This research introduces a novel framework using Bayesian Neural Ordinary Differential Equations (ODEs) and Gaussian Process (GP) priors to accurately predict vessel trajectories and quantify uncertainty from irregular AIS data, addressing the challenges of missing reports and co
Accurate vessel trajectory prediction from Automatic Identification System (AIS) data is critical for maritime situational awareness. To solve the challenges posed by irregular sampling and incomplete reports, this work proposes utilizing Bayesian Neural ODEs (NODEs) to model continuous-time vessel dynamics while inherently providing uncertainty estimates.
Traditional methods often struggle to encode the structural properties (like smoothness) of vessel movement. The authors address this limitation by developing a practical approach that imposes a GP-kernel-based prior directly on the vector field of the Neural ODE, augmenting the standard variational objective. To handle long and irregular trajectories efficiently, the method is combined with probabilistic multiple shooting, which allows for consistent, segment-wise inference. This provides a robust method for learning complex vessel motion and quantifying the reliability of trajectory forecasts.