Performance Analysis of the NAMF Test in Heterogeneous Non-Gaussian Radar Clutter Scenarios

An important issue in radar space-time adaptive processing (STAP) is the selection of training data for estimation of the unknown disturbance covariance matrix. In this paper, we address the training data selection issue and analyze the performance of normalized adaptive matched filter (NAMF) in heterogeneous non-Gaussian radar clutter scenarios. Simulations of the probability of detection (PD) versus signal-to- interference-plus-noise ratio (SINR) for the NAMF are conducted using Mountain Top data, which exhibit non-Gaussian statistics. To select the training data representative of the interference, we apply adaptive beamforming and show that PD versus SINR for NAMF is robust with respect to a wide range of variation in the beamformer output. To mitigate heterogeneous clutter, the self-censoring reiterative fast maximum likelihood (SCRFML) algorithm is employed to regularize the eigen spectrum underlying the unknown disturbance covariance matrix. We demonstrate that the NAMF detection performance can be significantly improved with the application of SCRFML.