Mobile tracking in unknown non-line-of-sight conditions

This paper studies the mobile tracking problem in mixed line-of-sight (LOS) and non-line-of-sight (NLOS) conditions, where the statistics of NLOS error are assumed unknown. Three different models are used to describe the NLOS errors. A Rao-Blackwellized particle filtering with parameter learning (RBPF-PL) is presented, in which the posterior of sight conditions is estimated by particle filtering while the mobile state and NLOS parameters are analytically computed. Simulation results are provided to evaluate the performance of RBPF-PL variants in different situations. Simulation show that unless it is known that NLOS noise has the same bias and variance in all the observations, the more complicated models should be employed as they work correctly even in NLOS model mismatch, with only slightly more computational complexity.