Comparison of sampling techniques for automatic detection of pulse signals with unknown time of arrival

Automatic detection of pulse signals with unknown arrival times requires sampling of signals and noise during the time interval that signals may occur. Three different sampling techniques were evaluated to determine noise false alarm rates and detection probability as a function of signal to noise ratio. Fixed interval sampling was found to be easiest to implement but had the lowest detection sensitivity. Integration during the range bin provided intermediate detection performance while selection of the maximum value during the range bin gave the best signal detectability indicating that peak selection is the preferred technique for automatic radar and sonar detectors.