Bayesian Algorithms for Mobile Terminal Positioning in Outdoor Wireless Environments

The ability to reliably and cheaply localize mobile terminals will allow users to understand and utilize the what, where and when of the surrounding physical world. Therefore, mobile terminal location information will open novel application opportunities in many areas. The mobile terminal positioning problem is categorized into three different types according to the availability of (1) initial accurate location information and (2) motion measurement data. Location estimation refers to the mobile positioning problem when both the initial location and motion measurement data are not available. If both are available, the positioning problem is referred to as position tracking. When only motion measurements are available the problem is known as global localization. These positioning problems were solved within the Bayesian filtering framework in order to work under a common theoretical context. Filter derivation and implementation algorithms are provided with emphasis on the radio mapping approach. The radio maps of the experimental area have been created by a 3D deterministic radio propagation tool with a grid resolution of 5 m. Real-world experiments were conducted in a GSM network, deployed in a semi-urban environment, in order to investigate the performance of the different positioning algorithms. A method is proposed to compute the Cram?r-Rao lower bound (CRLB) in order to asses the performance of the received signal strength (RSS) based location estimation algorithm (database correlation method). The fingerprinting databases are usually constructed using complex 3D radio propagation prediction tools. Thus, the RSS-location mapping function is neither continuous nor differentiable everywhere as required by the Cram?r-Rao bound calculations. The key approach is reconstructing the fingerprinting database using an empirical path loss formula that sufficiently characterizes the wireless propagation environment of the test area. The Cram?r-Rao lower bound is derived and calculated for the reconstructed database in the experimental area. Furthermore, the posterior Cram?r-Rao lower bound (PCRLB) is derived and computed in order to asses the performance of the position tracking algorithm.