Recursive Relaxation Algorithm For Identification Of Multiple Input Multiple Output Systems

The traditional system identification method based on least square method has the disadvantages of heavy computation and slow operation speed in a particular model. In this paper, the off-line identification based on the equation error autoregressive model is studied by using the relaxation algorithm which is proved the unbiasedness and effectiveness. Furthermore, on the basis of off-line identification parameters, the system is identified online by recursive relaxation algorithm that calculated by relaxation algorithm. The simulation results show that the recursive relaxation algorithm has good computational speed and identification accuracy in the multi-input-multi-output linear system.