Tube-MPC-based trajectory tracking control for robotic manipulators

To address the trajectory tracking control problem of robotic manipulators subject to input constraints and external disturbances, this paper proposes a Tube Model Predictive Control (Tube-MPC)-based strategy. First, MPC is utilized to compute nominal control inputs that strictly satisfy input constraints, providing a feasible baseline trajectory for the closed-loop system. Second, a nonlinear disturbance observer (NDO) is constructed to achieve online estimation of unknown lumped disturbances, and the tube radius is adjusted adaptively to alleviate the design conservatism caused by disturbance bounds. On this basis, a fixed-time sliding mode control (SMC) term is embedded into the tube error dynamics to suppress the influence of estimation errors and realize fast convergence independent of initial states. Lyapunov-based analysis is conducted to establish closed-loop stability and practical fixed-time convergence to a small neighborhood. Simulation results verify that, compared with conventional Tube-MPC, the proposed approach reduces the joint-tracking RMSE by 45.8%, shortens the convergence time by about 38.6%, and exhibits stronger robustness against time-varying torque disturbances.