Sonar detection performance prediction of a moving source in an uncertain environment

Abstract

This paper addresses the problem of passive sonar detection of a moving acoustic source when both the signal spatial wavefront and background noise field are uncertain. The classic sonar equation is derived for a stationary source-receiver geometry and known environment. For dynamic scenarios, however, array gain (AG) and transmission loss (TL) can fluctuate in time and frequency over the integration time and bandwidth used. This is caused by signal multipath group and phase velocity differences at the receiver array. Moreover, for adaptive methods, noise covariance matrix snapshot support is often limited. This paper presents methods for estimating TL and AG under these conditions from a moving source of opportunity. These estimates are then used to predict adaptive detection performance. Results for probability of detection versus signal-to-noise ratio using real horizontal array data from the Swellex-96 experiment are shown to compare favorably with theoretically predicted detection performance. [Work supported by ONR.]