SOFT ROBOTICS

Volume 1, Number 1, 2014

ª Mary Ann Liebert, Inc.
DOI: 10.1089/soro.2013.0001

PERSPECTIVE

Soft Robotics:
A Perspective—Current Trends and Prospects for the Future
Carmel Majidi

Abstract

Soft robots are primarily composed of easily deformable matter such as fluids, gels, and elastomers that match
the elastic and rheological properties of biological tissue and organs. Like an octopus squeezing through a
narrow opening or a caterpillar rolling through uneven terrain, a soft robot must adapt its shape and locomotion
strategy for a broad range of tasks, obstacles, and environmental conditions. This emerging class of elastically
soft, versatile, and biologically inspired machines represents an exciting and highly interdisciplinary paradigm
in engineering that could revolutionize the role of robotics in healthcare, field exploration, and cooperative
human assistance.

Introduction and its similarities to natural organisms, soft robots may be

considered a subdomain of the more general fields of soft-

C onventional robots and machines are made of rigid

materials that limit their ability to elastically deform and
adapt their shape to external constraints and obstacles. Al-
matter engineering or biologically inspired engineering.
However, whereas these existing fields can be defined by their
scientific foundations in soft-matter physics and biology, re-
though they have the potential to be incredibly powerful and spectively, the emerging field of soft robotics remains open
precise, these rigid robots tend to be highly specialized and and free of dogmatic restrictions to any constrained set of
rarely exhibit the rich multifunctionality of natural organisms. methods, principles, or application domains. Instead, soft
However, as the field of robotics continues to expand beyond robotics represents an exciting new paradigm in engineering
manufacturing and industrial automation and into the do- that challenges us to reexamine the materials and mechanisms
mains of healthcare, field exploration, and cooperative human that we use to make machines and robots so that they are
assistance, robots and machines must become increasingly more versatile, lifelike, and compatible for human interaction.
less rigid and specialized and instead approach the mechan-
ical compliance and versatility of materials and organisms
Compliance Matching
found in nature. As with their natural counterparts, this next
generation of robots must be elastically soft and capable of The promise of soft robots is perhaps best realized in en-
safely interacting with humans or navigating through tightly vironments and applications that require interaction with soft
constrained environments. Just as a mouse or octopus can materials and organisms and/or the artificial replication of
squeeze through a small hole, a soft robot must be elastically biological functionalities. For example, whereas industrial
deformable and capable of maneuvering through confined robots typically handle metals, hard plastics, semiconductors,
spaces without inducing damaging internal pressures and and other rigid materials, medical robots will primarily in-
stress concentrations. teract with soft materials such as natural skin, muscle tissue,
In contrast to conventional machines and robots, soft robots and delicate internal organs. Likewise, biologically inspired
contain little or no rigid material and are instead primarily robots for field exploration and disaster relief will often en-
composed of fluids, gels, soft polymers, and other easily de- counter easily deformable surfaces like sand, mud, and soft
formable matter. These materials exhibit many of the same soil. To prevent the robot from penetrating into the surface
elastic and rheological properties of soft biological matter and and causing damage or mechanical immobilization, the forces
allow the robot to remain operational even as it is stretched transferred between the robot and surface must be evenly
and squeezed. Because of the near absence of rigid materials distributed over a large contact area. This requires compliance

Soft Machines Lab, Department of Mechanical Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania.

5
6 MAJIDI

matching—that is, the principle that contacting materials medical implants and tissue growth. For joint replacements,
should share similar mechanical rigidity in order to evenly cardiac stents, and other medical implants, compliance
distribute internal load and minimize interfacial stress con- matching prevents stress concentrations and preserves the
centrations. natural distribution of internal forces and pressure.2 In tissue
One measure of material rigidity is the modulus of elas- growth and engineering, the relative elasticity of contacting
ticity, or Young’s modulus—a quantity that scales with the tissue can influence how tissue cells move, grow, and differ-
ratio of force to percent elongation of a prismatic (i.e., uniform entiate.3 Mismatches in elastic compliance can lead to dam-
cross section) bar that is stretched along its principal axis (Fig. aging stress concentrations, redistribute internal forces in a
1a).1 Young’s modulus is only defined for homogenous, way that leads to disuse atrophy of bone or tissue, or intro-
prismatic bars that are subject to axial loading and small de- duce rigid kinematic constraints that interfere with natural
formations (< 0.2% elongation for metals) and thus has limited motor function.
relevance to soft robots and other soft-matter technologies Compliance matching is particularly important in the
that have irregular (nonprismatic) shape and undergo large subdomain of wearable technologies for human motor assis-
elastic or inelastic deformations. Nonetheless, Young’s mod- tance. These soft robot technologies are wearable and contain
ulus is a useful measure for comparing the rigidity of the artificial muscles that match the compliance of natural muscle
materials that go into a soft robot. As shown in Figure 1b, and provide physical assistance to humans who have motor
most conventional robots are composed of materials such as impairments or are engaged in strenuous tasks. As with nat-
metals and hard plastics that have a modulus of greater than ural muscle, these artificial muscles must not only be capable
109 Pa = 109 N/m2. In contrast, most of the materials in natural of reversible shape change but also reversible changes in
organisms, such as skin and muscle tissue, have a modulus on elastic rigidity. For motor tasks that involve underactuated or
the order of 102–106 Pa. That is, the materials in natural or- passive dynamic motions, such as downhill walking, the as-
ganisms are 3–10 orders of magnitude less rigid than the sistive robot should be elastically soft and avoid interfering
materials in conventional robots. This dramatic mismatch in with the natural range of joint motion. For physically stren-
mechanical compliance is a big reason why rigid robots are uous motor tasks, the artificial muscle must supply mechan-
often biologically incompatible and even dangerous for inti- ical work and become rigid in order to support large forces. As
mate human interaction and rarely exhibit the rich multi- with natural muscle, the artificial muscle used in wearable
functionality and elastic versatility of natural organisms. soft robots should stiffen in order to prevent injury during
To prevent injury or robot immobility, the surface of soft collisions, absorb impacts, or to catch fast-moving objects.
robots must be adequately soft and deformable in order to
distribute forces over a large contact area and eliminate in-
Potential Applications
terfacial stress concentrations. For contact with human tissue
or organs, stress concentrations may cause physical discom- Because they are composed of materials that match the
fort and even physical injury. For a hard robot in contact with compliance of biological matter, soft robots are mechanically
a soft substrate, stress concentrations can cause the robot to biocompatible and capable of lifelike functionalities. These
puncture or ‘‘dig in’’ to the surface and become immobile. features will potentially lead to plenty of promising new
Compliance matching also has a critical role in areas such as technologies, from the aforementioned soft wearable robots

FIG. 1. (a) The elastic (Young’s) modulus scales with the ratio of the force F to the extension d of a prismatic bar with length
L0 and cross-sectional area A0. (b) Young’s modulus for various materials (adapted from Autumn et al.23).
SOFT ROBOTICS: A PERSPECTIVE 7

for human motor assistance and biologically inspired field multifunctional platforms that can be operated by nonspe-
robots for autonomous exploration to soft and lightweight cialists. Because they will physically interact with humans, co-
cooperative robots that safely interact with people (Fig. 2a). robots must be adequately soft and lightweight in order to
While these technological prospects are certainly exciting, how prevent injuries during collisions. Soft robot features such as
will soft robots specifically be used and what unique oppor- compliance matching and biocompatibility are especially
tunities will they create for society and industry in the future? important for applications in nursing and elderly care that
Perhaps the most immediate application of emerging soft require carrying, lifting, and other forms of intimate contact.
robot technologies will be in the domain of human motor With conventional machines and rigid robots, safe and com-
assistance and co-robotics. For example, a soft active ankle foot fortable human–machine interaction is possible but requires
orthotic (AFO) could help prevent foot dragging for patients precision sensing, fine motor control, and advanced feedback
that suffer gait abnormalities such as drop foot.4 When active, systems. While tractable in specialized applications, feedback-
the AFO would stiffen and supply mechanical work to the based compliance can be challenging in general-purpose
ankle to assist with lifting the foot. In its passive state, the AFO platforms, especially humanoid robots that must safely co-
would remain soft and allow the ankle joint to freely rotate. operate with humans in a broad range of medical, industrial,
Soft wearable robots could also assist with grasping and other and domestic tasks. In order to minimize demands on sens-
fine motor tasks in patients who have suffered stroke or ing, motors, and computation, future generations of co-robot
traumatic brain injury. As with the AFO, a soft hand orthotic platforms should be primarily composed of materials and
would contain artificial muscles that reversibly change shape machinery that are elastically soft and naturally match the
and elastic rigidity to alternately supply assistive mechanical compliance of human tissue. This same condition also applies
work and accommodate passive motion in the fingers and to soft prosthetics that are powered by artificial and natural
wrist. In addition to matching the natural compliance of hu- muscle and controlled through cognitive commands, body
man skin and tissue, the hand orthotic must be thin, com- gestures, and onboard sensing.
fortable, and lightweight. Such assistive technologies will Instead of conventional electric motors and hydraulics,
effectively function like a second skin5 that compensates for some existing soft human exoskeletons, robot arms, and hu-
missing or impaired motor function by cooperating with the manoids use pneumatic air muscles. The pneumatic air
body’s healthy tissue (Fig. 2b). By minimizing dependency on muscle, also known as a McKibben actuator, is a type of ar-
a physical therapist, these second-skin soft technologies can tificial muscle composed of an inflating balloon encased in a
give the patient greater physical independence and new op- braided shell of woven inextensible fibers. As compressed air
portunities to relearn or discover motor functions for grasping is delivered to the balloon, the braided shell constrains the
and gait. muscle to increase its diameter and shorten. In addition to
Assistive robots that cooperate with human partners—also shortening, the pressurized air muscle exhibits greater tensile
known as co-robots—will have an increasingly central role in rigidity—that is, more force is required to elongate the muscle
a broad range of social, scientific, and industrial activities. As by a prescribed amount. Pneumatic air muscles were origi-
with the personal computer, the universal integration of co- nally introduced by A.H. Morin and later adapted by J.L.
robots into society and industry will depend on robust and McKibben for applications in orthotics.6

FIG. 2. (a) Humanoid co-robot for elderly care; (b) ‘‘second skin’’ for human motor assistance.
8 MAJIDI

While promising, pneumatic air muscles rely on external have a unique role in any application that involves physical
pneumatic hardware such as valves, pumps, and compressors interaction with the human body or demands the levels of
in order to control the delivery of pressurized air. For a robot multifunctionality and elastic versatility observed in nature.
to be completely soft and autonomous, these artificial muscles However, just as conventional machines and robots are not
should be operated with soft or miniaturized pumps and always well suited for human–machine interaction, soft ro-
valves that can be embedded in the robot without introducing bots are fundamentally limited by their mechanical compli-
elastic rigidity. Soft or miniaturized pneumatic hardware are ance and will not be appropriate for applications requiring
also needed in recent bioinspired robots that use variations on high power or precision. For example, it is unlikely that soft
the pneumatic air muscle. These include the pneu-Net7 and machines composed entirely of fluids and elastomers would
Suzumori bending actuators for soft robot limb motion.8 By ever replace heavy-duty industrial robots. Likewise, on the
making these soft robots completely autonomous, they will be small scale, machine precision often requires rigid parts that
capable of crawling and swimming through tightly confined lock tightly in place and do not slacken or deform elastically
spaces that are impossible to navigate with rigid or tethered when loaded with surface tractions. Also, while natural neu-
robots. Specific applications include search operations for ral tissue is soft and capable of extraordinary computational
natural disaster relief, field operations for military recon- power, microengineered electronics are presently constructed
naissance, and pipe inspection for sewer maintenance (Fig. 3). from rigid materials with precisely spaced submicron fea-
Of course, soft robots are not limited to pneumatically tures. Until there is an elastically soft artificial brain, soft ro-
powered humanoids, orthoses, and prosthetics. As artificial bots will require rigid microprocessors for signal processing
muscle, skin, and nervous tissue technologies are further and actuator control.
miniaturized, soft robots will eventually be scaled down to
the size of small invertebrates, insects, and microorganisms.
Beyond Robotics
At these length scales, functionality will depend not only on
soft elasticity but also on the complex rheology of fluids, gels, Like its host platform, the artificial muscle, skin, and neural
and other inelastic soft matter. Beyond their potential role as tissue used in soft and bioinspired robots will be elastically
field robots for search missions and data collection, minia- soft and remain functional when deformed. In addition to
turized soft robots may also eventually be used for drug de- their potentially transformative role in robotics, these soft-
livery, minimally invasive surgery, and medical implants. Just matter technologies will also be used in personal electronics,
as with wearable robots, they should match the compliance of artificial organs, wearable computing, and other applications
internal organs and be capable of navigating through the that involve permanent or frequent contact with the human
body without damaging vascular walls and tissue. For ap- body. As the field of soft robotics grows, the supporting soft-
plications such as biopsy and angioplasty, they should also be matter technologies used in sensing, electronics, and actua-
capable of grasping tissue or anchoring to vascular walls tion will continue to mature and will eventually appear in
through mechanical interlocking or adhesion. Lastly, like a application domains. Likewise, the manufacturing methods
colony of ants or termites, a swarm of miniaturized soft robots used to produce soft robots will extend to other areas within
could be used in manufacturing applications to rapidly as- the field of soft-matter engineering and lead to new para-
semble structures from granular matter, burrow through soil digms in the rapid and high-volume production of rigidity-
to survey and extract natural resources, or to transport haz- tuning actuators, soft microfluidic circuits, and stretchable
ardous material. microelectronics.
These proposed applications represent only a few of the Stretchable microelectronics alone represents a cross-
myriad potential uses of soft robots. In general, soft robots cutting and high-impact technology that readily translates

FIG. 3. Soft field robot for military reconnaissance, natural disaster relief, and pipe inspection.
SOFT ROBOTICS: A PERSPECTIVE 9

into other applications. Just as electronic circuits and sensors reversibly tuning their shape and elastic rigidity, they can
must be capable of accommodating the elastic deformation of perform mechanical work and independently control the
a soft robot host, they must also be able to accommodate the distribution of internal load. Also, in contrast to combustion
stretching and bending that arise in wearable computing and engines and high-power electric motors, they can provide
smart textiles. Current approaches to stretchable electronics nonrepetitive and discontinuous actuation without sacrificing
include wavy circuits9,10 and soft microfluidics with conduc- efficiency. In addition to pneumatic air muscles, current
tive liquids.11 Wavy electronic circuits are typically composed technologies include dielectric elastomer actuators, ionic
of thin-film solid-state microelectronics bonded to a pre- polymer metal composites (IPMC), shape-memory alloys and
stretched sheet of soft silicone elastomer, for example, poly- polymers, and liquid-crystal elastomers.15 Along with these
dimethylsiloxane (PDMS). When the elastic sheet relaxes to its existing classes of actuators, future artificial muscles will not
original length, it causes the circuit to buckle into a wavy only power soft robots and assistive wearable technologies
pattern. The waves flatten out as the circuit is stretched and but also control the valves, pumps, and relays in soft-matter
allow for elongations that are well above the intrinsic strain microfluidics and be used in medical implants, minimally
limit of metals and semiconductors. In contrast, soft micro- invasive surgical tools, and diagnostic systems.
fluidic electronics contain no intrinsically rigid materials and Other potential spin-offs of soft robotics are technologies
are instead composed of microfluidic channels of conductive that use soft gels, colloidal substances, and rheologically
liquid, typically a liquid-phase gallium–indium alloy, em- complex fluids. In miniaturized soft robots, these materials
bedded in a silicone elastomer. As the surrounding elastomer may be used for pseudopod-like locomotion, adhesion, and
is stretched, the fluidic microchannels remain intact and de- grasping. In the case of magneto- and electrorheological flu-
form without losing conductivity. However, unlike wavy ids, which contain a high concentration of microparticles
electronics, soft microfluidic electronics are presently re- suspended in a carrier oil, such colloidal substances can also
stricted to liquid-phase metal alloys and electrolytic solutions be used for control valves in microfluidics16 or for rigidity-
that are conductive but do not have the semiconducting tunable artificial muscles.17 As the field of soft robotics ma-
properties required for transistor-based logic. tures, it will have an increasingly central role in efforts to
In addition to stretchable electronic sensors and circuits, identify new classes of gels and colloidal suspensions that
soft microfluidics have also been used in lab-on-a-chip tech- reversible changes their elastic, rheological, optical, and
nologies for applications such as biological cell sorting and morphological properties in response to external stimuli.
diagnostic assay analysis.12 Soft microfluidic circuits are cur- Imagine, for example, a wearable array of light-emitting gel
rently produced with soft lithography fabrication methods diodes that is elastically compatible with natural skin and
based on nonphotolithographic techniques such as replica functions as a wearable display or diagnostic tool for indi-
molding and microcontact printing.13 Silicone elastomers vidually stimulating photosensitive biological cells.
such as PDMS are embedded with microchannels of fluid that
flow under the influence of electrophoresis, electroosmotics,
Commercial Prospects
or peristalsis.14 Progress in soft microfluidics and soft lithog-
raphy microfabrication will lead to new families of valves, To be commercially viable, unique applications and func-
pumps, and relays to support the actuators and electronics tionalities are not enough—soft-matter robots and technolo-
used in soft robots. Likewise, soft robots will provide a new gies must also be inexpensive and mass-producible. Soft
source of technological demands that will continue to drive robots are currently produced with soft-lithography-
the nascent field of soft-matter engineering. manufacturing techniques that eliminate the need for slow
Artificial muscles represent another cross-cutting domain and costly clean-room fabrication and instead rely on replica
of soft robotics that will allow machines to be more light- molding and transfer printing. Templates and masters are
weight and elastically compatible with the human body. By fabricated with photolithography or rapid prototyping tools

FIG. 4. An overview of soft robotics, potential spin-off technologies, manufacturing methods, and commercial markets.
10 MAJIDI

such as laser micromachining, CNC milling, and 3D printing. applications, biohybridization can dramatically improve
These methods allow for inexpensive and easily customizable performance and overcome some of the fundamental barriers
fabrication and enable the manufacturing costs of soft robots encountered with synthetic soft materials. Consider, for ex-
to be competitive with conventional robots produced from ample, artificial muscles for reversible shape and rigidity
hard plastic. Moreover, the artificial muscles, skin, and control. Currently, most soft robots are electrically powered
nervous tissue that support robot functionality may even- with shape-memory alloys, IPMCs, and dielectric elastomer
tually be produced with stencil lithography, roll-to-roll actuators. For a robot to be untethered and autonomous, these
manufacturing, and direct ink-jet printing. For example, a actuators would require on-board electricity from an alkaline
millimeter-thin PDMS rubber embedded with a 0.1% volume or lithium-ion battery. However, the energy density of bat-
fraction of microfluidic liquid gallium–indium channels will teries is 10–100 times less than that of the sugars and fats
have a raw materials cost of approximately $100 per square used to power natural muscle. Therefore, replacing battery-
meter, approximately the same order of magnitude as the sale powered actuators and electronics with biohybrid materials
price of flexible copper circuit on polyester. that run on chemical fuel could lead to dramatically lighter
Apart from cost and manufacturing scalability, commercial and more autonomous soft robots.
viability also strongly depends on immediate consumer in-
terest. In the short-term, market demand will likely be driven Grand Challenges
by the medical robotics and gaming/entertainment industries
In closing, soft-matter engineering represents an exciting
(Fig. 4). Soft-matter sensors and electronics could be used in
new paradigm in robotics that has the potential to revolu-
gloves and orthoses that monitor hand gestures and joint
tionize its role in society and industry. In application domains
motion. In contrast to existing ‘‘dataglove’’ technologies,18
such as medical and personal co-robotics, soft-matter ma-
these wearable electronics would be inexpensive and com-
chines and robots allow for safe and biomechanically com-
posed almost entirely of soft elastic material. In the longer term,
patible interactions with humans. For field exploration and
soft wearable technologies may contain actuators and low-
disaster relief, soft robots can navigate challenging terrain and
power electrodes for user feedback and muscle stimulation.
penetrate tightly confined spaces by adapting their shape and
Also, as co-robots enter the marketplace, the artificial muscles
locomotion strategy in ways similar to natural organisms. At
and skin used in wearable technologies for medicine and
the small scale, miniature soft robots could function as arti-
gaming could eventually replace the rigid motors and sensors
ficial microorganisms in medical applications such as drug
in humanoids. ABI research predicts that the market for per-
delivery, angioplasty, and biopsy.
sonal robots may grow to $6.5 billion by 2017,19 a significant
As a field of academic research, soft robotics is highly in-
reduction from previous estimates but still a sizeable figure that
terdisciplinary and introduces several grand challenges that
could even be exceeded if robots are eventually designed to be
demand further scientific exploration. One of these is to in-
more lightweight, cheap, and safe for human contact.
troduce new classes of electrically and chemically powered
The 3D printing technology is another emerging market
soft-matter actuators that exhibit the shape and rigidity-
that is closely aligned with soft robotics. According to Lux
tunable properties of natural muscle tissue. Similarly, soft
Research, 3D printer sales may reach as high as $8.4 billion by
robotics requires artificial skin and neural tissue that are
2025.20 As with personal robotics, soft robotics may accelerate
elastically soft and can be embedded without introducing
this growth by expanding the market to include nonspecial-
kinematic constraints and rigidity. Also, as parallel efforts in
ists. An inexpensive 3D printer capable of producing soft-
synthetic biology and tissue engineering continue to advance,
matter electronics, machines, and robots would enable
there will be an increasing need for biocompatible technolo-
hobbyists, school robotics clubs, and artists to participate in
gies that support living cells and tissue. Lastly, commercial
the discovery of new soft robot functionalities and designs.
success depends on new innovation in soft lithography, 3D
printing, and other rapid prototyping technologies to mass
Living Robots? produce soft-matter machines and robots that are inexpensive
and satisfy market demand.
If soft robots achieve their extraordinary multifunctionality
with materials that match the elastic and rheological proper-
Author Disclosure Statement
ties of biological matter, why not just directly build them from
biological material? Hybridization of synthetic and biological No competing financial interests exist.
materials with tissue engineering and synthetic biology rep-
resents another emerging trend that will eventually lead to References
technologies that are more biocompatible and life-like. When
1. Gere JM, Timoshenko SP. Mechanics of Materials, 2nd edn.
applied to soft robotics, tissue engineering and synthetic bi-
Boston, MA: PWS-KENT, 1984.
ology could result in bending actuators that are powered with 2. Venkatraman S, Boey F, Lao LL. Implanted cardiovascular
natural muscle tissue21 or soft-matter circuits composed of polymers: natural, synthetic and bio-inspired. Prog Polymer
genes, protein, and bacteria.22 As with other soft-matter Sci 2008;33:853–874.
technologies, these biohybrid soft robots would contain vir- 3. Mammoto T, Ingber DE. Mechanical control of tissue and
tually no rigid materials and could be produced with soft organ development. Development 2010;137:1407–1420.
lithography or 3D printing. 4. Blaya JA, Herr H. Adaptive control of a variable-impedance
Like their engineering counterparts, natural muscles and ankle-foot orthosis to assist drop-foot gait. IEEE Trans
biological circuits have their own fundamental limitations Neural Syst Rehab Eng 2004;12:24–31.
and will eventually be used to complement, rather than re- 5. Goldfield EC, Park Y-L, Chen B-R, Hsu W-H, Young D,
place, nonbiological technologies. Nonetheless, for some Wehner M, et al. Bio-inspired design of soft robotic assistive
SOFT ROBOTICS: A PERSPECTIVE 11

devices: the interface of physics, biology, and behavior. Ecol 17. Majidi C, Wood RJ. Tunable elastic stiffness with micro-
Psychol 2012;24:300–327. confined magnetorheological domains at low magnetic field.
6. Daerden F, Lefeber D. Pneumatic artificial muscles: actua- Appl Phys Lett 2010;97:164104.
tors for robotics and automation. Euro J Mech Environ Eng 18. Dipietro L, Sabatini AM, Dario P. A survey of glove-based
2002;47:11–21. systems and their applications. IEEE Trans Systems Man
7. Ilievski F, Mazzeo AD, Shepherd RF, Chen X, Whitesides Cybernetics Part C Appli Rev 2008;38:461–482.
GM. Soft robotics for chemists. Angew Chem Int Ed 2011;50: 19. ABI Research. Consumer and Personal Electronics. New
1890–1895. York: Allied Business Intelligence, Inc., 2013.
8. Suzumori K, Iikura S, Tanaka H. Development of flexible 20. Lux Research. Building the Future: Assessing 3D Printing’s
microactuator and its applications to robotics mechanisms. Opportunities and Challenges. Boston, MA: Lux Research,
Proceedings of 1991 IEEE International Conference on Ro- Inc., 2013.
botics and Automation, Sacramento, CA. 21. Feinberg AW, Feigel A, Shevkoplyas SS, Sheehy S, White-
9. Lacour SP, Wagner S, Huang Z, Suo Z. Stretchable gold sides GM, Parker KK. Muscular thin films for building
conductors on elastomeric substrates. Appl Phys Lett 2003; actuators and powering devices. Science 2007;317:1366–
82:2404–2406. 1370.
10. Rogers JA, Someya T, Huang Y. Materials and mechanics for 22. Gardner TS, Cantor CR, Collins JJ. Construction of a ge-
stretchable electronics. Science 2010;327:1603–1607. netic toggle switch in Escherichia coli. Nature 2000;403:
11. Cheng S, Wu Z. Microfluidic electronics. Lab Chip 2012;12: 339–342.
2782–2791. 23. Autumn K, Majidi C, Groff RE, Dittmore A, Fearing R. Ef-
12. Quake SR, Scherer A. From micro- to nanofabrication with fective elastic modulus of isolated gecko setal arrays. J Exp
soft materials. Science 2000;290:1536–1539. Biol 2006;209:3558–3568.
13. Xia Y, Whitesides GM. Soft lithography. Angew Chem Int
Ed 1998;37:550–575.
14. Unger MA, Chou H-P, Thorsen T, Scherer A, Quake SR.
Address correspondence to:
Monolithic microfabricated valves and pumps by multilayer
soft lithography. Science 2000;288:113–116. Carmel Majidi, PhD
15. Madden JDW, Vandesteeg NA, Anquetil PAA, Madden Soft Machines Lab
PGA, Takshi A, Pytel RZ, et al. Artificial muscle technology: Department of Mechanical Engineering
physical principles and naval prospects. IEEE J Ocean Eng Carnegie Mellon University
2004;29:706–728. 5000 Forbes Avenue
16. Qian B, McKinley GH, Hosoi AE. Structure evolution in Pittsburgh, PA 15213
electrorheological fluids flowing through microchannels.
Soft Matter 2013;9:2889–2898. E-mail: cmajidi@andrew.cmu.edu
This article has been cited by:

1. Wehner Michael, Tolley Michael T., Mengüç Yiğit, Park Yong-Lae, Mozeika Annan, Ding Ye, Onal Cagdas, Shepherd Robert
F., Whitesides George M., Wood Robert J.. 2014. Pneumatic Energy Sources for Autonomous and Wearable Soft Robotics. Soft
Robotics 1:4, 263-274. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links]
2. 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]
3. Holland Dónal P., Park Evelyn J., Polygerinos Panagiotis, Bennett Gareth J., Walsh Conor J.. 2014. The Soft Robotics Toolkit:
Shared Resources for Research and Design. Soft Robotics 1:3, 224-230. [Abstract] [Full Text HTML] [Full Text PDF] [Full
Text PDF with Links]
4. Nurzaman Surya G., Iida Fumiya, Margheri Laura, Laschi Cecilia. 2014. Soft Robotics on the Move: Scientific Networks,
Activities, and Future Challenges. Soft Robotics 1:2, 154-158. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF
with Links]