A comparative study of FFT-summation and polyphase-FFT CFAR detectors

A priori knowledge of the signal channelization and bandwidth can be used to design efficient signal processing strategies for the detection of narrowband signals. Approaches based on digital filter banks are particularly attractive since a large number of channels can be searched in parallel. A simple and computationally efficient idea involves the use of an FFT that has been designed so that each FFT bin corresponds to a channel. The performance limitations of the FFT detector can be resolved by processing longer signal data records using the polyphase-FFT. An alternative idea for improving detection performance involves increasing the FFT length so that the signal power in each channel is obtained by summing the power computed for two or more FFT bins. This FFT-summation detector offers greater flexibility in the allowable channelization schemes and can provide performance characteristics similar to those of the polyphase-FFT detector.