Switching robust model-predictive-control strategy for constrained nonlinear system
约束非线性系统切换鲁棒预测控制

By adopting the piece-wise embedding scheme, we use a set of linear parameter-varying(LPV) models to approximate a nonlinear system subject to input and state constraints. For each embedding model, a parameter-dependent Lyapunov function is used to obtain the stable control law for the nonlinear system. When the LPV model is switched from one to the other, the switching law, which is obtained by using an off-line design approach, is used to guarantee the feasibility of the initial state at the switching point. In the last part of this paper, we prove the feasibility and stability of this switching control algorithm. The switching robust model-predictive-control strategy is less conservative than the traditional nonlinear predictive control and will greatly reduce the online computational burden.