An Underwater Threat Detection System Employing Low-Cost Sensing and Advanced Algorithms

A vast array of high-value installations are situated within or near waterways. With such a large number of vulnerable targets, current threat detection schemes are too costly to provide sufficient levels of protection. At the same time, with 95% of material goods arriving via ports and waterways, any new system must offer sufficient performance to prove effective in such an active environment. A new paradigm in identification and localization of targets of interest is the use of large numbers of low-cost distributed sensors rather than single high-fidelity sensors. This proposal seeks to provide an autonomous system that mates arrays of underwater transducers to novel fusion, tracking, and target classification algorithms. Our sensing approach is based on the concept of active forward scattering. Sets of measurements from the sensor array are passed to a central control center, which can be located on-or off-site and uses COTS computer hardware. The particle filter will be used for fusion and tracking due to the ease with which it can manage nonlinear measurements and negative information. A Bayesian network will be used for target classification because the algorithm can incorporate an array of target attributes including signature, velocity, bearing, maneuverability, etc.