arXiv:2510.03534v5 Announce Type: replace-cross 
Abstract: We study the problem of long-term (multiple days) mapping of a river plume using multiple autonomous underwater vehicles (AUVs), focusing on the Douro river representative use-case. We propose an energy - and communication - efficient multi-agent reinforcement learning approach in which a central coordinator intermittently communicates with the AUVs, collecting measurements and issuing commands. Our approach integrates spatiotemporal Gaussian process regression (GPR) with a multi-head Q-network controller that regulates direction and speed for each AUV. Simulations using the Delft3D ocean model demonstrate that our method consistently outperforms both single- and multi-agent benchmarks, with scaling the number of agents both improving mean squared error (MSE) and operational endurance. In some instances, our algorithm demonstrates that doubling the number of AUVs can more than double endurance while maintaining or improving accuracy, underscoring the benefits of multi-agent coordination. Our learned policies generalize across unseen seasonal regimes over different months and years, demonstrating promise for future developments of data-driven long-term monitoring of dynamic plume environments.

تم إجراء دراسة حول رسم خرائط طويلة الأمد لتيار نهر دورو باستخدام عدة مركبات تحت الماء ذاتية القيادة (AUV). تقدم الدراسة نهجًا فعالًا من حيث الطاقة للتعلم المعزز متعدد الوكلاء، حيث يتواصل منسق مركزي بشكل متقطع مع AUVs لجمع البيانات وإصدار الأوامر، مما يظهر أداءً متفوقًا في المحاكاة مقارنةً بالمعايير الحالية.

Se ha llevado a cabo un estudio sobre la cartografía a largo plazo del afluente del río Duero utilizando múltiples vehículos submarinos autónomos (AUV). La investigación introduce un enfoque de aprendizaje por refuerzo multiagente eficiente en energía, donde un coordinador central se comunica intermitentemente con los AUV para recopilar datos y emitir comandos, demostrando un rendimiento superior en simulaciones en comparación con los puntos de referencia existentes.

Une étude a été réalisée sur la cartographie à long terme du panache du fleuve Douro en utilisant plusieurs véhicules sous-marins autonomes (AUV). La recherche propose une approche d'apprentissage par renforcement multi-agents économe en énergie, où un coordinateur central communique de manière intermittente avec les AUV pour collecter des données et émettre des commandes, montrant une performance supérieure dans les simulations par rapport aux références existantes.

A study has been conducted on the long-term mapping of the Douro River plume utilizing multiple autonomous underwater vehicles (AUVs). The research introduces an energy-efficient multi-agent reinforcement learning approach, where a central coordinator communicates intermittently with the AUVs to collect data and issue commands, demonstrating superior performance in simulations compared to existing benchmarks.

Long-Term Mapping of the Douro River Plume with Multi-Agent Reinforcement Learning

One More Thing in AI – Your Shortcut to AI Mastery

Long-Term Mapping of the Douro River Plume with Multi-Agent Reinforcement Learning

Was this article worth reading? Share it

One More Thing in AI

LucidQuery AI

AIvilization

Attentive AI

Metaflow AI

VUME

Ready to build your own newsroom?