Image processing and agent-based framework for the geolocation of emitters

The research presented in this thesis is about a task of geolocation of radio frequency emitters. In this research the problem of geolocation of non-collaborative emitter was addressed. This thesis presents the novel algorithm for the RF emitter geolocation based on the image process technique known as Hough Transform. The comparison of this algorithm with traditional approaches to geolocation showed a number of benefits, like robustness, accuracy and advanced fusion capability. The application of the Hough Transform to data fusion allowed to use the modern concepts of agentbased fusion and cluster level fusion, thus moving the solution of the problem of the geolocation to upper level of fusion hierarchy. The work on Hough Transform lead to a comparison of the Bayesian and non-Bayesian approaches in solving the task of geolocation. Exploitation of the comparison lead to the derivation of a generalized estimator. This estimator highlighted a number of mathematical functions which can be exploited for geolocation and data fusion. These functions has been tested for the purpose of data fusion in geolocation and it was found that Hough Transform is a useful alternative approach for the data fusion for geolocation of RF emitter.