. Wood, D. Rus, and
S. Kim, “Meshworm: A Peristaltic Soft Robot With An-
tagonistic Nickel Titanium Coil Actuators,” IEEE/ASME
Transactions on Mechatronics, vol. 18, pp. 1485-1497,
2013.
[21] H. T. Lin, G. G. Leisk, and B. Trimmer, “GoQBot: a
caterpillar-inspired soft-bodied rolling robot,” Bioinspir
Biomim, vol. 6, p 026007, 2011. [click]
[22] C. Laschi, M. Cianchetti, B. Mazzolai, L. Margheri, M.
Follador, and P. Dario, “Soft robot arm inspired by the
octopus,” Advanced Robotics, vol. 26, pp. 709-727, 2012.
[click]
[23] H. Schulte Jr, “The characteristics of the McKibben artifi-
cial muscle,” The Application of External Power in Pros-
thetics and Orthotics (Washington, DC: Nat. Acad. Sci.-
Nat. Res. Council), vol., 1961.
[24] C.-P. Chou and B. Hannaford, “Measurement and mod-
eling of McKibben pneumatic artificial muscles,” IEEE
Transactions on Robotics and Automation, vol. 12, pp. 90-
102, 1996.
[25] M. T. Tolley, R. F. Shepherd, B. Mosadegh, K. C. Gal-
loway, M. Wehner, M. Karpelson, R. J. Wood, and G.
M. Whitesides, “A Resilient, Untethered Soft Robot,” Soft
Robotics, vol. 1, pp. 213-223, 2014.
[26] R. F. Shepherd, F. Ilievski, W. Choi, S. A. Morin, A. A.
Stokes, A. D. Mazzeo, X. Chen, M. Wang, and G. M.
Whitesides, “Multigait soft robot,” Proc Natl Acad Sci U
S A, vol. 108, pp. 20400-20403, 2011. [click]
[27] B. Mosadegh, P. Polygerinos, C. Keplinger, S. Wennst-
edt, R. F. Shepherd, U. Gupta, J. Shim, K. Bertoldi, C.
J. Walsh, and G. M. Whitesides, “Pneumatic networks for
soft robotics that actuate rapidly,” Advanced Functional
Materials, vol. 24, pp. 2163-2170, 2014. [click]
[28] S. A. Morin, R. F. Shepherd, S. W. Kwok, A. A. Stokes, A.
Nemiroski, and G. M. Whitesides, “Camouflage and dis-
play for soft machines,” Science, vol. 337, pp. 828-832,
2012. [click]
[29] D. O. Cagdas and R. Daniela, “Autonomous undulatory
serpentine locomotion utilizing body dynamics of a flu-
idic soft robot,” Bioinspiration & Biomimetics, vol. 8, p
026003, 2013. [click]
[30] E. Steltz, A. Mozeika, N. Rodenberg, E. Brown, and H.
M. Jaeger, “JSEL: Jamming skin enabled locomotion,”
Proc. of IEEE/RSJ International Conference on Intelligent
Robots and Systems, pp. 5672-5677, 2009.
[31] C. D. Onal, X. Chen, G. M. Whitesides, and D. Rus, “Soft
mobile robots with on-board chemical pressure genera-
tion,” Proc. of International Symposium on Robotics Re-
search, pp. 1-16, 2011.
[32] M. T. Tolley, R. F. Shepherd, M. Karpelson, N. W. Bartlett,
K. C. Galloway, M. Wehner, R. Nunes, G. M. White-
sides, and R. J. Wood, “An untethered jumping soft robot,”
Proc. of IEEE/RSJ International Conference on Intelligent
Robots and Systems, pp. 561-566, 2014.
[33] K. Suzumori, S. Iikura, and H. Tanaka, “Applying a flex-
ible microactuator to robotic mechanisms,” Control Sys-
tems, IEEE, vol. 12, pp. 21-27, 1992.
[34] W. McMahan, V. Chitrakaran, M. Csencsits, D. Dawson, I.
D. Walker, B. A. Jones, M. Pritts, D. Dienno, M. Grissom,
and C. D. Rahn, “Field trials and testing of the OctArm
continuum manipulator,” Proc. of IEEE International Con-
ference on Robotics and Automation, 2006. ICRA 2006.,
pp. 2336-2341, 2006.
[35] F. Ilievski, A. D. Mazzeo, R. F. Shepherd, X. Chen, and G.
M. Whitesides, “Soft robotics for chemists,” Angew Chem
Int Ed Engl, vol. 50, pp. 1890-1895, 2011. [click]
[36] R. Deimel and O. Brock, “A novel type of compliant and
underactuated robotic hand for dexterous grasping,” The
International Journal of Robotics Research, vol., 2015.
[37] S. Sanan, P. S. Lynn, and S. T. Griffith, “Pneumatic tor-
sional actuators for inflatable robots,” Journal of Mecha-
nisms and Robotics, vol. 6, pp. 031003-031003, 2014.
[38] A. A. Stokes, R. F. Shepherd, S. A. Morin, F. Ilievski,
and G. M. Whitesides, “A hybrid combining hard and soft
robots,” Soft Robotics, vol. 1, pp. 70-74, 2014. [click]

12 Chiwon Lee, Myungjoon Kim, Yoon Jae Kim, Nhayoung Hong, Seungwan Ryu, H. Jin Kim, and Sungwan Kim
[39] K. Suzumori, S. Endo, T. Kanda, N. Kato, and H.
Suzuki, “A bending pneumatic rubber actuator realizing
soft-bodied manta swimming robot,” Proc. of IEEE In-
ternational Conference on Robotics and Automation, pp.
4975-4980, 2007.
[40] R. K. Katzschmann, A. D. Marchese, and D. Rus, “Hy-
draulic autonomous soft robotic fish for 3D swimming,”
Experimental Robotics, pp. 405-420, 2016. [click]
[41] P. Polygerinos, Z. Wang, K. C. Galloway, R. J. Wood,
and C. J. Walsh, “Soft robotic glove for combined as-
sistance and at-home rehabilitation,” Robotics and Au-
tonomous Systems, vol. 73, pp. 135-143, 2015. [click]
[42] K. J. Kim and S. Tadokoro, “Electroactive polymers for
robotic applications,” Artificial Muscles and Sensors (291
p.), Springer: London, United Kingdom, vol., 2007.
[43] E. N. Gama Melo, O. F. Aviles Sanchez, and D. Amaya
Hurtado, “Anthropomorphic robotic hands: a review,” In-
geniería y Desarrollo, vol. 32, pp. 279-313, 2014.
[44] Y. Bar-Cohen, “EAP as artificial muscles: progress and
challenges,” Smart Structures and Materials, pp. 10-16,
2004. [click]
[45] Y. Nakabo, T. Mukai, K. Ogawa, N. Ohnishi, and K. Asaka,
“Biomimetic soft robot using artificial muscle,” Proc. of
IEEE/RSJ Int. Conf. Intelligent Robots and Systems, in tu-
torial, WTP3 Electro-Active Polymer for Use in Robotics,
2004.
[46] K. Ogawa, Y. Nakabo, T. Mukai, K. Asaka, and N. Ohnishi,
“A snake-like swimming artificial muscle,” Proc. of The
2nd Conf. on Artificial Muscles, Osaka, 2004.
[47] Y. Nakabo, T. Mukai, and K. Asaka, “A multi-DOF robot
manipulator with a patterned artificial muscle,” Proc. of
The 2nd Conf. on Artificial Muscle, Osaka, 2004.
[48] Y. Nakabo, T. Mukai, and K. Asaka, “Kinematic modeling
and visual sensing of multi-DOF robot manipulator with
patterned artificial muscle,” Proc. of IEEE International
Conference on Robotics and Automation, ICRA 2005, pp.
4315-4320, 2005.
[49] S. Guo, T. Fukuda, and K. Asaka, “A new type of fish-
like underwater microrobot,” IEEE/ASME Transactions on
Mechatronics, vol. 8, pp. 136-141, 2003.
[50] M. Mojarrad and M. Shahinpoor, “Biomimetic robotic
propulsion using polymeric artificial muscles,” Robotics
and Automation, 1997. Proceedings., 1997 IEEE Interna-
tional Conference on, pp. 2152-2157, 1997.
[51] D. M. Vogt, Y.-L. Park, and R. J. Wood, “Design and char-
acterization of a soft multi-axis force sensor using embed-
ded microfluidic channels,” IEEE Sensors Journal, vol. 13,
pp. 4056-4064, 2013.
[52] C. Majidi, R. Kramer, and R. Wood, “A non-differential
elastomer curvature sensor for softer-than-skin electron-
ics,” Smart Materials and Structures, vol. 20, p 105017,
2011. [click]
[53] R. K. Kramer, C. Majidi, R. Sahai, and R. J. Wood, “Soft
curvature sensors for joint angle proprioception,” Proc. of
IEEE/RSJ International Conference on Intelligent Robots
and Systems, pp. 1919-1926, 2011.
[54] R. D. P. Wong, J. D. Posner, and V. J. Santos, “Flexible
microfluidic normal force sensor skin for tactile feedback,”
Sensors and Actuators A: Physical, vol. 179, pp. 62-69,
2012. [click]
[55] J. T. Muth, D. M. Vogt, R. L. Truby, Y. Mengüç, D. B.
Kolesky, R. J. Wood, and J. A. Lewis, “Embedded 3D
printing of strain sensors within highly stretchable elas-
tomers,” Advanced Materials, vol. 26, pp. 6307-6312,
2014. [click]
[56] R. K. Kramer, C. Majidi, and R. J. Wood, “Masked deposi-
tion of Gallium-Indium alloys for liquid-embedded elas-
tomer conductors,” Advanced Functional Materials, vol.
23, pp. 5292-5296, 2013. [click]
[57] W. Felt and C. D. Remy, “Smart braid: air muscles that
measure force and displacement,” Proc. of IEEE/RSJ In-
ternational Conference on Intelligent Robots and Systems,
pp. 2821-2826, 2014.
[58] W. Felt, K. Y. Chin, and C. D. Remy, “Contraction sensing
with smart braid McKibben muscles,” IEEE/ASME Trans-
actions on Mechatronics, vol. 21, pp. 1201-1209, 2016.
[59] M. Shahinpoor, Y. Bar-Cohen, J. Simpson, and J. Smith,
“Ionic polymer-metal composites (IPMCs) as biomimetic
sensors, actuators and artificial muscles-a review,” Smart
Materials and Structures, vol. 7, p R15, 1998. [click]
[60] M. Mojarrad and M. Shahinpoor, “Ion-exchange-metal
composite sensor films,” Smart Structures and Materials’
97, pp. 52-60, 1997.
[61] Y.-L. Zheng, X.-R. Ding, C. C. Y. Poon, B. P. L. Lo, H.
Zhang, X.-L. Zhou, G.-Z. Yang, N. Zhao, and Y.-T. Zhang,
“Unobtrusive sensing and wearable devices for health in-
formatics,” IEEE Transactions on Biomedical Engineer-
ing, vol. 61, pp. 1538-1554, 2014.
[62] K. C. Galloway, K. P. Becker, B. Phillips, J. Kirby, S.
Licht, D. Tchernov, R. J. Wood, and D. F. Gruber, “Soft
robotic grippers for biological sampling on deep reefs,”
Soft Robotics, vol. 3, pp. 23-33, 2016. [click]
[63] E. Brown, N. Rodenberg, J. Amend, A. Mozeika, E. Steltz,
M. R. Zakin, H. Lipson, and H. M. Jaeger, “Universal
robotic gripper based on the jamming of granular material,”
Proceedings of the National Academy of Sciences, vol. 107,
pp. 18809-18814, 2010. [click]
[64] J. R. Amend, E. Brown, N. Rodenberg, H. M. Jaeger, and
H. Lipson, “A positive pressure universal gripper based on
the hamming of granular material,” IEEE Transactions on
Robotics, vol. 28, pp. 341-350, 2012.
[65] J. L. Tangorra, A. P. Mignano, G. N. Carryon, and J.
C. Kahn, “Biologically derived models of the sunfish for
experimental investigations of multi-fin swimming,” 2011
IEEE/RSJ International Conference on Intelligent Robots
and Systems, pp. 580-587, 2011.
[66] A. D. Marchese, C. D. Onal, and D. Rus, “Autonomous
soft robotic fish capable of escape maneuvers using fluidic
elastomer actuators,” Soft Robotics, vol.1, pp. 75-87, 2014.
[click]

Soft Robot Review 13
[67] R. Stopforth and G. Bright, “MechaBird: a biological in-
spired mechatronics bird for the evaluation of flight charac-
teristics,” Proc. of IEEE International Conference on Au-
tomation Science and Engineering (CASE), pp. 335-341,
2015.
[68] D. G. Caldwell, G. A. Medranocerda, and M. Goodwin,
“Control of pneumatic muscle actuators,” IEEE Control
Systems Magazine, vol. 15, pp. 40-48, 1995.
[69] D. G. Caldwell, G. A. Medrano-Cerda, and M. Goodwin,
“Characteristics and adaptive control of pneumatic muscle
actuators for a robotic elbow,” Robotics and Automation,
1994. Proceedings., 1994 IEEE International Conference
on, pp. 3558-3563, 1994.
[70] M. T. Gillespie, C. M. Best, and M. D. Killpack, “Si-
multaneous position and stiffness control for an inflatable
soft robot,” Proc. of IEEE International Conference on
Robotics and Automation (ICRA), pp. 1095-1101, 2016.
[71] J. A. Rogers, T. Someya, and Y. Huang, “Materials and
mechanics for stretchable electronics,” Science, vol. 327,
pp. 1603-1607, 2010. [click]
[72] K. H. Low, J. Yang, A. P. Pattathil, and Y. Zhang, “Initial
prototype design and iInvestigation of an undulating body
by SMA,” Proc. of IEEE International Conference on Au-
tomation Science and Engineering, pp. 472-477, 2006.
[73] Z. Chen, T. I. Um, J. Zhu, and H. Bart-Smith, “Bio-inspired
robotic cownose ray propelled by electroactive polymer
pectoral fin,” ASME 2011 International Mechanical Engi-
neering Congress and Exposition, pp. 817-824, 2011.
[74] A. A. Stokes, R. F. Shepherd, S. A. Morin, F. Ilievski,
and G. M. Whitesides, “A hybrid combining hard and soft
robots,” Soft Robotics, vol. 1, pp. 70-74, 2013. [click]
[75] M. Cianchetti, T. Ranzani, G. Gerboni, I. D. Falco, C.
Laschi, and A. Menciassi, “STIFF-FLOP surgical manip-
ulator: Mechanical design and experimental characteriza-
tion of the single module,” 2013 IEEE/RSJ International
Conference on Intelligent Robots and Systems, pp. 3576-
3581, 2013.
[76] R. Merz, F. B. Prinz, K. Ramaswami, M. Terk, and L.
Weiss, “Shape deposition manufacturing,” Proc. of the
Solid Freeform Fabrication Symposium, 1994.
[77] Y. L. Park, K. Chau, R. J. Black, and M. R. Cutkosky,
“Force sensing robot fingers using embedded fiber bragg
grating sensors and shape deposition manufacturing,” Proc.
of IEEE International Conference on Robotics and Au-
tomation, pp. 1510-1516, 2007.
[78] S. Kim, J. E. Clark, and M. R. Cutkosky, “iSprawl: design
and tuning for high-speed autonomous open-loop running,”
The International Journal of Robotics Research, vol. 25,
pp. 903-912, 2006. [click]
[79] D. Santos, B. Heyneman, S. Kim, N. Esparza, and M. R.
Cutkosky, “Gecko-inspired climbing behaviors on verti-
cal and overhanging surfaces,” Robotics and Automation,
2008. ICRA 2008. IEEE International Conference on, pp.
1125-1131, 2008.
[80] J. G. Cham, S. A. Bailey, J. E. Clark, R. J. Full, and M.
R. Cutkosky, “Fast and robust: hexapedal robots via shape
deposition manufacturing,” The International Journal of
Robotics Research, vol. 21, pp. 869-882, 2002.
[81] K. Hosoda, Y. Tada, and M. Asada, “Anthropomorphic
robotic soft fingertip with randomly distributed receptors,”
Robotics and Autonomous Systems, vol. 54, pp. 104-109,
2006. [click]
[82] E. Hawkes, B. An, N. M. Benbernou, H. Tanaka, S. Kim,
E. D. Demaine, D. Rus, and R. J. Wood, “Programmable
matter by folding,” Proceedings of the National Academy
of Sciences, vol. 107, pp. 12441-12445, 2010. [click]
[83] O. P.-A. Néstor, Y. M. Kevin, C. G. Kevin, D. G. Jack, and
J. W. Robert, “First controlled vertical flight of a biologi-
cally inspired microrobot,” Bioinspiration & Biomimetics,
vol. 6, p 036009, 2011. [click]
[84] K. Peterson, P. Birkmeyer, R. Dudley, and R. S. Fearing,
“A wing-assisted running robot and implications for avian
flight evolution,” Bioinspiration & Biomimetics, vol. 6, p
046008, 2011. [click]
[85] T. Umedachi, V. Vikas, and B. A. Trimmer, “Highly de-
formable 3-D printed soft robot generating inching and
crawling locomotions with variable friction legs,” Proc. of
IEEE/RSJ International Conference on Intelligent Robots
and Systems, pp. 4590-4595, 2013.
[86] M. Giorelli, F. Renda, G. Ferri, and C. Laschi, “A feed-
forward neural network learning the inverse kinetics of a
soft cable-driven manipulator moving in three-dimensional
space,” Proc. of IEEE/RSJ International Conference on In-
telligent Robots and Systems, pp. 5033-5039, 2013.
[87] S. Frank, A. T. Barry, and R. Jason, “Modeling locomotion
of a soft-bodied arthropod using inverse dynamics,” Bioin-
spiration & Biomimetics, vol. 6, p 016001, 2011. [click]
[88] A. D. Marchese, K. Komorowski, C. D. Onal, and D. Rus,
“Design and control of a soft and continuously deformable
2D robotic manipulation system,” Proc. of IEEE Interna-
tional Conference on Robotics and Automation (ICRA), pp.
2189-2196, 2014.
[89] A. D. Marchese and D. Rus, “Design, kinematics, and con-
trol of a soft spatial fluidic elastomer manipulator,” The In-
ternational Journal of Robotics Research, 2015.
[90] F. J. Chen, S. Dirven, W. L. Xu, and X. N. Li, “Soft Actu-
ator Mimicking Human Esophageal Peristalsis for a swal-
lowing robot,” IEEE/ASME Transactions on Mechatronics,
vol. 19, pp. 1300-1308, 2014.
[91] F. Renda, M. Cianchetti, M. Giorelli, A. Arienti, and C.
Laschi, “A 3D steady-state model of a tendon-driven con-
tinuum soft manipulator inspired by the octopus arm,”
Bioinspiration & Biomimetics, vol. 7, p 025006, 2012.
[click]
[92] M. Cianchetti, M. Follador, B. Mazzolai, P. Dario, and C.
Laschi, “Design and development of a soft robotic octopus
arm exploiting embodied intelligence,” Robotics and Au-
tomation (ICRA), 2012 IEEE International Conference on,
pp. 5271-5276, 2012.

14 Chiwon Lee, Myungjoon Kim, Yoon Jae Kim, Nhayoung Hong, Seungwan Ryu, H. Jin Kim, and Sungwan Kim
[93] C. Laschi and M. Cianchetti, “Soft robotics: new perspec-
tives for robot bodyware and control,” Frontiers in Bioengi-
neering and Biotechnology, vol. 2, 2014.
[94] A. Goh, “Back-propagation neural networks for modeling
complex systems,” Artificial Intelligence in Engineering,
vol. 9, pp. 143-151, 1995. [click]
[95] M. Giorelli, F. Renda, G. Ferri, and C. Laschi, “A feed
forward neural network for solving the inverse kinetics of
non-constant curvature soft manipulators driven by cables,”
Proc. of ASME Dynamic Systems and Control Conference,
pp. 823-834, 2013.
[96] M. Giorelli, F. Renda, M. Calisti, A. Arienti, G. Ferri, and
C. Laschi, “Neural network and Jacobian method for solv-
ing the inverse statics of a cable-driven soft arm with non-
constant curvature,” IEEE Transactions on Robotics, vol.
31, pp. 823-834, 2015.
[97] J. Rieffel, F. Saunders, S. Nadimpalli, H. Zhou, S. Has-
soun, J. Rife, and B. Trimmer, Evolving Soft Robotic Loco-
motion in PhysX, ACM, Montreal, pp. 2499-2504, Canada,
2009.
[98] C. Majidi, “Soft robotics: a perspective-current trends and
prospects for the future,” Soft Robotics, vol. 1, pp. 5-11,
2014. [click]
[99] N. Correll, Ç. D. Önal, H. Liang, E. Schoenfeld, and D.
Rus, Soft Autonomous Materials-Using Active Elasticity
and Embedded Distributed Computation, Springer Berlin
Heidelberg, Berlin, Heidelberg, 2014.
[100] W. Wei, L. Jang-Yeob, R. Hugo, S. Sung-Hyuk, C.
Won-Shik, and A. Sung-Hoon, “Locomotion of inchworm-
inspired robot made of smart soft composite (SSC),” Bioin-
spiration & Biomimetics, vol. 9, p 046006, 2014. [click]
[101] M. Cianchetti, M. Calisti, L. Margheri, M. Kuba, and
C. Laschi, “Bioinspired locomotion and grasping in wa-
ter: the soft eight-arm OCTOPUS robot,” Bioinspiration &
Biomimetics, vol. 10, p 035003, 2015. [click]
[102] M. Calisti, A. Arienti, F. Renda, G. Levy, B. Hochner, B.
Mazzolai, P. Dario, and C. Laschi, “Design and develop-
ment of a soft robot with crawling and grasping capabili-
ties,” Proc. of IEEE International Conference on Robotics
and Automation (ICRA), pp. 4950-4955, 2012.
[103] F. Renda, F. Giorgio-Serchi, F. Boyer, and C. Laschi,
“Locomotion and elastodynamics model of an underwater
shell-like soft robot,” Proc. of IEEE International Confer-
ence on Robotics and Automation (ICRA), pp. 1158-1165,
2015.
[104] Y. Sugiyama and S. Hirai, “Crawling and jumping by a
deformable robot,” The International Journal of Robotics
Research, vol. 25, pp. 603-620, 2006. [click]
[105] R. F. Shepherd, A. A. Stokes, J. Freake, J. Barber, P.
W. Snyder, A. D. Mazzeo, L. Cademartiri, S. A. Morin,
and G. M. Whitesides, “Using explosions to power a soft
robot,” Angewandte Chemie International Edition, vol. 52,
pp. 2892-2896, 2013. [click]
[106] N. W. Bartlett, M. T. Tolley, J. T. B. Overvelde, J. C.
Weaver, B. Mosadegh, K. Bertoldi, G. M. Whitesides, and
R. J. Wood, “A 3D-printed, functionally graded soft robot
powered by combustion,” Science, vol. 349, pp. 161-165,
2015. [click]
[107] R. V. Martinez, J. L. Branch, C. R. Fish, L. Jin, R. F.
Shepherd, R. M. D. Nunes, Z. Suo, and G. M. Whitesides,
“Robotic tentacles with three-dimensional mobility based
on flexible elastomers,” Advanced Materials, vol. 25, pp.
205-212, 2013. [click]
[108] B. Mazzolai, L. Margheri, M. Cianchetti, P. Dario, and
C. Laschi, “Soft-robotic arm inspired by the octopus: II.
from artificial requirements to innovative technological so-
lutions,” Bioinspiration & Biomimetics, vol. 7, p 025005,
2012. [click]
[109] R. K. Katzschmann, A. D. Marchese, and D. Rus, “Au-
tonomous object manipulation using a soft planar grasp-
ing manipulator,” Soft Robotics, vol. 2, pp. 155-164, 2015.
[click]
[110] C. Lee, W. J. Park, M. Kim, S. Noh, C. Yoon, C. Lee,
Y. Kim, H. H. Kim, H. C. Kim, and S. Kim, “Pneumatic-
type surgical robot end-effector for laparoscopic surgical-
operation-by-wire,” Biomed Eng Online, vol. 13, p 130,
2014.
[111] A. Jiang, E. Secco, H. Wurdemann, T. Nanayakkara,
P. Dasgupta, and K. Athoefer, “Stiffness-controllable
octopus-like robot arm for minimally invasive surgery,”
Workshop on New Technologies for Computer/Robot As-
sisted Surgery, 2013.
[112] M. Cianchetti, T. Ranzani, G. Gerboni, T. Nanayakkara,
K. Althoefer, P. Dasgupta, and A. Menciassi, “Soft robotics
technologies to address shortcomings in today’s mini-
mally invasive surgery: the STIFF-FLOP approach,” Soft
Robotics, vol. 1, pp. 122-131, 2014. [click]
[113] S. Sareh, A. Jiang, A. Faragasso, Y. Noh, T. Nanayakkara,
P. Dasgupta, L. D. Seneviratne, H. A. Wurdemann, and K.
Althoefer, “Bio-inspired tactile sensor sleeve for surgical
soft manipulators,” Proc. of IEEE International Confer-
ence on Robotics and Automation (ICRA), pp. 1454-1459,
2014.
[114] T. Deng, H. Wang, W. Chen, X. Wang, and R. Pfeifer,
“Development of a new cable-driven soft robot for car-
diac ablation,” Proc. of IEEE International Conference on
Robotics and Biomimetics (ROBIO), pp. 728-733, 2013.
[115] P. Maeder-York, T. Clites, E. Boggs, R. Neff, P. Polygeri-
nos, D. Holland, L. Stirling, K. Galloway, C. Wee, and C.
Walsh, “Biologically inspired soft robot for thumb reha-
bilitation,” Journal of Medical Devices, vol. 8, p 020933,
2014. [click]
[116] P. Polygerinos, K. C. Galloway, E. Savage, M. Herman,
K. O. Donnell, and C. J. Walsh, “Soft robotic glove for
hand rehabilitation and task specific training,” Proc. of
IEEE International Conference on Robotics and Automa-
tion (ICRA), pp. 2913-2919, 2015.
[117] H. K. Yap, J. C. H. Goh, and R. C. H. Yeow, Design
and Characterization of Soft Actuator for Hand Rehabilita-
tion Application, Springer International Publishing, Cham,
2015.
[118] H. K. Yap, L. Jeong Hoon, F. Nasrallah, J. C. H. Goh, and
R. C. H. Yeow, “A soft exoskeleton for hand assistive and
rehabilitation application using pneumatic actuators with
variable stiffness,” Proc. of IEEE International Conference
on Robotics and Automation (ICRA), pp. 4967-4972, 2015.

Soft Robot Review 15
[119] Y. S. Song, Y. Sun, R. v. d. Brand, J. v. Zitzewitz, S.
Micera, G. Courtine, and J. Paik, “Soft robot for gait re-
habilitation of spinalized rodents,” Proc. of IEEE/RSJ In-
ternational Conference on Intelligent Robots and Systems,
pp. 971-976, 2013.
[120] A. T. Asbeck, R. J. Dyer, A. F. Larusson, and C. J.
Walsh, “Biologically-inspired soft exosuit,” Rehabilitation
Robotics (ICORR), 2013 IEEE International Conference
on, pp. 1-8, 2013.
[121] P. Yong-Lae, C. Bor-rong, O. P.-A. Néstor, Y. Diana, S.
Leia, J. W. Robert, C. G. Eugene, and N. Radhika, “Design
and control of a bio-inspired soft wearable robotic device
for ankle–foot rehabilitation,” Bioinspiration & Biomimet-
ics, vol. 9, p 016007, 2014. [click]
[122] J. A. Gallego, E. Rocon, J. Ib, x00E, x00F, ez, J. L.
Dideriksen, A. D. Koutsou, R. Paradiso, M. B. Popovic,
J. M. Belda-Lois, F. Gianfelici, D. Farina, D. B. Popovic,
M. Manto, T. D. Alessio, and J. L. Pons, “A soft wear-
able robot for tremor assessment and suppression,” Proc. of
IEEE International Conference on Robotics and Automa-
tion (ICRA), pp. 2249-2254, 2011.
[123] Y. Sun, C. M. Lim, H. H. Tan, and H. Ren, “Soft oral in-
terventional rehabilitation robot based on low-profile soft
pneumatic actuator,” Proc. of IEEE International Confer-
ence on Robotics and Automation (ICRA), pp. 2907-2912,
2015.
Chiwon Lee received the B.S. degree
in Mechanical & Aerospace Engineering
and Ph.D. degree in the Interdisciplinary
Program for Bioengineering from Seoul
National University (SNU), in 2011 and
2015, respectively. He is a senior re-
searcher with the Institute of Medical and
Biological Engineering, SNU since Sept.
2015. He is currently conducting re-
searches about medical robot, machine learning, and 3D printing
technology for medical applications.
Myungjoon Kim received the B.S. de-
gree in Electronics Engineering from Ts-
inghua University, Beijing China, in 2012.
He is currently pursuing a Ph.D. degree
in the Interdisciplinary Program for Bio-
engineering from Seoul National Univer-
sity, Seoul Korea. His main research area
is development of medical robot system.
Yoon Jae Kim received the B.S. degree
in Mechanical & Aerospace Engineering
from Seoul National University in 2014.
He is currently pursuing a Ph.D. degree in
the Interdisciplinary Program for Bioengi-
neering from Seoul National University,
Seoul Korea. His main research interests
include rehabilitation robots and biosignal
processing.
Nhayoung Hong received the B.S. degree
in Bio-Mechatronics from Sungkyunkwan
University in 2016. She is currently pursu-
ing a Ph.D. degree in the Interdisciplinary
Program for Bioengineering from Seoul
National University. Her research inter-
ests include surgical robots and biomedical
control.
Seungwan Ryu received the B.S. degree
in mechanical engineering from Hanyang
University, Seoul, Korea, in 2013, and the
M.S degree from Seoul National Univer-
sity, Seoul, Korea, in 2015. He is currently
pursuing the Ph.D. degree in the Depart-
ment of Mechanical and Aerospace Engi-
neering at Seoul National University. His
research interests include control of un-
manned micro aerial vehicles and biomimetic mobile robots, and
application of flapping wing mechanisms.
H. Jin Kim received her B.S. degree
from Korean Advanced Institute of Tech-
nology in 1995, and M.S. and Ph.D. de-
grees from the University of California,
Berkeley, USA, in 1999 and 2001, re-
spectively, all in mechanical engineering.
From 2002 to 2004, she was a postdoc-
toral researcher in Electrical Engineering
and Computer Sciences at University of
California, Berkeley. In 2004, she joined the School of Me-
chanical and Aerospace Engineering at Seoul National Univer-
sity, where she is currently a Professor. Her research interests
are intelligent control of robotic systems, motion planning and
vision-based navigation.
Sungwan Kim received his B.S. degree
in Electronics Engineering and M.S. de-
gree in Control & Instrumentation Engi-
neering from Seoul National University
(SNU), Seoul Korea in 1985 and 1987, and
his Ph.D. degree in Electrical Engineering
from University of California at Los Ange-
les in 1993, respectively. He is a professor
with the Department of Biomedical Engi-
neering, SNU College of Medicine since 2010. Prior to joining
to the SNU, he worked as a Senior Aerospace Engineer at Na-
tional Aeronautics and Space Administration (NASA) Langley
Research Center, Hampton, Virginia, USA. He is an Associate
Fellow of the AIAA and a Senior Member of the IEEE.