Abstract: A trajectory tracking method is proposed for three-order uncertain nonholonomic mobile robot with actuator dynamic model. This method is based on the backstepping technique and dual reinforcement learning adaptive fuzzy logic system. The shunting control technique is used to prevent the velocity jump at the beginning of system for the motion controller. The reinforcement learning adaptive fuzzy logic control is employed for both body dynamics and actuator dynamics, such that the unknown system uncertainties of parameters and non-parameters can be compensated optimally. Moreover, in order to enhance the control performance, the robust compensators are applied to eliminate the impact of unknown external disturbance and approximation error of fuzzy system. The stability and ultimately bounded convergence of the control system are proven by Lyapunov theory. The validity of this method is verified by the simulation results.

Key words: automatic control technology, wheeled mobile robot, trajectory tracking, adaptive fuzzy control, reinforcement learning, shunting model, nonholonomic system