Active array target localization using time reversal signal processing.

摘要

Source detection and localization using sensor arrays are of considerable interest in classical array signal processing, wireless communications, radar systems, and sonar applications. Most radar systems are designed under the line-of-sight (LOS) condition and multipath (the propagation phenomenon that results in the transmitted signal reaching the receiver via multiple alternate paths) has a negative impact on radar resolution and its sensitivity in detecting and localizing the target. Considerable research attention has been devoted to address the problem of multipath in radar and sonar applications.;Rather than treating multipath as a detrimental effect, this thesis introduces time reversal (TR) to treat multipath positively for enhancing the performance of the target detection and localization algorithms. For this purpose, the TR matched filter based range estimation and wideband TR Direction-of-Arrival (DOA) estimation algorithms are formulated in a multipath environment and are compared with their conventional counterparts. The proposed TR localization framework is further extended from the traditional phased array radars to the MUltiple Input MUltiple Output (MIMO) radars. From a theoretical standpoint, this research derives the Cramer-Rao Bounds (CRBs) for the proposed TR localization algorithms taking advantage of the benefits of the spatial/multipath diversity in the time reversal DOA and range observations. The contribution of multipath to both the TR and conventional CRBs is analyzed through the impact of temporal processing on the quality of different types of estimators. To the best of our knowledge, this is the first instance of applying TR to the DOA estimation and of deriving the corresponding lower bounds in multipath environments. The emergence of software-driven waveform generators with radar provides us with the ability to modify the transmitted waveform to match the environment and makes TR reshaping in radar a practically feasible approach. The proposed TR/MIMO radar framework provides the signal processing community with a novel adaptive technique that has a built in ability to adapt the transmitted waveform to the multipath environment and, therefore, enhance the performance of the localization algorithms. Experimental simulations based on the finite difference, time domain electromagnetic simulations verify the improvement that TR array processing offers over its traditional counterparts.