Analyses of the trimmed-mean switching CFAR detector in non-homogenous gaussian clutter

The Switching Constant False Alarm Rate (S-CFAR) processor takes into account the statistics of the sample in the test cell for reference sample selection. This processor, which is based on two Cell-Averaging (CA-) CFAR detectors, is robust in homogenous and non-homogenous Gaussian clutter. However, it suffers from a considerable CFAR loss, when the number of interferences exceeds certain value. In the other hand, it is known that the Trimmed-Mean (TM-) CFAR detector improves the detection performances in non-homogeneous clutter because it allows the censer of the unwanted cells from both lower and upper ends after ordering them in ascending order. To overcome the previous problem, in this paper, we propose an improved version of the S-CFAR processor based on both S-CFAR and TM-CFAR detectors, namely, the Trimmed-Mean Switching (TMS-) CFAR processor. The proposed detector shows the better performance in homogenous and non homogenous environment especially in the case of both the clutter edge case and the interfering targets exist.