Narrowband interference rejection using small FFT block sizes

The performance of fast Fourier transform (FFT)-based excision processing (suppression of narrowband interference collocated in frequency with direct-sequence spread-spectrum signals) using block sizes which have time durations less than a bit time is examined. The FFT block size is chosen to provide frequency resolution consistent with the interference environment, and it is smaller than the number of samples taken during the data bit period. This technique is favorable in situations where the interference is sparse and the processing gains are large. Rectangular windowing with nonoverlapped processing and Hann and Hamming windowing with 50% overlapped processing are compared. Results for single continuous wave interference signals at jammer-to-signal ratios of 0 to 100 dB are presented. The performance for this method of processing is comparable to that for systems where the block size is coincident with the data bit duration, but the processing load has been reduced.<>