Selected Topics In Direct Geolocation Of Radio Transmitters & Passive Targets

This dissertation is dedicated to the exploration of various direct positioning algorithms for radio transmitters and passive target geolocation. Contrary to the traditional “two-step” approach, the “direct positioning” approach states that the radio transmitter’s position can be extracted directly from the raw samples of the radio transmitter signals collected by the system sensors, without explicitly going through an estimation of position-related parameters such as time-delay, angular or amplitude information. In this work, the concept of direct positioning is applied to various models and consistently outperforms the traditional two-step position estimators, while tightly attaining the theoretical performance bounds. In the sequel, we explore 3 models for radio transmitters and passive target geolocation. The first model discussed in chapter 3, harnesses the transmit signal diversity of MIMO Radar systems to enhance passive-target position estimation via direct estimation algorithms. The algorithms are developed for both known and unknown transmitted signal waveforms. The known transmitted waveform is applicable to the classical radar model, while the unknown waveform assumption is applicable to passive MIMO radars utilizing “signals of opportunity” for target localization. The second model addressed in chapter 4, explores the performance of radio transmitter geolocation using a so-called “single platform geolocation” (SPG). The geolocation system in this case consists of only a single sensor (or a single array of sensors). The transmitter localization is accomplished using “in-band aiding” enabled by passive signal reflectors or transponders that are placed at known locations. Algorithms that enable direct geolocation of radio transmitters emitting waveforms, which are either known or unknown to the receiver, are derived and analyzed. The third model, described in chapter 5, addresses the problem of emitter geolocation in the presence of local scattering. In contrast to the first two models, which rely on the classical assumption of the emitting source being a point in the 3-D space, this model assumes that the emitting source is surrounded by a large number of closely-spaced local scatterers. The local-scattering assumption which often characterizes geolocation of radio transmitters in indoor, urban or suburban scenarios, imposes certain challenges on the geolocation task. The geolocation system in this model consists of multiple receiving sensor-arrays deployed in various geographical locations. The emitter signals impinging upon the arrays are associated with random angular information that is distributed according to some known statistical distribution. The angular distribution model known as the Gaussian angle of arrival (GAA) is employed. Following the GAA assumption, a spatio-temporal model is developed and various algorithms for direct geolocation under local scattering are derived and analyzed. For all the models outlined, theoretical analyzes, as well as simulated data are provided.