The FFT majority and median CFAR detectors have been proposed for applications where impulsive noise severely degrades the performance of the FFT summation CFAR detector. Although based on different principles, these two FFT CFAR detectors are actually identical if the number of data blocks, L, is odd. When the number of data blocks L is even, these two detectors are no longer identical and hence require different treatment. This paper provides a comparative study of the FFT majority and median CFA R detectors for an additive white Gaussian noise channel assuming an even number of data blocks. The main results of this paper complement other recent papers and technical reports on FFT CFAR detectors and are useful for selecting the appropriate FFT CFA R detector to meet application requirements.

Additional Metadata
Keywords CFAR detector, Constant false alarm rate detector, FFT filter bank, Majority CFAR detector, Median CFAR detector, Probability of detection
Persistent URL dx.doi.org/10.1109/CCECE.2006.277338
Conference 2006 Canadian Conference on Electrical and Computer Engineering, CCECE'06
Citation
Wang, S. (Sichun), Inkol, R. (Robert), & Rajan, S. (2007). Performance comparison of the FFT filter bank-based majority and median CFAR detectors. In Canadian Conference on Electrical and Computer Engineering (pp. 36–40). doi:10.1109/CCECE.2006.277338