Adaptive ocean sampling for acoustic ocean uncertainty reduction

Abstract

For many oceanic environments, acoustic propagation is critically dependent upon the local environment through boundary interaction and volume variability and fluctuations. In shallow water, near the continental shelves, ocean dynamics and variability can lead to significantly different propagation regimes on timescales of hours and spatial scales of kilometers. Optimal adaptive sampling, with assimilation of oceanographic variables into a dynamic ocean model, can be used to reduce the model forecast uncertainty and improve the confidence in acoustic predictions. Oceanographic uncertainty and acoustic sensitivity, however, are not always well correlated. In this paper, an acoustic uncertainty cost function will be used to drive the adaptive sampling of an oceanographic AUV in coastal water. Applications to an experiment off the coast of Taiwan (quantifying, predicting, and exploiting uncertainty) will be performed, with model-data comparisons of acoustic propagation for moving sources and drifting receivers. During the test, acoustic adaptive sampling computations were performed in real time, oceanographic measurements were made, assimilated into a dynamic ocean model and acoustic transmissions were recorded. [Work Supported by ONR.]