Trajectory Tracking of Robotic Manipulators with Constraints Based on Model Predictive Control

This paper presents a model predictive control scheme for robotic manipulator in trajectory tracking in the presence of input constraints, which provides convergent tracking of reference trajectories and robustness to model mismatch. Firstly, the dynamic model of n-link robotic manipulator is linearized and discretized using Taylor approximation, based on which the constrained optimization question is converted to a quadratic programming problem. Then future output of system is predicted and the optimum control problem is solved online according to current state and previous input, while terminal constraint is included to reduce the tracking error. Finally, the convergence of the proposed control scheme is proved in simulation with the UR5 model and its robustness to model mismatch is verified by comparison with classical predictive control method.