Multiple models adaptive control based on online optimization

Aiming at the limitation of traditional multiple models adaptive control, such as a large number of sub-models, a multiple model adaptive control method based on online optimization is presented. The whole controlled system is divided into basic operating condition level and control model level. Multiple models and corresponding controllers are automatically built by online learning, and the built dynamic model bank is optimized so as to reduce both the sub-models in guantity and the computational load. The stability of the closed-loop system's and its asymptotical convergence of tracking errors can be guaranteed. Simulation results show the efficiency of the proposed method.