A switching robust model predictive control approach for nonlinear systems

This work considers enhancing the stability and improving the economic performance of nonlinear model predictive control in the presence of disturbances or model uncertainties. First, a robust control Lyapunov function (RCLF)-based predictive control strategy is proposed. Second, the approximate dynamic programming (ADP) is employed to further improve regulation performance. Finally, the ADP and RCLF-MPC are combined to provide a switching control scheme, which is illustrated on a CSTR example to show its effectiveness.

• Construction of robust control Lyapunov function for nonlinear affine system.
• Robust control Lyapunov function is embedded into model predictive control.
• Approximate dynamic programming improves regulation performance.
• The approach can regulate nonlinear affine systems under bounded uncertainties.