Gaussian Process-Based Adaptive Sliding Mode Tracking Control for Robotic Manipulators

A Gaussian process (GP) based adaptive sliding mode control algorithm is proposed for tracking control of robotic manipulators with dynamical uncertainties. Gaussian process is used to obtain a data-driven model which can compensate for the unknown dynamics of the system. The sliding mode controller with parameter adaptive law is designed to bring the trajectory to the normalized tracking and to force the trajectory close to the sliding surface. Stability analysis based on the Lyapunov theorem is given to guarantee the tracking control performance. Numerical simulations on a 2-DOF robotic manipulator verify the effectiveness of the proposed controller.