Oceansub-bottom sensing using randommatrixmodels

Cross-correlations of diffuse ocean noise provide travel time information between different sensors, which correspond to the various propagation paths in the environment. We use random matrix theory to model the diffuse component of the ocean noise field and distinguish statistically, the outliers which are due to anthropogenic sources (such as ships). This approach enables improving the “diffusivity” of the noise field by selective removal of outlier eigenvalues. We demonstrate its application in sensing the ocean sub-bottom structure using an L-shaped array (horizontal and vertical line arrays).