Cyclostationarity for ship detection using passive sonar: progress towards a detection and identification framework
David G. Hanson, Jérôme Antoni, Graham Brown, Ross Emslie · 2009
As the blades of a propeller pass through the water they produce characteristic amplitude modulated random noise signals which can be detected using sonar. It has recently been shown that the cyclostationary properties of this signal can be exploited to detect the presence of the propeller craft in significant extraneous noise. A detection technique based on the Cyclic Modulation Spectrum was shown to offer advantages over existing detection techniques in that no user interaction was required to design band pass filters, and superior frequency resolution was available to more accurately identify shaft and propeller pass frequencies. This technique has subsequently been developed to further exploit the cyclostationary properties of the signal by designing statistical thresholds which support automatic detection. This paper provides an overview of the progress of the cyclostationary detection work presented to date, and introduces a further development: exploiting cyclostationarity to determine the range, heading and speed of the surface ship. This concept is based on array processing using the cyclic autocorrelation function. The performance of this technique is demonstrated using simulation and the work is placed in the context of an overall detection and identification framework.