Model Predictive Control for Cooperative Grasping with Dual Rigid-Soft Hybrid Robots

This paper focuses on cooperative grasping technology of dual rigid-soft hybrid robots, each comprising a six-degree-of-freedom rigid robotic manipulator a two-degree-of-freedom soft robot. The modeling of the rigid robot utilizes the Denavit-Hartenberg parameter method, while the soft robot is modeled using the Piecewise Constant Curvature approach. Ultimately, the system model based on the anlytical Jacobian is obtained. A model predictive controller tailored to this model is developed to precisely manipulate the terminal motion of the hybrid robots. For the grasping task, posture constraints required at the end-effectors of both robots are incorporated into the control scheme. The effectiveness of this controller is validated through a practical grasping experiment.