Model and Control of Hybrid Hard-soft Robots Using Model Predictive Control

Abstract

The compliance of soft robots contributes to excellent superiority over rigid robots in touching screens and human-machine interaction. Therefore, this paper designs a hybrid hard-soft touch robot that can be easily modeled and keeps softness and body compliance. We built a hybrid hard-soft robot that contains a six-axis robot and a soft tube attached at the end. A hybrid hard-soft touch robot means it combined a rigid part for moving and a soft part for touching. Based on the piecewise constant curvature model, the touch robot's kinematics model is established through the spinor theory and geometric method. Due to the redundancy of the robot's spatial motion, a predictive control method is proposed to solve the problem of redundant degrees of freedom. And it is used for trajectory tracking control. To complete the closed-loop control of the entire robot system, force feedback control and visual servo control are proposed. The experiment shows that the hybrid hard-soft robot has a fast response, and can adaptively perform pressing operations.