ML and Bayesian TOA location estimators for NLOS environments

Abstract

A major concern of cellular and PCS-based wireless location systems is the effect of the harsh propagation environment, particularly non-line-of-sight (NLOS) propagation. In order to improve location accuracy under such conditions, this paper proposes a novel location technique that estimates the true, or line-of-sight (LOS), ranges based on NLOS range measurements. The algorithms, designed for NLOS environments, make use of well-known multipath scattering models (ring/disk of scatterers and Gaussian distributed scatterers) in order to incorporate the effects of NLOS propagation. From PDFs of the TOAs derived using the models, maximum likelihood expectation-maximization and Bayesian estimators are applied to multipath TOA measurements at each BS to estimate the LOS distance between the MS and the BS. Simulated results show that the algorithms provide considerable improvement over traditional location algorithms.