Implementation of Particle Filtering in TDOA Positioning

Abstract

This paper describes the implementation of particle filtering (PF) estimation method in TDOA positioning to mitigate the effect of NLOS signal propagation on the TDOA measurements. The measurement errors were modelled according to the circular disk scatter model (CDSM) ranging from 0 to 600 m. In this paper, we consider static PF which uses one TDOA measurement to estimate one fixed MT position. The network layout is set up with five base stations (BS) that resolve to a total of ten measured TDOAs in every PF iteration. The performance of the static PF was compared to two basic estimation methods namely robust intersection estimation (RIE) and linear least square (LLS). The simulation results show the stability of static PF over a wide range of measurement errors and giving an almost constant estimation error at various CDSM radiuses. Static PF outperforms RIE and LLS with the estimation error of less than 40 m and 60 m for 67% and 90% of the time respectively.