Section: New Results

Dynamic control of soft robots

The objective is to design a closed-loop strategy to control the dynamics of soft robots. We model the soft robot using the Finite Element Method, which leads to work with large-scale systems that are difficult to control. No unified framework exist to control these robots, especially when considering their dynamics. The main contribution of our work is a reduced order model-based control law, that consists in two main features: a reduced state feedback tunes the performance while a Lyapunov function guarantees the stability of the large-scale closed-loop systems. The method is generic and usable for any soft robot, as long as a FEM model is obtained. Simulation and real robots experiments show that we can control and reduce the settling time of the soft robot and make it converge faster without oscillations to a desired position. It can make the robot converge faster and with reduced oscillations to a desired equilibrium state in the robot's work-space. These results have been presented at the European Control Conference [24] and accepted for publication in Robotics and Automation Letters [8].