Hybrid location estimation by fusing WLAN signals and inertial data

摘要

Radio frequency (RF) signal propagation suffers from time-varying fading effects, and thus radio map-based localization systems are hard to hold the expected accuracy. Base stations (BS)-based architectures show us the probable solutions to overcome the negative impacts by producing adaptive radio maps. In this chapter, the adaptive approach that is presented in our previous work is adopted. To further mitigate the impacts of dynamic environments, we propose a hybrid location estimation method that fuses WLAN signals and inertial data through the sequential importance resampling (SIR) Particle Filter (PF) algorithm. Experimental results suggest that the hybrid method can provide more accurate location tracking, compared to previous algorithms, such as K weighted nearest neighbors (KWNN), initial radio map-based PF, adaptive radio map-based PF, pedestrian dead reckoning (PDR). And it nearly costs equivalent computational time, compared to those radio map-based PF algorithms.