基于统计假设检验的传感器网络目标探测机制

多传感器组网对目标探测是当前军事领域的研究热点,也是传感器网络的基本功能。但是,现有的基于目标输出信号阈值机制难以精确描述目标发现事件。根据数理统计理论将目标探测问题建模为统计推断过程,本文提出一种基于统计假设检验的传感器网络目标探测机制,通过样本训练确定检验参数,采用数据融合方法优化探测结果。实验显示,该方法能有效解决探测过程的噪声干扰问题,提高目标捕捉的正确率。

A Target Detection Mechanism in Sensor Networks Based on Statistical Hypothesis Testing

It is a hot research topic in the current military field that target detection is executed with many sensors in a networking manner. Target detection is also the fundamental function of a sensor network. However it is difficult for the current mechanism to describe the target discovery event which is based on an output signal threshold of the target. According to the mathematical statistics theory, the problem of target detection is modeled as a statistical inference process in the paper. A target detection mechanism based on the hypothesis testing method is proposed. In this mechanism the testing parameters can be determined by a sample training process, and moreover a data fusion method is adopted for the detection result optimization. The experimental results show that this mechanism can avoid the noise interference in the detection process, and improve the exactness performance of target catch in sensor networks.