Distribution Selection Using Integral Goodness of Fit Test in Radar Sea Clutter Simulation for CFAR Detection

Selecting the most suitable statistical distribution for sea clutter simulation plays an important role in constant false alarm rate (CFAR) radar target detection. However, conventional distance-based goodness of fit (GoF) tests are unable to quantitatively evaluate the deviation of detection thresholds between the specified distribution and real-world sea clutter data. To address this, this paper proposes an integral goodness of fit (Gof)-based distribution selection technique. The IGoF test integrates the difference between complementary cumulative distribution functions (CCDFs) of different distributions and real-world data over the range that exceeds the detection threshold corresponding to a specified false alarm rate (FAR). Experimental results on measured sea clutter data demonstrate the usefulness of the distribution selection technique.