Subspace predictive controlle method with updating variable forgetting factor

Abstract

According to the general form of updating data with forgetting factor, several conclusions are obtained in updating I/O Hankel matrices. Based on these conclusions, new subspace identification and relative SPC methods are proposed with variable forgetting factor for improving the performance in tracking the time-varying information. And the forgetting factor's update is realized by the distance between expected outputs and measured outputs. Finally, the efficiency of modified method is illustrated with a simulation example.