传感器网络中的时变信号跟踪分布式估计器  

The Distributed Estimator for Time-varying Signal Tracking in Sensor Networks

作　　者：马玉清[1] 李琳[2] MA Yuqing;LI Lin(School of Information Engineering,Anhui Business And Technology College,Hefei 231131,China;School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]安徽工商职业学院信息工程学院,安徽合肥231131 [2]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《探测与控制学报》2019年第4期84-91,共8页Journal of Detection & Control

基　　金：安徽省自然科学研究项目资助(KJ2017B001)

摘　　要：针对传感器网络中时变信号的跟踪估计及其他估计器存在的不足,提出了新的分布式估计器。首先,该估计器的每个节点测量一个时变带噪信号,并计算出其局部估计值作为其自身和它的邻居的测量值和估计值的加权和;其次,通过一个合适的帕累托优化问题来估计和更新它的权值,以使估计误差的方差和均值最小化;还对分布式估计器在估计偏差和估计误差性能方面进行了研究,给出了偏差的上限;估计器不依赖于中心协调,且参数优化和估计都分布在节点上。仿真结果表明,提出的分布式估计器相比于现有的分布式跟踪估计器,能更好地跟踪传感器网络中由噪声所损坏的未知时变信号,并能得到更好的估值结果。Aiming at the tracking and estimation of time-varying signal in sensor network and the shortcomings of other estimators,the novel distributed estimator was proposed.Firstly,each node in this estimator measured a time-varying noisy signal and computed its local estimate as a weighted sum of its own and its neighbors'measurements and estimates.Secondly,a suitable Pareto optimization issue was used to estimate and update its weights to minimize both the variance and the mean of the estimation error.The performance of the distributed estimator was investigated in terms of estimation bias and estimation error.Moreover,an upper bound of the bias was provided.The estimator did not rely on a central coordination,parameter optimization and estimation were distributed across the nodes.Simulation results showed that the proposed distributed estimator could track unknown time-varying signals damaged by noise and obtain estimation results.

关 键 词：传感器网络 时变信号 跟踪估计 帕累托优化 成本函数 均方误差 

分 类 号：TP301[自动化与计算机技术—计算机系统结构] TN911[自动化与计算机技术—计算机科学与技术]