Autonomy and Route Optimization Lead AI Research Boom
The first applications of AI in ocean science and maritime engineering date back to the 2000s, and a paper recently published in Ocean Engineering tracks the subsequent growth in the number of scientific studies charting its advance.
Since 2018, the number of studies using AI techniques has been doubling each year, reaching 1,329 in 2022. This growth corresponds with developments in autonomous vessel and vehicle navigation, route optimization and wave climate characterization.
The authors explain that AI spans a wide range of technological developments and can be broadly categorized as:
• Artificial Narrow Intelligence (ANI) intelligent systems that can perform specific tasks,
• Artificial General Intelligence (AGI) technologies equivalent to human cognitive capabilities, and
• Artificial Super Intelligence (ASI) technologies surpassing human performance.
“As AI systems evolve towards AGI capabilities, the maritime sector may benefit from intelligent agents that support high-level strategic decision- making. These systems could offer context-aware insights, adapt to unstructured problems and preserve ethical standards, safety protocols and human oversight in increasingly autonomous marine operations.”
Meanwhile, Artificial Narrow Intelligence (ANI) predominates. It includes Optimization Algorithms (OA), Fuzzy Logic (FL) and Machine Learning (ML).
Optimization Algorithms optimize processes, models and trajectories and minimize errors. In ocean and maritime engineering, the most common optimization algorithms are Genetic Algorithm (GA), Genetic Programming (GP), Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO).
Fuzzy Logic is a reasoning method that handles imprecise and uncertain information by allowing variables to take values between completely true and completely false.
Machine Learning can be classified into Supervised Learning (SL), Unsupervised Learning (UL) and Reinforcement Learning (RL).
Supervised Learning is the predominant form of ML in ocean and maritime engineering. Key methodologies include Artificial Neural Networks (ANNs) that emulate the structure of the brain, Support Vector Machines (SVM) and Support Vector Regression (SVR).
Unsupervised Learning is less common, with the most popular technique being K-Means Clustering which clusters data without prior knowledge of input labels.
With Reinforcement Learning, an agent learns optimal behavior through trial-and-error interactions with a dynamic environment. It is used for solving problems involving sequential decision-making in changing or real-time environments.
Looking ahead, the authors see the growing importance of Explainable AI (XAI) and digital twins for transparent and trustworthy deployment in safety-critical systems.
Advancing artificial intelligence in ocean and maritime engineering: Trends, progress, and future directions was published in Ocean Engineering 339 (2025).