Ph.D. Tezi Görüntüleme

      The position of electromagnetic, acoustic, or seismic source can be determined by spatially separated synchronous receivers. Hyperbolic lines of positions are defined by using the time differences of arrivals of transmitted signals and the target is positioned at the intersection point. In this PhD thesis, high accurate and time difference of arrival (TDOA) based novel positioning methods are proposed. In the first section, the coefficient matrix which is used in time difference of arrival averaging (TDOAA) method is reproduced by a simple graph traversal algorithm and also showed that this matrix can be obtained not only using three TDOA groups but also four, five, and their combinations. In the second section, two novel algorithms based on TDOAA are proposed. In the first method, the averaging is combined with particle swarm optimization (PSO) and a significant decrease obtained in the positioning error by showing that the Cramer-Rao lower bound (CRLB) that is computed by independent set can be passed using the full TDOA set. In the second algorithm, the linear least square (LLS) and weighted linear least square (WLLS) techniques are used together and the traditional theoretical limit is passed by using the independent set. In the third section, a novel positioning method independent of the propagation speed and based on PSO is proposed. In the proposed method, the propagation speed has been canceled out by dividing the time differences each other and the TDOA ratios are obtained. Then, the target coordinates are found independent of the propagation speed by using these coordinates and PSO. In the fourth section, in order to increase the positioning accuracy further, optimum receiver arrays are dynamically oriented regarding to the target coordinates and the significant performance increase has been obtained. In the last section, the position of an acoustic source has been localized and performances of proposed methods compared with theoretical limit by using three different receiver geometries and six target points.