CFAR Adaptive TVAR versus Diagonally Loaded AMF Detectors

We consider adaptive detection for interferences that are described by a time-varying autoregressive model of order m, TVAR(m), where the training sample size T is less than the dimension of the adaptive filter (antenna array) M. We discuss the "practical CFARness" of this detector and formulate conditions on its efficiency (CFAR stands for constant false-alarm rate). The detection performance of the TVAR(m)-based adaptive matched filter (AMF) detector is compared with the performance of the diagonally loaded AMF (LAMF) detector. We demonstrate that the relative efficiency of these detectors strongly depends on the properties of the eigenspectrum of the interference covariance matrix.