Advances in Soft Robots Control

Concepción A. Monje1, Helmut Hauser2, Cecilia Laschi3

  • 1University Carlos III of Madrid
  • 2University of Bristol
  • 3National University of Singapore

Details

09:00 - 18:00 | Mon 4 Nov | L1-R3 | MoW-R3.1

Session: Advances in Soft Robots Control

Abstract

Introduction: The emerging field of soft robotics is nowadays looking at innovative ways to create and apply robotic technology in our lives. It is a relatively new domain in the field of robotics, but one that has a lot of potential to change how we relate with robots and also how they are used. The term "soft robot" describes a system that is inherently soft yielding a complex dynamics and a passive compliance similar to the biological counterpart. As this was a new design paradigm for the hardware, methods or algorithms to prescribe the robotic system a certain dynamics changed as well. Motivation: Classical control approaches in robotics are nonlinear model based. However, the highly complex and nonlinear models necessary for a soft robotic system make this approach a difficult task and therefore seem to come to a limit in the presence of a soft robot. Therefore, other methods have been applied seemingly being more useful in this context, such as learning-based control algorithms, model-free approaches like bang bang control, control algorithms motivated by neuroscience, or morphological computation. These methods add new perspectives to the well known model-based approach. We want to provide an inter- and cross-disciplinary platform to discuss techniques, conventional as well as novel, that are currently applied and developed and discuss limitations, potentials and future directions. Contribution: The workshop will bring together experts in the above mentioned control methods as well as people with neuro-scientific and biology background in order to discuss the following issues: -How are control loops set up in biology (from a biological and neuroscientifical point of view)? -Requirements of each method in order to apply them on a real system. -Is there a theoretical background, proving that the method works?; passivity, robustness, Lyapynov stability, others. -Limitation of the control approaches. -Are there specific tasks / trajectories / applications for which each method is applicable? Goals of the workshop: -Answer the question: "Do we have to rethink the basic approach in robot control, which is model based, when it comes to controlling soft robotic systems?". -Identify the best experimental testbeds for soft robot control evaluation. -Contribute to the field by giving the audience a structured review of existing control methods for these systems, including the above-mentioned bullets. -A Special Issue in International Journal of Humanoid Robotics will be published as an outcome of the workshop summarizing and categorizing the presented methods.