A Cascade Multiple-Model Predictive Controller of Nonlinear Systems by Integrating Stability and Performance

Abstract

To deal with strong nonlinearity in nonlinear systems, a new method called cascade multiple-model predictive controller based on gap metric and stability margin, is proposed. The gap metric is utilized to describe the nonlinear system by a linear model bank. It is possible to select nominal local models from the linear model bank by an algorithm based on the gap metric and stability margin to avoid the redundancy of the local controllers. By scheduling proportional controller for each nominal local model, the robust stability is guaranteed whereas there will be no guarantee for the desired performance. Then, by designing a model predictive controller in the cascade structure, the closed loop performance is improved. The achievements of the proposed method are illustrated by computer simulations.