Numerically Efficient Digital Matched Filter for Periodic and Windowed Periodic Radar Waveforms

摘要

An algorithm is presented for implementing a digital matched filter for periodic and windowed periodic radar waveforms. The algorithm is based on circular convolutional techniques. For moderate to high duty cycles d (radar on time to pulse repetition interval (PRI) ratio), the circular convolutional algorithm uses 50% fewer complex multiply operations (CMOPs) than does a fast convolution implementation. In fact, if BPRI is the radar waveform bandwidth-PRI product, then the circular convolutional technique is numerically more efficient if d > 1/sq. rt(8(BPRI)). Furthermore, the circular convolution technique is numerically more efficient when the pulse compression ratio p > sq. rt. (BPRI/8).