摘要
针对现有的目标跟踪分簇算法没有从根本上解决参与跟踪的节点数量过多,导致整个无线传感器网络(WSN)能耗的增加问题,提出一种基于Fisher信息矩阵的改进卡尔曼滤波的目标跟踪分簇方法 (Fisher Matrix for Kalman Filter,FMKF),用于针对性的选择节建立跟踪簇。该算法利用随机矢量估计的克拉美罗下界获得未知噪声的统计特性,优化卡尔曼滤波器的误差协方差。在无线传感器网络动态分簇时,创新的使用信息判据作为标准,并且加入节点剩余能量判据。仿真结果显示,FMKF算法与控制簇的激活半径算法和无分簇算法相比,FMKF算法可以在减少跟踪节点的数量的同时提高跟踪精度。
To reduce the working sensor nodes caused by the lack of efficient criterion in clustering-based wireless sensor network,an improved Kalman filtering based on Fisher information matrix for target tracking clustering method is proposed. In the process of filtering,the Cramer-Rao low bound of random vector needed to be estimated in this algorithm is used to calculate its error of mean square. Then,it can optimize the error covariance of Kalman Filter. Information criterion is taken as the main basis for dynamic clustering in sensor network,while residual energy is regarded as the assistant criterion. So,the most suitable sensor nodes are activated as working node in tracking cluster. Comparing to the method based on controlled radius and non-clustering information matrix filter, the simulation shows that FMKF algorithm can reduce the number of working nodes greatly and improve the tracing accuracy.
作者
王锋
郭滨
白雪梅
陈帅坤
WANG Feng GUO Bin BAI Xuemei CHEN Shuaikun(School of Electronic and Information Technology, Changchun University of Science and Technology, Changchun 130022)
出处
《长春理工大学学报（自然科学版）》
2017年第3期103-107,共5页
Journal of Changchun University of Science and Technology(Natural Science Edition)
关键词
改进卡尔曼滤波
目标跟踪
克拉美罗界
分簇
improved kalman filter
target tracking
cramer-Rao low bound
clustering
