Research On Wireless Mobile Network-aided Positioning Algorithms

This dissertation focuses on the positioning techniques and methods, which has various and important applications in the cellular wireless networks, wireless LANs and ad hoc sensor networks. Corresponding to different network topology and different real application scenarios, we construct mathematical models from various angles of view, analyze their detail theory foundation and propose several positioning algorithms with theory and practical value. Simulation results reveal the excellent performance of our proposed algorithms.First, we survey the state-of-the-art for the network-aided positioning design, challenges and techniques, with special focus on cellular network, WLAN and sensor network based design. Different network topologies pose various technical challenges in mobile positioning, which lead us to investigate the differences and relationships between them. We also provide the new opportunities and directions for future research to implement ubiquitous positioning.Non-line of sight propagation is the "killer issue" for the Time of Arrival (TOA) based cellular wireless location systems. From robust estimation, statistics and information theory, we propose a bootstrapping M-estimator and a minimum entropy estimator for the mobile positioning in Chapter 2. Compared with the minimum least squares error rule based Triangle location, simulation results show that these two positioning algorithms effectively suppress the NLOS error with high robustness and accuracy.Without any prior information, the Maximum Likelihood Estimation (MLE) is best, however, the real cellular network can learn some prior location information by itself or network management. Considering this case, we propose a cellular network-aided positioning model based on Least Square Support Vector Machine (LS-SVM) from the view of statistics learning theory and data fusion in Chapter 3. Simulation results show that our algorithm is more effective than the Triangle location algorithm, especially for NLOS propagation environments.