A nonlinear model predictive controller for kinematically redundant space manipulators

This study investigates the use of nonlinear model predictive control (NMPC) for the approach of a space redundant manipulator to an un-cooperative target satellite in space. The objective is to evaluate the performance of the predictive controller for the approaching task and to investigate the need and feasibility of incorporating constraints into the controller. Firstly, the nonlinear dynamic model of an n-link manipulator is recalled, which then is linearized and decoupled by feedback. Secondly, a nonlinear model predictive control scheme, implemented with an optimized dynamic model and running within small sampling period, is presented. The derived nonlinear predictive control law uses a quadratic performance index of the predicted tracking error and the predicted control effort. The constrained predictive controller solves a quadratic programming problem at every sampling interval as receding horizon. The real-time implementation is based on Simulink with the model predictive controller and computed torque controller (CTC). Simulations performed by using a 7 degree-of-freedom (DOF) redundant manipulator mounted on a 6 DOF spacecraft prove the effectiveness of the proposed control method. The NMPC and the widely used CTC are compared. Tracking performance and robustness under external disturbance or errors in the model are evaluated. Asymptotic error tracking and constraint handling results particularly demonstrate the effectiveness and potential of the nonlinear model predictive controller for space redundant manipulators.