This thesis presents mobile positioning and tracking with non-line of sight (NLOS) mitigation using time difference of arrival (TDOA) in biased extended Kalman filter (BEKF) in indoor dense multipath Ultra-Wideband (UWB) environment. The most serious issues which render to influence accuracy for the time-based location system is NLOS problem. Kalman filters (KFs) are used for smoothing range measurement data, and a method with sliding window is proposed to process range data for calculating standard deviation in a hypothesis testing and then identifying NLOS scenarios. When the measured arrival time has been converted to range difference, the biased extended Kalman filter is proposed to mitigate the NLOS error in the certain base stations (BSs) for mobile station (MS) positioning and trajectory tracking. From the simulation results in the indoor positioning environment with measurement and NLOS error, the sliding window algorithm and biased extended Kalman filter have higher accuracy than other related methods for NLOS identification and mitigation in positioning.

In the thesis, we propose a non-line of sight (NLOS) mitigation approach based on the interacting multiple model (IMM) algorithm. The IMM-based structure, composed of a biased Kalman filter (BKF) and a Kalman filter with NLOS-discarding process (KF-D), is capable of mitigating the ranging error caused by the NLOS effects, and therefore improving the performance and accuracy in wireless location systems. The NLOS effect on signal transmission is one of the major factors that affect the accuracy of the time-based location systems. Effective NLOS identification and mitigation usually count on pre-determined statistic distribution and hypothesis assumption in the signals. Because the variance of the NLOS error is much large than that of measurement noise, hypothesis testing on the LOS/NLOS status can be formulated.The BKF combines the sliding window and decides the status by using hypothesis testing. The calculated variance and the detection result are used in switching between the biased and unbiased modes in the Kalman filter. In the contrast, the KF-D scheme identifies the NLOS status and tries to eliminate the NLOS effects by directly using the estimated results from the LOS stage. The KF-D scheme can achieve reasonably good NLOS mitigation if the estimates in the LOS status are obtained. Due to the discarding process, changes of the state vector within the NLOS stage are possibly ignored, and will cause larger errors in the state estimates. The BKF and KF-D can make up for each other by formulating the filters in an IMM structure, which could tune up the probabilities of BKF and KF-D. In our approach, the measured data are smoothed by sliding window and a BKF. The variance of data and the hypothesis test result are passed to the two filters. The BKF switches between the biased/unbiased modes by using the result. The KF-D may receive the estimated value from BKF based on the results. The probability computation unit changes the weights to get the estimated TOA values. With the simulations in ultra-wideband (UWB) signals, it can be seen that the proposed IMM-based approach can effectively mitigate the NLOS effects and increase the accuracy in wireless position.

Ultra Wideband (UWB) signal can offer an accurate location service in wireless sensor networks because its high range resolution. Target tracking by multiple sensors can provide better performance, but the centralized algorithms are not suitable for wireless sensor networks. In additional, the non line of sight (NLOS) propagation error leads to severe degradation of the accuracy in location systems. In this thesis, NLOS identification and mitigation technique utilizing modified biased Kalman filter (KF) is proposed to reduce the NLOS time of arrival (TOA) errors in UWB environments. We combine the modified biased Kalman filter with sliding window to identify and mitigate different degree of NLOS errors immediately. In order to deal with the influence of inaccurate NLOS angle of arrival (AOA) measurements, we also had a discussion on AOA selection and fusion methods. In the distributed location structure, we used the extended Information filter (EIF) to process the formulated time difference of arrival (TDOA) and AOA measurements for the target positioning and tracking. Instead of using extended Kalman filter, extended Information filter can assimilate selected AOA easily without dynamic dimensions. The sensors are divided into different groups for distributed TDOA/AOA location to reduce computation and then each group can assimilate information from other groups easily to maintain precise location. The simulation results show that the proposed architecture can mitigate NLOS errors effectively and improve the accuracy of target positioning and tracking from distributed location and data fusion in wireless sensor networks.

We consider time-of-arrival based robust geolocation in harsh line-of-sight/non-line-of-sight environments. Herein, we assume the probability density function (PDF) of the measurement error to be completely unknown and develop an iterative algorithm for robust position estimation. The iterative algorithm alternates between a PDF estimation step, which approximates the exact measurement error PDF (albeit unknown) under the current parameter estimate via adaptive kernel density estimation, and a parameter estimation step, which resolves a position estimate from the approximate log-likelihood function via a quasi-Newton method. Unless the convergence condition is satisfied, the resolved position estimate is then used to refine the PDF estimation in the next iteration. We also present the best achievable geolocation accuracy in terms of the Cramér-Rao lower bound. Various simulations have been conducted in both real-world and simulated scenarios. When the number of received range measurements is large, the new proposed position estimator attains the performance of the maximum likelihood estimator (MLE). When the number of range measurements is small, it deviates from the MLE, but still outperforms several salient robust estimators in terms of geolocation accuracy, which comes at the cost of higher computational complexity.