Details

Abstrakt in Englisch

Indoor positioning is a service that is in high demand. Its goal is to locate a large number of users inside a building, and is widely known as local positioning system (LPS). It is particularly effective in large densely crowded places, such as hospitals, airports, parking areas, and on public transportation hubs. Furthermore, accurate indoor positioning is an essential technique in a so-called “smart factor”. Several commercial radio frequency (RF) based systems achieve a high level of accuracy under outdoor conditions, however they fail when tested indoors. Developing a single system that is suitable for a range of different conditions and purposes is a big challenge facing developers of LPSs. This research presents a multi-band scalable positioning system that not only provides accurate means of navigation, but also seamlessly expands the maximum number of reference and mobile nodes. Time synchronization is carried out wirelessly, minimizing the hardware required to set up the system easily. Because of its high precision and immunity to RF interference, a multi-band frequency modulated continuous wave (FMCW) radar system has been designed to handle core time measurements. Several techniques and algorithms are specifically optimized to follow the special requirements of the FMCW radar system design. The system relies on commercial crystal oscillators (XOs) to generate its reference clock at each of its nodes. Due to the different manufacturing and temperature changes, each XO generates a slightly different frequency that has a great influence on the accuracy of the positioning estimate. A scalable FMCW based synchronization protocol is proposed to estimate the XO offset and compensate for it. The accuracy of the positioning system was enhanced by a factor of at least 3 in various conditions after compensating for the errors in measurements that are due to the frequency offset. In order to withstand complex indoor conditions, a line-of-sight (LOS) searching protocol together with a particle filter (PF) tracking algorithm is considered. Measurements were carried out to evaluate the system in several practical conditions and to compare it with current state-of-the-art (SOA) techniques. The positioning outcomes were compared to a reference optical system with a precision of about 1 mm. The proposed system was evaluated with different configurations in two practical scenarios. The first experiment was performed outdoors in relaxed channel conditions, while a second indoor scenario was used to demonstrate the impact of the dual-band and the scalability of the system. The evaluation scenarios included areas with both LOS and non-line-of-sight (NLOS) conditions. A root mean square (RMS) positioning error of less than 17 cm and 30 cm was achieved in a coverage area of around 500 m2, in both outdoor and indoor conditions respectively. The system was optimized for minimum positioning error with maximum effective coverage area. To the best of the author’s knowledge, the proposed system outperforms the current commercial systems not only in the aspect of positioning error, but also in the coverage efficiency of the reference stations, which defines the minimum number of reference stations that are required to cover a specific area. The thesis leads the reader from the system concept through the hardware realization, passing by the associated algorithms to the practical applications and challenges. The text also introduces its readers to the details of different practical system level techniques and emergent filtering algorithms that are especially optimized for the multi-band FMCW radar system.