小规模传感器网络远程目标探测系统的建模与性能分析

摘要

该文针对传感器网络目标探测的实际问题，构建了远程目标探测模型。在Neyman-Pearson准则下，提出一种小规模传感器网络对远程目标探测的融合规则(FRSS)，旨在通过数据融合方式提高目标的探测距离。推导出远程目标探测融合统计量 Counting 统计量；通过随机化检测方法构建了门限求解模型，推导出系统检测性能的闭式表达式。最后，采用解析方法分析了理想信道、非理想信道下FRSS的检测性能，并采用Monte-Carlo方法对比分析了FRSS规则，Chair-Varshney融合规则，Bayes融合规则的探测性能。仿真结果表明：在远程目标探测模型下，FRSS规则探测性能略有下降，但是该规则在融合过程中需要较少的先验信息，大大减少了数据传输量；相比单节点探测系统，基于FRSS规则的融合系统探测性能显著提高。%In view of practical application of the target detection based on sensor networks, the remote target detection model is established. This paper addresses a novel Fusion Rule for Small-scale Sensor networks (FRSS) under the Neyman-Pearson criteria. The detection range or detection performance is improved by the way of data fusion. The fusion statistic (Counting statistic) is derived. Besides, the threshold of fusion center solving model is constructed through the randomized test and the closed-form expression of the detection performance is given. The performance of the target detection system under the ideal channel between local sensor nodes and fusion center and non-ideal channel using BPSK modulation is evaluated by numerical approach. Moreover, the Monte-Carlo approach is used to analyze comparatively the detection performance of FRSS rule and the previous Chair-Varshney rule, Bayes rule. The simulation results show that the proposed FRSS rule exhibits slight decline considering the detection performance compared with the other two rules. However, the FRSS rule requires less prior information and greatly reduces the amount of data transmission. The detection performance of FRSS is extremely improved compared with the single sensor node.