Abstract
The estimation of ocean dynamics is a key challenge for applications ranging from climate modeling to ship routing. State-of-the-art methods relying on satellite-derived altimetry data can hardly resolve spatial scales below ∼100 km. In this work we investigate the relevance of AIS data streams as a new mean for the estimation of the surface current velocities. Using a physics-informed observation model, we propose to solve the associated the ill-posed inverse problem using a trainable variational formulation. The latter exploits variational auto-encoders coupled with neural ODE to represent sea surface dynamics. We report numerical experiments on a real AIS dataset off South Africa in a highly dynamical ocean region. They support the relevance of the proposed learning-based AIS-driven approach to significantly improve the reconstruction of sea surface currents compared with state-of-the-art methods, including altimetry-based ones.
Highlights
Currents from AutomaticIdentification System (AIS) Maritime TrafficOver the last decades, in a context of globalization, we have observed an exponential growth of maritime traffic
Drawing inspiration from a 4D-Var data assimilation formulation [7], we cast the estimation of sea surface current from AIS data streams as a minimization problem that involves a physics-informed observation term coupled with a trainable ordinary differential equation (ODE) prior
We investigate the use of AIS maritime data and deep learning methods for the reconstruction of sea surface current
Summary
In a context of globalization, we have observed an exponential growth of maritime traffic. Drawing inspiration from a 4D-Var data assimilation formulation [7], we cast the estimation of sea surface current from AIS data streams as a minimization problem that involves a physics-informed observation term coupled with a trainable ordinary differential equation (ODE) prior. This formulation relates to variational auto-encoders (VAEs) and exploits external data to regularize the considered ill-posed inverse problem. The proposed learning scheme applies directly to AIS data streams with no requirement for a groundtruthed dataset for sea surface currents As such, it is regarded as a non-supervised approach.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.