A predictive control strategy based on fuzzy system

In this paper a new predictive control strategy based on fuzzy system is proposed. An improved nearest neighborhood clustering algorithm is presented to establish fuzzy system. The algorithm is used to identify predictive model on-line. And the nonderivative optimization method is used to optimize the parameter of the controller. The algorithm can solve the problem of predictive control of nonlinear system, which has the control signal constraints of range and changing rate. Computer simulation results demonstrate the calculating speed, system's stability and the ability to restrain disturbance.