Performance assessment of CFAR processors in Pearson-distributed clutter

An analytical study on the performance of three existing classes of constant false alarm rate (CFAR) detectors operating in heavy tailed distributed data is presented. In particular, we study the performance of the cell averaging, order statistics, and p-percent truncated mean CFAR processors when the output measurements of the square-law detector can be modeled as positive alpha-stable (PαS) random variables with a shape parameter (characteristic exponent) equal to 0.5. We derive the exact expressions for the detection and false alarm probabilities of the detectors, and compare their performance by means of their corresponding receiver operating characteristics.