A Robust Mobile Location Estimator in NLOS Environments using Hybrid Filtering

One of the major difficulties to performing accurate mobile terminal location is the non line-of-sight (NLOS) condition, caused by blocking of the signal direct path by obstacles. A tracking approach based on the Kalman filtering framework has been proposed to mitigate NLOS error, but it strongly depends on proper state model characterization of both the terminal dynamics and the NLOS error evolution. This paper presents a new location estimation scheme for a time of arrival system, which is robust to some modelling mismatches, using a sequential Monte Carlo method (particle filtering) interacting with a Kalman filter to propagate the estimates in time. The performance evaluation of the proposed approach is carried out considering different scenarios and compared to other alternatives.