Enhancing passive automation performance using an acoustic propagation simulation

During an at-sea encounter, signatures of interest can exhibit characteristics that differ from those observed in previously recorded data. These differences can occur due to variations in a number of factors including encounter geometry, propagation channel, and receiving sensor configuration. This paper presents a simulation technique that imposes low-frequency propagation effects on a time-domain signal using a normal mode method. High-quality, time-varying recorded signatures are used as inputs into the algorithm, which outputs band-limited time-series data for a selected geometry and environment. The output time-series are phased to simulate the time-varying pressure amplitudes that would be received by a towed array or any multielement passive sensor configuration operating in a realistic multipath environment. These capabilities enable the simulation of signatures of interest as captured under a broad range of littoral conditions by various passive sensors. These simulated data are used to augment scarce signature collections for training and assessing the performance of passive sonar automation.