.png){kind=icon}
.png){kind=icon}
Return
Energy-efficient Navigation in Unknown Static Flow Fields using Reinforcement Learning
.png){kind=icon}
PRE
AI
DOI：10.1109/TLA.2026.11435595.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Energy efficiency is becoming increasingly critical in the Maritime Industry due to expectations of cost reduction, environmental sustainability, regulatory compliance, trade volume growth, and safety enhancement. In this context, Reinforcement Learning (RL) methods have gained significant attention for solving control and optimization problems. Recent advancements have demonstrated effectiveness in diverse domains, including logistics, scheduling, and resource allocation. Hence, in this paper, we address the problem of energy-efficient navigation of vessels in unknown static flow fields using RL. We propose a reward function that minimizes control effort under flow influence and evaluate two tabular and two function-approximation methods across different scenarios with the state-of-the-art literature. Finally, we provide a zero-shot sim-to-sim analysis to evaluate the impact of environmental uncertainty on our proposed method.
Keywords:
Navigation
Energy efficiency
Oceans
Vehicle dynamics
Q-learning
Aerospace electronics
Uncertainty
Safety
Robustness
Planning
Autonomous Agents
Reinforcement Learning
Intelligent Transportation
Journal
IF:
1.3
Papers:
164
Citations:
1.8K