Incoherent sub-band averaging for improved target detection and Doppler estimation in linearly frequency modulated continuous active sonar

Abstract

Linear frequency modulated (LFM) continuous active sonar (CAS) waveforms show promise for use in target tracking given that waveforms can be split into sub-waveforms (sub-bands), thereby increasing the target refresh rate. However, reducing the duration and bandwidth of the sub-band decreases the SNR/SRR, adversely affecting detection. We present a target detection technique in which sub-bands are averaged incoherently while exploiting the range bias error (commonly observed in large duration LFM CAS waveforms with significant Doppler) to significantly improve detection. Sub-band averaging, known to reduce the false alarm rate, also mitigates channel coherence losses while maintaining detectability in CAS. A method for performing incoherent averaging over many possible Doppler shifts and identifying contacts via clustering in the 3-dimensional range-bearing-Doppler parameter space will be described. Finally, the promising results obtained with this technique on data acquired by the 2016 Littoral CAS Multi-National Joint Research Project sea trial will be shown.