基于Hsia算法的Hammerstein模型辨识  

An Identification Method for Hammerstein Model based on Hsia Algorithm

作　　者：沈同全[1] 孙逢春[1] 程夕明[1] 

机构地区：[1]北京理工大学机械与车辆工程学院,北京100081

出　　处：《系统仿真学报》2007年第23期5373-5375,共3页Journal of System Simulation

基　　金：国家高技术研究发展计划(863)(2003AA501800)

摘　　要：针对采用广义最小二乘法(GLS)辨识Hammerstein模型时反复滤波,计算效率不高的问题,提出了利用Hsia算法进行Hammerstein模型辨识的策略。该法不需滤波,克服最小二乘解有偏性的同时,消除了Hammerstein模型中非线性增益的阶次对计算效率降低的作用,编程简单,计算快速。仿真及试验研究证明该法省时明显,值得推广。When Hammerstein model is identified based on the least squares algorithm （GLS）, both input and output data are filtered repeatedly and its computation efficiency is very low. A novel identification method for Hammerstein model based on Hsia algorithm was put forward. The bias of the least square solution is solved and the reducing effect for computing efficency from high order of the nonlinear system is got rid of. And also filters are unnecessary here. Hsia algorithm proposed here is simple and feasible, which is verified by simulation and test study.

关 键 词：非线性动态过程 HAMMERSTEIN模型 Hsia算法 系统辨识 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]