基于观测数据删减及量化新息的无线传感器网络目标跟踪  

Target tracking in wireless sensor networks based on censoring and quantized innovations

作　　者：杨小军[1] 张亚粉[1] 

机构地区：[1]长安大学信息工程学院,西安710064

出　　处：《控制与决策》2015年第5期951-954,共4页Control and Decision

基　　金：国家自然科学基金项目(61473047;60975028);陕西省科学技术研究发展计划项目(2014K05-26);中国博士后科学基金项目(2013M540725)

摘　　要：针对资源受限的无线传感器网络,提出一种基于数据删减及量化新息的目标跟踪方法.利用融合中心接收到的量化新息以及数据删减过程传递的信息对目标状态进行估计.每个传感器节点利用容积卡尔曼滤波执行数据删减过程,融合中心执行一个辅助粒子滤波器.为了节省节点的能量和带宽,将所选择的观测数据的新息符号发送到融合中心,融合中心将数据丢失过程所包含的信息加以利用,提高了目标跟踪精度.仿真结果表明了该方法的有效性.The target tracking method based on censoring and quantization is proposed in the resource-constrained wireless sensor networks. Both the quantized innovations and the information conveyed by missing data due to data censoring are used to estimate the state of the target. Each local sensor maintains a cubature Kalman filter（CKF） to perform the data censoring. The fusion center runs an auxiliary particle filter（APF）. To save the energy and bandwidth of the node, the innovations of the selected data are quantized and transmitted to the fusion center. The fusion center makes use of the information contained in the missing data which leads to the improvement of the estimation accuracy. The simulation results show the effectiveness of the proposed method.

关 键 词：目标跟踪 粒子滤波器 数据删减 数据量化 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]