Microwave Imaging using Time Reversal Techniques and Resolution Enhancement in Dispersive Media

Abstract

TR-based techniques have recently been proposed for the microwave imaging of stationary and moving targets. Classical TR-based array processing methods perform poorly for imaging of non-pointlike targets. This thesis proposes a novel TR-MUltiple SIgnals Classification (MUSIC)-based algorithm for the Ultra WideBand (UWB) through-the-wall imaging of multiple extended moving targets. This algorithm performs spatiotemporal windowing on the received signals, to exploit their temporal and spatial diversity. Then, it applies a novel process to identify the signal subspace of each window. Finally, it produces the final radar image by selecting the strongest results obtained by different temporal and spatial windows and frequencies. This thesis applies the proposed algorithm to a simulated practical scenario and achieves the detection of five overlapping human-like targets moving behind a brick wall, whereas the state-of-the-art windowed UWB-MUSIC method detects only two targets. Dispersive media cause frequency-dependent additional attenuation onto electromagnetic waves propagating through them. Consequently, the resolution of the TR imaging is degraded in such media. This thesis also introduces a new algorithm for the resolution enhancement of UWB TR radar imaging in dispersive environments. This algorithm takes into account the frequency-dependent complex permittivity of the propagation medium across the entire bandwidth of the excitation pulse. Using this complex permittivity, it constructs a Continuous Wavelet Transform (CWT)-based model of the attenuation, to create inverse filters in the wavelet domain, which compensate for the effects of the attenuation. This algorithm also introduces a smart wavelet scaling concept to minimise undesired noise amplification. This thesis applies this proposed algorithm to a practical scenario and enhances the resolution of UWB microwave TR imaging of a simulated brain tumour inside the Digital Human Phantom (DHP), whilst the existing work fails to detect the tumour.

Date of Award	1 Aug 2022
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Laith Danoon (Supervisor) & Fumie Costen (Supervisor)