Space time adaptive processing in multichannel passive radar

Nowdays, passive bistatic radar (PBR) systems have become a subject of intensive research, owing essentially to its unique features, such as low probability of interception, small size and low cost. Passive radar is a concept where illuminators of opportunity are used. In a bistatic passive radar the main challenges are: estimating the reference signal which is required for detection, mitigating the direct signal, multipath and clutter echoes on the surveillance channel and finally achieving a sufficient SINR to detect targets. This thesis is concerned with the definition and application of adaptive signal processing techniques to a multichannel passive radar receiver. Adaptive signal processing techniques are well known for active pulse radars. A PBR system operates in a continuous mode, therefore the received signal is not avalaible in the classical array elements-slow time-range domain such as in active pulse radar. A major component of this research focuses on demonstrating the applicability of traditional adaptive algorithms, developed in the active radar contest, with passive radar. Firstly a new detailed formulation of the sub optimum “batches algorithm”, used to evaluate the cross correlation function, is proposed. Then innovative 1D temporal adaptive processing techniques are defined extending the matched filter concept to an adaptive matched filter formulation. Afterwards a new spatial adaptive technique, based on the application of the adaptive digital beamforming after the matched filter, is investigated. Finally both 1D spatial and temporal adaptive techniques are extended to 2D space-time adaptive processing techniques. Specifically we demonstrate the applicability of STAP processing to a passive bistatic radar and we show how the classical STAP algorithms, developed for active radar systems, can be applied to a PBR system. The new defined passive radar signal processing architectures are compared with the standard approaches and the effectiveness of the proposed techniques is demonstrated considering both simulated and real data.