摘要
针对无线传感器网络中基于 RSS 的多目标定位具有天然稀疏性的问题,提出了基于自适应网格的多目标定位算法,将多目标定位问题分解为大尺度网格定位和自适应网格定位 2 个阶段。大尺度网格定位阶段根据序贯压缩感知原理确定最优观测次数,再利用lp (0<p< 1)最优化重构出存在目标的初始网格;自适应网格定位阶段根据压缩感知原理自适应划分初始网格,再利用 lp 最优化重构出目标的精确位置。仿真结果表明,相较于传统的基于压缩感知的多目标定位算法,所提算法在目标个数未知的场景下具有更高的定位精度和更低的定位时延,且更适合大规模无线传感器网络的多目标定位问题。
The RSS-based multi-target localization has the natural property of the sparsity in wireless sensor networks. A multi-target localization algorithm based on adaptive grid in wireless sensor networks was proposed, which divided the multi-target localization problem into two phases: large-scale grid-based localization and adaptive grid-based localization. In the large-scale grid-based localization phase, the optimal number of measurements was determined due to the sequential compressed sensing theory, and then the locations of the initial candidate grids were reconstructed by applying lp (0<p<1) optimization. In the adaptive grid-based localization phase, the initial candidate grids were adaptively partitioned according to the compressed sensing theory, and then the locations of the targets were precisely estimated by applying lp optimization once again. Compared with the traditional multi-target localization algorithm based on compressed sensing, the simulation results show that the proposed algorithm has higher localization accuracy and lower localization delay without foreknowing the number of targets. Therefore, it is more appropriate for the multi-target localization problem in the large-scale wireless sensor networks.
作者
王天荆
李秀琴
白光伟
沈航
WANG Tianjing;LI Xiuqin;BAI Guangwei;SHEN Hang(School of Computer Science and Technology,Nanjing Tech University,Nanjing 211816,China)
出处
《通信学报》
EI
CSCD
北大核心
2019年第7期197-207,共11页
Journal on Communications
基金
国家自然科学基金资助项目(No.61501224,No.61502230,No.61602235,No.61802176)
江苏省自然科学基金资助项目(No.BK20161007,No.BK20150960)
江苏省研究生科研与实践创新计划基金资助项目(No.SJCX18_0339)~~
关键词
无线传感器网络
多目标定位
压缩感知
序贯压缩感知
自适应网格
wireless sensor network
multi-target localization
compressed sensing
sequential compressed sensing
adaptive grid
