A Probabilistic Expert System Approach for Sea Mine Burial Prediction

Knowledge of the extent of burial of bottom sitting sea mines is critical to mine detection due to the significantly degraded capabilities of mine-hunting systems when the mines are buried. To provide an enhanced capability for predicting mine burial in support of U.S. Navy mine countermeasure (MCM) operations, an expert system approach to predicting sea mine burial has been developed. This expert system serves as a means to synthesize previous and current research on sea mine burial due to impact upon deployment and subsequently due to scour, the two dominant burial mechanisms in littoral waters. Prediction systems for impact and scour burial have been implemented as simple Bayesian networks whose probabilistic basis provides means of accounting for the inherent uncertainties associated with mine deployment methods, simplified physics-based burial models, and environmental variability. Examples of burial predictions and comparisons to results from field experiments are illustrated. In addition, a proposed risk metric is developed and applied to provide a geospatial mapping of mine burial probability.