Invited Talk Title: Rethink How We Control Soft Robots to Explore the World
Abstract: Soft robotics focuses on the deformable and non-rigid character of our world. These soft-bodied robots are made of deformable materials that greatly surpass the limited degrees of freedom of rigid robots and potentially offer inherently safe and adaptable ways of achieving versatile forms in locomotion and manipulation. A soft robot can homogeneously combine actuation, sensing, and structural complexity within the same component of the robot. In this talk, I will provide an overview of several efficient modeling and control approaches that bring us closer to the development of intelligent robotic “brains”. We are aiming for an approach that exploits the unique properties of soft robots and the deformable world we life in. I will cover in this talk my lab’s recent developments in overcoming the challenges in modeling and control of soft robotic systems. A focus will be on explaining and comparing our lab’s modeling methods, which include: minimal parameter modeling, reduced order modeling, variational deep Koopman operator modeling, differentiable projective dynamics, and physics-informed model learning. Based on these modeling techniques, I will then introduce frameworks that perform model-based control of multi-segment pneumatic soft robotic arms. Finally, I will show how we use our modeling approach to co-optimize the designs of geometry and control of our soft robotic systems.
Bio: Robert is a professor of robotics at ETH. He takes inspiration from living creatures and develops robots with deformable properties that adapt better to their environment when compared to traditional robots. His lab designs, fabricates, models, and controls soft robotic systems such as robotic fish for ocean exploration, arms for dynamic manipulation tasks, and drones for swift grasping. Robert’s work has appeared in leading academic journals, such as Science Robotics, and has been featured in major outlets, including New York Times. Robert is one of 20 TED Fellows in 2022. Robert is a member of the ETH AI Center, the Max Planck ETH Center for Learning Systems (CLS), and the ETH Competence Center for Materials and Processes (MaP). Robert is an Area Chair for Robotics Science and Systems (RSS), an Associate Editor for IROS, an Associate Editor for RA-L, and a reviewer for leading peer-reviewed journals, including Science and Nature.
