基于二元传感器网络的多源定位研究

Multi-source localization with binary sensor networks

针对多源定位模型计算比较复杂的情况,基于Neyman-Pearson准则对二元传感器网络的多源探测模型进行了研究,然后在2个信号源的情况下,提出利用Fisher准则将传感器分为两部分,每部分传感器与相应信号源对应,并在此基础上提出利用加权减负加正(WSNAP,weighted subtract on negative add on positive)算法对多信号源进行定位计算。仿真结果表明:Fisher准则能以较高的正确率的将报警传感器分为两部分;与质心算法和加正(AP,add positive)算法相比较,所提出的方法计算复杂度较低、定位精度更高,并利用数据库对文中的结论进行了验证。

A new multi-source detection model was proposed based on Neyman-Pearson criterion to reduce the computa-tional complexity caused in the multi-source localization.The Fisher criterion was employed to divide sensors into two parts,where two sources were present and each part corresponds to one of the sources.The WSNAP（weighted subtract on negative add on positive） multi-source location algorithm was applied to localize the multiple sources.The simulation results show that Fisher criterion is able to divide the alarmed sensor into two parts with relatively higher accuracy.The proposed WSNAP has better estimation accuracy than AP（add positive） algorithm and CE（centroid estimator） algorithm under the circumstance of lower computation complexity.Finally,the results are verified using the database of distributed wireless sensor networks.