Energy saving by coherent sequential detection of radar signals with unknown Doppler shift

Abstract

A coherent sequential test is discussed. Instead of forming a bank of filters, a combination of the autocorrelation estimates (ACE) of the received signal is used to form the ACE test function. These estimates are computed by a running accumulation of signal products and therefore the test function is updated with each new signal. The proposed coherent sequential test is compared with the incoherent one by a simulation study. The gain is expressed by the saving in signal-to-noise ratio. The gain increases with the mean test length. The performance of the ACE sequential test was also compared with the fixed-sample-size filter-bank test: the gain is about 4.8 dB. This compares well with gain values by sequential detection in the incoherent case. The test function is also developed for target detection with multiple-range elements. The Doppler frequency of the target echoes may vary during the dwell time, e.g. tangential flying targets show an approximately linear frequency variation. An advantage of the ACE test function is a robust behaviour in the presence of frequency variations. It therefore allows long target dwell times, assisting energy management in the detection of targets with low RCS.