Data-Driven Modeling and Control for Maritime Robotics

Since much of the surface of the Earth is covered in water, maritime activities play a significant role in global economies, providing avenues for trade, food, and energy acquisition. The utilization of autonomous and teleoperated systems is vital for global maritime efforts since they can be designed to perform specialized tasks in environments for which humans are unsuited. To enable autonomous operation and facilitate underwater teleoperation, accurate models are required to design localization and control systems that enable the system to perform the desired tasks. However, acquiring accurate models of systems in underwater environments is challenging using only first-principles physics-based models due to the complexity of modelling hydrodynamic forces acting on the system. For this reason, dynamic models for underwater vehicles are often simplified by assuming the vehicle only moves at low speed and there is information regarding the objects to be manipulated. The simplifications made for the dynamic model also place design constraints on the controller. Utilizing data-driven modelling methods for underwater systems can generate accurate models that do not rely on these simplifications, allowing more complex controllers to be designed to increase capabilities of the system.

This workshop will bring together experts in the fields of data-driven model and control and maritime systems and applications. The workshop will take the format of a collection of invited talks, a lightning presentation and poster session, and panel discussion. We are accepting extended abstract submissions for those interested in presenting in the lightning talks and poster sessions. Submissions on modeling and control methods not applied to maritime systems but are transferrable to maritime systems will also be considered.

The workshop will take place at the CASE 2025 conference in Los Angeles, California, USA on August 21, 2025.