非线性系统的参数空间分割辨识方法

NONLINEAR SYSTEM IDENTIFICATION WITH PARAMETER-SPACE SEGMENTING

本文提出了一种新的非线性系统 Volterra级数模型辨识方法 ,为非线性系统辨识中的“维数灾难”问题提供了一种满意的解决 .算法中参数空间分割和模型辨识同时完成 ,降维依据采用输出拟合结果的均方误差 ,最终得到输出拟合均方误差意义上的准最优解 .本算法也可以作为非线性系统模型的结构辨识算法 ,并可以直接推广应用于其它很大一类非线性系统模型 .仿真试验结果表明 ,算法计算量小 ,精度高 ,并具有较好的稳定性 ,可以应用于在线实时辨识 .

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.