基于N-P决策融合准则的传感器网络最优检测机制  

N-P fusion based optimal detection scheme in wireless sensor networks

作　　者：曹义亲[1] 薛海英[1] 袁朝晖[1] 

机构地区：[1]华东交通大学软件学院,江西南昌330013

出　　处：《计算机工程与设计》2014年第1期97-101,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61163056);江西省科技支撑计划基金项目(20123BBE50093);江西省教育厅科技基金项目(GJJ12305);江西省教育厅科技基金项目(GJJ12324);江西省自然科学基金项目(20114BAB211018

摘　　要：针对分布式传感器网络检测系统中的最优决策融合问题,引入了奈曼-皮尔逊判定准则,在此基础上推导出由虚警率条件约束的最大检测准确概率的局部判决阀值和全局最优判决阀值的求解方法。设计了基于该准则的最优检测阀值求解算法和全局检测机制。仿真结果表明,在相同的参数设置下,该机制能在减小系统通信开销和达到节点负载均衡的同时,达到与数值融合机制相当的性能水平。与已有的决策融合方法相比,能提高平均10%的系统检测准确概率。To find out the optimal decision fusion rules in distributed wireless sensor networks, Neyman-Pearson rule is imported into the detection systems. The local and system detection threshold are derived respectively to guarantee the optimal detection probabilities while satisfying the minimal false alarm rate constraint. Based on these thresholds, an optimal detection algorithm and a fusion scheme are proposed. Simulation results show both of them can reach similar performance with the data fusion scheme while reducing communication cost and balancing the system workload, and nearly 10% improvements are obtained com pared with the states of arts.

关 键 词：分布式传感器网络 目标检测 决策融合 N-P准则 最优阀值 

分 类 号：TP393.01[自动化与计算机技术—计算机应用技术]