Simulation and performance analysis of New Logarithm-based CFAR processors in compound Gaussian clutter

This paper addresses Constant False Alarm Rate (CFAR) detection in compound Gaussian clutter. The proposed detectors, namely: Logarithmic Cell Averaging Greatest Of CFAR (Log CAGO-CFAR) and Logarithmic Cell Averaging Smallest Of CFAR (Log CASO-CFAR) are based on estimating the clutter’s texture component from the leading and the lagging windows that surround the cell under test (CUT) by computing the mean of the Logarithms of clutter samples of each window, then selecting the Maximum (GO) and the Minimum (SO) respectively. Clutter levels are evaluated by taking the exponential of the Logarithms in order to convert the estimators to the intensity domain. The latter is multiplied by a constant threshold factor in order to maintain a constant Probability of False Alarm (Pfa), the product is compared to the CUT as part of the decision rule. Expressions of the Logarithmic estimation are derived for each detector from clutter samples vectors. Detection schemes and expressions of decision rules are also presented. Detection performances of both detectors are analyzed by means of Monte Carlo simulations considering various parameters such as: reference window size, correlation coefficient and (Pfa).