Two Dimensional Optimal Linear Detector for Slowly Fluctuating Radar Signals in Compound Gaussian Clutter

Abstract

In this paper, we propose a two dimensional (2-D) optimal linear detector (OLD) for radar target detection in compound Gaussian clutter and obtain an explicit relation of its coefficients for slowly fluctuating targets. We assume that the samples of signal and clutter are correlated in both range and azimuth directions, and the target detection in each radar cell is implemented by a 2-D samples collection of the received signal. In most conventional detectors, in each pulsation interval, the samples of echo of each radar cell are passed through a matched filter along the range, and a pre-detection is performed; then the binary results are integrated for successive echoes of that cell (along the azimuth). In fact, by applying the binary integration, we ignore the considerable correlation among 2-D data in the azimuth direction. In the proposed 2-D OLD detector, the correlation of signal and clutter in both range and azimuth directions is considered, aiming to improve the detection performance. Our simulations confirm that, this detector outperforms the conventional one dimensional OLD, as well as AND, OR, and ”κ out of n” binary integrators.