Soft Robots in the Real World

 Modelling Challenges for Soft Robots towards Real-World Applications [W5a] 

Soft robotics is a growing research field which aims at incorporating softness in robotic bodies or in novel end effectors, enabling safe and adaptive interactions. Soft robotics is bio-inspired, since it tries to reproduce the capabilities of certain animals such as worms, snakes, or the octopus to move without having a rigid skeleton or exoskeleton, exploiting their softness to squeeze and adapt to unstructured environments. The stiffness characteristic of traditional industrial robots was desirable because it enabled fast, reliable, and precise tasks such as those required in factory lines. Conversely, soft robotics finds application in tasks where safety and adaptability to unstructured environments is of paramount importance. Such tasks include delicate food handling, medical procedures, and assistive tasks, among others.  Despite the great potential shown by soft robots, their applicability in real-world application is still limited.

The compliance which characterizes soft robots, besides granting the desired properties described above, also introduces challenges from the point of view of modeling and control.

The hysteresis of the materials and the high redundancy due to the virtually infinite number of degrees of freedom (DoF) of soft robots make them hard to model with high accuracy.

However, the long-term success for the practical application of soft robotic systems is dependent up on the development of real-time dynamic models that can be embedded in controllers to facilitate fast, reliable, accurate, and energy-efficient control. This is non-trivial because: (i) unlike rigid robots, elastic deformation of soft robots results in virtually infinite degrees-of- freedom (DoF) motions (bending, extension, contraction, torsion, buckling, etc.), (ii) the material properties exhibit non-linear characteristics such as compliance and hysteresis, (iii) interfaces between materials and friction can make models predictions inaccurate. Notably, in recent years great progresses have been achieved in developing dynamic models approximating the behavior of continuously deformable soft robots, but their complexity introduces new challenges for their use in real-time applications.

From the perspective of control of soft robots, the highly complex and nonlinear models necessary for a soft robotic system make control a difficult task and classical control approaches seem to come to a limit in the presence of a soft robot. Therefore, other control methods must be applied seemingly being more useful in this context, such as robust control, learning-based control algorithms, model-free approaches, control algorithms motivated by neuroscience, or morphological computation. These methods add new perspectives to the well-known model-based approach.

This workshop will focus on application-driven soft robotic systems and will discuss main issues preventing their commercial use, with a focus on modelling and control challenges. We want to provide an inter- and cross-disciplinary platform to discuss modelling and control techniques, conventional as well as novel, that are being currently developed and applied to soft robots with a focus on the translational challenges for an immediate application in industry, healthcare, inspection, agriculture, etc.

 Field Deployable Soft Robotics: Challenges and Opportunities [W5b]

Since the beginning of the field, the promise of soft robotics has built a reputation as being able to provide unique solutions to problems requiring navigation in complex environments, and safe interaction with fragile objects. Aquatic and terrestrial creatures such as octopuses, worms, caterpillars and rays serve as model organisms to describe the potential of soft robots. Their sophisticated structures enable them to move through small openings, conform to terrain, grasp arbitrarily shaped objects, and generate multiple methods of locomotion; features with obvious advantages in agriculture, environmental monitoring, oceanic and space exploration.

Soft robotics has undergone significant growth in the past decade and now contains a sizable research community, dedicated streams at major robotics conferences, and several dedicated conferences like Robosoft.

Despite the major interest in the field, and its relative maturity, few of its major developments or achievements have ever made it out of the lab and into commercial usage. Passive, pneumatic and granular jamming soft grippers are now commercially available and used in manufacturing pick-and-place and robotic harvesting. It is a rare example of research translation, driven by scale the convenience of integrating soft end-effectors into existing workflows.

The current state of soft robotics research is similar to conventional robotics 20-30 years ago, where robot arms had become common in factories, but mobile robotics was in its infancy. Today legged autonomous robotic platforms like Boston Dynamics’ Spot, Anybotics’ Anymal, and CSIRO Data61/ Nexxus’ Magneto are being deployed for mapping, monitoring and inspection in numerous industries.

The aim of this workshop is to bring together the soft robotics community at its flagship conference to chart a roadmap towards field deployment of soft robots. Through a set of presentations, interactive discussions, and tutorial sessions, it will bring together representatives of relevant research areas and industry to identify challenges faced in deploying and commercialising soft robots for field applications.