Smartphone-based WiFi FTM Fingerprinting Approach with Map-aided Particle Filter

Abstract

Smartphone-based WiFi ranging positioning based on fine time measurement (FTM) always collapses in real-life scenarios. In this work, a novel map-aided particle filter (PF)-based WiFi FTM fingerprinting approach is proposed to address the poor performance of the WiFi FTM ranging positioning. Different from manually collecting fingerprints, this approach utilizes the theoretical received signal strength and geometric distances between the access points and reference points as the fingerprints, which means less labour-intensive work. For accurate WiFi position estimation, a map-aided PF is designed to find the optimal position. Extensive experiments are carried out in the non-line-of-sight (NLoS) and mixed line-of-sight/non-line-of-sight (LoS/NLoS) environments, and the testing results show that the accuracy and stability of FTM fingerprinting are improved by using the mixed RSS and ranging data fingerprints. The minimal mean location errors (MEs) of the PF-based WiFi FTM fingerprinting in NLoS and mixed LoS/NLoS conditions are 1.70 m and 1.85 m, respectively. Compared to the classic weighted least square method, the MEs are reduced by 54.91% and 45.43%, respectively. The testing results demonstrate that the PF-based FTM fingerprinting is an effective approach that provides satisfactory localization results in real-life indoor environments.