基于离散萤火虫压缩感知重构的无线传感器网络多目标定位  

Multiple target localization in WSNs via CS reconstruction based on discrete GSO algorithm

作　　者：刘洲洲[1,2] 王福豹[1] 

机构地区：[1]西北工业大学电子信息学院,陕西西安710072 [2]西安航空学院,陕西西安710077

出　　处：《光学精密工程》2014年第7期1904-1911,共8页Optics and Precision Engineering

基　　金：国家自然科学基金资助项目(No.61103242)

摘　　要：研究了压缩感知(CS)理论在无线传感器网络(WSNs)多目标定位中的应用。提出一种基于离散萤火虫算法的压缩感知重构方法,并设计了具体算法实现流程,该算法摆脱了传统压缩感知重构算法对稀疏度K的依赖且能够准确地重构出原始信号。基于此,将新的压缩感知重构算法应用于WSNs目标定位,建立了WSNs系统模型,构造了合理的测量矩阵和稀疏矩阵,并分析了测量矩阵与重构结果之间的关系,最终实现了WSNs多目标定位。仿真结果表明该方法在稀疏信号重构性能及多目标定位精度方面具有较好效果,定位精度优于贪婪匹配跟踪(GMP)算法、正交匹配追踪(OMP)算法和最大似然估计(MLE)算法,且用于WSNs定位的传感器节点数目减少了20%,抗噪性达到了20dB。The application of Compressed Sensing（CS）theory to multiple target localization in Wireless Sensor Networks（WSNs）was explored.A CS reconstruction method based on Discrete Glowworm Swarm Optimization（DGSO）algorithm was proposed and the algorithm processing was designed.Different from the traditional reconstruction method,the DGSO algorithm is independent on the sparse K and can accurately reconstruct the original signal.The improved CS was applied to the multiple target localization in WSNs,and the WSNs application model was established.Then,a reasonable measuring matrix and a sparse matrix were constructed,and the relationship between measuring matrix and reconstructed results was analyzed.Finally,the multiple target localization was achieved in WSNs.The simulation results show that this method has better effect in the sparse signal reconstruction and multi target locating precision,and its location precision is better than those ofGreedy Matching Pursuit（GMP）,Orthogonal Matching Pursuit（OMP）and mum Likelihood Estimation（MLE）.Moreover,it reduces the network communication data amounts,extends the lifetime of WSNs.The number of sensor nodes for localization of WSNs is reduced by 20%,and the anti-noise can reach to 20dB.

关 键 词：无线传感器网络 压缩感知 多目标定位 重构方法 离散萤火虫优化算法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]