Impact of processing approach on SNR performance for passive bistatic noise radar

In passive bistatic noise radar, cross-correlation is a popular processing choice for range compression of received signals. However, when using cross-correlation, the noise floor is dependent upon the characteristics of the clutter in the scene. For similar computational requirements, least-mean-square (LMS) processing can yield signal-to-noise ratio (SNR) performance which is independent of the clutter characteristics, making it a superior choice relative to cross-correlation. The superior SNR performance of LMS relative to cross-correlation may be even more striking when the transmitted noise waveform is not spectrally uniform. This paper will study the results achieved in a zero-Doppler range profile containing a single point scatterer and multiple point scatterers when cross-correlation, whitened cross-correlation, and LMS processing are used with a transmitted waveform whose spectrum is flat (white noise) and notched (colored noise).