Nonlinear model predictive tracking control of nonholonomic wheeled mobile robot using modified C/GMRES algorithm

Abstract

This paper investigates the trajectory tracking control of the nonholonomic wheeled mobile robot (NWMR) using nonlinear model predictive control (NMPC). By defining the tracking error in the NWMR's local coordinate, the tracking problem is formulated into standard NMPC framework. However, the standard NMPC algorithm is computationally demanding so that it is difficult for the real-time implementation of the NMPC algorithm on NWMR systems. To overcome this limitation, a continuation/generalized minimal residual (c/GMRES) algorithm is adopted. To handle the actuator saturation for the NWMR systems, a modified c/GMRES method is developed based on the logarithmic barrier function so that inequality constraints in NMPC can be handled. Finally, two reference trajectories are given to evaluate the proposed algorithm. The simulation results show the efficiency and effectiveness of the proposed algorithm for the trajectory tracking problem.