Optimal detectors for nonfluctuating targets under spiky K-distributed clutter in radar applications

The optimum parametric detector structure for nonfluctuating target models in spiky K-distributed clutter is proposed and analyzed by computer simulations. The performance of this new detector (the clipping detector) is considerably better than the performance of the linear and quadratic detectors. Monte Carlo simulations have been used for the estimation of false alarm and detection probabilities, nevertheless importance sampling technique has been used for the estimation of very low false alarm probabilities (e.g., less than 10^-5). Also, a detailed discussion of the simulation results will be provided and some conclusions compiled at the end of the paper. Furthermore, results for Swerling II target model are presented.