NONLINEAR SYSTEM IDENTIFICATION WITH PARAMETER-SPACE SEGMENTING

A new algorithm of nonlinear system identification for Volterra series model is presented in this paper.The algorithm provides a satisfactory solution to the dimension disaster of nonlinear system identification.Parameter space segmenting and identification is completed synchronously in the algorithm.The target of P Space segment is MSE of output approach and a near optimal solution is achieved finally.The algorithm can be used as model structure identification and popularized to many other kinds of nonlinear system model.It is indicated by the result of simulation that the new method is of very limited calculation time,high accuracy and robustness.Therefore it can be used to identify the real plant on line.