Rejection of radar clutter having autoregressive spectra

Performances of clutter rejection schemes have all along been evaluated assuming the clutter to have a Gaussian spectrum. Recent studies have however shown that the clutter may often be an AR or ARMA process having non-Gaussian spectra. We therefore evaluate the performance of different MTI schemes for clutter modeled as AR process. Both optimum as well as binomial weighted MTI schemes are considered, and depending on the type of clutter(s), a cascade of one or more of the three structures - MTI, AMTI, CI - are investigated. It is shown that the improvement factor for AR clutter is always considerably below that for clutter with Gaussian spectra except for certain cases involving the two-pulse canceller. The results are verified using computer simulation and explained using theoretical models.