A DNN Framework for Learning Lagrangian Drift With Uncertainty
This work addresses uncertainty in ocean drift predictions for search and rescue operations, but it is incremental as it builds on existing data-driven methods with limited generalization to unseen conditions.
The paper tackles the problem of uncertain Lagrangian drift reconstructions, such as for objects lost at sea, by introducing a data-driven framework that models probabilistic drift without specific assumptions, achieving good agreement with numerical simulations in controlled scenarios but showing limitations with imperfect inputs.
Reconstructions of Lagrangian drift, for example for objects lost at sea, are often uncertain due to unresolved physical phenomena within the data. Uncertainty is usually overcome by introducing stochasticity into the drift, but this approach requires specific assumptions for modelling uncertainty. We remove this constraint by presenting a purely data-driven framework for modelling probabilistic drift in flexible environments. Using ocean circulation model simulations, we generate probabilistic trajectories of object location by simulating uncertainty in the initial object position. We train an emulator of probabilistic drift over one day given perfectly known velocities and observe good agreement with numerical simulations. Several loss functions are tested. Then, we strain our framework by training models where the input information is imperfect. On these harder scenarios, we observe reasonable predictions although the effects of data drift become noticeable when evaluating the models against unseen flow scenarios.