首页 | 本学科首页   官方微博 | 高级检索  
     

带丢包一致性滤波算法研究
引用本文:王帅,杨文,侍洪波.带丢包一致性滤波算法研究[J].自动化学报,2010,36(12):1689-1696.
作者姓名:王帅  杨文  侍洪波
作者单位:1.华东理工大学信息科学与工程学院自动化系 上海 200237
摘    要:提出了一种基于一致性的分布式滤波算法, 针对实际应用中存在的网络丢包问题, 重点研究了有丢包时的分布式滤波算法, 通过理论分析给出了估计误差系统收敛的充分条件. 应用数值仿真将本 文提出的算法与已有的经典滤波算法分别在理想状况与有丢包状况时进行比较, 研究表明本算法在丢包时具有较优的滤波效果. 并进一步研究了一致性步长对估计误差协方差的影响, 发现存在估计误差达到最小值的最优步长. 最后, 研究了丢包率对算法的影响, 发现起``领导'作用的传感器在滤波时发挥重要作用, 可通过控制这些传感器的丢包率来减小丢包对整个网络系统的影响.

关 键 词:分布式滤波    一致性算法    卡尔曼滤波    传感器网络    多智能体系统
收稿时间:2010-4-26
修稿时间:2010-7-1

Consensus-based Filtering Algorithm with Packet-dropping
WANG Shuai,YANG Wen,SHI Hong-Bo.Consensus-based Filtering Algorithm with Packet-dropping[J].Acta Automatica Sinica,2010,36(12):1689-1696.
Authors:WANG Shuai  YANG Wen  SHI Hong-Bo
Affiliation:1.School of Information Science and Engineering, East China University of Science and Technology, Shanghai 200237
Abstract:This paper introduces a consensus-based distributed filtering algorithm. Aiming at the problem of network packet-dropping in practical applications, we focus on the distributed filtering algorithm with packet-dropping. By theoretical analysis, we give a sufficient condition for the convergence of the estimation error system. We compare our algorithm with a classical filtering one by some simulations in ideal and packet-dropping cases, respectively. The results show that our algorithm does better filtering in the packet-dropping case. Furthermore, we study the influence of the consensus step-size on the estimation error covariance, and find out that an optimal step-size may lead to the minimum of estimation error covariance. Finally, we study the influence of the packet-dropping rate on our algorithm, and find that the leaders of the sensors play an important role in the filtering. We can decrease the influence of the packet-dropping on the whole network system by controlling the packet-dropping rate of the leaders.
Keywords:Distributed filtering  consensus algorithm  Kalman filtering  sensor network  multi-agent system
点击此处可从《自动化学报》浏览原始摘要信息
点击此处可从《自动化学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号