Interactive-Multiple-Model Based Robust Target Tracking In LOS/NLOS Hybrid Environments

In modern wireless communications networks, services and applicationsassociated with position are used more and more frequent, gradually changingpeople’s lives, and effectively promoting the development of the target trackingtechnology, making its have more broad application prospects.The target trackingtechnology in Line-of-sight(LOS) and None-line-of-sight(NLOS) hybrid environmentis tudied in this thesis. To attack parameter estimation and solution of trackingequations problem in LOS/NLOS hybrid environment, mixed parameter estimationand robust tracking algorithm are proposed in this thesis. Simulation results show thatthe proposed algorithms are applied to LOS/NLOS hybrid environment. Besides, ithas higher positioning accuracy, better tracking stability and realizability.The relevant measurement parameters involved in the target tracking technologyand some classical algorithms are introduced at first.And the tracking algorithm basedon Kalman filter is analyzed,with giving the standard Kalman filter filtering principleand iterative process. By optimizing the iterative process of standard Kalman filterand known the NLOS error distribution situation, the algorithm can effectively inhibitthe NLOS outlier influence. But with the changing of NLOS error in LOS/NLOShybrid environment, a common approach based on Kalman filter is unstable evendivergence.And then a robust interactive multiple model approach based on extendedH-infinity Filter (EHF) and extended Kalman filter (EKF) is put forward, designed themodel set that making up of line-of-sight (LOS) and non-line-of-sight (NLOS) todescribe the motion of the MS. The LOS and NLOS transmission channels betweenthe base station and target are matched by EKF and EHF, respectively. Then, themode transmission between EKF and EHF is described by a Markov process.Simulation results show that the proposed algorithm can effectively inhibit theinfluence of non-line-of-sight error and has a higher positioning accuracy, reflecting abetter positioning performance and real-time in LOS/NLOS hybrid environment.Based on this algorithm this thesis further proposed time of arrival(TOA) andReceived signal strength Indication(RSSI) mixed parameter estimation based on theEHF and EKF, combined the advantages of measurement parameters alone,effectively improve the tracking performance. In conclusion, we have been trying to summarize the theoretical knowledge andthe algorithms of the research on target tracking, and made a prospect for furtherresearch.