ArticlePDF Available

Abstract and Figures

A pneumatically powered, fully untethered mobile soft robot is described. Composites consisting of silicone elastomer, polyaramid fabric, and hollow glass microspheres were used to fabricate a sufficiently large soft robot to carry the miniature air compressors, battery, valves, and controller needed for autonomous operation. Fabrication techniques were developed to mold a 0.65-meter-long soft body with modified Pneu-Net actuators capable of operating at the elevated pressures (up to 138 kPa) required to actuate the legs of the robot and hold payloads of up to 8 kg. The soft robot is safe to interact with during operation, and its silicone body is innately resilient to a variety of adverse environmental conditions including snow, puddles of water, direct (albeit limited) exposure to flames, and the crushing force of being run over by an automobile.
Content may be subject to copyright.
Autonomous Soft Robotic Fish Capable of Escape
Maneuvers Using Fluidic Elastomer Actuators
Andrew D. Marchese,
Cagdas D. Onal,
and Daniela Rus
In this work we describe an autonomous soft-bodied robot that is both self-contained and capable of rapid,
continuum-body motion. We detail the design, modeling, fabrication, and control of the soft fish, focusing on
enabling the robot to perform rapid escape responses. The robot employs a compliant body with embedded
actuators emulating the slender anatomical form of a fish. In addition, the robot has a novel fluidic actuation
system that drives body motion and has all the subsystems of a traditional robot onboard: power, actuation,
processing, and control. At the core of the fish’s soft body is an array of fluidic elastomer actuators. We design
the fish to emulate escape responses in addition to forward swimming because such maneuvers require rapid
body accelerations and continuum-body motion. These maneuvers showcase the performance capabilities of
this self-contained robot. The kinematics and controllability of the robot during simulated escape response
maneuvers are analyzed and compared with studies on biological fish. We show that during escape responses,
the soft-bodied robot has similar input–output relationships to those observed in biological fish. The major
implication of this work is that we show soft robots can be both self-contained and capable of rapid
body motion.
Body compliance is a salient feature in many natural
systems. Compliant bodies offer inherent robustness to
uncertainty, adaptability to environmental variation, and the
capacity to redirect and distribute applied forces. In an effort
to make machines more capable, our aim is to exploit this
principle and design softness into robots.
In this work, we advance soft robotics by providing
a method for creating and controlling autonomous, self-
contained, soft-bodied systems. Specifically, we introduce a
novel self-contained fluidic actuation system and control al-
gorithms used to deliver continuum motion in soft robots. We
demonstrate this soft actuation in a case study by building an
autonomous soft-bodied robotic fish powered by an onboard
energy source (Fig. 1). The fish is novel in that it uses a soft
continuum body and an innovative fluidic actuation system
for the soft body and has onboard autonomy. All power, ac-
tuation, and computational systems are located onboard. The
continuum body has an embedded flexible spine and em-
bedded anatomically proportioned musclelike actuators. The
robot is capable of forward swimming and performing agile
maneuvers, scaled versions of an escape response.* A fish
was chosen as a case study because it naturally exhibits
continuum-body curvature, rapid motion during an escape
a compliant posterior that bends under hydro-
dynamic resistance,
and an anterior suitable for housing ri-
gid supporting hardware.
We have evaluated the forward swimming and escape re-
sponse maneuver of this soft robot in a suite of experiments.
Extensive kinematic data have been collected on the escape
response, and we compare the performance of the robot with
various studies on biological fish. We show that our robotic
system, although on a different time scale, is able to emulate
the basic structure of an escape response and that the
Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, Cambridge, Massachusetts.
Department of Mechanical Engineering, Worcester Polytechnic Institute, Worcester, Massachusetts.
*Escape response maneuvers are characterized by rapid body
accelerations over very short durations and that often involve the
body initially bending into a ‘‘C’’ shape.
Among vertebrates, these
are some of the most rapid maneuvers
and subject of frequent
study. The extremely agile behavior exhibited by fish during escape
response maneuvers is central to predator–prey interactions,
accordingly escape response performance carries marked ecological
Recently, the hydrodynamics of the maneuver have
been explored in great detail.
Volume 1, Number 1, 2014
ªMary Ann Liebert, Inc.
DOI: 10.1089/soro.2013.0009
performed maneuvers have a similar input–output relation-
ship as observed in biological fish.
Performance and autonomy are competing goals in fluid-
powered soft robots. Some fluid-powered soft machines show
promising capabilities such as walking
and leaping
are primarily driven by cumbersome external hardware lim-
iting their practical use. Conversely, there are instances of
self-contained fluidic soft robots;
however, because of
the constraints imposed by bringing all supporting hardware
onboard, performance of these robots is severely limited
when compared with rigid-bodied robots. The primary
technical challenge addressed by this work is the advance-
ment of soft-bodied robots to simultaneously be capable
of rapidly achieving continuum-body motion and be self-
contained. We illustrate our proposed technical approach by
designing and building a soft robot fish capable of emulating
the escape response of fish because this maneuver ex-
emplifies rapid and continuum-body motion and exhibits the
highest accelerations seen in fish.
The soft robotic fish exhibits continuum motion that
conventional rigid-bodied robotic fish cannot achieve. For
instance, although many notable robotic fish exist,
these prior robotic systems have bodies composed of rigid
segments connected by fixed joints and are consequently
incapable of reproducing the body kinematics observed
during agile escape response maneuvers. Previous at-
tempts to recreate an escape response used a body com-
posed of multiple position-controlled, rigid links;
however, such fully actuated, rigid-bodied systems inher-
ently fail to capture the continuum motion of the escape
response maneuver.
We build on several prior works that aim to create ro-
botic fish using biologically inspired flexible posteriors. Self-
propelling flexible foils driven by an external robotic actuator
have been studied by Lauder and colleagues.
Valdivia y
Alvarado and Youcef-Toumi used a compliant body in the
design of a robotic fish to mimic the swimming kinematics of
a natural fish.
Similarly, the robot fish FILOSE
has a
compliant posterior and serves as a test bed for fishlike
sensing and locomotion. Both of these systems are cable-
driven and actuated with an onboard servomotor but lack
autonomy and require an external power supply. Recently,
researchers have developed a cable-driven, flexible spring-
steel spine to model escape response behavior;
however, in
this system the motor, control system, and power supply are
external to the apparatus, and its motion is constrained. Long
et al. have developed a flexible biomimetic vertebral column
used to propel an autonomous surface-swimming robot.
The vehicle can also perform an escape response.
Again, a
single servomotor is used to actuate the compliant spine.
Although this system is autonomous, relative to the afore-
mentioned work, only a small portion of the body is flexible,
namely, its posterior tail, and because its large anterior is a
surface vessel, the system is limited to surface swimming.
Notably, the above-mentioned compliant-bodied robotic fish
operate on the principle of a passive, flexible mechanism
FIG. 1. Details of a soft-
bodied robotic fish. Top:A
dorsal view of the fish
showing (A) rigid anterior,
(B) center of mass, (C) an-
terior trunk musclelike actu-
ator pair, (D) inextensible
vertebrate-like constraint, (E)
posterior trunk actuator pair,
and (F) passive caudal fin.
Center: A cross-sectional
rendering of the mechanism
showing (G) fluidic elasto-
mer channels grouped into
antagonistic actuator, (H)
flexible constraint layer, and
(I) pressurized elastomer
channels in agonistic actua-
tor. Bottom: An exploded
view of the robot detailing
( J) silicone skin, (K) com-
munication and control elec-
tronics, (L) compressed gas
cylinder and regulator, (M)
flow control valves, (N) ac-
tuator access port, (O) plastic
fuselage, (P) videography
markers, and (Q) silicone
elastomer trunk.
driven by a traditional electromechanical actuator, and they
were primarily designed to understand the hydrodynamics of
the flexible body. However, in this work our primary goal was
to develop a fluidic actuation system that is embedded within
the flexible body, yielding a compliant and active body within
a completely self-contained system.
We have also built on prior work that introduced compliant
active-bodied robotic swimmers. Shen et al. have used an
oscillating strip of ionic polymer–metal composite as the
posterior trunk of a dolphinlike robot.
This is a free-
swimming robot, but again limited by an external tether.
Perhaps the closest precursor to our work is the Airacuda fish
developed by Festo.
This robot has a flexible body and is
driven by fluidic actuators. Similar to our system, the fluidic
and electronic components are located in the fish’s rigid an-
terior, and its actuators extend along the length of its flexible
trunk. However, this system differs considerably from ours in
design. It is composed of a plastic skeleton covered by flex-
ible skin with two actuators along the anteroposterior axis,
whereas we have a body composed almost entirely of soft
rubber with numerous actuators embedded in the dorsoven-
tral orientation along the anteroposterior axis (see the Ac-
tuation section in Materials and Methods). Another
difference is that Airacuda can do static diving and swim-
ming using the onboard pneumatic actuation system, whereas
the focus of the fish presented here was on forward swimming
and planar escape response maneuvers.
The work presented in this article differs from this prior
work in design, fabrication, and control, to enable new au-
tonomous capabilities for soft robotic fish. Specifically, the
main contributions of this article include the following:
A novel fluidic soft actuation system capable of rapidly
achieving continuum-body motion
A method for designing, fabricating, and controlling
autonomous, self-contained soft-bodied robots
A self-contained soft robot device that embodies our
approach to soft robots and emulates forward swim-
ming and planar escape maneuvers of biological fish,
along with experimental evaluations of the robot
Materials and Methods
System overview
A defining characteristic of the soft-bodied robotic fish
is the separation of actuation power from the rest of the
system—specifically, the utilization of mechanical energy in
the form of pressurized fluid instead of electrical energy to
power actuation. The body of the robotic fish (Fig. 1C–E, G–
I, and Q) is entirely composed of fluidic elastomer actuators
which are directly powered by pressurized
fluid and accordingly no energy conversion takes place at the
actuators. However, in order to control the fluidic system,
supporting valve hardware is also incorporated into the ar-
chitecture to electrically address and isolate the mechanical
actuation system.
The soft robot has onboard all the subsystems of a con-
ventional robot: an actuation system, power system, driving
electronics, and computation and control systems. These
systems (Fig. 1K–N) are stored in the fish’s rigid anterior
region (A), a region with minimal contribution to body cur-
vature during escape responses.
Technological advance-
ments in these subsystems enable autonomous operation of a
soft-bodied robot underwater.
FEA technology forms the core of the soft-bodied robotic
fish. FEAs are elastomer modules that bend under fluid
pressure. Bending is accomplished using a two-layer bimorph
structure. Pressurized gas expands fluidic channels embedded
within an elastomer layer, and a second inextensible but
flexible layer functions to constrain the axial tension gener-
ated by the expanding channels along one side. This trans-
forms lateral stress in the elastomer into a bending moment.
Moreover, three layers can be used to form a bidirectional
bending FEA: an inextensible constraining layer sandwiched
between both an agonistic and antagonistic expanding layer,
as has been demonstrated in Ref.
This bidirectional FEA
structure is fundamental to the soft robotic fish (see Fig. 1G,
agonistic layer; H, constraining layer; and I, antagonistic
layer). However, in this work we have advanced FEA tech-
nology by abandoning a simple rectangular shape and cre-
ating FEAs that conform to the complex anatomical shape of
a fish. The structure and operating principles of a tapered
bidirectional FEA are illustrated schematically in Figure 2.
In their model, Onal and colleagues
describe the total
bending angle hof a rectangular FEA using both the physical
properties of the channels and internal actuator pressure P
h¼2ntan 1we(r)
Here, nis the number of channels, wis channel width, eis
material strain and a nonlinear function of material stress r,
and, lastly, h
and h
are the constant heights of the actuator
and channels, respectively. However, because our actuator
differs considerably from a rectangular actuator with uniform
channels, we have developed a new model. By extending the
model presented in Ref.
to include variable channel height as
well as radial stress (i.e., normal to the inextensible constraint
layer), we can statically model the nonuniform bending of a
tapered FEA (Fig. 2). Specifically, the accumulated angle
along the length of the actuator after a given embedded
channel n,representedash
, can be estimated as a function of
both the physical properties of the preceding channels and P
aiFi, (2)
cos 1b
Atan 1w
Here, iis the channel index, a
and F
are construction
angles for the indexed channel, wand w
ˆrepresent the initial
and deformed widths of the indexed channel, and hand h
represent the initial and deformed heights of the indexed
channel. These parameters are illustrated in Figure 2C. It is
important to note that this simplifying static model assumes
that channels deform purely by extending their side and top
walls, and that these wall stresses are based on initial channel
geometry. In reality, the wall stresses change as the channel
surfaces deform. For this reason, this analytic model is most
valid for small deformations, that is, when pressure is low and
the actual stresses approximate those calculated from initial
channel geometry. The model also ignores external forces
such as the compressive forces generated by the antagonistic
half of the actuator. In Figure 2D we show a predicted
bending of the fish’s anterior actuator overlaid on top of the
actuator’s actual bending. Here, h
was predicted to be 52
degrees but measured approximately 45 degrees. Table 1 lists
actuator-specific parameter values for this experiment.
This static analysis suggests that by independently varying
the height of embedded channels, net complex curvatures of
the body can be achieved. A curvature profile can be me-
chanically ‘‘programmed’’ into the body geometry of the fish.
Such construction serves to simplify computational control
inputs. For instance, a single binary control input can be used
to drive the robot’s body through a complex kinematic profile
(Fig. 7).
The robotic fish used in this case study employs four sili-
cone FEAs that are molded to replicate the slender anatomy
of a natural fish, creating an actuated but continuously de-
formable body. The actuated body spans 43% to 100% of
the robot’s overall fork length (30.5 cm) (Fig. 1, top). Em-
bedded fluidic channels are grouped into two independently
actuated pairs: an agonistic and antagonistic anterior trunk
pair ranging from 45% to 70% of fork length (Fig. 1C), and a
posterior trunk pair ranging from 70% to 90% of fork length
(Fig. 1E). Separating the agonistic and antagonistic channel
groups is an inextensible but flexible constraining layer in-
troduced along the fish’s posterior midline (Fig. 1D and H).
This layer enables channel stresses to generate body curvature,
analogous to, though inverted from, the process by which
muscles generate a bending moment about the vertebrate
column in a fish.
A fabrication process was developed to first cast and then
combine components of the fish’s soft body. The process is
illustrated in Figure 3. First, each half of the body is cast
from silicone rubber (Mold Star 15; Smooth-on, Easton,
PA) by a two-part mold. The top mold piece creates the
embedded channels of both the anterior and posterior ac-
tuator grouping, and the bottom piece creates the anatomical
FIG. 2. Schematic repre-
sentation of a tapered bidi-
rectional FEA in cross
section. (A) The three-layer
structure: symmetric agonis-
tic (1) and antagonistic (3)
expanding layers sandwiching
an inextensible but flexible
constraining layer (2). Here,
embedded channel groupings
are in a depressurized state.
(B) Pressurized gas (red) ex-
panding the agonistic channel
group. Because of the con-
straining layer, fluid pressure
induces a bending moment
producing curvature. (C)
Model parameters. (D) Pre-
dicted curvature of the fish’s
anterior actuator overlaid atop
the actuator’s actual defor-
mation. FEA, fluidic elasto-
mer actuator.
Table 1. Physical Parameters of the Soft
Anterior Actuator
Parameter Value
55.8 kPa
18.6 mm
11.8 mm
2.5 mm
1.8 mm
w2.5 mm
fish shape (Fig. 3[1a]). In parallel, a connector piece is cast
with holes to serve as access ports to each channel grouping
mate with the fish’s rigid anterior (Fig. 3[1b]). In addition, a
thin constraint layer is cast with an embedded 0.5mm
Acetal film that provides inextensibility and also forms the
caudal fin (Fig. 3[1c]). Next, the four pieces undergo post-
casting preparation and are sequentially bonded to each
other using a thin layer of silicone (Fig. 3[2]). Once cured,
the body is ready for attachment to the rigid anterior
(Fig. 3[3]).
As mentioned, fluid energy is used to actuate the robot’s
body. Work is done on the body by the onboard power sys-
tem. When the body is actuated, a portion of this work
is stored as potential energy within the elastomer and com-
pressible fluid, W
, and the remainder is dissipated be-
cause of friction inside and outside the body, W
order to characterize the relative amounts of work required to
actuate the body at various rates, a pressure–volume analysis
technique was used similar to that used to measure the work
of breathing.
While submerged in water, the agonistic an-
terior actuator was repeatedly filled to a target volume, V
the flow rate into the actuator was varied, ranging from a
baseline flow of 5 L/m to a maximum flow of 50 L/m. Both
the change in internal actuator pressure, P
, and change in
volume, v, were measured. We assume that at the baseline
flow rate, there are no resistive losses and all energy delivered
to the actuator is stored elastically,
WElastic ¼ZvD
PaBaseline (v)dv:(4)
At each flow rate above baseline, i, the resistive component
of work done on the body can be approximated as
WResistive ¼ZvD
(Pai(v)PaBaseline (v))dv:(5)
Figure 4 shows the pressure–volume profiles at various flow
rates as well as an illustration of the work calculations. Table 2
lists the total, elastic, and resistive components of work done on
the soft body in water at various flow rates. Here, work is
calculated according to Equations 4 and 5 not measured di-
rectly. The resistive component of work increases as the actu-
ator is driven at higher rates; however, even at the highest rate of
actuation, nearly 78% of delivered energy is stored elastically.
Power supply
In order to drive locomotion, there must be an apparatus
onboard for supplying fluidic power. In an autonomous soft
robot application, it is desirable to maximize the energy
density ( MJ
L) of the power supply. As a solution, we expand on
our approach in Ref.
and again use an 8 g CO
gas cylinder
(PN 80121; Leland Ltd., Inc., South Plainfield, NJ) housing
fluid at high pressure and low volume (Fig. 5A). In this form,
a relatively large amount of fluidic energy can be stored in a
volume suitable for storage onboard the robot.
The total potential energy stored within, E, and energy
density, D, of the high-pressure container can be theorized by
assuming that the energy release is isothermal and by ne-
glecting fluid phase changes,
Here, the supply pressure P
is a nonlinear function of
volume v, and this relation is defined using the Van der Waals
FIG. 3. Illustration of the soft fish body fabrication process. First, two halves of the body (1a), a connector piece (1b), and
a constraining layer (1c) are all cast from silicone by two-part molds. Next, these four pieces are sequentially bonded
together using a thin layer of silicone (2). Lastly, once cured, the fish body is ready for operation (3).
gas law. The volume v
is the fluid’s volume corresponding to
the vapor pressure of the gas, P
. The volume v
the fluid’s volume at standard temperature and pressure. The
volume occupied by the container is V. Table 3 contains ap-
plication-specific parameters for the power supply. According
to Equation 6, the estimated energy density of this power
supply is 0.1 ( MJ
L), and the estimated total energy stored is 1.86
kJ. Although a typical lithium polymer battery has an energy
density of 0.9–1.5 ( MJ
L), electrical energy is not suitable for
directly powering FEAs; in such a case, the energy must first
be converted to the fluid domain using supporting hardware,
occupying additional volume and imposing energy losses. This
motivates our use of a fluidic power supply.
Gas delivery
Both an onboard gas regulation mechanism
shown in Figure
5B and delivery system (Fig. 5C and D) are required to dis-
tribute the stored fluid to the robot’s actuators. First, the fluid is
regulated to a suitable driving pressure, P
, an order of mag-
nitude below supply pressure. Subsequently, gas flows through
a series of control valves and into the body actuators. The
network of valvesand fluid pathways between the regulator and
actuators resist fluid flow. Consequently, inlet flow Qto the
actuators is driven by DP, the pressure gradient between P
the actuator’s internal pressure P
. To ensure that P
is of
sufficient magnitude to drive the body through its target kine-
matic profile, a relationship between Qand DPis theorized.
From our analysis in Figure 4, inlet flows of between 30
and 50 L/m are required for several hundred milliseconds.
Within the gas delivery system, Reynolds numbers are in
excess of 3 ·10
, above the critical Reynolds number of
2.3 ·10
, indicating turbulent flow. Assuming fully devel-
oped turbulent flow as well as constant fluid density, vis-
cosity, and temperature, we can estimate the DPrequired to
drive Qas follows:
where qis the fluid’s density, gthe gravity, and h
the total
head loss of the fluid delivery system, defined as,
where iindexes the nserial pathways within the fluid de-
livery system, Land dare the length and diameter of the
pathway, K
the loss coefficients of a pathway containing m
losses, and fthe friction factor as defined by the Haaland
Specifically, we calibrate Equation 8 to characterize the case
of maximum excursion. Here, the anterior agonistic control
valve is fully open, the antagonistic valve is fully occluded, and
for simplicity we ignore the posterior actuator pair. The fluid
pathway has three main parts: (1) the section between the reg-
ulator and valve orifice, (2) the valve, and (3) the valve to the
actuator inlet. Pathways 1 and 3 are designed to minimize re-
sistance and accordingly contribute little to h
. However, con-
straints on valve performance and size force pathway 2 to be of
high resistance. Finite element analysis was used to estimate
losses (SK) (see Table 3 for parameter estimates). As a result of
this analysis, DPis required to be between 0.06 and 0.10 MPa,
and considering the range of P
in Figure 4 and commercial
availability, a regulator providing a P
of 0.29 MPa was chosen.
The above analysis suggests that the gas delivery system is
responsible for considerable resistive energy losses. Given an
anterior actuator frequency of f, we can estimate the losses of
this subsystem over a single bidirectional actuation cycle:
EResistive ¼(PdPa)Q1
f:On the basis of experimental data
detailed in Figure 4, we can expect losses of around 50 J per
cycle while driving the anterior actuators during forward
Processing and control
The robot contains an onboard microprocessor
wireless communication module
that enable it to both pro-
cess external inputs and execute control policies. This com-
putational system interfaces directly with onboard hardware
such as control valves and any available sensors. Two pro-
portional valves** are used to control the pressurization of
FIG. 4. Pressure–volume profiles of fluid used to fill the
anterior agonist actuator at various flow rates. By integrating
the change in pressure as a function of displaced volume, the
elastic and resistive components of work done on the actu-
ator are determined.
Table 2. Elastic and Resistive Components
of Work Done on the Soft Body in Water
flow (L/m)
work ( J)
work (%)
work (%)
5 2.89 100 0
10 3.11 93.5 6.5
20 3.36 87.1 12.9
30 3.60 81.9 18.1
40 3.69 80.6 19.4
50 3.88 77.7 22.3
0.29 MPa regulator (PN 50044; Leland Ltd., Inc.).
Atmega 644P (Atmel Corporation, San Jose, CA).
XBee-PRO 900 (Digi International Inc., Minnetonka, MN).
**PN 921-111051-000 (Parker Precision Fluidics).
the agonistic and antagonistic anterior actuators, and two
exhaust valves
control depressurization. Two solenoid
control pressurization and depressurization of the
posterior actuator pair. Depressurization of the body actua-
tors is primarily passive. The exhaust and solenoid valve
orifices resist fluid outflow, and the actuators act as fluid
energy storage devices, analogous to a capacitor in the
electrical domain. If we ignore the fluid’s negligible inertia,
we can represent actuator depressurization as an undriven
first-order system, analogous to a capacitor discharging
through a resistor. The instant the exhaust valve is open, the
pressure gradient driving exhaust flow is equal to P
, and the
flow is at a maximum. Subsequently, the flow exponentially
drops to zero. Although the computational system can pro-
vide onboard autonomy, the robot wirelessly communicates
with a host PC to receive user commands. The fish is nega-
tively buoyant and held at a fixed submergence by means of a
floating support.
The control policy of fully actuated fish locomotion can be
described using seven parameters, T
, and
. The anterior control valves are driven using a square-
wave input, and agonistic and antagonistic valves are ener-
gized sequentially ensuring while one actuator is pressurizing
the other is exhausting. Here, T
and T
are the open periods
of the anterior agonistic and antagonistic actuator propor-
tional control valves, respectively. The corresponding orifice
magnitudes M
and M
of these control valves are defined as a
percent of maximum available flow. The parameter uis the
phase delay between the anterior and posterior actuator pairs,
and T
and T
are the open periods of the posterior agonistic
and antagonistic actuator control valves, respectively, and
their magnitudes are fixed. However, the body can also be
driven as an underactuated system where fewer than the total
number of available actuators are used. For forward loco-
motion, this abbreviated policy can be parameterized using
only T
, and M
. For escape responses, this policy can
be further reduced to using only T
and M
. During escape
responses, the posterior agonistic actuator is also used, but T
is set equal to T
As mentioned, the fish’s body can function as an under-
actuated system, where a single control input can excite
multiple modes of motion. Using just the anterior actuator
pair as opposed to all four available actuators for forward
swimming allows the fish to conserve the limited fluid energy
stored onboard. Similar to the work of Valdivia y Alvarado
and Youcef-Toumi,
our fish can use anterior trunk actuators
to excite movement in the entire compliant body, producing
fast forward swimming. During the experiment detailed in
Figure 6, the fish operated using an open-loop controller, and
its objective was to swim in a straight line. Here, the posterior
trunk actuator pair was passive, the anterior trunk actuators
were periodically driven at 1.67 Hz, and linear velocities
of 150 mm/s, or 0.44 body lengths/s, were attained. Ap-
proximately 30 tail beats were available from the 8 g CO
cylinder under these experimental conditions. Forward lo-
comotion performance parameters were measured using a
FIG. 5. Details of gas storage,
regulation, and release mechanism.
The mechanism consists of a high-
pressure CO
gas cylinder (A), a
passive mechanical regulator (B),
an interface manifold (C), propor-
tional control valves (D), and ex-
haust valves (E). On the left, red
arrows illustrate gas flow during
actuator pressurization; on the
right, black arrows illustrate gas
flow during depressurization.
Table 3. Robot Parameters Used in Modeling
Fluid Energy Supply and Fluid Delivery System
Parameter Value
42.7 mL
4.37 L
)6.08 MPa
V18.6 mL
q1.83 kg/m
4.72 mm
2.75 mm
4.72 mm
61 mm
3.5 mm
83 mm
PN PND-05A-12 (Parker Precision Fluidics).
PN 914-232123-000 (Parker Precision Fluidics).
high-speed video camera and a mirror mounted at 45 degrees
underneath the robot. Video was recorded at 300 fps, and nine
colored points along the robot’s dorsal midline were digitized
during postprocessing.
The forward-swimming results show that the self-
contained actuation system is capable of sustained operation,
exemplified by forward fishlike locomotion. However, the
swimming gait we present here is certainly suboptimal. A
comprehensive sweep over driving pressures, tail frequen-
cies, and body stiffness would allow optimization of forward
Escape response
A critical behavior exhibited among fish is an escape re-
sponse, which involves rapidly bending the fish’s body to
large angles in order to accelerate away from adverse stimuli.
Experiments were conducted to investigate the escape re-
sponse of our soft-bodied robotic fish and compare its ma-
neuvers to that of natural fish. A total of 28 escape response
experiments are reported and were carried out in a 76-cm-
long by 76-cm-wide by 61-cm-tall water tank filled to 41 cm.
The robot was completely submerged and held at an initial
fixed submergence by means of a 172-cm-long nylon string.
Escape responses were generated using an open-loop con-
troller. Escape response performance parameters were mea-
sured using the same high-speed videography techniques as
listed for the forward locomotion experiment; however, here
the camera was positioned overhead. Heading angle is de-
fined in degrees as the rotational displacement of the rigid
anterior midline. Escape angle is defined as the resulting
heading angle after the maneuver when the angular rate drops
to zero. Linear escape velocity is defined as the Cartesian
velocity of the center of mass 500 ms after the onset of the
Fish 30–40 cm in length (rainbow trout and northern pike)
have been shown to perform type I, or single-bend, escape
responses where the body initially bends into an ‘‘S’’ shape
and subsequently into a ‘‘C’’ shape with heading angles
greater than 90 degrees.
Our robot exhibits fishlike mech-
anisms to accelerate (see the exemplary single-bend escape
response in Figures 7A–D and 8A). We create single-bend
responses by using only the anterior and posterior agonistic
body actuators. Here, from 0 to 150 ms the body’s midline
assumes an ‘‘S’’ shape with the caudal fin curving in the
opposite direction of the trunk. From 150 to 500 ms the
body’s midline assumes a ‘‘C’’ shape as the caudal fin, pe-
duncle region, and trunk exhibit the same direction of cur-
vature. The robot reaches a maximum heading angle of 100
degrees at approximately 520 ms. The aforementioned natu-
ral fish also exhibit type II, or double-bend, escape responses
characterized by the fish’s anterior region bending away from
the posterior midline’s primary direction of curvature as the
body assumes a ‘‘C’’ shape.
In Figures 7E–H and 8B, the
robot performs a double-bend escape response using both
anterior and posterior agonistic and antagonistic body
FIG. 6. Experimental results of the robotic fish during
forward swimming. The top panel shows the digitized av-
erage body midline position moving as a function of time.
The bottom panel details the corresponding linear velocity
of the center of mass as a function of time. During this
experiment, tail stroke frequency is 1.67 Hz, and a velocity
of approximately 150 mm/s is attained.
FIG. 7. Sequences depicting the soft robotic fish performing both a single-bend (A–D) and double-bend (E–H) escape
response. The single-bend response requires only agonistic actuator effort. The double-bend response requires sequential
agonistic and antagonistic actuator efforts, causing a significant decrease in heading angle and ultimately resulting in lower
escape angles than single-bend responses. Actuator effort durations of 160 ms were used in both escape responses.
FIG. 8. Escape response kinematics of the soft-bodied robotic fish. Panel (A) details kinematics of a typical single-bend
escape response for the robotic fish; similarly, panel (B) details a double-bend escape response. The top portions of the
panels show the digitized body midline (red) overlaid every 10 ms from the first detectable motion to the end of maneuver.
The middle portions show the corresponding angular velocity of the head along with actuator effort (agonistic is positive;
antagonistic is negative). At the bottom is the resulting center-of-mass velocity for each maneuver.
actuators. Here, the ‘‘C’’ shape is assumed at 170 ms and is
quickly preceded by the body straightening. A maximum
heading angle of 72 degrees is obtained at 410 ms. In general,
the robotic fish exhibits higher escape angles in single-bend
than in double-bend escape responses. Figure 9 portrays both
a single- and double-bend escape response of an angelfish
(body length of 7.3 cm) exhibiting a similar kinematic pattern
to our robot; that is, the angelfish is also shown to have higher
escape angles in single-bend than in double-bend re-
Foreman and Eaton
document that single-bend maneu-
vers in fish result in higher escape angles and slower center-
of-mass motion, while double-bend maneuvers result in
lower escape angles and faster center-of-mass motion. Con-
sequently, although antagonistic effort is not required for
center-of-mass motion, it is responsible for amplifying the
center-of-mass motion and a period of negative angular rate
in natural fish. An important result is that our robot exhibits
similar behavior: In double-bend responses, escape angle is
lower and linear escape velocity higher than in single-bend
responses. In double-bend escape responses having antago-
nistic actuator activity (Fig. 8B), the escape angle is reduced,
as there is significant angular rate in the negative (antagonistic)
direction starting at approximately 380 ms (60 ms after the
completion of antagonistic activity). Onset of significant
center-of-mass motion occurred at 160 ms, approximately
synchronized with the onset of antagonistic activity.
A point of contrast is that in natural fish the head and tail
both move toward each other in the first stage of the escape
response (Fig. 9). However, because the center of mass of our
robot is in the anterior head region, head movement is
greatest during the second stage of the escape response.
Eaton, Lee, and Foreman’s direction-change hypothesis in
natural fish
is consistent with the behavior of our robotic
fish. Specifically, a combination of agonistic and antagonistic
actuator inputs can independently influence both escape
angle and linear escape velocity during the robot’s escape
response. A series of experiments were carried out to inves-
tigate the effect of two parameters, T
and M
, on the escape
angle and linear escape velocity of the robotic fish (Fig. 10).
As a result we found that T
has marked influence on escape
angle. Mean escape angle at T
equal to 100 ms was 26.6
degrees, whereas mean escape angle was 81.4 degrees at T
equal to 160 ms. Also, M
provides control over escape angle.
Increasing M
decreased escape angle (Fig. 10A). Unlike
FIG. 9. Fast-start kinemat-
ics of an angelfish. At the top
is the body midline plotted
for a single-bend (A) and a
double-bend (B) fast-start. At
the bottom is the corre-
sponding angular velocity
profile for the double-bend
fast-start. These figures are
reproduced with permission
from Domenici and Blake
and Domenici and Blake,
in all tests.
escape angle, T
has minimal influence on linear escape ve-
locity, mean values of 193.1 and 207.2 mm/s for 100 and
160 ms, respectively. However, M
provided control over
linear escape velocity in that increasing effort level expo-
nentially increased velocity (Fig. 10B). These findings indi-
cate that through a combination of agonistic and antagonistic
actuator efforts, escape angle and linear escape velocity can
be independently altered. Given a fixed maximal agonistic
effort level, altering T
and M
can independently influence
both the resulting escape angle and escape velocity.
The robotic fish provides an instantiation of our approach
to creating autonomous soft-bodied robots capable of rap-
idly achieving continuum-body motion. In this system, soft
musclelike actuators generate curvature in a continuously
deformable, vertebrate-like body. Novel, form-independent
actuator technology as well as miniaturization of supporting
hardware enable the robot to take on the fundamental ana-
tomical structure of a fish while being self-contained and
The programmability of our system allows repeatable
evaluation of the robot’s escape response maneuvers. By
directly controlling the duration and magnitude of agonistic
and antagonistic actuator efforts and measuring the resulting
escape response performance, we conclude that agonistic
duration has strong authority over escape angle and minimal
authority over linear escape velocity. Antagonistic magni-
tude has nonlinear control authority over both escape angle
and escape velocity. Increasing agonistic effort duration al-
lows for greater angular displacement of the head to be
reached during the first stage of the response. Soon after this
duration of time, the robot’s head begins to decelerate. The
longer the agonistic effort is applied, the greater the escape
angle. When antagonistic effort magnitude increases, there is
more energy to decelerate the turn and consequently escape
angle lowers. The antagonistic effort also provides the pro-
pulsive stroke, so greater antagonistic effort yields higher
escape velocity.
Evidence suggests that a similar input–output relationship
holds in biological escape response behavior. Consistent with
our robotic system, Foreman and Eaton
presented the di-
rection change concept, where they show that escape re-
sponse heading angle is a function of the relative magnitudes
and timings of agonist and antagonist muscle contractions.
¨hl and Schuster
investigated the predictive start of
hunting archer fish and showed that in the underlying C-start
behavior, escape angle and escape velocity need to be de-
coupled. Also in line with our robotic system, Tytell and
found that, among investigated variables, stage-one
duration correlated most strongly with escape angle and that
antagonistic effort magnitude correlated most strongly with
escape velocity in biological fish.
Our findings also suggest that despite the apparent com-
plexity of the maneuver, it is feasible that a robotic system
with limited onboard computational power could determine
required escape response control parameters in real time and
with no a priori planning. Because the robot’s escape re-
sponse performance outputs, angle and velocity, are inher-
ently decoupled by the physical form of the body and
structure of the maneuver, control parameters could poten-
tially be computed onboard. For instance, the maneuver may
start in immediate response to a perceived external stimulus
with a predetermined maximal agonistic effort. As this effort
occurs, the desired escape angle and velocity may be com-
puted relative to the perceived stimuli. Accordingly, the
control variables T
and M
can be determined given a
mapping similar to that provided in Figure 10.
A special thanks to Robert Katzschmann from the Dis-
tributed Robotics Laboratory and the Soft Robotics reviewers
for their extensive feedback. This work was done in the
Distributed Robotics Laboratory at MIT with support from
the National Science Foundation (grant numbers NSF
IIS1226883 and NSF CCF1138967) and National Science
FIG. 10. Input–output relationship of escape response
maneuvers in the robotic fish. (A) Escape angle as a function
of antagonistic actuator effort. (B) Escape velocity as a
function of antagonistic actuator effort. In both cases, equal
duration agonistic and antagonistic efforts of 100 and
160 ms were used (blue and red lines, respectively). Data
points represent mean values (n=4 and n=3 for 100 and
160 ms scenarios, respectively, for a total of 28 tests), and
error bars represent standard deviations.
Foundation Graduate Research Fellowship Program (primary
award number 1122374). We are grateful for this support.
Author Disclosure Statement
The authors declare no competing financial interests exist.
1. Domenici P, Blake RW. The kinematics and performance of
fish fast-start swimming. J Exp Biol 1997;200:1165–1178.
2. Borazjani I, Sotiropoulos F, Tytell ED, Lauder GV. Hy-
drodynamics of the bluegill sunfish C-start escape response:
three-dimensional simulations and comparison with ex-
perimental data. J Exp Biol 2012;215:671–684.
3. Wakeling JM, Johnston IA. Body bending during fast-starts
in fish can be explained in terms of muscle torque and
hydrodynamic resistance. J Exp Biol 1999;202:675–682.
4. Jayne BC, Lauder GV. Red and white muscle activity
and kinematics of the escape response of the bluegill
sunfish during swimming. J Comp Physiol A 1993;173:
5. Webb PW, Skadsen JM. Strike tactics of Esox. Can J Zool
6. Walker JA, Ghalambor CK, Griset OL, McKenney D,
Reznick DN. Do faster starts increase the probability of
evading predators? Funct Ecol 2005;19:808–815.
7. Gibb AC, Swanson BO, Wesp HM, Landels C, Liu C.
Development of the escape response in teleost fishes: do
ontogenetic changes enable improved performance? Phsiol
Biochem Zool 2006;79:7–19.
8. Domenici P, Turesson H, Brodersen J, Bronmark C. Pre-
dator-induced morphology enhances escape locomotion in
crucian carp. Proc R Soc Ser B 2008;275:195–201.
9. Bergstrom CA. Fast-start swimming performance and re-
duction in lateral plate number in threespine stickleback.
Can J Zool 2002;80:207–213.
10. Shepherd RF, Ilievski F, Choi W, Morin SA, Stokes AA,
Mazzeo AD, Chen X, Wang M, Whitesides GM. Multigait
soft robot. Proc Natl Acad Sci USA 2011;108:20400–
11. Shepherd RF, Stokes AA, Freake J, Barber J, Snyder PW,
Mazzeo AD, Cademartiri L, Morin SA, Whitesides GM.
Using explosions to power a soft robot. Angew Chem
12. Onal CD, Rus D. Autonomous undulatory serpentine lo-
comotion utilizing body dynamics of a fluidic soft robot.
Bioinspir Biomim 2013;8:026003.
13. Onal CD, Chen X, Whitesides GM, Rus D. Soft mobile
robots with on-board chemical pressure generation. Inter-
national Symposium on Robotics Research (ISRR), Flag-
staff, AZ, August 28–September 1, 2011.
14. Marchese AD, Onal CD, Rus D. Towards a self-contained
soft robotic fish: on-board pressure generation and em-
bedded electro-permanent magnet valves. Exp Rob 2013;
15. Liu J, Hu H. Biological inspiration: from carangiform fish
to multi-joint robotic fish. J Bionic Eng 2010;7:35–48.
16. Barrett DS, Triantafyllou MS, Yue DKP, Wolfgang MJ,
Grosenbaugh MA. Drag reduction in fish-like locomotion. J
Fluid Mech 1999;392:183–212.
17. Triantafyllou M, Triantafyllou G. An efficient swimming
machine. Sci Am 1995;272:64–70.
18. Zhong Y, Chong CW, Zhou C, Seet G, Low KH. Perfor-
mance predict model for a body and caudal fin (bcf) bio-
mimetics fish robot. IEEE/ASME International Conference
on, July 14–17, 2009. Adv Intell Mechatronics 2009;1230–
19. Wen L, Wang TM, Wu GH, Liang JH. Hydrodynamic in-
vestigation of a self-propelled robotic fish based on a force-
feedback control method. Bioinspir Biomim 2012;7:
20. Lui J, Hu H. Mimicry of sharp turning behaviours in a
robotic fish. Proceedings of the 2005 IEEE International
Conference on Robotics and Automation (ICRA), 2005.
21. Xu JX, Ren Q, Gao W, Niu XL. Mimicry of fish swimming
patterns in a robotic fish. 2012 IEEE International Sym-
posium on Industrial Electronics (ISIE), 2012.
22. Su Z, Yu J, Tan M, Zhang J. A closed-loop method to
generate fast C-start for a robotic fish. 2011 International
Conference on Mechatronics and Automation (ICMA),
Beijing, China, 2011.
23. Lauder GV, Flammang B, Alben S. Passive robotic models
of propulsion by the bodies and caudal fins of fish. Integr
Comp Biol 2012;52:576–587.
24. Alben S, Witt C, Baker TV, Anderson E, Lauder GV.
Dynamics of freely swimming flexible foils. Phys Fluids
25. Valdivia y Alvarado P, Youcef-Toumi K. Design of ma-
chines with compliant bodies for biomimetic locomotion
in liquid environments. J Dyn Sys Meas Control 2006;
26. El Daou H, Salumae T, Toming G, Kruusmaa M. Bio-
inspired compliant robotic fish: design and experiments.
IEEE International Conference on Robotics and Automa-
tion (IEEE ICRA 2012), St. Paul, Minnesota, 2012.
27. El Daou H, Salumae T, Ristolainen A, Toming G, Listak
M, Kruusmaa M. A bio-mimetic design and control of a
fish-like robot using compliant structures. The 15th IEEE
International Conference on Advanced Robotics (ICAR
2011), Tallinn, 2011.
28. Feng C, Bonafilia BR, Modarres-Sadeghi Y, Triantafyllou
MS. The mechanics of fast-start performance of pike
studied using a mechanical fish. Proceedings of the ASME
2011 International Mechanical Engineering Congress &
Exposition IMECE2011, Denver, Colorado, 2011.
29. Long JH, Koob T, Schaefer J, Summers A, Bantilan K,
Grotmol S, Porter M. Inspired by sharks: a biomimetic
skeleton for the flapping, propulsive tail of an aquatic robot.
Mar Technol Soc J 2011;45:119–129.
30. Long JH, Krenitsky NM, Roberts SF, Hirokawa J, de
Leeuw J, Porter ME. Testing biomimetic structures in
bioinspired robots: how vertebrae control the stiffness of
the body and the behavior of fish-like swimmers. Integr
Comp Biol 2011;51:158–175.
31. Shen Q, Wang T, Liang J, Wen L. Hydrodynamic perfor-
mance of a biomimetic robotic swimmer actuated by ionic
polymer–metal composite. Smart Mater Struct 2013;22:
32. Festo. Airacuda, 2006. Available at:
en_corp/9761.htm (accessed Dec. 31, 2013).
33. Correll N, Onal CD, Liang H, Schoenfeld E, Rus D. Soft
autonomous materials—using active elasticity and embed-
ded distributed computation. Springer Tracts Adv Rob
34. Marchese AD, Onal CD, Rus D. Soft robot actuators using
energy-efficient valves controlled by electropermanent
magnets. IEEE/RSJ International Conference on Intelligent
Robots and Systems (IROS), San Francisco, CA, 2011.
35. Spierts IL, Van Leeuwen JL. Kinematics and muscle dy-
namics of c-start and s-starts of carp (Cyprinus carpio l.). J
Exp Biol 1999;202:393–406.
36. Otis AB. The work of breathing. Physiol Rev 1954;34:
37. Haaland SE. Simple and explicit formulas for the friction
factor in turbulent. J Fluids Eng 1983;105:8990.
38. Harper DG, Blake RW. Fast-start performance of rainbow
trout Salmo gairdneri and northern pike Esox lucius. J Exp
Biol 1990;150:321–342.
39. Domenici P, Blake RW. The kinematics and performance
of the escape response in the angelfish (Pterophyllum ei-
mekei). J Exp Biol 1991;156:187–205.
40. Foreman MB, Eaton RC. The direction change concept for
reticulospinal control of goldfish escape. J Neurosci 1993;
41. Eaton RC, Lee RKK, Foreman MB. The Mauthner cell and
other identified neurons of the brainstem escape network of
fish. Prog Neurobiol 2001;63:467–485.
42. Wo
¨hl S, Schuster S. The predictive start of hunting archer
fish: a flexible and precise motor pattern performed with the
kinematics of an escape C-start. J Exp Biol 2007;210:311–
43. Tytell ED, Lauder GV. The C-start escape response of
Polypterus senegalus: bilateral muscle activity and vari-
ation during stage 1 and 2. J Exp Biol 2002;205:2591–
Address correspondence to:
Andrew D. Marchese
Department of Electrical Engineering
and Computer Science
Massachusetts Institute of Technology
32 Vassar Street, Room 32-376
Cambridge, MA 02139
This article has been cited by:
1. Lee Belding, Bilge Baytekin, Hasan Tarik Baytekin, Philipp Rothemund, Mohit S. Verma, Alex Nemiroski, Dan Sameoto, Bartosz
A. Grzybowski, George M. Whitesides. 2018. Slit Tubes for Semisoft Pneumatic Actuators. Advanced Materials 6, 1704446.
2. Federico Renda, Francesco Giorgio-Serchi, Frederic Boyer, Cecilia Laschi, Jorge Dias, Lakmal Seneviratne. A Multi-soft-body
Dynamic Model for Underwater Soft Robots 143-160. [Crossref]
3. Humna Hasan, Muhammad Hassan Safdar, Sana Zahid, Maria Bibi, Alvina Gul. Robotics and High-Throughput Techniques
155-166. [Crossref]
4. Dean Keithly, John Whitehead, Alexandra Voinea, Diego Horna, Samantha Hollenberg, Mason Peck, James Pikul, Robert F.
Shepherd. 2018. A cephalopod-inspired combustion powered hydro-jet engine using soft actuators. Extreme Mechanics Letters
. [Crossref]
5. Panagiotis Polygerinos, Nikolaus Correll, Stephen A. Morin, Bobak Mosadegh, Cagdas D. Onal, Kirstin Petersen, Matteo
Cianchetti, Michael T. Tolley, Robert F. Shepherd. 2017. Soft Robotics: Review of Fluid-Driven Intrinsically Soft Devices;
Manufacturing, Sensing, Control, and Applications in Human-Robot Interaction . Advanced Engineering Materials 19:12, 1700016.
6. Vo Doan Tat Thang, Tan Melvin Y.W., Bui Xuan Hien, Sato Hirotaka. An Ultralightweight and Living Legged Robot. Soft
Robotics, ahead of print. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links] [Supplemental Material]
7. Benjamin Gorissen, Dominiek Reynaerts, Satoshi Konishi, Kazuhiro Yoshida, Joon-Wan Kim, Michael De Volder. 2017. Elastic
Inflatable Actuators for Soft Robotic Applications. Advanced Materials 29:43, 1604977. [Crossref]
8. Zhongying Ji, Changyou Yan, Bo Yu, Xiaolong Wang, Feng Zhou. 2017. Multimaterials 3D Printing for Free Assembly
Manufacturing of Magnetic Driving Soft Actuator. Advanced Materials Interfaces 4:22, 1700629. [Crossref]
9. Baytekin Bilge, Cezan S. Doruk, Baytekin H. Tarık, Grzybowski Bartosz A.. Artificial Heliotropism and Nyctinasty Based on
Optomechanical Feedback and No Electronics. Soft Robotics, ahead of print. [Abstract] [Full Text HTML] [Full Text PDF] [Full
Text PDF with Links] [Supplemental Material]
10. Mac Murray Benjamin C., Futran Chaim C., Lee Jeanne, O'Brien Kevin W., Amiri Moghadam Amir A., Mosadegh Bobak,
Silberstein Meredith N., Min James K., Shepherd Robert F.. Compliant Buckled Foam Actuators and Application in Patient-
Specific Direct Cardiac Compression. Soft Robotics, ahead of print. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text
PDF with Links] [Supplemental Material]
11. Zhong Shen, Junhan Na, Zheng Wang. 2017. A Biomimetic Underwater Soft Robot Inspired by Cephalopod Mollusc. IEEE
Robotics and Automation Letters 2:4, 2217-2223. [Crossref]
12. Sheila Russo, Tommaso Ranzani, Conor J. Walsh, Robert J. Wood. 2017. An Additive Millimeter-Scale Fabrication Method for
Soft Biocompatible Actuators and Sensors. Advanced Materials Technologies 2:10, 1700135. [Crossref]
13. Alexander D. Valentine, Travis A. Busbee, John William Boley, Jordan R. Raney, Alex Chortos, Arda Kotikian, John Daniel
Berrigan, Michael F. Durstock, Jennifer A. Lewis. 2017. Hybrid 3D Printing of Soft Electronics. Advanced Materials 29:40,
1703817. [Crossref]
14. Rahim Mutlu, Charbel Tawk, Gursel Alici, Emre Sariyildiz. A 3D printed monolithic soft gripper with adjustable stiffness
6235-6240. [Crossref]
15. Sun Yi, Yap Hong Kai, Liang Xinquan, Guo Jin, Qi Peng, Ang Marcelo H. Jr., Yeow Chen-Hua. 2017. Stiffness Customization
and Patterning for Property Modulation of Silicone-Based Soft Pneumatic Actuators. Soft Robotics 4:3, 251-260. [Abstract] [Full
Text HTML] [Full Text PDF] [Full Text PDF with Links] [Supplemental Material]
16. Jusufi Ardian, Vogt Daniel M., Wood Robert J., Lauder George V.. 2017. Undulatory Swimming Performance and Body Stiffness
Modulation in a Soft Robotic Fish-Inspired Physical Model. Soft Robotics 4:3, 202-210. [Abstract] [Full Text HTML] [Full Text
PDF] [Full Text PDF with Links] [Supplemental Material]
17. George Thuruthel Thomas, Falotico Egidio, Manti Mariangela, Pratesi Andrea, Cianchetti Matteo, Laschi Cecilia. 2017. Learning
Closed Loop Kinematic Controllers for Continuum Manipulators in Unstructured Environments. Soft Robotics 4:3, 285-296.
[Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links]
18. Felt Wyatt, Chin Khai Yi, Remy C. David. 2017. Smart Braid Feedback for the Closed-Loop Control of Soft Robotic Systems.
Soft Robotics 4:3, 261-273. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links] [Supplemental Material]
19. Pham Huy Nguyen, Saivimal Sridar, Wenlong Zhang, Panagiotis Polygerinos. 2017. Design and control of a 3-chambered fiber
reinforced soft actuator with off-the-shelf stretch sensors. International Journal of Intelligent Robotics and Applications 1:3, 342-351.
20. JunZhi Yu, Li Wen, ZiYu Ren. 2017. A survey on fabrication, control, and hydrodynamic function of biomimetic robotic fish.
Science China Technological Sciences 60:9, 1365-1380. [Crossref]
21. Zhongkai Zhang, Thor Morales Bieze, Jeremie Dequidt, Alexandre Kruszewski, Christian Duriez. Visual servoing control of soft
robots based on finite element model 2895-2901. [Crossref]
22. Kristin M. de Payrebrune, Oliver M. O’Reilly. 2017. On the development of rod-based models for pneumatically actuated soft
robot arms: A five-parameter constitutive relation. International Journal of Solids and Structures 120, 226-235. [Crossref]
23. Nobuhiro Takahashi, Hideki Koike. SHIN-TAI: A method for controlling characteristics of a humanoid robot's body using
artificial muscles and fats 1400-1405. [Crossref]
24. Arijit Ghosh, ChangKyu Yoon, Federico Ongaro, Stefano Scheggi, Florin M. Selaru, Sarthak Misra, David H. Gracias. 2017.
Stimuli-Responsive Soft Untethered Grippers for Drug Delivery and Robotic Surgery. Frontiers in Mechanical Engineering 3. .
25. Elliot W. Hawkes, Laura H. Blumenschein, Joseph D. Greer, Allison M. Okamura. 2017. A soft robot that navigates its
environment through growth. Science Robotics 2:8, eaan3028. [Crossref]
26. Evgeniy Boyko, Moran Bercovici, Amir D. Gat. 2017. Viscous-elastic dynamics of power-law fluids within an elastic cylinder.
Physical Review Fluids 2:7. . [Crossref]
27. Hisham M.C.M. Anver, Rahim Mutlu, Gursel Alici. 3D printing of a thin-wall soft and monolithic gripper using fused filament
fabrication 442-447. [Crossref]
28. Liu Hanlin, Taylor Bevan, Curet Oscar M.. 2017. Fin Ray Stiffness and Fin Morphology Control Ribbon-Fin-Based Propulsion.
Soft Robotics 4:2, 103-116. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links]
29. Matia Yoav, Elimelech Tsah, Gat Amir D.. 2017. Leveraging Internal Viscous Flow to Extend the Capabilities of Beam-Shaped
Soft Robotic Actuators. Soft Robotics 4:2, 126-134. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links]
30. Wei Wang, Jindong Liu, Guangming Xie, Li Wen, Jianwei Zhang. 2017. A bio-inspired electrocommunication system for small
underwater robots. Bioinspiration & Biomimetics 12:3, 036002. [Crossref]
SCALES FOR 21ST CENTURY SOFT MATERIALS TECHNOLOGIES. Rubber Chemistry and Technology 90:2, 207-224.
32. Paul Phamduy, Miguel Angel Vazquez, Changsu Kim, Violet Mwaffo, Alessandro Rizzo, Maurizio Porfiri. 2017. Design and
characterization of a miniature free-swimming robotic fish based on multi-material 3D printing. International Journal of Intelligent
Robotics and Applications 1:2, 209-223. [Crossref]
33. Saurab Verma, Jian-Xin Xu. 2017. Data-Assisted Modeling and Speed Control of a Robotic Fish. IEEE Transactions on Industrial
Electronics 64:5, 4150-4157. [Crossref]
34. Dylan Drotman, Saurabh Jadhav, Mahmood Karimi, Philip deZonia, Michael T. Tolley. 3D printed soft actuators for a legged
robot capable of navigating unstructured terrain 5532-5538. [Crossref]
35. M. Calisti, G. Picardi, C. Laschi. 2017. Fundamentals of soft robot locomotion. Journal of The Royal Society Interface 14:130,
20170101. [Crossref]
36. Yufei Hao, Tianmiao Wang, Ziyu Ren, Zheyuan Gong, Hui Wang, Xingbang Yang, Shaoya Guan, Li Wen. 2017. Modeling
and experiments of a soft robotic gripper in amphibious environments. International Journal of Advanced Robotic Systems 14:3,
172988141770714. [Crossref]
37. Tiefeng Li, Guorui Li, Yiming Liang, Tingyu Cheng, Jing Dai, Xuxu Yang, Bangyuan Liu, Zedong Zeng, Zhilong Huang,
Yingwu Luo, Tao Xie, Wei Yang. 2017. Fast-moving soft electronic fish. Science Advances 3:4, e1602045. [Crossref]
38. F Giorgio-Serchi, A Arienti, F Corucci, M Giorelli, C Laschi. 2017. Hybrid parameter identification of a multi-modal underwater
soft robot. Bioinspiration & Biomimetics 12:2, 025007. [Crossref]
39. Yi Sun, Xinquan Liang, Hong Kai Yap, Jiawei Cao, Marcelo H. Ang, Raye Chen Hua Yeow. 2017. Force Measurement Toward
the Instability Theory of Soft Pneumatic Actuators. IEEE Robotics and Automation Letters 2:2, 985-992. [Crossref]
40. Espen Knoop, Andrew Conn, Jonathan Rossiter. 2017. VAM: Hypocycloid Mechanism for Efficient Bioinspired Robotic Gaits.
IEEE Robotics and Automation Letters 2:2, 1055-1061. [Crossref]
41. Erik H. Skorina, Ming Luo, Weijia Tao, Fuchen Chen, Jie Fu, Cagdas D. Onal. 2017. Adapting to Flexibility: Model Reference
Adaptive Control of Soft Bending Actuators. IEEE Robotics and Automation Letters 2:2, 964-970. [Crossref]
42. Jenett Benjamin, Calisch Sam, Cellucci Daniel, Cramer Nick, Gershenfeld Neil, Swei Sean, Cheung Kenneth C.. 2017. Digital
Morphing Wing: Active Wing Shaping Concept Using Composite Lattice-Based Cellular Structures. Soft Robotics 4:1, 33-48.
[Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links] [Supplemental Material]
43. Cosimo Della Santina, Cristina Piazza, Gian Maria Gasparri, Manuel Bonilla, Manuel Giuseppe Catalano, Giorgio Grioli, Manolo
Garabini, Antonio Bicchi. 2017. The Quest for Natural Machine Motion: An Open Platform to Fast-Prototyping Articulated
Soft Robots. IEEE Robotics & Automation Magazine 24:1, 48-56. [Crossref]
44. Chiwon Lee, Myungjoon Kim, Yoon Jae Kim, Nhayoung Hong, Seungwan Ryu, H. Jin Kim, Sungwan Kim. 2017. Soft robot
review. International Journal of Control, Automation and Systems 15:1, 3-15. [Crossref]
45. Shuo Li, Huichan Zhao, Robert F. Shepherd. 2017. Flexible and stretchable sensors for fluidic elastomer actuated soft robots.
MRS Bulletin 42:02, 138-142. [Crossref]
46. Ellen T. Roche, Markus A. Horvath, Isaac Wamala, Ali Alazmani, Sang-Eun Song, William Whyte, Zurab Machaidze,
Christopher J. Payne, James C. Weaver, Gregory Fishbein, Joseph Kuebler, Nikolay V. Vasilyev, David J. Mooney, Frank A. Pigula,
Conor J. Walsh. 2017. Soft robotic sleeve supports heart function. Science Translational Medicine 9:373, eaaf3925. [Crossref]
47. Fionnuala Connolly, Conor J. Walsh, Katia Bertoldi. 2017. Automatic design of fiber-reinforced soft actuators for trajectory
matching. Proceedings of the National Academy of Sciences 114:1, 51-56. [Crossref]
48. Jaeyoun Kim. Soft Robotic Micro-Tentacle: A Case Study 39-58. [Crossref]
49. Jaeyoun Kim. Introduction 1-3. [Crossref]
50. John Morrow, Samantha Hemleben, Yigit Menguc. 2017. Directly Fabricating Soft Robotic Actuators With an Open-Source 3-
D Printer. IEEE Robotics and Automation Letters 2:1, 277-281. [Crossref]
51. Jaeyoun Kim. Enabling Technologies 11-38. [Crossref]
52. Jaeyoun Kim. Towards Full-Scale Integration and Beyond 79-91. [Crossref]
53. Zhicong Deng, Martin Stommel, Weiliang Xu. Soft Robotics Technology and a Soft Table for Industrial Applications 397-409.
54. Liangliang Zhu, Yunteng Cao, Yilun Liu, Zhe Yang, Xi Chen. 2017. Architectures of soft robotic locomotion enabled by simple
mechanical principles. Soft Matter 13:25, 4441-4456. [Crossref]
55. T. J. Wallin, J. H. Pikul, S. Bodkhe, B. N. Peele, B. C. Mac Murray, D. Therriault, B. W. McEnerney, R. P. Dillon, E. P.
Giannelis, R. F. Shepherd. 2017. Click chemistry stereolithography for soft robots that self-heal. Journal of Materials Chemistry
B 5:31, 6249-6255. [Crossref]
56. Jaeyoun Kim. Current Progress 59-78. [Crossref]
57. Ross W. Jaggers, Stefan A. F. Bon. 2017. Temporal and spatial programming in soft composite hydrogel objects. J. Mater. Chem.
B 5:36, 7491-7495. [Crossref]
58. Langquan Shui, Liangliang Zhu, Zhe Yang, Yilun Liu, Xi Chen. 2017. Energy efficiency of mobile soft robots. Soft Matter 13:44,
8223-8233. [Crossref]
59. Cecilia Laschi, Barbara Mazzolai, Matteo Cianchetti. 2016. Soft robotics: Technologies and systems pushing the boundaries of
robot abilities. Science Robotics 1:1, eaah3690. [Crossref]
60. Ross Dylan, Nemitz Markus P., Stokes Adam A.. 2016. Controlling and Simulating Soft Robotic Systems: Insights from a
Thermodynamic Perspective. Soft Robotics 3:4, 170-176. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with
Links] [Supplemental Material]
61. Nemitz Markus P., Mihaylov Pavel, Barraclough Thomas W., Ross Dylan, Stokes Adam A.. 2016. Using Voice Coils to Actuate
Modular Soft Robots: Wormbot, an Example. Soft Robotics 3:4, 198-204. [Abstract] [Full Text HTML] [Full Text PDF] [Full
Text PDF with Links] [Supplemental Material]
62. V.S.C. Chillara, L.M. Headings, M.J. Dapino. 2016. Multifunctional composites with intrinsic pressure actuation and prestress
for morphing structures. Composite Structures 157, 265-274. [Crossref]
63. Tomoya Kimura, Ziqiao Jin, Ryuma Niiyama, Yasuo Kuniyoshi. Evolving soft robots to execute multiple tasks with Combined-
CPPN-NEAT 409-414. [Crossref]
64. Yu Zhou, Hu Jin, Chunshan Liu, Erbao Dong, Min Xu, Jie Yang. A novel biomimetic jellyfish robot based on a soft and smart
modular structure (SMS) 708-713. [Crossref]
65. Paul Phamduy, Jayhwan Cheong, Maurizio Porfiri. 2016. An Autonomous Charging System for a Robotic Fish. IEEE/ASME
Transactions on Mechatronics 21:6, 2953-2963. [Crossref]
66. Seonghwan Jeong, Jihong Lee. 2016. Modified Swimming Pattern to Control Propulsive Force for Biomimetic Underwater
Articulated Robot. Journal of Korea Robotics Society 11:4, 285-292. [Crossref]
67. Josie Hughes, Utku Culha, Fabio Giardina, Fabian Guenther, Andre Rosendo, Fumiya Iida. 2016. Soft Manipulators and Grippers:
A Review. Frontiers in Robotics and AI 3. . [Crossref]
68. H. Thien, M. Stommel, F. Le Daheron, A. Le Page, Z. Deng, W. L. Xu. Embedded infrared imaging to measure the deformation
of a soft robotic actuator 1-6. [Crossref]
69. Robert MacCurdy, Jeffrey Lipton, Shuguang Li, Daniela Rus. Printable programmable viscoelastic materials for robots 2628-2635.
70. Jun Shintake, Herbert Shea, Dario Floreano. Biomimetic underwater robots based on dielectric elastomer actuators 4957-4962.
71. Robert K Katzschmann, Austin de Maille, David L Dorhout, Daniela Rus. Cyclic hydraulic actuation for soft robotic devices
3048-3055. [Crossref]
72. Elisa Donati, Martin Worm, Stefano Mintchev, Marleen van der Wiel, Giovanni Benelli, Gerhard von der Emde, Cesare Stefanini.
2016. Investigation of Collective Behaviour and Electrocommunication in the Weakly Electric Fish, Mormyrus rume , through a
biomimetic Robotic Dummy Fish. Bioinspiration & Biomimetics 11:6, 066009. [Crossref]
73. Junzhi Yu, Jundong Xiao, Xiangbin Li, Weibing Wang. 2016. Towards a miniature self-propelled jellyfish-like swimming robot.
International Journal of Advanced Robotic Systems 13:5, 172988141666679. [Crossref]
74. Mutlu Rahim, Alici Gursel, in het Panhuis Marc, Spinks Geoffrey M.. 2016. 3D Printed Flexure Hinges for Soft Monolithic
Prosthetic Fingers. Soft Robotics 3:3, 120-133. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links]
75. Yap Hong Kai, Ng Hui Yong, Yeow Chen-Hua. 2016. High-Force Soft Printable Pneumatics for Soft Robotic Applications. Soft
Robotics 3:3, 144-158. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links] [Supplemental Material]
76. Jun Yuan, Junzhi Yu, Zhengxing Wu, Min Tan. 2016. Precise planar motion measurement of a swimming multi-joint robotic
fish. Science China Information Sciences 59:9. . [Crossref]
77. Francesco Giorgio-Serchi, Andrea Arienti, Cecilia Laschi. 2016. Underwater soft-bodied pulsed-jet thrusters: Actuator modeling
and performance profiling. The International Journal of Robotics Research 35:11, 1308-1329. [Crossref]
78. Jordan R. Raney, Neel Nadkarni, Chiara Daraio, Dennis M. Kochmann, Jennifer A. Lewis, Katia Bertoldi. 2016. Stable
propagation of mechanical signals in soft media using stored elastic energy. Proceedings of the National Academy of Sciences 113:35,
9722-9727. [Crossref]
79. Michael Wehner, Ryan L. Truby, Daniel J. Fitzgerald, Bobak Mosadegh, George M. Whitesides, Jennifer A. Lewis, Robert
J. Wood. 2016. An integrated design and fabrication strategy for entirely soft, autonomous robots. Nature 536:7617, 451-455.
80. Hu Jin, Erbao Dong, Min Xu, Chunshan Liu, Gursel Alici, Yang Jie. 2016. Soft and smart modular structures actuated by shape
memory alloy (SMA) wires as tentacles of soft robots. Smart Materials and Structures 25:8, 085026. [Crossref]
81. Tao Wang, Hongliang Ren. Reduction of power consumption for fluidic soft robots using energy recovery technique 1403-1408.
82. Zuo Cui, Hongzhou Jiang. 2016. Design, analysis, and simulation of a planar serial–parallel mechanism for a compliant robotic
fish with variable stiffness. Advances in Mechanical Engineering 8:8, 168781401666092. [Crossref]
83. Rahim Mutlu, S. Kumbay Yildiz, Gursel Alici, Marc in het Panhuis, Geoff M. Spinks. Mechanical stiffness augmentation of a
3D printed soft prosthetic finger 7-12. [Crossref]
84. Qi Shen, Sarah Trabia, Tyler Stalbaum, Viljar Palmre, Kwang Kim, Il-Kwon Oh. 2016. A multiple-shape memory polymer-metal
composite actuator capable of programmable control, creating complex 3D motion of bending, twisting, and oscillation. Scientific
Reports 6:1. . [Crossref]
85. Andrew D. Marchese, Russ Tedrake, Daniela Rus. 2016. Dynamics and trajectory optimization for a soft spatial fluidic elastomer
manipulator. The International Journal of Robotics Research 35:8, 1000-1019. [Crossref]
86. Feng Ni, Andreas Henning, Kai Tang, Lilong Cai. Soft damper for quick stabilization of soft robotic actuator 466-471. [Crossref]
87. Saurab Verma, Khalid Abidi, Jian-Xin Xu. Terminal Sliding Mode Control for speed tracking of a Carangiform robotic fish
345-350. [Crossref]
88. Joo Chuan Yeo, Hong Kai Yap, Wang Xi, Zhiping Wang, Chen-Hua Yeow, Chwee Teck Lim. 2016. Flexible and Stretchable
Strain Sensing Actuator for Wearable Soft Robotic Applications. Advanced Materials Technologies 1:3, 1600018. [Crossref]
89. Junzhi Yu, Cheng Zhang, Lianqing Liu. 2016. Design and Control of a Single-Motor-Actuated Robotic Fish Capable of Fast
Swimming and Maneuverability. IEEE/ASME Transactions on Mechatronics 21:3, 1711-1719. [Crossref]
90. Andrew D. Marchese, Daniela Rus. 2016. Design, kinematics, and control of a soft spatial fluidic elastomer manipulator. The
International Journal of Robotics Research 35:7, 840-869. [Crossref]
91. Jihong Yan, Hongbing Dong, Xinbin Zhang, Jie Zhao. A three-chambed soft actuator module with omnidirectional bending
motion 505-510. [Crossref]
92. Selim Ozel, Erik H. Skorina, Ming Luo, Weijia Tao, Fuchen Chen, Yixiao Pan, Cagdas D. Onal. A composite soft bending
actuation module with integrated curvature sensing 4963-4968. [Crossref]
93. Martin Stommel, Weiliang Xu. 2016. Optimal, Efficient Sequential Control of a Soft-Bodied, Peristaltic Sorting Table. IEEE
Transactions on Automation Science and Engineering 13:2, 858-867. [Crossref]
94. Wei-Kuo Yen, Jenhwa Guo. Wall following control of a robotic fish using dynamic pressure 1-7. [Crossref]
95. Tonazzini Alice, Sadeghi Ali, Mazzolai Barbara. 2016. Electrorheological Valves for Flexible Fluidic Actuators. Soft Robotics 3:1,
34-41. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links] [Supplemental Material]
96. Youngsu Cha, Jeffrey Laut, Paul Phamduy, Maurizio Porfiri. 2016. Swimming Robots Have Scaling Laws, Too. IEEE/ASME
Transactions on Mechatronics 21:1, 598-600. [Crossref]
97. Kyu-Jin Cho, Robert Wood. Biomimetic Robots 543-574. [Crossref]
98. Weijia Tao, Erik H. Skorina, Fuchen Chen, Jennifer McInnis, Ming Luo, Cagdas D. Onal. Bioinspired design and fabrication
principles of reliable fluidic soft actuation modules 2169-2174. [Crossref]
99. Yang Zhao, Masaaki Fukuhara, Takahiro Usami, Yogo Takada. 2015. Performance of Very Small Robotic Fish Equipped with
CMOS Camera. Robotics 4:4, 421-434. [Crossref]
100. Sanlin S. Robinson, Kevin W. O’Brien, Huichan Zhao, Bryan N. Peele, Chris M. Larson, Benjamin C. Mac Murray, Ilse M. Van
Meerbeek, Simon N. Dunham, Robert F. Shepherd. 2015. Integrated soft sensors and elastomeric actuators for tactile machines
with kinesthetic sense. Extreme Mechanics Letters 5, 47-53. [Crossref]
101. Benjamin C. Mac Murray, Xintong An, Sanlin S. Robinson, Ilse M. van Meerbeek, Kevin W. O'Brien, Huichan Zhao, Robert
F. Shepherd. 2015. Poroelastic Foams for Simple Fabrication of Complex Soft Robots. Advanced Materials 27:41, 6334-6340.
102. William C. Witt, Li Wen, George V. Lauder. 2015. Hydrodynamics of C-Start Escape Responses of Fish as Studied with Simple
Physical Models. Integrative and Comparative Biology 55:4, 728-739. [Crossref]
103. Hernando Leon-Rodriguez, Viet Ha Le, Seong Young Ko, Jong-Oh Park, Sukho Park. Ferrofluid soft-robot bio-inspired by
Amoeba locomotion 1833-1838. [Crossref]
104. Bryan N Peele, Thomas J Wallin, Huichan Zhao, Robert F Shepherd. 2015. 3D printing antagonistic systems of artificial muscle
using projection stereolithography. Bioinspiration & Biomimetics 10:5, 055003. [Crossref]
105. Ming Luo, Yixiao Pan, Erik H Skorina, Weijia Tao, Fuchen Chen, Selim Ozel, Cagdas D Onal. 2015. Slithering towards
autonomy: a self-contained soft robotic snake platform with integrated curvature sensing. Bioinspiration & Biomimetics 10:5,
055001. [Crossref]
106. Qi Shen, Viljar Palmre, Tyler Stalbaum, Kwang J. Kim. 2015. A comprehensive physics-based model encompassing variable surface
resistance and underlying physics of ionic polymer-metal composite actuators. Journal of Applied Physics 118:12, 124904. [Crossref]
107. Alouges François, DeSimone Antonio, Giraldi Laetitia, Zoppello Marta. 2015. Can Magnetic Multilayers Propel Artificial
Microswimmers Mimicking Sperm Cells?. Soft Robotics 2:3, 117-128. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text
PDF with Links]
108. Gilad Refael, Amir Degani. Momentum-driven single-actuated swimming robot 2285-2290. [Crossref]
109. Bianca S. Homberg, Robert K. Katzschmann, Mehmet R. Dogar, Daniela Rus. Haptic identification of objects using a modular
soft robotic gripper 1698-1705. [Crossref]
110. N. W. Bartlett, M. T. Tolley, J. T. B. Overvelde, J. C. Weaver, B. Mosadegh, K. Bertoldi, G. M. Whitesides, R. J. Wood. 2015.
A 3D-printed, functionally graded soft robot powered by combustion. Science 349:6244, 161-165. [Crossref]
111. M. Rolf, K. Neumann, J.F. Queißer, R.F. Reinhart, A. Nordmann, J.J. Steil. 2015. A multi-level control architecture for the
bionic handling assistant. Advanced Robotics 29:13, 847-859. [Crossref]
112. Rahim Mutlu, Gursel Alici, Marc in het Panhuis, Geoff Spinks. Effect of flexure hinge type on a 3D printed fully compliant
prosthetic finger 790-795. [Crossref]
113. Jingjin Xie, James B. McGovern, Rutvij Patel, Woobiehn Kim, Saugata Dutt, Aaron D. Mazzeo. 2015. Elastomeric Actuators on
Airfoils for Aerodynamic Control of Lift and Drag. Advanced Engineering Materials 17:7, 951-960. [Crossref]
114. Huichan Zhao, Yan Li, Ahmed Elsamadisi, Robert Shepherd. 2015. Scalable manufacturing of high force wearable soft actuators.
Extreme Mechanics Letters 3, 89-104. [Crossref]
115. Daniela Rus, Michael T. Tolley. 2015. Design, fabrication and control of soft robots. Nature 521:7553, 467-475. [Crossref]
116. Saurab Verma, Jian-Xin Xu, Qin-Yuan Ren. Agile gait generation for an anguilliform robotic fish 1-6. [Crossref]
117. Andrew D. Marchese, Russ Tedrake, Daniela Rus. Dynamics and trajectory optimization for a soft spatial fluidic elastomer
manipulator 2528-2535. [Crossref]
118. Hong Kai Yap, Jeong Hoon Lim, Fatima Nasrallah, James C. H. Goh, Raye C. H. Yeow. A soft exoskeleton for hand assistive
and rehabilitation application using pneumatic actuators with variable stiffness 4967-4972. [Crossref]
119. Matia Yoav, Gat Amir D.. 2015. Dynamics of Elastic Beams with Embedded Fluid-Filled Parallel-Channel Networks. Soft Robotics
2:1, 42-47. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links]
120. Marchese Andrew D., Katzschmann Robert K., Rus Daniela. 2015. A Recipe for Soft Fluidic Elastomer Robots. Soft Robotics 2:1,
7-25. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links]
121. M. Stommel, W. L. Xu. Qualitative control of soft robotic peristaltic sorting tables 82-87. [Crossref]
122. A. M. Ajith, Sourabh Moitra, A. P. Sudheer. Parameter optimization and experimentation of the undulating fin of a knife fish
robot 1-6. [Crossref]
123. Bo Liu, Shiwu Zhang, Fenghua Qin, Jie Yang. 2014. Fluid–structure interaction study on the performance of flexible articulated
caudal fin. Advanced Robotics 28:24, 1665-1676. [Crossref]
124. Stephen Ornes. 2014. Inner Workings: A soft robot that swims like a fish. Proceedings of the National Academy of Sciences 111:50,
17688-17688. [Crossref]
125. Wei Wang, Jang-Yeob Lee, Hugo Rodrigue, Sung-Hyuk Song, Won-Shik Chu, Sung-Hoon Ahn. 2014. Locomotion of
inchworm-inspired robot made of smart soft composite (SSC). Bioinspiration & Biomimetics 9:4, 046006. [Crossref]
126. S.B. Elbaz, A.D. Gat. 2014. Dynamics of viscous liquid within a closed elastic cylinder subject to external forces with application
to soft robotics. Journal of Fluid Mechanics 758, 221-237. [Crossref]
127. Tolley Michael T., Shepherd Robert F., Mosadegh Bobak, Galloway Kevin C., Wehner Michael, Karpelson Michael, Wood Robert
J., Whitesides George M.. 2014. A Resilient, Untethered Soft Robot. Soft Robotics 1:3, 213-223. [Abstract] [Full Text HTML]
[Full Text PDF] [Full Text PDF with Links] [Supplemental Material]
128. Stephen A. Morin, Yanina Shevchenko, Joshua Lessing, Sen Wai Kwok, Robert F. Shepherd, Adam A. Stokes, George M.
Whitesides. 2014. Using “Click-e-Bricks” to Make 3D Elastomeric Structures. Advanced Materials 26:34, 5991-5999. [Crossref]
... Robots such as the 16g DASH robot [42] leverage lightweight materials and actuators to achieve high-speed running and resistance to damage from collisions and falls. Others use soft-robotic and compliant structures to build robots that are resistant to crushing and can navigate tight spaces [43][44][45]. Other researchers have incorporated bio-inspired collapsible wing features into flapping robots [24] that are able to dampen collisions with obstacles. ...
Full-text available
Mobile millimeter and centimeter scale robots often use smart composite manufacturing (SCM) for the construction of body components and mechanisms. The fabrication of SCM mechanisms requires laser machining and laminating flexible, adhesive, and structural materials into small-scale hinges, transmissions, and, ultimately, wings or legs. However, a fundamental limitation of SCM components is the plastic deformation and failure of flexures. In this work, we demonstrate that encasing SCM components in a soft silicone mold dramatically improves the durability of SCM flexure hinges and provides robustness to SCM components. We demonstrate this advance in the design of a flapping-wing robot that uses an underactuated compliant transmission fabricated with an inner SCM skeleton and exterior silicone mold. The transmission design is optimized to achieve desired wingstroke requirements and to allow for independent motion of each wing. We validate these design choices in bench-top tests, measuring transmission compliance, kinematics, and fatigue. We integrate the transmission with laminate wings and two types of actuation, demonstrating elastic energy exchange and limited lift-off capabilities. Lastly, we tested collision mitigation through flapping-wing experiments that obstructed the motion of a wing. These experiments demonstrate that an underactuated compliant transmission can provide resilience and robustness to flapping-wing robots.
Complete compliance of soft robots is a trending research topic. One significant aspect is the control element. In the current work, a flexible electro-rheological (ER) valve development is chosen due to the advantage of its dynamic power density, convenient electronic control mode, and no mechanical moving parts. These properties enable the ER valve to be miniaturized and easily integrated into a fluid-driven soft robot. However, the integration effect of the existing flexible ER valve configuration in soft robots is not ideal. In this study, a novel concentric semi-cylindrical flexible giant ER (GER) valve with a dual flowing direction is proposed for a soft robot to enable a larger deformation bending motion with multiple motion modes. The resulting valve has an overall volume of 50 mm³, a mass of 0.88 g, and a maximum controlling pressure of 84 kPa with a pressure change rate of 13.9, which is suitable for most soft robot applications. The pressure difference calculation method for the concentric semi-cylindrical GER valve is discussed. The effects of size, voltages, and flow rates on the valve’s static and dynamic performances, as well as the characteristics of the two pressure regulation modes, are investigated through simulations and experiments. The seamless integration and flexibility of the proposed valve are demonstrated by controlling a typical network actuator with six motion modes. Finally, a soft robot composed of two actuators is fabricated and realized the grasping of multiple everyday objects using different motion modes.
Multi-chamber soft pneumatic actuators (m-SPAs) have been widely used in soft robotic systems to achieve versatile grasping and locomotion. However, existing m-SPAs have slow actuation speed and are either limited by a finite air supply or require energy-consuming hardware to continuously supply compressed air. Here, we address these shortcomings by introducing an internal exhaust air recirculation (IEAR) mechanism for high-speed and low-energy actuation of m-SPAs. This mechanism recirculates the exhaust compressed air and recovers the energy by harnessing the rhythmic actuation of multiple chambers. We develop a theoretical model to guide the analysis of the IEAR mechanism, which agrees well with the experimental results. Comparative experimental results of several sets of m-SPAs show that our IEAR mechanism significantly improves the actuation speed by more than 82.4% and reduces the energy consumption per cycle by more than 47.7% under typical conditions. We further demonstrate the promising applications of the IEAR mechanism in various pneumatic soft machines and robots such as a robotic fin, fabric-based finger, and quadruped robot. Corresponding author(s) Email:
Full-text available
Despite tremendous progress in the development of untethered soft robots in recent years, existing systems lack the mobility, model‐based control, and motion planning capabilities of their piecewise rigid counterparts. As in conventional robotic systems, the development of versatile locomotion of soft robots is aided by the integration of hardware design and control with modeling tools that account for their unique mechanics and environmental interactions. Here, a framework for physics‐based modeling, motion planning, and control of a fully untethered swimming soft robot is introduced. This framework enables offline co‐design in the simulation of robot parameters and gaits to produce effective open‐loop behaviors and enables closed‐loop planning over motion primitives for feedback control of a frog‐inspired soft robot testbed. This pipeline uses a discrete elastic rods (DERs) physics engine that discretizes the soft robot as many stretchable and bendable rods. On hardware, an untethered aquatic soft robot that performs frog‐like rowing behaviors is engineered. Hardware validation verifies that the simulation has sufficient accuracy to find the best candidates for sets of parameters offline. The simulator is then used to generate a trajectory library of the robot's motion in simulation that is used in real‐time closed‐loop path following experiments on hardware. A discrete elastic rods (DERs) soft robotics physics simulator is used to specify design parameters and gaits for soft, untethered frog‐like robots. It is then used to generate a trajectory library for fast online closed‐loop path following.
Soft climbing/crawling robots have been attracting increasing attention in the soft robotics community, and many prototypes with basic locomotion have been implemented. Most existing soft robots achieve locomotion by planar bending deformation and lack sufficient mobility. Enhancing the mobility of soft climbing/crawling robots is still an open and challenging issue. To this end, we present a novel pneumatic leech-like soft robot, Leechbot, with both bending and stretching deformation for locomotion. With a morphological structure, the robot consists of a three-chambered actuator in the middle for the main motion, two chamber-net actuators that act as ankles, and two suckers at the ends for anchoring on surfaces. The peristaltic motion for locomotion is implemented by body stretching, and direction changing is achieved by body bending. Due to the novel design and two deformation modes, the robot can make turns and transit between different surfaces; the robot, hence, has excellent mobility. The development of the robot prototype is presented in detail in this paper. To control its motion, tests were carried out to determine the relationship between step length and air pressure as well as the relationship between motion speed and periodic delay time. A kinematic model was established, and the kinematic mobility and surface transitionability were analyzed. Gait planning based on the inflating sequence of the actuating chambers is presented for straight crawling, turn making, and transiting between surfaces and was verified by a series of experiments with the prototype. The results show that a high mobility in soft climbing/crawling robots can be achieved by a novel design and by proper gait planning.
The soft actuators integrated with the reversible proton exchange membrane fuel cell (RPEMFC) structure exhibit the potential to serve in portable and untethered soft robots with advantages of electric actuation, and compact and lightweight structure. But their performances in actuation strain, load capacity, controllability, and robustness are expected to be further improved. Here, a soft electrochemical actuator (SECA) that is integrated with an electrochemical reactor (ECR) designed based on the RPEMFC structure is reported. The ECR works as a pneumatic source that provides regulated pneumatic pressure through the water electrolysis reaction and hydrogen-oxygen fuel cell reaction. With the designed water-containing structure, the ECR has an improved performance of robustness to pose variations. The SECA can reach controllable deformation by the proposed actuation and deformation control methods which require a low-voltage (< 3 V) and simple electrical system. With the new structural design, the SECA achieves excellent performances of large actuation deformation (150% actuation strain) and a high load-to-weight ratio (25:1) with a lightweight, compact, and easy-to-fabricate structure.
This paper presents the design and analysis of a novel end effector used in the integrated limb mechanism that enables manipulation and locomotion tasks to be performed with a single limb. A Fin Ray structure-based two-finger gripper design is incorporated into the end effector with a novel flexible tendon design that wraps around the base of each finger. A Finite Element Analysis (FEA) study was performed to optimize the Fin Ray structure-based finger design by varying its physical parameters and quantifying its performance. Simulation results were discussed with several trends to be used to design the structure for specific tasks. A preliminary prototype was developed, and a range of experimental tests were carried out to validate the FEA study.KeywordsIntegrated limb mechanismEnd-effector designFin Ray structureFEA analysis
This article presents a novel soft robotic gripper with a high payload capacity based on the layer jamming technology. Soft robots have a high adaptability, however suffer a low payload capacity. To overcome these conflicting challenges, here we introduce a 3D printed multi-material gripper that integrates jamming layers for enhancing payload capacity. By inflating the internal air chamber with positive pressure, the finger can be actuated to a large bending angle for adapting complex shapes. Layers of jamming sheets are bounded on the finger structure and are then sealed inside a vacuum bag. When a high payload is desired, air inside the vacuum bag is drawn out and a negative air pressure is applied to the jamming layers, which leads to the gripper locked at the actuated shape. To evaluate the performance of the gripper, we conducted extensive tests including actuation, stiffness variation, typical payload capacity, and adaptability. The results show that our gripper is not only highly adaptable just like most soft grippers but also more importantly capable of grasping heavy (about 6–10 kg) objects comparable to rigid-body counterparts.
Untethered operation remains a fundamental challenge in soft robotics. Soft robotic actuators are generally unable to produce the forces required for carrying essential power and control hardware on-board. Moreover, current untethered soft robots often have low operating times given soft actuators' limited efficiency and lifetime. Here, we 3D print cylindrical handed shearing auxetics (HSAs) from single-cure polyurethane resins for use as scalable, motorized soft robotic actuators for untethered machines. Mechanical characterization of individual HSAs confirms their auxetic behaviors and suitability as actuators. HSA pairs of opposite handedness are assembled to form multi-degree-of-freedom legs for untethered quadrupeds. We explore several leg designs to understand the role of length and auxetic pattern density on overall motion and blocked force generated. Finally, we demonstrate untethered locomotion with two soft robotic quadrupeds. We find that our taller soft robot is capable of walking at 2 body lengths per min (BL min-1) for 65 min, all while carrying a payload of at least 1.5 kg. We compare our soft robots' capabilities to those of previously reported untethered, terrestrial systems and find that our motorized HSAs lead to the second highest operating time with an above average velocity. We anticipate that these methods will open new avenues for designing untethered soft robots with the robustness, operating times, and payload capacities required for future fundamental investigations in embodied intelligence and adaptive, physical learning.
Full-text available
In this study, we describe the most ultralightweight living legged robot to date that makes it a strong candidate for a search and rescue mission. The robot is a living beetle with a wireless electronic backpack stimulator mounted on its thorax. Inheriting from the living insect, the robot employs a compliant body made of soft actuators, rigid exoskeletons, and flexure hinges. Such structure would allow the robot to easily adapt to any complex terrain due to the benefit of soft interface, self-balance, and self-adaptation of the insect without any complex controller. The antenna stimulation enables the robot to perform not only left/right turning but also backward walking and even cessation of walking. We were also able to grade the turning and backward walking speeds by changing the stimulation frequency. The power required to drive the robot is low as the power consumption of the antenna stimulation is in the order of hundreds of microwatts. In contrast to the traditional legged robots, this robot is of low cost, easy to construct, simple to control, and has ultralow power consumption.
Combustion in soft and compliant materials may enable new methods for underwater locomotion. Inspired by the motion of the soft bodied cephalopods, we developed a combustion-powered hydro-jet engine (CPHJE) using an expandable silicone bladder. The CPHJE used high energy density methane combustion to expand a silicone bladder and accelerate water into a hydro-jet that propelled the CPHJE. Cephalopods demonstrate the epitome of underwater soft robot capabilities. Their use of hydrojetting as a mode of high-speed underwater motion resembles high-power combustion events. This paper describes a soft combustion engine that propels a ∼0.7 kg body, with hydrodynamic surfaces in the likeness of an ellipsoid shaped squid, at instantaneous velocities of at least 0.8 m s⁻¹ and single-pulse average velocities of 0.1 m s⁻¹. High-speed videography informs a model for the CPHJE’s motion in time; the imaging also reveals that the hydrojetting generates a ring vortex. Tracking of this vortex results in an estimate that the working fluid is ejected in excess of 6.7 m s⁻¹. We demonstrate the CPHJE motion is both repeatable and able to produce multiple actuations with the same bladder. These results demonstrate a new method of underwater propulsion using combustion in soft and compliant bladders.
Robots are playing a vital role as a connection between the spheres of biology and physics. This new field covers the multifarious disciplines under one heading "Biorobotics," thus providing biologists with tools to evaluate biological algorithms for engineering applications. Biorobots have enabled scientists to understand intricate relationship between animals and environment. From understanding brain functions to providing answers to the intriguing questions of evolution, biorobotics have helped biologists to widen their knowledge of the sphere. Biorobotics is considered to be a promising new field inspiring biologists for novel researches. High-throughput technology provides a feasible option for the automation of experiments in the field of Robotics. High throughput has also enabled biologists to meet the great challenge of scientific research, e.g., understand the causes and consequences of disease and the underlying basic mechanisms. It has also invited researchers from interdisciplinary sciences for the purpose of inducing bio-inspired capabilities in the robotics.
Although plants are typically not considered an inspiration for designing motile robots, they do perform a variety of intricate motion patterns, including diurnal cycles of sun tracking (heliotropism) and leaf opening (nyctinasty). In real plants, these motions are controlled by complex, feedback-based biological mechanisms that, to date, have been mimicked only in computer-controlled artificial systems. This work demonstrates both heliotropism and nyctinasty in a system in which few simple, but strategically positioned thermo-responsive springs and lenses form a feedback loop controlling these motions and substantiating a behavioral analogy to "plants." In particular, this feedback allows the "artificial plant" to reach and stabilize at a metastable position in which the solar flux on the "plants" and the solar power "leaves" are maximized. Unlike many soft robotic systems, our "plants" are completely autonomous, in that, they do not require any external controls or power sources. Bioinspired designs such as this could be of interest for soft robotic systems in which materials alone-rather than power-consuming electronic circuitry-control the motions.
We introduce the use of buckled foam for soft pneumatic actuators. A moderate amount of residual compressive strain within elastomer foam increases the applied force ∼1.4 × or stroke ∼2 × compared with actuators without residual strain. The origin of these improved characteristics is explained analytically. These actuators are applied in a direct cardiac compression (DCC) device design, a type of implanted mechanical circulatory support that avoids direct blood contact, mitigating risks of clot formation and stroke. This article describes a first step toward a pneumatically powered, patient-specific DCC design by employing elastomer foam as the mechanism for cardiac compression. To form the device, a mold of a patient's heart was obtained by 3D printing a digitized X-ray computed tomography or magnetic resonance imaging scan into a solid model. From this model, a soft, robotic foam DCC device was molded. The DCC device is compliant and uses compressed air to inflate foam chambers that in turn apply compression to the exterior of a heart. The device is demonstrated on a porcine heart and is capable of assisting heart pumping at physiologically relevant durations (∼200 ms for systole and ∼400 ms for diastole) and stroke volumes (∼70 mL). Although further development is necessary to produce a fully implantable device, the material and processing insights presented here are essential to the implementation of a foam-based, patient-specific DCC design.
The performance of mobile soft robots were usually characterized by their locomotion/velocity efficiency, whereas the energy efficiency is a more intrinsic and fundamental criterion for the performance evaluation of independent...