Trajectory-Tracking for a Mobile Robot Using Robust Predictive Control and Adaptive Control

This paper proposes a control scheme of trajectory tracking for the mobile robots by combining robust predictive control in velocity level and adaptive control in acceleration level. In velocity level, a constrained quadratic programming (QP) problem can be obtained through the proposed model predictive control (MPC) approach. A primal-dual neural network (PDNN) is used to solve the QP problem over a finite receding horizon. And then, an adaptive controller employing the neural network technology is presented to approximate the uncertain robotic dynamics in acceleration level. Also, in order to deal with the saturation, an auxiliary system is proposed and designed. Besides, we apply a disturbance observer to rejecting the unknown disturbance online for achieving the trajectory tracking. Finally, the experimental studies are carried out on the mobile robots.