A Representative Modeling Approach to Sampling Dynamic Ocean Structures

The primary goal to this research is to develop a principled approach to sampling in the context of a spatiotemporally dynamic feature. We examine the concepts of registration and association of sparse data in non-static environments with the aim of developing generative models and low-order descriptions to improve localization relative to the dynamic feature, which in turn will improve methods for adaptive sampling. The motivation for this work is in the study of coherent structures within the ocean, e.g., ocean fronts and Lagrangian Coherent Structures (LCS), and how intelligent sampling can help provide an understanding of these complex phenomena. We additionally examine the way that dynamic ocean phenomenon not only provide the features of interest to sample, but also provide the information to generate an environmental representation that drives intelligent adaptation for sampling. The proposed work lies at the intersection of robotics, AI, nonlinear dynamics, perception, network sciences, and ocean science.