InterMap-NRTI: A Novel Radio Tomographic Imaging Positioning Model Based on the Intersection Map of Wireless Links

Device-free localization (DFL) is a potential positioning technique that can localize targets without requiring any device. Radio tomographic imaging (RTI) is one of the most favorable DFL methods and has many advantages over other methods. In multi-target localization indoors, target-affected wireless signal is inevitably contaminated with non-negligible interference of noise and complex multipath. Existing RTI positioning models utilize contaminated signals for multi-target localization, resulting in low positioning performance. To decrease the effect of contaminated signals, we propose a novel RTI positioning model based on the intersection map of wireless links (InterMap-NRTI). InterMap-NRTI utilizes the statistical method to determine wireless links mainly affected by targets and the interpolation method to complement the wireless links that are interrupted due to hardware malfunction and targets' obstruction. The next step is to formulate the definitions to extract high-intersections closest to the target based on the intersection map of target-affected links. Subsequently, we establish a novel positioning model based on signal strengths of wireless links and high-intersections in each cluster determined by DBSCAN algorithm using a saddle surface model. Finally, Hooke-Jeeves method is used to search for the optimal solution of the positioning model. The experiment shows that InterMap-NRTI's accuracy and identification accuracy rate outperforms other methods by an average of more than 30% and more than 14%, respectively, and InterMap-NRTI has a good tolerance for scenario layout.