Abstract
Detection and localization performance is analyzed for an underwater surveillance system utilizing low frequency passive acoustic sensors. Signal to noise ratio drives system performance, so knowledge of both signal and noise characteristics within a specific environment is required. This effort identifies the ambient noise level as a critical parameter in bounding system performance and therefore concentrates on modeling of surface noise propagation. Two juxtapositional environments were chosen. In the Florida Straits, a complex bathymetric and acoustical environment emphasizes distant noise sources. In the Adriatic sea, a relatively flat basin with little acoustic variability, a simpler performance model is justified.The ambient noise field is treated as that due to a distribution of uncorrelated, vertically oriented surface dipoles whose intensities are taken to be functions of local wind speed, rainfall rate and shipping density. A hybrid model of propagation from these dipoles is developed that incorporates local, high-angle paths using a Green's function solution as well as low-angle paths using a parabolic equation solution. This novel ambient noise model provides a unique capability with its three-dimensional noise propagation within a range dependent environment. Specific environmental factors are investigated to determine their impact on noise propagation. These factors are examined to determine under what circumstances more distant noise sources significantly impact local ambient noise levels.Using the acoustic model to predict signal and noise levels within a specific environment allows for estimation of detection and localization performance measures. Measures of effectiveness are developed and estimated for a hypothetical target moving through a particular configuration of sensors. Detection performance is presented in the form of probability of detection and detection range. Localization performance is presented in the form of bearing errors.