IGWL: The Intersection Graph of Wireless Links Based Multi-target Localization in UHF RFID Scenarios

Radio tomographic imaging (RTI) is one of the most favorable device-free localization methods with many advantages. However, since its model is approximate and ill-posed, it has low multi-target localization accuracy in indoor environments. To overcome the problem, we propose an IGWL scheme, the intersection graph of wireless links based multi-target localization, which directly utilizes target-affected links to localize targets. Firstly, the intersections of target-affected links are extracted from all intersections of all wireless links according to the first Fresnel zone. Then, analyze the distribution of extracted intersection graph and determine the intersections with high respect to targets. After, the gap statistic method is used to determine the number of targets. Finally, we modify the K-means algorithm to acquire the locations of targets. The simulation shows that the localization accuracy of the proposed IGWL scheme is significantly higher than the traditional RTI.