ROBUST ADAPTIVE CONTROL BASED ON MACHINE LEARNING AND NTSMC FOR WORKPIECE SURFACE-GRINDING ROBOT

In this paper, a robust adaptive trajectory tracking control method is presented for the workpiece surface-grinding robot. The workpiece surface-grinding robot is a highly nonlinear complex system, and it is difficult to describe its dynamic characteristics accurately. The system dynamics can be identified by the appropriate machine learning method. The adaptive law is proposed to adjust the neural network weights. To avoid the long convergence time and control singularity of the system, the nonsingular terminal sliding mode control (NTSMC) is employed to solve the disturbance, joint friction, and approximation error of the adaptive machine learning. The characteristics of the presented control scheme are illustrated through simulations and experiments, in which the convergence time decreases from 0.8s to 0.6s, and the amplitude of static error decreases from 0.025 rad to 0.02 rad.