Precision Trajectory Tracking Control for Space Manipulators in on-Orbit Assembly Considering Truss Flexibility and Dynamic Coupling

The trajectory tracking control of space manipulators during on-orbit truss assembly tasks faces significant challenges due to structural flexibility, multi-source disturbances, and strong dynamic coupling effects. To address these issues, a radial basis function neural network-enhanced active disturbance rejection control (RBF-ADRC) strategy for trajectory tracking is proposed. First, a coupled dynamic model of a satellite base-flexible truss-manipulator system is established using the Lagrange formulation. Then, an ADRC controller is designed, where an extended state observer (ESO) is employed to estimate and compensate for disturbances in real time. Meanwhile, an RBF neural network is introduced to adaptively tune the nonlinear state error feedback (NLSEF) gains according to system state errors, thereby improving the controller adaptability under model uncertainties. Finally, numerical simulations are conducted to evaluate the effectiveness of the proposed method. The results show that the proposed controller achieves precise trajectory tracking, with the maximum joint tracking error maintained within 0.007 rad, which significantly outperforms classical ADRC (0.011 rad) and PID control (0.028 rad). The proposed method enhances tracking accuracy and improves the stability and reliability of on-orbit assembly operations.