Performance comparison of nonlinear filters for indoor WLAN positioning

Abstract

Indoor WLAN positioning should be modeled as a nonlinear and non-Gaussian dynamic system due to the complex indoor environment, radio propagation and motion behaviour. The aim of this paper is to analyze different filtering strategies for real life indoor WLAN positioning systems. The performance criteria for the comparison are the mean of localization errors and computational complexity. Three nonlinear filters are analyzed: Fourier density approximation (FF), particle filter (PF) and grid-based filter (GF), which are representatives for deterministic and random density approximation approaches. Our experimental results help to choose the appropriate filtering techniques under different resource limitations.