Multi-Source Localization Using Time of Arrival Self-Clustering Method in Wireless Sensor Networks

The localization of multiple signal sources based on Time Of Arrival (TOA) measurements in wireless sensor networks is investigated in this paper. When the signal sources cannot be distinguished by their signatures or other unique characteristics, the correspondence between the sources and the TOA measurements at different sensors is unknown, which makes the multi-source localization problem quite challenging. A self-clustering measurement combination method is proposed for the problem. The source location estimate obtained by the hyperbolic localization algorithm is used as the clustering pattern, and the scatter of the patterns of different subsets of TOA measurements is defined as a criterion function, which is extremized by the combination of TOA measurements from the same source. A three-step heuristic clustering algorithm is pursued to resolve the TOA ambiguity, and its mean square error performance and computational complexity are also analyzed. The simulation and experiment indicate that the presented method has higher location accuracy and lower complexity compared with the existing methods.