Predicting the movement of vessels across the oceans remains one of the most challenging problems in transportation science. Ships do not move freely across a blank canvas; currents, winds, weather patterns, and the rules of physics shape their paths. Yet, many existing forecasting models still treat mobility as a sequence of past positions, ignoring surrounding forces.
I am excited to share that my project, “Incorporating Contextual and Physical Principles for Enhancing Short- and Long-Term Mobility Forecasting,” has been awarded funding through a Discovery Grant from NSERC. Over the next five years, this work at Dalhousie University’s MAPS Lab will focus on fundamentally rethinking how we predict the motion of vessels in dynamic, ever-changing marine environments.
The project will combine scientific knowledge of physical oceanography with machine learning, aiming to create predictive models that respect the realities of marine mobility. Instead of relying solely on historical data patterns, our models will embed environmental context and physical constraints directly into the forecasting process. This means forecasting where a vessel might go and how its route will be shaped by currents, winds, and encounters with other moving entities sharing the ocean space.
We hope to deliver more reliable, interpretable, and trustworthy trajectory predictions by bridging machine learning with the laws of motion and environmental forces. These tools will support maritime safety, environmental conservation, ocean management, and national security.
This grant represents a significant step toward the long-term vision of developing mobility forecasting models that are not only statistically powerful but also physically meaningful. I am deeply grateful for this opportunity and look forward to sharing the advances and insights from this new journey.
Stay tuned for updates!
