DEM-assisted Geo-location Of Communication Signals Using TDOA Measurements

The geo-location of a radio frequency emitter is a problem of practical interest. The problem has been studied from many angles, for example developing analytic solutions to the problem, using multi-linear algebra to correct for time difference of arrival (TDOA) inconsistencies, trilateration, or using statistical techniques such as clustering or maximum likelihood estimation. With the increasing development of technology, higher positioning accuracy is attached greater importance in present research and corresponding applications. The exact localization of reconnaissance objectives will change the means of warfare and enhance the timeliness of operational command. Existing means of ground reconnaissance mainly are aerial reconnaissance and space reconnaissance which are influenced by topography. The means of DEM-assisted geo-location of reconnaissance objectives are realistic with greater improvement of source location accuracy. Hence, the research on the DEM-assisted geo-location of reconnaissance objectives has very far-reaching significance.The thesis addresses the problem of geo-location of a source whose altitude is unknown in the presence of random errors in receiver locations and TDOA/FDOA measurements by proposing novel LSSQP methods using TDOA measurements and TDOA/FDOA measurements, respectively. More importantly, a new localization algorithm based on Niche Genetic Algorithm (NGA) and an improved TDOA estimate method based on MUSIC algorithm are proposed, respectively. In addition, this paper performs an analysis and provides an efficient algorithm for locating source using time difference of arrival measurements. The dissertation is organized as follows.Firstly, a novel LSSQP algorithm using TDOA/FDOA measurements which effectively integrate the digital elevation information is proposed. The analysis starts with the Cramer-Rao lower bound (CRLB) with receiver location errors, and derives the increase in mean square error (MSE) in source location estimate with the geometrical constraint if the source altitude is assumed accurate but in fact has error. And then, we derive the MSE of source location with the geometrical constraint without the altitude error. A LSSQP algorithm based on sequential quadratic programming (SQP) method is then proposed for this constrained nonlinear optimization problem to reduce the estimation error, and it is shown analytically, without any approximation, to achieve the CRLB accuracy for an unknown altitude source.Secondly, this paper proposes a modified iterative algorithm using TDOA measurements, for estimating the location of an unknown altitude source with the introduction of digital elevation model (DEM). The proposed algorithm, with the constraint of the WGS-84 oblate spheroid earth model which can be modified in each iteration till convergence, can produce a more accurate source location estimate. The introduction of DEM effectively reduces the degradation of localization accuracy of an unknown altitude source. The theoretical analysis shows that the performance of the proposed algorithm surpasses that of others. The simulation results indicate that it is computationally efficient and significantly improves the source location accuracy. With this algorithm, the localization accuracy of an unknown altitude source on the land approaches that of a source on sea level.Thirdly, the LSSQP algorithm is applied to the passive geo-location using TDOA measurements. The search efficiency of the LSSQP algorithm which utilizes the constraint of the WGS-84 oblate spheroid earth model and combines SQP method with least square estimate is significantly improved. A comparison is drawn between the MSE in source location estimate with and without the geometrical constraint. And then, another comparison is drawn between the MSE of source location with the geometrical constraint with and without the altitude error. In addition, the disable threshold of WGS-84 oblate spheroid earth model and the analysis for the effect of the source altitude on the performance of localization algorithm are given. The error analysis and simulation results indicate the computationally efficient and superior performance of the proposed method as compared to the existing closed-form method and Taylor-series iterative based approach for the source geo-location.Fourthly, the paper presents a new TDOA location algorithm based on niche genetic algorithm for geo-location of an unknown altitude source. The algorithm which effectively integrates the digital elevation information skillfully combines the niche technology with the optimum constraint of minimizing the CRLB. With the introduction of the digital elevation which corrects the fitness of individuals, the individual achieves local optimum value along the direction of optimal evolution. Furthermore, the niche technology enables this algorithm to achieve global searching and solve easy trapping into local optimum value problem. Numerical simulations show that the proposed algorithm has higher accuracy and rapid convergence.Fifthly, time delay estimation is a basic technique for numerous applications where there is a need to localize and track a radiating source. The generalized cross-correlation (GCC) method which performs reasonably well when reverberation or noise is not too high is summarized. But the performance of the GCC method is poor when the transmit signals are unknown and reference signal is unavailable. Hence, for the TDOA estimation in passive positioning system, an improved MUSIC-based TDOA estimate method is proposed over multipath fading channels when noise is white Gaussian noise. Simulation result shows that the high-resolution algorithm features higher computational efficiency. Finally, the author presents a summarization of the dissertation, introducing both the originality and deficiency of the current research. A discussion of the future prospect of the research is also given at the end of the dissertation.