Performance of the Developed Versions of CFAR Schemes Processing Non-Coherently Integrated M-Pulses in the Presence of Outliers

The CFAR processors have been utilized in radar systems where the clutter environment is partially unknown and/or has varying statistical properties. In such situations, the performance of the fixed-threshold detector deteriorates significantly, and the CFAR detector is designed to be invariant to changes in the clutter intensity. The CA and OS procedures are the most popular ones in the CFAR world. The CA is the optimum processor in terms of detection probability in homogeneous background while the OS algorithm has its immunity against outlying targets whenever their number is within an allowable range. To benefice the merits of these two algorithms, new versions, which combine the basic of these techniques in estimating the unknown noise power level, have been recently proposed. These versions have improved detection performance either they operate in homogeneous or in non-homogeneous background environments.This paper is devoted to the detection analysis of these versions when the radar receiver contains a noncoherent integrator amongst its fundamental elements. Exact formulas for their false alarm and detection probabilities are derived, in the absence as well as in the presence of interferers. The primary as well as the secondary outlying targets is assumed to be fluctuating in accordance with SWII model. Our obtained results indicate that while the SO version performs nearly as good as the conventional OS scheme in the presence of outlying targets, all the developed versions perform much better than OS in homogeneous situations. In comparison with the CA detector, the modified versions perform better in an ideal operating environment, and behave much better when the background environment is contaminated with a number of spurious targets.