A NEW APPROACH FOR COHERENT RADAR DETECTION IN PSEUDO- GAUSSIAN INTERFERENCE

Abstract

In this paper a new algorithm is presented for coherent radar detection of targets, when distribution of the interference (clutter and noise) is non-Gaussian. The Neyman-Pearson criterion is used for optimal detection and several successive received samples from a radar range cell are used for detection in the same cell. On the basis of recent empirical evidences, the interference distribution is described statistically by the correlated pseudo-Gaussian distribution which is also called SIRP. The joint pdf of the pseudo-Gaussian interference distribution is too complicated to lead to the optimum detector (or ALR detector). Therefore, a convenient approximation of its analytical solution is utilized. The obtained AALR detector outperforms the GLR detector. Moreover, the proposed algorithm is similar to the GLR detector. Since, the AALR detector-has been derived from the analytical solution of the ALR detector, its similarity to the GLR algorithm confirms the proper performance of the GLR detector. The performance of the proposed detector is also compared to the OLD and ECD detectors. Computer simulations confirm superiority of the ALR detector to the GLR detector while both are significantly better than the OLD and ECD detectors. The proposed detector completely prefers to the other detectors when the deviation of the interference distribution from Gaussian assumption is considerably high.