See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/346266320

Soft Robots

Article · February 2021

DOI: 10.1007/978-3-642-41610-1_146-2

CITATIONS READS

0 1,131

3 authors, including:

Cosimo Della Santina Antonio Bicchi

Delft University of Technology IIT Genova and Univ. Pisa
73 PUBLICATIONS   732 CITATIONS    619 PUBLICATIONS   19,311 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

WALK-MAN View project

Robotics: dynamics and control View project

All content following this page was uploaded by Cosimo Della Santina on 25 February 2021.

The user has requested enhancement of the downloaded file.

 Noname manuscript No.
(will be inserted by the editor)

Soft Robots

Cosimo Della Santina · Manuel G. Catalano ·

Antonio Bicchi

Received: date / Accepted: date

1 Synonyms

Flexible Robots, Compliant Robots, Elastically Actuated Robots

2 Definition

Soft robots are robotic systems with purposefully designed compliant elements em-
bedded into their mechanical structure.

3 Overview

The physical characteristics of animals’ bodies are substantially different from those
of classic robots. Elastic tendons, ligaments and muscles enable animals to robustly
interact with the external world and perform dynamic tasks (Fig. 1). The function of
elastic elements in natural bodies is discussed in [69], and summarized in table 1.
As opposed to natural bodies, the physical structure of traditional robots has gen-
erally been very stiff. This has reflected the preeminent design goal of accuracy and

Cosimo Della Santina

Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, 32 Vassar
St., Cambridge, MA 02139, USA
E-mail: [email protected]
Manuel G. Catalano
Fondazione Istituto Italiano di Tecnologia, via Morego, 30, 16163 Genova, Italy
E-mail: [email protected]
Antonio Bicchi
Fondazione Istituto Italiano di Tecnologia, via Morego, 30, 16163 Genova, Italy
&
Università di Pisa, Largo Lucio Lazzarino 1, 56126 Pisa, Italy
E-mail: [email protected]
2 Cosimo Della Santina et al.

(a) (b) (c)

Fig. 1 Compliance is ubiquitous in natural bodies, and appears in disparate forms. The only rigid body
part of an octopus (a) is its beak, which is about 1 inch wide, so that this invertebrate mollusc can squeeze
in small crevices to prey. The compliant body of vertebrates such as humans (b) enables them to perform
dynamic motions with high resilience to impacts. Some vertebrates have continuously deformable parts,
such as e.g. elephants (c) whose continuously deformable trunk (c) is dexterously used to reach and grasp
objects placed in narrow and/or distant locations.

non-deformability under loads. In recent years, attention has shifted from heavy-
industrial applications to human-centered and service-oriented goals . This produced
in turn a shift in the design approach, from the traditional “design for accuracy, con-
trol for safety” to “design for safety, control for performance”. As part of the efforts to
guarantee safety and comfort to humans co-existing with robots, the co-design of the
physical and cognitive parts of a robot system was advanced, which made possible
the advent of ”soft robotics”. Inspired by the natural example, in soft robots elastic
elements are purposefully introduced into their physical structure, to obtain “natural”
motions which embody in the very morphology of the robot the intelligent principles
of motor control in humans and animals.
Two main branches exist in soft robotic research according to their main source
of inspiration. The first type, inspired by the vertebrate muscle-skeletal system [2],
typically has compliant elements in the actuation or in the transmission elements, and
is referred to as articulated soft robots. The other branch, inspired by invertebrates (or
boneless animal body parts) contains continuously deformable elements, and are of-
ten referred to as continuum soft robots1 . Fig. 2 shows some notable examples of soft
robots of different types. In the evolution of robotics, articulated soft robots have their
ancestors in flexible joint robots [17], where however elasticity was mostly regarded
as a parasitic effect rather than purposefully designed and introduced. Continuum soft
robots have roots in the early 90’s [81, 71], although they grew very popular since the
2000’s. They are often characterized by innovative designs and materials, which re-
quire inventive control solutions. Their precursors in robotic research are flexible link
robots [48] and hyper-redundant manipulators [15]. Some other characteristics that
differentiate soft robots from their predecessors are that soft robots purposefully un-
dergo large-scale deformation with nonlinear characteristics; have several degrees of
freedom and are underactuated; perform significant, non-trivial computations with
their physical structure [63].

1 The term continuous robots is also used, albeit less frequently

Soft Robots 3

Table 1 Different uses of compliance in the natural body. The pictures represent a simplified natural
articulation, with only one elastic tendon, a single muscle and a concentrated inertial mass. The same
concepts however extend to invertebrate bodies. The colored areas indicate where energy is stored, while
arrows show energy flows. This table is partially adapted from [69].
Graphical representation Used in
Adapting to uncertainties,
steady state force generation,
sustaining low frequency
perturbations.
Low frequency

Adaptability

Jumping,
incline running,
ballistic feeding,
Power amplification

fast acceleration,
object throwing.

Landing,
reacting to
unexpected impacts,
vibration absorption,
High frequency

Power absorption

decline running.

Running,
walking,
Power conservation

crawling,
swimming,
periodic pick and place.
4 Cosimo Della Santina et al.

Fig. 2 Soft robots can be broadly classified in continuum and articulated classes, depending on the type
of natural inspiration. The first class includes robots inspired by more ancient biological species, such as:
unicellular organisms [61, 43]; plants [75, 38]; jellyfish [91, 29]; echinoderms [78, 85]; molluscs [54, 92,
30]; worms [7, 76], and fish [46, 62]. The second class includes robots inspired by more recent species,
such as amphibians [41]; reptiles [49, 73]; birds [26] (and www.agilityrobotics.com), and mammals [8, 60,
45, 88, 77, 39, 44, 6, 51, 80].

4 Key Research Findings

Soft robotics is a vast research field with important applications, which pose challeng-
ing problems ranging from modeling to design and control. Research in these aspects
is thriving and a brief account of the main research topics and results is provided
below.

4.1 Modeling

The dynamics of an articulated soft robot can be written by the standard ordinary
differential equations for a Lagrangian system as

B(p) p̈ + h(p, ṗ) + d(p, ṗ) = A(p) τ − T (p) , (1)

where p ∈ RN describes the configurations of the N inertial bodies in the system,

and τ ∈ Rm is a vector of actuator and external forces. A(p) ∈ RN×m is an input
connection matrix modeling the transmission system acting between actuators and
robot configurations. B(p) ∈ RN×N is the inertia matrix, c(p, ṗ) ∈ Rn collects Coriolis
and centrifugal terms, and d(p, ṗ) ∈ Rn is a generic nonlinear damping term. The term
T (p) ∈ RN = ∂V∂ (p)
p represents the gravitational and elastic forces, corresponding to
Soft Robots 5

the potential energy function V (p). A reformulation of (1) in the port-Hamiltonian

formalism is sometimes useful, as discussed in [56].
When the interaction between links and motors is fully mediated by the elastic
elements, the configuration vector p is often explicitly split into robot and motor
configuration, the first often being referred as q and the latter as θ (see e.g. [1]).
Despite the apparent simplicity of the model of articulated soft robots, much re-
search is being directed to tailoring them to the many different types of actuator ar-
rangements (series and parallel elastic actuators, variable stiffness actuators etc., see
Sect. 4.2 below) and giving them forms so that they can be identified and controlled
effectively [4], [9].
Modeling soft continuum robots is challenging due to the continuous nature of
their states, whose description requires partial rather than ordinary differential equa-
tions . As an example, the classic Euler-Bernoulli slender beam equation describing
the horizontal displacement w : [0, xmax ] → R of an inextensible planar beam is
!
∂ 2w ∂4 ∂w ∂2 ∂ 2w
µ(x) 2 + ξ 4 = A(x)τ(x,t) − 2 E(x)I(x) 2 , (2)
∂t ∂ x ∂t ∂x ∂x

where x ∈ [0, xmax ] is the horizontal coordinate; µ : [0, xmax ] → R+ is the mass per
unit length in x; E : [0, xmax ] → R+ and I : [0, xmax ] → R+ are the elastic modulus and
the second moment of inertia, respectively, of the cross-section in x, and ξ is a damp-
ing coefficient. Finally, τ : [0, xmax ] × R → R denotes actuator and external forces.
Suitable boundary conditions (in the general form of ordinary differential equations)
must be added at the interfaces of the continuous body with its surroundings (other
bodies, joints, or actuators).
General formulations of the elastic theory, which are particularly relevant for soft
robotics [86] are the Kirchhoff-Clebesch-Love [25], and Cosserat [37] rod theories.
Although very well studied in the literature, practical use of continuum models is
challenging, since analytical solutions exist only in very specific conditions.

The adaptation of discretization techniques from continuum mechanics to soft

robotics has highlighted that accurate dynamic models can be derived by space dis-
cretization of the continuum dynamics. Relevant examples are Ritz-Galerkin projec-
tion in polynomial subspaces [74], finite element modeling [13], and direct space
discretization of the rod model [68, 35]. Such approximations again lead to dynamic
forms equivalent to the one in (1), albeit possibly with large state-space dimensions.

4.2 Design

There are two main components that make up the body of classic rigid robots: struc-
tural elements (links) and actuation sources (motors). In soft robots, elements for
energy storing and damping are also introduced. These parts roughly correspond to
artificial counterparts of animal bones and ligaments, muscles, and tendons, although
it is not easy to establish a one-to-one mapping . Figure 3 presents some of the tech-
nologies used in the design of soft robots. Standard torsion and compression springs,
6 Cosimo Della Santina et al.

(a) Torsion spring (b) Compression spr. (c) PAM (d) Tendon

(e) Vacuum (f) Fluid pressure (g) Dielectric elastomer (h) Shape memory mat.

Fig. 3 Examples of elastic components used in soft robots. Panels (a) and (b) show a torsion [11] and a
compression spring. Panel (c) shows a McKibben actuator [90], and an example of a pneumatic artificial
muscle (PAM). A tendon-driven continuum soft limb [54] is presented in panel (d). Panels (e,f) depict
vacuum [70] and pressure [20] driven fluid actuators. Panels (g,h) show examples of smart materials that
can be used as soft actuators: a dielectric elastomer [5], and a shape memory polymer [55] respectively.

as in Figure 3 (a,b), are commonly used in articulated soft robots to store elastic en-
ergy. Examples of robots using these constitutive elements are [49, 45, 88, 77, 39, 44,
51, 80].
In many cases, however, the components of soft robots have more than one func-
tion, i.e. both a source of actuation and an energy storage element with damping. For
instance, pneumatic artificial muscles [90], an example of which is shown in Fig. 3
(c) carry out actuation and elasticity simultaneously, with a degree of damping. Their
model is therefore not dissimilar from models of animal muscles [83]. Examples of
robots using these constitutive elements are [60, 6].
There are several possible interconnection layouts between elastic, structural el-
ements, and actuation sources. Some of the most important ones are depicted in Fig.
4. As an example, panel (a) shows a series elastic actuator [65]. Here a compliant
element (with stiffness k in the figure) is used to decouple the prime mover (with
configuration p1 ) from the output shaft (with configuration p2 ). Another common
layout is the parallel elastic actuator [58]. Here, elasticity is added in parallel to the
actuation source. Fig. 4 (b) shows a two-link robot with parallel elastic actuation.
Panel (c) presents an interesting class of soft robotics actuation mechanisms, i.e.
variable stiffness actuators. The example depicted here is an agonist-antagonist actu-
ator [89], as this layout directly reproduces a pair of agonist-antagonist muscles [28].
When the two actuators move in coordination in the same direction, the output shaft
also moves in that direction if free. However, when the actuators act proportionately
in opposite directions (co-contraction), the output shaft remains still. Provided that
Soft Robots 7

(a) SEA (b) PEA

(c) VSA: agonist-antagonist

Fig. 4 Three common examples of how elastic elements can be introduced into robot mechanics. Panel (a)
depicts a simple variable stiffness actuator - or VSA [89]- with an antagonistic architecture. Two bodies
are serially connected to a third one through nonlinear springs. There are several possible equilibrium
configurations with the same p3 but different p1 and p2 . Each of them is characterized by a different
stiffness measured at p3 . Panel (b) shows a series elastic actuator, or SEA [65]. Two bodies free of rotating
around their centers, are connected through a (linear) torsional spring, which couples their motion. In
parallel elastic actuators, or PEA [58], the softness is connected in parallel to the actuation source. Panel
(c) shows a robot with two revolute joints, each of which with parallel elastic actuation.

the elastic elements are correctly designed, for example with nonlinear springs, the
stiffness of the output link is varied, so that under an external load the link deflection
is different. Variable stiffness actuation can also be obtained in many other different
arrangements, including in continuum soft robots, as discussed in [57]. Series and
parallel schemes can be combined in various ways to obtain more complex intercon-
nections, such as in the micro-macro actuator discussed in [94].
In continuum soft robots the whole robot is soft, so instead of adding separate ele-
ments, elasticity must be actually built into the structure and the actuation. Soft fluidic
elastomers, which can be either vacuum-powered [70], or pressure-powered [20], are
often used to achieve this. Some examples are shown in Fig. 3 (e,f), respectively. The
advantages of this technology are high specific power and versatility. However, this
entails a fluid power source, which is typically rather cumbersome and rigid. Smart
materials offer an alternative, such as dielectric elastomers [10], and shape memory
polymers [55]. Fig. 3 (g,h) shows two examples of smart materials that can be em-
ployed for building soft actuators. Both fluidic elastomers and smart materials can
also be used as a means of building artificial soft muscles for articulated soft robots.
Tendons (Fig. 3 (d) ) can also be used for transmitting the action from standard
sources, without negatively impacting on the soft nature of the system. This solution
8 Cosimo Della Santina et al.

Fig. 5 The introduction of elastic elements in soft robots produces a physical feedback control loop. This
low level controller is designed to obtain the desired closed loop behavior. The design of the standard
controller must be carried out taking into account the hierarchical nature of the architecture.

is used both in continuum [54] and in articulated [23] soft robots. Despite their im-
portance in defining the impedance of the robot, the design of sources of damping is
a relatively less studied topic: see [12, 50].
The development of compliant mechanisms goes beyond robotics, with appli-
cations spanning the various sub-fields of mechanical engineering - see [42]. for a
review.
Finally, deformable sensors and electronics merit a special mention given their
central role in building the soft robot body, especially for the continuum kind. The
integration of comprehensive sensing into continuum soft robots alongside structures
with high degrees of freedom is challenging and requires specialized methods, ma-
terials, and equipment [53, 87]. It is fundamental that the sensors used are tolerant
to large strains, which are reproducible and stable in the long-term, and which can
be integrated in sufficient quantities - this can be difficult to achieve. As an alter-
native, exogenous sensing approaches, such as motion capture vision systems and
magnetic tracking, have been widely used in research labs for demonstrating proofs
of concepts. An interesting approach towards fully soft logic circuitry is to use fluids
rather than electrons to implement logic ports [66]. A discussion about stretchable
electronics can be found in [24].

4.3 Control

Equation (1) highlights that introducing elastic elements entails algebraic physical
feedback T (q). This aspect is discussed in [34] and [58]. Fig. 5 summarizes this idea
and highlights the hierarchical nature of the resulting control problem. This physical
feedback action has characteristics that differentiate it from a standard control loop
implemented with actuators, sensors and a microcontroller. To be physically imple-
mentable the field has to be integrable, i.e. a function V (q) : Rn → R should exist
such that T (q) = ∂V∂ (q)
q . Moreover, the feedback cannot produce energy. Thus V (q)
Soft Robots 9

has to be semi-positive defined. On the other hand, apart from the obvious benefit
of not requiring any computation power to be realized. physical feedback has two
other major advantages. Firstly, elastic elements can operate at high frequencies, and
this property is central in shock absorption, as shown by Fig. 6. Secondly, the elastic
elements do not require any energy to operate.
Let us now discuss the design of the actual controller. Articulated soft robotics
has benefited from its close relationship with flexible joint robots, by inheriting a
vast literature dealing with model based control, for example: PD control [19], feed-
back linearization [79], and impedance control [3]. With the development of effective
models for these systems, model-based control has also become a viable option for
controlling continuum soft robots, both for quasi-static [27], and dynamic [22] tasks.
Although they are very effective in achieving their control goals, these techniques
are often applied disregarding the hierarchical nature of the problem discussed in the
previous subsection. This can lead to serious problems, such as the alteration of the
compliance purposefully introduced in the design phase, and the consequent robot
stiffening [21]. Recent work has tackled the problem of designing control strategies
overcoming this issue. The design of feedback controllers that maintains the physical
properties of soft robots can be achieved by structure-preserving controllers, such as
dynamic gravity cancellation [18], and ESP control [47]. An alternative can be found
in animals, whose brain generates motor commands by exploiting learned anticipa-
tory actions, which is feedforward in control language [21].
Another major challenge in soft robots control is to find ways of exploiting elas-
ticity effectively. Optimal control can be used to create meaningful strategies to
achieve this goal. An example is [31], where this technique is used to produce ex-
plosive motions. On the other hand, the capacity to periodically store and release
energy is studied through the theory of nonlinear oscillations [32], with promising
applications in soft locomotion [45, 51].
While the lack of tractable models initially prevented the direct application of
some of the results discussed above to the control of continuum soft robots, it also
drove researchers to find innovative solutions to solve the control problem. Very in-
teresting examples of this line of research are learning-based techniques, for example
[84]. Alternatively, evolutionary robotics techniques can be used to co-evolve the
body and brain of soft robots [14]. Continuum soft robots have also been studied
from the point of view of morphological computation, as discussed in [40]. For more
details on the control of continuum soft robot see [33].

5 Examples of Application

Thanks to their intrinsic compliance, soft robots can perform real world tasks that
could be hardly executed by classic rigid robots. Fig. 7 shows examples of robust
and safe interactions enabled by softness [54, 23], and how this property can be used
in working environments [82]. Thanks to their compliance, soft robots adapt to the
task uncertainties in a way that it is intuitive and predictable for the human. This
makes soft robots ideal co-workers [52, 36], and platforms to be tele-operated [59].
Biomedical and assistive robotics are another natural application field for soft robots.
10 Cosimo Della Santina et al.

(a) (b) (c)

Fig. 6 Soft robots are robust to violent interactions. In panel (a) the robot body supports the weight of a
car [85]. In panels (b) and (c) an articulated [36] and a continuum [76] soft robots are violently hit without
breaking.

(a) (b) (c) (d)

(e) (f)

Fig. 7 Soft robots at work. Panels (a) and (b) demonstrate direct and safe interactions between human
and robot [54, 23]. This enables the use of soft robots as co-workers, as demonstrated in panel (c) with
the Coman robot [52]. Soft robots can also directly substitute humans. In panel (d) the humanoid robot
David [36] drills a wall. In panel (e) the soft robot Walkman [59] operates in a post-earthquake scenario,
controlled by an operator safely standing outside the damaged area. Both configuration and compliance
are here mapped to the robot. In panel (f) the Giacometti continuum soft arm [82] is used for industrial
inspection.

Some examples [72, 64, 67, 23] are depicted in Fig. 8. Finally, the ability of adapting
to extreme stress conditions (see e.g. Fig. 6) makes exploration of harsh environments
Soft Robots 11

(a) (b) (c) (d)

Fig. 8 Soft robots as medical devices. Panel (a) shows a soft robotic sleeve assisting cardiovascular func-
tion in an in vivo porcine model of heart failure [72]. Panel (b) shows a soft hand exoskeleton made of
continuously deformable parts [64]. Panel (c) depicts a soft robot designed for minimally invasive surgery
[67]. In panel (c) and articulated soft prosthesis is used to perform a precision task [23].

an important field of application for soft robots, as shown in Fig. 9, and discussed in
[46, 44, 38].

6 Future Directions for Research

Despite the common natural inspiration and their similarities, continuum and artic-
ulated soft robots emerged at different times and have generally followed separate
paths.
However, given that nature is typically not completely continuous or discrete, also
soft robots are beginning to be designed with both kinds of components. Examples
in this direction are tensegrity robots [93], and rigid-soft hybrids [16]. In fact, most
continuum soft robots already include rigid parts, either as structural elements or as
an actuation source (see Fig. 2). Although less common, there are examples of artic-
ulated soft robots including continuum parts [49]. Nonetheless, we are still awaiting
a real integration between the two technologies , and there is still no complete under-
standing of the pros and cons of using continuum or discrete components.
From a theoretical point of view a common ground between the two fields is cur-
rently being built through the development of dynamic models (see control section).
Open issues are to what extent the techniques and principles developed for articulated
soft robots can be extended to the control of continuously deformable soft robots, and
in turn which methods developed for soft-bodied robots can be applied in articulated
soft robotics.

References

1. Alin Albu-Schäffer and Antonio Bicchi. Actuators for soft robotics. In Springer Handbook of
Robotics, pages 499–530. Springer, 2016.
12 Cosimo Della Santina et al.

(a) (b) (c)

(d) (e) (f) (g) (h) (i)

Fig. 9 Soft robots in exploration. In panel (a), a soft fish swims near to real real ones [46]. In panel
(b) a soft quadruped walks in a disaster area [44]. In panel (c) a soft robot inspects a burning wood
(www.agilityrobotics.com). Finally, panels (d-i) show a photo-sequence of a continuum soft robot sneaking
between two sticky surfaces [38].

2. Alin Albu-Schaffer, Oliver Eiberger, Markus Grebenstein, Sami Haddadin, Christian Ott, Thomas
Wimbock, Sebastian Wolf, and Gerd Hirzinger. Soft robotics. IEEE Robotics & Automation Maga-
zine, 15(3), 2008.
3. Alin Albu-Schäffer, Christian Ott, and Gerd Hirzinger. A unified passivity-based control framework
for position, torque and impedance control of flexible joint robots. The international journal of
robotics research, 26(1):23–39, 2007.
4. Alin Albu-Schäffer, Sebastian Wolf, Oliver Eiberger, Sami Haddadin, Florian Petit, and Maxime
Chalon. Dynamic modelling and control of variable stiffness actuators. In Robotics and Automa-
tion (ICRA), 2010 IEEE International Conference on, pages 2155–2162. IEEE, 2010.
5. Oluwaseun A Araromi, Irina Gavrilovich, Jun Shintake, Samuel Rosset, Muriel Richard, Volker Gass,
and Herbert R Shea. Rollable multisegment dielectric elastomer minimum energy structures for a
deployable microsatellite gripper. IEEE/ASME Transactions on Mechatronics, 20(1):438–446, 2015.
6. Yuki Asano, Kei Okada, and Masayuki Inaba. Design principles of a human mimetic humanoid:
Humanoid platform to study human intelligence and internal body system. Science Robotics,
2(13):eaaq0899, 2017.
7. Alexander S Boxerbaum, Kendrick M Shaw, Hillel J Chiel, and Roger D Quinn. Continuous wave
peristaltic motion in a robot. The international journal of Robotics Research, 31(3):302–318, 2012.
8. Jonas Buchli, Fumiya Iida, and Auke Jan Ijspeert. Finding resonance: Adaptive frequency oscilla-
tors for dynamic legged locomotion. In Intelligent robots and systems, 2006 ieee/rsj international
conference on, pages 3903–3909. IEEE, 2006.
9. Gabriele Buondonno and Alessandro De Luca. Efficient computation of inverse dynamics and feed-
back linearization for vsa-based robots. IEEE robotics and automation letters, 1(2):908–915, 2016.
Soft Robots 13

10. Federico Carpi, Danilo De Rossi, Roy Kornbluh, Ronald Edward Pelrine, and Peter Sommer-Larsen.
Dielectric elastomers as electromechanical transducers: Fundamentals, materials, devices, models
and applications of an emerging electroactive polymer technology. Elsevier, 2011.
11. Giorgio Carpino, Dino Accoto, Fabrizio Sergi, Nevio Luigi Tagliamonte, and Eugenio Guglielmelli.
A novel compact torsional spring for series elastic actuators for assistive wearable robots. Journal of
Mechanical Design, 134(12):121002, 2012.
12. Manuel Catalano, Giorgio Grioli, Manolo Garabini, Felipe Weilemann Belo, Andrea Di Basco, Nikos
Tsagarakis, and Antonio Bicchi. A variable damping module for variable impedance actuation. In
Robotics and Automation (ICRA), 2012 IEEE International Conference on, pages 2666–2672. IEEE,
2012.
13. Jean Chenevier, David González, J Vicente Aguado, Francisco Chinesta, and Elı́as Cueto. Reduced-
order modeling of soft robots. PloS one, 13(2):e0192052, 2018.
14. Nick Cheney, Josh Bongard, Vytas SunSpiral, and Hod Lipson. Scalable co-optimization of morphol-
ogy and control in embodied machines. Journal of The Royal Society Interface, 15(143):20170937,
2018.
15. Gregory S Chirikjian. Hyper-redundant manipulator dynamics: A continuum approximation. Ad-
vanced Robotics, 9(3):217–243, 1994.
16. Utku Culha, Josie Hughes, Andre Rosendo, Fabio Giardina, and Fumiya Iida. Design principles for
soft-rigid hybrid manipulators. In Soft Robotics: Trends, Applications and Challenges, pages 87–94.
Springer, 2017.
17. Alessandro De Luca and Wayne Book. Robots with flexible elements. In Springer Handbook of
Robotics, pages 287–319. Springer, 2008.
18. Alessandro De Luca and Fabrizio Flacco. Dynamic gravity cancellation in robots with flexible trans-
missions. In Decision and Control (CDC), 2010 49th IEEE Conference on, pages 288–295. Citeseer,
2010.
19. Alessandro De Luca, Bruno Siciliano, and Loredana Zollo. Pd control with on-line gravity compen-
sation for robots with elastic joints: Theory and experiments. Automatica, 41(10):1809–1819, 2005.
20. Raphael Deimel and Oliver Brock. A novel type of compliant and underactuated robotic hand for
dexterous grasping. The International Journal of Robotics Research, 35(1-3):161–185, 2016.
21. Cosimo Della Santina, Matteo Bianchi, Giorgio Grioli, Franco Angelini, Manuel Catalano, Manolo
Garabini, and Antonio Bicchi. Controlling soft robots: balancing feedback and feedforward elements.
IEEE Robotics & Automation Magazine, 24(3):75–83, 2017.
22. Cosimo Della Santina, Robert K Katzschmann, Antonio Bicchi, and Daniela Rus. Model-based dy-
namic feedback control of a planar soft robot: trajectory tracking and interaction with the environment.
The International Journal of Robotics Research, page 0278364919897292, 2019.
23. Cosimo Della Santina, Cristina Piazza, Gian Maria Gasparri, Manuel Bonilla, Manuel Giuseppe Cata-
lano, Giorgio Grioli, Manolo Garabini, and Antonio Bicchi. The quest for natural machine motion:
An open platform to fast-prototyping articulated soft robots. IEEE Robotics & Automation Magazine,
24(1):48–56, 2017.
24. Michael D Dickey. Stretchable and soft electronics using liquid metals. Advanced Materials,
29(27):1606425, 2017.
25. Ellis Harold Dill. Kirchhoff’s theory of rods. Archive for History of Exact Sciences, 44(1):1–23, 1992.
26. Courtney E Doyle, Justin J Bird, Taylor A Isom, Jason C Kallman, Daman F Bareiss, David J Dun-
lop, Raymond J King, Jake J Abbott, and Mark A Minor. An avian-inspired passive mechanism for
quadrotor perching. IEEE/ASME Transactions on Mechatronics, 18(2):506–517, 2013.
27. Christian Duriez. Control of elastic soft robots based on real-time finite element method. In Robotics
and Automation (ICRA), 2013 IEEE International Conference on, pages 3982–3987. IEEE, 2013.
28. Anatol G Feldman and Mindy F Levin. The equilibrium-point hypothesis–past, present and future. In
Progress in motor control, pages 699–726. Springer, 2009.
29. Jennifer Frame, Nick Lopez, Oscar Curet, and Erik D Engeberg. Thrust force characterization of
free-swimming soft robotic jellyfish. Bioinspiration & biomimetics, 13(6):064001, 2018.
30. Jan Fras, Yohan Noh, Mateusz Macias, Helge Wurdemann, and Kaspar Althoefer. Bio-inspired oc-
topus robot based on novel soft fluidic actuator. In 2018 IEEE International Conference on Robotics
and Automation (ICRA), pages 1583–1588. IEEE, 2018.
31. Manolo Garabini, Andrea Passaglia, Felipe Belo, Paolo Salaris, and Antonio Bicchi. Optimality
principles in variable stiffness control: The vsa hammer. In Intelligent Robots and Systems (IROS),
2011 IEEE/RSJ International Conference on, pages 3770–3775. IEEE, 2011.
32. Gianluca Garofalo and Christian Ott. Energy based limit cycle control of elastically actuated robots.
IEEE Transactions on Automatic Control, 62(5):2490–2497, 2017.
14 Cosimo Della Santina et al.

33. Thomas George Thuruthel, Yasmin Ansari, Egidio Falotico, and Cecilia Laschi. Control strategies for
soft robotic manipulators: A survey. Soft robotics, 5(2):149–163, 2018.
34. Ambarish Goswami, Michael A Peshkin, and James Edward Colgate. Passive robotics: An explo-
ration of mechanical computation. In Robotics and Automation, 1990. Proceedings., 1990 IEEE
International Conference on, pages 279–284. IEEE, 1990.
35. Stanislao Grazioso, Giuseppe Di Gironimo, and Siciliano Bruno. A geometrically exact model for
soft continuum robots: The finite element deformation space formulation. Soft Robotics, 2018.
36. Markus Grebenstein, Alin Albu-Schäffer, Thomas Bahls, Maxime Chalon, Oliver Eiberger, Werner
Friedl, Robin Gruber, Sami Haddadin, Ulrich Hagn, Robert Haslinger, et al. The dlr hand arm sys-
tem. In Robotics and Automation (ICRA), 2011 IEEE International Conference on, pages 3175–3182.
IEEE, 2011.
37. Albert Edward Green, PM Naghdi, and ML Wenner. On the theory of rods. i. derivations from the
three-dimensional equations. Proc. R. Soc. Lond. A, 337(1611):451–483, 1974.
38. Joseph D Greer, Tania K Morimoto, Allison M Okamura, and Elliot W Hawkes. A soft, steerable
continuum robot that grows via tip extension. Soft robotics, 2018.
39. Duncan W Haldane, Mark M Plecnik, Justin K Yim, and Ronald S Fearing. Robotic vertical jumping
agility via series-elastic power modulation. Science Robotics, 1(1), 2016.
40. Helmut Hauser, Auke J Ijspeert, Rudolf M Füchslin, Rolf Pfeifer, and Wolfgang Maass. Towards a
theoretical foundation for morphological computation with compliant bodies. Biological cybernetics,
105(5-6):355–370, 2011.
41. Hugh Herr and Robert G Dennis. A swimming robot actuated by living muscle tissue. Journal of
neuroengineering and rehabilitation, 1(1):6, 2004.
42. Larry L Howell, Spencer P Magleby, and Brian M Olsen. Handbook of compliant mechanisms. John
Wiley & Sons, 2013.
43. Chaolei Huang, Jiu-an Lv, Xiaojun Tian, Yuechao Wang, Yanlei Yu, and Jie Liu. Miniaturized swim-
ming soft robot with complex movement actuated and controlled by remote light signals. Scientific
reports, 5:17414, 2015.
44. Marco Hutter, Christian Gehring, Dominic Jud, Andreas Lauber, C Dario Bellicoso, Vassilios Tsounis,
Jemin Hwangbo, Karen Bodie, Peter Fankhauser, Michael Bloesch, et al. Anymal-a highly mobile
and dynamic quadrupedal robot. In 2016 IEEE/RSJ International Conference on Intelligent Robots
and Systems (IROS), pages 38–44. IEEE, 2016.
45. Fumiya Iida, Jürgen Rummel, and Andre Seyfarth. Bipedal walking and running with spring-like
biarticular muscles. Journal of biomechanics, 41(3):656–667, 2008.
46. Robert K Katzschmann, Joseph DelPreto, Robert MacCurdy, and Daniela Rus. Exploration of under-
water life with an acoustically controlled soft robotic fish. Science Robotics, 3(16):eaar3449, 2018.
47. Manuel Keppler, Dominic Lakatos, Christian Ott, and Alin Albu-Schäffer. Elastic structure preserving
(esp) control for compliantly actuated robots. IEEE Transactions on Robotics, 34(2):317–335, 2018.
48. Chang Tai Kiang, Andrew Spowage, and Chan Kuan Yoong. Review of control and sensor system of
flexible manipulator. Journal of Intelligent & Robotic Systems, 77(1):187–213, 2015.
49. Sangbae Kim, Matthew Spenko, Salomon Trujillo, Barrett Heyneman, Daniel Santos, Mark R
Cutkosky, et al. Smooth vertical surface climbing with directional adhesion. IEEE Transactions
on robotics, 24(1):65–74, 2008.
50. Matteo Laffranchi, Nikos G Tsagarakis, and Darwin G Caldwell. Compact arm: a compliant manipu-
lator with intrinsic variable physical damping. In Robotics: Science and Systems, volume 8, page 225,
2013.
51. Dominic Lakatos, Kai Ploeger, Florian Loeffl, Daniel Seidel, Florian Schmidt, Thomas Gumpert,
Freia John, Torsten Bertram, and Alin Albu-Schäffer. Dynamic locomotion gaits of a compliantly
actuated quadruped with slip-like articulated legs embodied in the mechanical design. IEEE Robotics
and Automation Letters, 3(4):3908–3915, 2018.
52. Jessica Lanini, Hamed Razavi, Julen Urain, and Auke Ijspeert. Human intention detection as a mul-
ticlass classification problem: Application in physical human–robot interaction while walking. IEEE
Robotics and Automation Letters, 3(4):4171–4178, 2018.
53. Christina Larson, B Peele, S Li, S Robinson, M Totaro, L Beccai, B Mazzolai, and R Shepherd. Highly
stretchable electroluminescent skin for optical signaling and tactile sensing. science, 351(6277):1071–
1074, 2016.
54. Cecilia Laschi, Matteo Cianchetti, Barbara Mazzolai, Laura Margheri, Maurizio Follador, and Paolo
Dario. Soft robot arm inspired by the octopus. Advanced Robotics, 26(7):709–727, 2012.
55. Andreas Lendlein and Steffen Kelch. Shape-memory polymers. Angewandte Chemie International
Edition, 41(12):2034–2057, 2002.
Soft Robots 15

56. Alessandro Macchelli, Claudio Melchiorri, and Stefano Stramigioli. Port-based modeling and simula-
tion of mechanical systems with rigid and flexible links. IEEE transactions on robotics, 25(5):1016–
1029, 2009.
57. Mariangela Manti, Vito Cacucciolo, and Matteo Cianchetti. Stiffening in soft robotics: a review of
the state of the art. IEEE Robotics & Automation Magazine, 23(3):93–106, 2016.
58. Uwe Mettin, Pedro X La Hera, Leonid B Freidovich, and Anton S Shiriaev. Parallel elastic actuators
as a control tool for preplanned trajectories of underactuated mechanical systems. The international
journal of robotics research, 29(9):1186–1198, 2010.
59. Francesca Negrello, Alessandro Settimi, Danilo Caporale, Gianluca Lentini, Mattia Poggiani, Dim-
itrios Kanoulas, Luca Muratore, Emanuele Luberto, Gaspare Santaera, Luca Ciarleglio, et al. Walk-
man humanoid robot: Field experiments in a post-earthquake scenario. IEEE Robotics & Automation
Magazine, 2018.
60. Ryuma Niiyama, Akihiko Nagakubo, and Yasuo Kuniyoshi. Mowgli: A bipedal jumping and landing
robot with an artificial musculoskeletal system. In Robotics and Automation, 2007 IEEE International
Conference on, pages 2546–2551. IEEE, 2007.
61. Stefano Palagi, Andrew G Mark, Shang Yik Reigh, Kai Melde, Tian Qiu, Hao Zeng, Camilla Parmeg-
giani, Daniele Martella, Alberto Sanchez-Castillo, Nadia Kapernaum, et al. Structured light enables
biomimetic swimming and versatile locomotion of photoresponsive soft microrobots. Nature materi-
als, 15(6):647, 2016.
62. Sung-Jin Park, Mattia Gazzola, Kyung Soo Park, Shirley Park, Valentina Di Santo, Erin L Blevins,
Johan U Lind, Patrick H Campbell, Stephanie Dauth, Andrew K Capulli, et al. Phototactic guidance
of a tissue-engineered soft-robotic ray. Science, 353(6295):158–162, 2016.
63. Rolf Pfeifer, Max Lungarella, and Fumiya Iida. The challenges ahead for bio-inspired’soft’robotics.
Communications of the ACM, 55(11):76–87, 2012.
64. Panagiotis Polygerinos, Zheng Wang, Kevin C Galloway, Robert J Wood, and Conor J Walsh. Soft
robotic glove for combined assistance and at-home rehabilitation. Robotics and Autonomous Systems,
73:135–143, 2015.
65. Gill A Pratt and Matthew M Williamson. Series elastic actuators. In Intelligent Robots and Systems
95.’Human Robot Interaction and Cooperative Robots’, Proceedings. 1995 IEEE/RSJ International
Conference on, volume 1, pages 399–406. IEEE, 1995.
66. Daniel J Preston, Philipp Rothemund, Haihui Joy Jiang, Markus P Nemitz, Jeff Rawson, Zhigang Suo,
and George M Whitesides. Digital logic for soft devices. Proceedings of the National Academy of
Sciences, 116(16):7750–7759, 2019.
67. Tommaso Ranzani, Giada Gerboni, Matteo Cianchetti, and A Menciassi. A bioinspired soft manipu-
lator for minimally invasive surgery. Bioinspiration & biomimetics, 10(3):035008, 2015.
68. Federico Renda, Frédéric Boyer, Jorge Dias, and Lakmal Seneviratne. Discrete cosserat approach for
multisection soft manipulator dynamics. IEEE Transactions on Robotics, 2018.
69. Thomas J Roberts and Emanuel Azizi. Flexible mechanisms: the diverse roles of biological springs
in vertebrate movement. Journal of Experimental Biology, 214(3):353–361, 2011.
70. Matthew A Robertson and Jamie Paik. New soft robots really suck: Vacuum-powered systems em-
power diverse capabilities. Science Robotics, 2(9):eaan6357, 2017.
71. Graham Robinson and J Bruce C Davies. Continuum robots-a state of the art. In Proceedings 1999
IEEE international conference on robotics and automation (Cat. No. 99CH36288C), volume 4, pages
2849–2854. IEEE, 1999.
72. 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, et al. Soft robotic sleeve
supports heart function. Science Translational Medicine, 9(373):eaaf3925, 2017.
73. David Rollinson, Yigit Bilgen, Ben Brown, Florian Enner, Steven Ford, Curtis Layton, Justine Rem-
bisz, Mike Schwerin, Andrew Willig, Pras Velagapudi, et al. Design and architecture of a series elastic
snake robot. In Intelligent Robots and Systems (IROS 2014), 2014 IEEE/RSJ International Conference
on, pages 4630–4636. IEEE, 2014.
74. SM Hadi Sadati, S Elnaz Naghibi, Ian D Walker, Kaspar Althoefer, and Thrishantha Nanayakkara.
Control space reduction and real-time accurate modeling of continuum manipulators using ritz and
ritz–galerkin methods. IEEE Robotics and Automation Letters, 3(1):328–335, 2018.
75. Ali Sadeghi, Alice Tonazzini, Liyana Popova, and Barbara Mazzolai. A novel growing device inspired
by plant root soil penetration behaviors. PloS one, 9(2):e90139, 2014.
76. Sangok Seok, Cagdas D Onal, Robert Wood, Daniela Rus, and Sangbae Kim. Peristaltic locomo-
tion with antagonistic actuators in soft robotics. In Robotics and Automation (ICRA), 2010 IEEE
International Conference on, pages 1228–1233. IEEE, 2010.
 16 Cosimo Della Santina et al.

77. Maziar Ahmad Sharbafi, Christian Rode, Stefan Kurowski, Dorian Scholz, Rico Möckel, Katayon
Radkhah, Guoping Zhao, Aida Mohammadinejad Rashty, Oskar von Stryk, and Andre Seyfarth. A
new biarticular actuator design facilitates control of leg function in biobiped3. Bioinspiration &
biomimetics, 11(4):046003, 2016.
78. Robert F Shepherd, Filip Ilievski, Wonjae Choi, Stephen A Morin, Adam A Stokes, Aaron D Mazzeo,
Xin Chen, Michael Wang, and George M Whitesides. Multigait soft robot. Proceedings of the national
academy of sciences, 108(51):20400–20403, 2011.
79. Mark Spong, Khashayar Khorasani, and Petar Kokotovic. An integral manifold approach to the feed-
back control of flexible joint robots. IEEE Journal on Robotics and Automation, 3(4):291–300, 1987.
80. Alexander T Spröwitz, Alexandre Tuleu, Mostafa Ajallooeian, Massimo Vespignani, Rico Möckel,
Peter Eckert, Michiel D’Haene, Jonas Degrave, Arne Nordmann, Benjamin Schrauwen, et al. Oncilla
robot: A versatile open-source quadruped research robot with compliant pantograph legs. Frontiers
in Robotics and AI, 5, 2018.
81. Koichi Suzumori, Shoichi Iikura, and Hirohisa Tanaka. Development of flexible microactuator and its
applications to robotic mechanisms. In Proceedings. 1991 IEEE International Conference on Robotics
and Automation, pages 1622–1627. IEEE, 1991.
82. M Takeichi, K Suzumori, G Endo, and H Nabae. Development of giacometti arm with balloon body.
IEEE Robotics and Automation Letters, 2(2):951–957, 2017.
83. Dean C Taylor, James D Dalton JR, Anthony V Seaber, and William E Garrett JR. Viscoelastic
properties of muscle-tendon units: the biomechanical effects of stretching. The American journal of
sports medicine, 18(3):300–309, 1990.
84. Thomas George Thuruthel, Egidio Falotico, Federico Renda, and Cecilia Laschi. Learning dynamic
models for open loop predictive control of soft robotic manipulators. Bioinspiration & biomimetics,
12(6):066003, 2017.
85. Michael T Tolley, Robert F Shepherd, Bobak Mosadegh, Kevin C Galloway, Michael Wehner, Michael
Karpelson, Robert J Wood, and George M Whitesides. A resilient, untethered soft robot. Soft robotics,
1(3):213–223, 2014.
86. Deepak Trivedi, Amir Lotfi, and Christopher D Rahn. Geometrically exact models for soft robotic
manipulators. IEEE Transactions on Robotics, 24(4):773–780, 2008.
87. Ryan L Truby, Cosimo Della Santina, and Daniela Rus. Distributed proprioception of 3d configu-
ration in soft, sensorized robots via deep learning. In Robotics and Automation (ICRA), 2020 IEEE
International Conference on. IEEE, 2020.
88. Nikos G Tsagarakis, Stephen Morfey, Gustavo Medrano Cerda, Li Zhibin, and Darwin G Caldwell.
Compliant humanoid coman: Optimal joint stiffness tuning for modal frequency control. In Robotics
and Automation (ICRA), 2013 IEEE International Conference on, pages 673–678. IEEE, 2013.
89. Bram Vanderborght, Alin Albu-Schäffer, Antonio Bicchi, Etienne Burdet, Darwin G Caldwell, Raf-
faella Carloni, MG Catalano, Oliver Eiberger, Werner Friedl, Ganesh Ganesh, et al. Variable
impedance actuators: A review. Robotics and autonomous systems, 61(12):1601–1614, 2013.
90. Björn Verrelst, Ronald Van Ham, Bram Vanderborght, Frank Daerden, Dirk Lefeber, and Jimmy Ver-
meulen. The pneumatic biped lucy actuated with pleated pneumatic artificial muscles. Autonomous
Robots, 18(2):201–213, 2005.
91. Alex Villanueva, Colin Smith, and Shashank Priya. A biomimetic robotic jellyfish (robojelly) actuated
by shape memory alloy composite actuators. Bioinspiration & biomimetics, 6(3):036004, 2011.
92. Michael Wehner, Ryan L Truby, Daniel J Fitzgerald, Bobak Mosadegh, George M Whitesides, Jen-
nifer A Lewis, and Robert J Wood. An integrated design and fabrication strategy for entirely soft,
autonomous robots. Nature, 536(7617):451, 2016.
93. Davide Zappetti, Stefano Mintchev, Jun Shintake, and Dario Floreano. Bio-inspired tensegrity soft
modular robots. In Conference on Biomimetic and Biohybrid Systems, pages 497–508. Springer, 2017.
94. Michael Zinn, Oussama Khatib, Bernard Roth, and J Kenneth Salisbury. Playing it safe [human-
friendly robots]. IEEE Robotics & Automation Magazine, 11(2):12–21, 2004.

View publication stats