Predictive control research and its application for a class of uncertain nonlinear systems

An analytic model predictive control method is proposed to deal with a class of bounded uncertain nonlinear systems. Firstly, it can be redefined as the nominal system with uncertain model form, and the optimal control law with uncertainties is obtained. Derivation shows that the obtained control law is bounded and the analytic model predictive control that eliminates uncertainties is then able to acquire utilizing the boundary value. Apply Routh-Hurwitz criterion theory to prove the stability of the closed-loop system, therefore, provide the selection conditions of controller parameter. The analytic model of the original system predictive control in the process of solving the uncertainty problem is then converted into a deterministic problem. The controller can be realized without extra compensation of uncertainties in the process of controlling nonlinear systems. Finally, the article is able to introduce the design of analytic model predictive control which were achieved by application in the ship rudder roll stabilization controller design. The simulation results show that, under two different sea conditions, the proposed method has good robustness and control effects, comparing to the model that does not consider the uncertain predictive control and generalized predictive control of the analytic model. And it is proved that the control method has important theoretical significance and engineering application values.