Target tracking in mixed LOS/NLOS environments

• Main Author: Yi, Lili
• Other Authors: Lin Zhiping
• Format: Theses and Dissertations
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/62117
• Institution: Nanyang Technological University

In this thesis, we discuss the problem of tracking the geographic position of a moving target in mixed line-of-sight and non-line-of-sight (LOS/NLOS) environments using measurements including time-of-arrival (TOA), angle-of-arrival (AOA) and time-difference-of-arrival (TDOA). While many tracking algorithms are available for accurately tracking a moving target in the LOS environment, it is desirable to develop reliable tracking algorithms for accurately tracking a moving target in the mixed LOS/NLOS environments since purely LOS environment seldom exists in practice, particularly in urban areas. When NLOS errors exist, traditional tracking techniques such as least squares (LS), Kalman filter (KF) and extended Kalman filter (EKF) will not work well. Thus, new tracking algorithms are required to mitigate or to remove the NLOS errors to improve the tracking accuracy. In this thesis, we first propose a new idea called individual measurement detection (IMD), which is one of the central ideas in this thesis. An IMD based EKF tracking strategy in conjunction with the IMD method is then applied to track a moving target with improved tracking performance. This approach is further extended to the case of robust EKF (rEKF) using TOA measurements. This tracking algorithm turns out to work better than the EKF tracking strategy for exponential NLOS errors. To further improve the tracking accuracy in mixed LOS/NLOS environments especially in severe NLOS conditions, we propose an individual TOA measurement estimation and LOS measurement detection algorithm, which is labeled as IMED. In this approach, each TOA measurement collected at a certain time step is treated individually to estimate a pseudo-measured position of the moving target. Then these pseudo-measured positions are passed to a detector to identify the LOS ones. The average of selected LOS pseudo-measured positions is then used into a KF. The developed tracking algorithms outperform various robust competing estimations found in the literature while no prior knowledge of the NLOS error statistics is required. The IMD based EKF and the IMD based rEKF tracking approaches are then used into TDOA based tracking problems. With the assistance of the road constraints, which are used as pseudo measurements, the tracking performance is improved with respect to these without road constraints. To further improve the tracking accuracy of the proposed IMED algorithm, some AOA measurements are incorporated into the IMED algorithm together with TOA measurements. The joint TOA/AOA measurements estimation and LOS measurement detection algorithm is improved with better performance than using TOA measurements only.