摘要
针对非线性系统辨识问题,提出了一种基于Volterra级数模型的非线性系统的全解耦RLS自适应辨识算法。按照Volterra伪线性组合结构,采用RLS自适应辨识和约束优化理论,导出了具有分块对角形输入相关矩阵的全解耦Volterra标准方程,据此设计了一种基于Volterra级数模型的全解耦的RLS自适应辨识算法。该算法与部分解耦的RLS自适应算法相比,显著提高了辨识过程的收敛速度和精度。仿真结果验证了该方法的有效性。
Aiming at the identification problem for the nonlinear system, a fully decoupled RLS adaptive identification algorithm for nonlinear system based on Volterra series is presented in this paper. According to the pseudo-linear combination structure of Volterra series, by applying the principle of RLS adaptive identification and constrained optimization theory, a fully decoupled Volterra normal equation with block diagonal input correlation matrix is educed. A fully decoupled RLS adaptive identification algorithm based on Volterra series is then designed. The algorithm can remarkably improve the convergence speed and precision of the identification process compared with the partially decoupled RLS adaptive identification algorithm. Simulation results indicate that the proposed method in the paper is efficient.
出处
《系统仿真学报》
CAS
CSCD
2004年第4期807-809,共3页
Journal of System Simulation
基金
国家重点基础研究发展规划(973)项目(2001CB309403)
国家自然科学基金资助项目(60304004)
中国博士后基金项目(2003033512)。
