Space-time adaptive processing techniques for multichannel mobile passive radar

Passive radar technology has reached a level of maturity for stationary sensor operations, widely proving the ability to detect, localize and track targets, by exploiting different kinds of illuminators of opportunity. In recent years, a renewed interest from both the scientific community and the industry has opened new perspectives and research areas. One of the most interesting and challenging ones is the use of passive radar sensors onboard moving platforms. This may offer a number of strategic advantages and extend the functionalities of passive radar to applications like synthetic aperture radar (SAR) imaging and ground moving target indication (GMTI). However, these benefits are paid in terms of motion-induced Doppler distortions of the received signals, which can adversely affect the system performance. In the case of surveillance applications, the detection of slowly moving targets is hindered by the Doppler-spread clutter returns, due to platform motion, and requires the use of space-time processing techniques, applied on signals collected by multiple receiving channels. Although in recent technical literature the feasibility of this concept has been preliminarily demonstrated, mobile passive radar is still far from being a mature technology and several issues still need to be addressed, mostly connected to the peculiar characteristics of the passive bistatic scenario. Specifically, significant limitations may come from the continuous and time-varying nature of the typical waveforms of opportunity, not suitable for conventional space-time processing techniques. Moreover, the low directivity of the practical receiving antennas, paired with a bistatic omni-directional illumination, further increases the clutter Doppler bandwidth and results in the simultaneous reception of non-negligible clutter contributions from a very wide angular sector. Such contributions are likely to undergo an angle-dependent imbalance across the receiving channels, exacerbated by the use of low-cost hardware. This thesis takes research on mobile passive radar for surveillance applications one step further, finding solutions to tackle the main limitations deriving from the passive bistatic framework, while preserving the paradigm of a simple system architecture. Attention is devoted to the development of signal processing algorithms and operational strategies for multichannel mobile passive radar, focusing on space-time processing techniques aimed at clutter cancellation and slowly moving target detection and localization. First, a processing scheme based on the displaced phase centre antenna (DPCA) approach is considered, for dual-channel systems. The scheme offers a simple and effective solution for passive radar GMTI, but its cancellation performance can be severely compromised by the presence of angle-dependent imbalances affecting the receiving channels. Therefore, it is paired with adaptive clutter-based calibration techniques, specifically devised for mobile passive radar. By exploiting the fine Doppler resolution offered by the typical long integration times and the one-to-one relationship between angle of arrival and Doppler frequency of the stationary scatterers, the devised techniques compensate for the angle-dependent imbalances and prove largely necessary to guarantee an effective clutter cancellation. Then, the attention is focused on space-time adaptive processing (STAP) techniques for multichannel mobile passive radar. In this case, the clutter cancellation capability relies on the adaptivity of the space-time filter, by resorting to an adjacent-bin post-Doppler (ABPD) approach. This allows to significantly reduce the size of the adaptive problem and intrinsically compensate for potential angle-dependent channel errors, by operating on a clutter subspace accounting for a limited angular sector. Therefore, ad hoc strategies are devised to counteract the effects of channel imbalance on the moving target detection and localization performance. By exploiting the clutter echoes to correct the spatial steering vector mismatch, the proposed STAP scheme is shown to enable an accurate estimation of target direction of arrival (DOA), which represents a critical task in system featuring few wide beam antennas. Finally, a dual cancelled channel STAP scheme is proposed, aimed at further reducing the system computational complexity and the number of required training data, compared to a conventional full-array solution. The proposed scheme simplifies the DOA estimation process and proves to be robust against the adaptivity losses commonly arising in a real bistatic clutter scenario, allowing effective operation even in the case of a limited sample support. The effectiveness of the techniques proposed in this work is validated by means of extensive simulated analyses and applications to real data, collected by an experimental multichannel passive radar installed on a moving platform and based on DVB-T transmission.