Adaptive signal processing and mapping in Doppler radar

Radar detection of targets moving at ground level or at low altitude requires some specific signal processing schemes. These include clutter filtering and adaptive thresholding to maintain a constant false alarm rate. The clutter is then assumed to both locally homogeneous and extended. On practice, these hypotheses are only partially true because the surveyed area contains clutter edges (forest edges, coasts, etc.). Signal processors ignoring discontinuities and parameter variations of the model encounter false alarm in the neighbourhood of clutter edges and loss of detection sensibility. A method is presented which minimizes the limitations described above by precisely analysing the clutter characteristics in real time. Cancellation of the clutter signal by a filter adapted in real time to each antenna sweep and to each radar resolution cell is performed. This filter is designed to have a transfer function almost equal to the inverse of the clutter spectral density as observed in a selected area