Study On Control Of Nonlinear Systems Based On Multi-Model Adaptive Method

Multiple-Model Adaptive (MMAC) is an effective approach to solve complicated problems such as nonlinear, uncertainties and operating points variations etc. However the problem exists in Multi-Model Control method is that, although it has been ever-increasingly used in practice, little has been done on the theoretical studies, especially, nonlinear control approach has not been employed for this study yet. The study of this paper is based on the nonlinear Multi-Model Adaptive Control method, which combines predictive control, fuzzy control with multiple model structure to control some nonlinear systems with complexity and strong coupling. The main theoretical results can be extracted in a summary form. (1)The sufficient conditions for the feasibility of nonlinear system's MMAC are revealed. The method that the multi-models represent nonlinear system are proposed. In addition, feedforward pole controller was designed and subsequently applied in paper manufacturing industry. (2)After the reasonable layout of the reference trajectory of the model, the nonlinear witching multi-model predictive control is proposed. To solve the problem of perturbation of witching multi-model predictive control, weighted multi-model fuzzy predictive control is put forward. (3)The design method of T-S fuzzy controller is given and we applied the genetic algorithm to optimize the parameters of membership function of fuzzy controller. Multi-model solutions for chaos control are studied, chaotic trajectory is directed by improved genetic algorithm that could create the law of the multi-model solution.