摘要
本文利用变点统计学的思想,讨论离散动态时变系统的多模型参数辨识问题,并给出了基于黄金分割优化思想的最小二乘多模型参数辨识算法以及计算机仿真结果。仿真结果显示出本文所提出的算法对多模型参数以及变点的良好估计特性。
The multi-model parameter identification problems for discrete dynamic time-varying systems are discussed based on the ideas of changing-point statistics in this paper.The multi-mod- el least squares parameter identification algorithms are presented with the optimal searching tech- niques of Fibonacci method(golden section method),and then.the computer simulations are car- ried out and the good simulation results show us that the algorithms proposed in this paper can give.effectively.good estimated values of the changing-points and parameters of the multi-model time-varving systems.
关键词
系统辨识
多模型系统
时变系统
control theory
system identification
multi-model systems
changing-point statistics
Fibonacci method
