Robust Target Tracking in Mixed LOS/NLOS Environments Using a Probabilistic Multihypothesis Tracker

Tracking algorithms generally perform well in line-of-sight (LOS) environments but face significant challenges in non-LOS (NLOS) scenarios due to multipath effects that degrade tracking accuracy. To address these challenges, we propose a robust tracking algorithm that integrates interacting multiple models (IMMs) with a probabilistic multihypothesis tracker (PMHT). This innovative approach improves upon the state of the art by effectively distinguishing between LOS and NLOS measurements and optimally fusing their information. Hypothesis testing, combined with IMM model probabilities, is used to classify measurements as either LOS or NLOS, enabling the algorithm to handle the unique characteristics of each type effectively. LOS measurements are processed using the Hungarian algorithm for data association and are updated through LOS-extended Kalman filter (EKF) to ensure high accuracy. In parallel, NLOS measurements are combined with predicted pseudomeasurements using the PMHT algorithm, generating composite measurements and covariances that account for NLOS-related uncertainties. Finally, the algorithm fuses LOS and NLOS estimates using IMM mode probabilities, yielding a unified and optimal state estimate capable of addressing the challenges posed by mixed LOS/NLOS environments. Simulation results validate the algorithm's effectiveness in mitigating NLOS-induced errors, demonstrating robust tracking performance even in scenarios with severely mixed LOS and NLOS conditions.