A class of suboptimum methods for space-time adaptive processing using empirical characteristic function

In this paper, a novel class of suboptimum methods for space-time adaptive processing (STAP) is presented. The newly proposed algorithm uses some special feature of the empirical characteristic function (ECF) in Fourier domain rather than predefined data probability distribution structures in order to constitute the STAP weights. Robustness against the statistical uncertainties of the input observations and great reduction in STAP covariance matrix dimensions are major benefits of our new method. This study can find applications in moving target indication in presence of highly correlated non-Gaussian interferences such as K-distributed observations. Performance of the method is assessed in comparison with those of a few conventional approaches via Monte Carlo simulations.