引用本文: | 赵文虓,陈翰馥.随机系统的递推辨识: 从个例到一般框架[J].控制理论与应用,2014,31(7):962~973.[点击复制] |
ZHAO Wen-xiao,Chen Han-Fu.Recursive identification of stochastic systems: from individual system to a general framework[J].Control Theory and Technology,2014,31(7):962~973.[点击复制] |
|
随机系统的递推辨识: 从个例到一般框架 |
Recursive identification of stochastic systems: from individual system to a general framework |
摘要点击 3002 全文点击 2872 投稿时间:2014-02-24 修订日期:2014-05-20 |
查看全文 查看/发表评论 下载PDF阅读器 |
DOI编号 10.7641/CTA.2014.40134 |
2014,31(7):962-973 |
中文关键词 Hammerstein系统 Wiener系统 非线性ARX系统 PageRank 随机逼近 |
英文关键词 Hammerstein system Wiener system nonlinear ARX system PageRank stochastic approximation |
基金项目 国家重点基础研究发展计划“973”计划资助项目(2014CB845301); 国家自然科学基金资助项目(61104052, 61273193, 61227902, 61134013, 61120106011). |
|
中文摘要 |
本文考察Hammerstein系统、Wiener系统和非线性带外源输入的自回归系统(autoregressive system with exogenous input, ARX)等常见的随机非线性系统的递推辨识和因特网PageRank的分布式、随机化算法. 对非线性系 统分别构造递推辨识算法, 证明了估计的强一致性; 对因特网PageRank的分布式、随机化算法, 给出估计的强一致 性和收敛速度; 在此基础上, 总结了这类问题的统一处理框架–将辨识(估计) 问题转化为函数求根、进而基于随机 逼近构造算法得到强一致的递推辨识; 最后, 通过数值例子来验证算法的有效性. |
英文摘要 |
We consider the recursive identification algorithms for the stochastic nonlinear systems, such as Hammerstein system, Wiener system, and nonlinear autoregressive system with exogenous input (ARX) system and the distributed randomized PageRank algorithm (DRPA). The strong consistency of estimates given by the identification algorithms and DRPA is established. Based on the stochastic approximation algorithm with expanding truncations, a unified framework for solving these kinds of problems is introduced. Numerical examples are given to testify the performance of the proposed algorithms. |