A Weighted APIT Localization Algorithm based on Vector Similarity

Abstract

To reduce incorrect decisions of PIT test (Point-In-Triangulation Test) and localization error of APIT algorithm, a weighted APIT algorithm is proposed based on vector similarity. Firstly, for better distance precision, the estimated distances between nodes computed from RSSI attenuation model are corrected by the least square method. Secondly, a center of gravity method is proposed for PIT test, which decreases InToOut and OutToIn error. Finally, a weight is introduced based on vector similarity for the different reliability of triangular regions. The simulation result shows that the positioning error of ASG-APIT algorithm decreased by 43%, 17% and 13% respectively, compared with APIT, MIX-APIT and RAPIT algorithm.