Adaptive Backstepping Tracking Control of Space Manipulator Based on Neural Network

In this paper, an adaptive backstepping method is presented for a space manipulator in the presence of parametric uncertainties, external disturbance, and actuator saturation. First, a trajectory tracking error system is established. Then, a radial basis function neural network is employed to approximate the lumped uncertainties. The weight coefficients and upper bounds of the approximation errors in the neural network are estimated through adaptive laws. Subsequently, based on the backstepping method, the adaptive tracking controller is designed, where the approximation results are used to compensate for the adverse effect of uncertainties. Lyapunov theory guarantees that the states of the closed-loop system and the estimation errors of weight coefficients and upper bounds converge into an ellipsoid centered at the origin. Finally, a numerical simulation illustrates the effectiveness of the proposed controller.