引用本文:梁啸海,闻继伟,栾小丽,刘飞.双边轮询协议下的时变线性系统分布式滤波[J].控制理论与应用,2025,42(7):1407~1416.[点击复制]
LIANG Xiao-hai,WEN Ji-wei,LUAN Xiao-li,LIU Fei.Distributed filtering of time-varying linear systems under two-sided round-robin protocol[J].Control Theory & Applications,2025,42(7):1407~1416.[点击复制]
双边轮询协议下的时变线性系统分布式滤波
Distributed filtering of time-varying linear systems under two-sided round-robin protocol
摘要点击 3866  全文点击 187  投稿时间：2023-01-15  修订日期：2025-04-15
查看全文  查看/发表评论  下载PDF阅读器   HTML
DOI编号  10.7641/CTA.2024.30021
  2025,42(7):1407-1416
中文关键词  时变线性系统  滤波网络  分布式滤波  双边轮询协议  拓扑切换  H∞性能
英文关键词  time-varying linear systems  filtering network  distributed filtering  two-sided round-robin protocol  topology switching  H∞ performance
基金项目  国家自然科学基金项目(62073154, 61991402)资助.
作者单位E-mail
梁啸海 江南大学物联网工程学院 lxh20172017@163.com 
闻继伟* 江南大学 物联网工程学院 wjw8143@jiangnan.edu.cn 
栾小丽 江南大学物联网工程学院  
刘飞 江南大学物联网工程学院  
中文摘要
      针对时变线性系统, 提出滤波网络的双边轮询协议, 研究其在随机拓扑切换下的分布式滤波问题. 首先, 对 滤波网络建立双边轮询协议, 将对观测对象的测量和网络内部智能节点之间的交互通信同时纳入轮询协议, 即每一 个采样时刻只有一个滤波节点占用通信资源, 减少滤波网络资源消耗, 解决带宽有限和数据冲突问题; 其次, 采用 Markov链描述滤波网络的拓扑切换行为, 通过构造拓扑依赖的Lyapunov函数给出滤波误差系统均方稳定的充分条 件, 利用差分线性矩阵不等式在线求解滤波器参数; 最后, 通过数值及实例仿真验证所设计滤波方法的可行性和有 效性.
英文摘要
      For time-varying linear systems, the distributed filtering is addressed for filtering networks with two-sided round-robin protocol (TSRRP) under random topology switching. Firstly, TSRRP is established for the filtering network, incorporating both the measurements for the plant and the interactive communications among the intelligent nodes within the filter network. That is to say only one filtering node occupies the communication resources at each sampling moment, which reduces the resource consumption of the filtering network and solves the problems caused by limited bandwidth and data conflicts. Secondly, a Markov chain is used to describe the topology switching of the filter network, and sufficient conditions for the exponentially mean square stability of the filtering error dynamics are given by constructing topologydependent Lyapunov functions. The filter parameters are solved online by using differential linear matrix inequalities. Finally, the feasibility and applicability of the designed filtering is verified by numerical simulations and practical examples.