Urban Location Estimation for Mobile Cellular Networks: A Fuzzy-Tuned Hybrid Systems Approach

In this paper, we present a plausible approach using time-of-arrival (TOA) measurements for enhancing accuracy and robustness of urban location estimation in mobile cellular networks. The mobile localization problem is cast into the state estimation of a fuzzy-tuned hybrid system. First, we propose a Markov-transitioned fuzzy-tuned hybrid framework for modeling the dynamics of a mobile station (MS), received line-of-sight (LOS)/or non-LOS (NLOS) range measurements, and NLOS bias variations for each base station. The proposed framework also incorporates fuzzy-logic rules for adaptively tuning process noise covariances to model the effects of both mobility variations of the MS. Second, we derive a selective fuzzy-tuned interacting multiple-model (SFT-IMM) algorithm based on the proposed framework. The proposed algorithm can lead to notable performance gains because it can more accurately identify LOS and NLOS conditions and selectively perform fuzzy tuning of process noise covariances. Simulations confirm the performance advantages of the proposed algorithm over other methods such as the Kalman filter and the IMM.