.... These strategies can be broadly classified into four categories: geometric method, optimization algorithm, virtual field method, and artificial intelligence method (Zhang et al., 2021). The Distance at Closest Point of Approach (DCPA) and the Time to Closest Point of Approach (DCPA) calculation methods are typical...

... ABSTRACT Based on the dynamics analysis of the deep-sea mining vehicle, this study proposes a new deep-sea mining full-coverage path planning method, which combines the bio-excitation network algorithm and the improved A* algorithm, and is verified by simulation experiments to have a superior...

... related to its sealing performance, this paper establishes the optimal design method of the two-channel connector based on Kriging method and NSGA-II algorithm (non-dominated sorting genetic algorithm), and establishes the sealing structure model of the two-channel connector combined with related products...

... genetic algorithm. Both of the solution exhibit reasonable trends along the input parameters. The results from the two different surrogate models are then compared and their differences are discussed. INTRODUCTION As the global demand for renewable energy grows, the development of offshore wind...

... mechanisms of corrugation geometric parameters on mechanical responses; a surrogate model is constructed to replace computationally expensive numerical simulations, enabling collaborative optimization of key parameters via multi-objective algorithms to identify Pareto optimal solutions balancing strength...

... ABSTRACT This study proposes an intelligent optimization method for submersible pump impellers and inducers, leveraging deep reinforcement learning (DRL), computational fluid dynamics (CFD), and the non-dominated sorting genetic algorithm (NSGA-II) for multi-objective optimization...

... ABSTRACT Collision avoidance is one of the most critical issues in ensuring a ship's safety. Ship maneuvering for collision avoidance first demands high real-time performance and rapid response from algorithms. In practical applications, arrival time constraints also need to be taken care...

... ABSTRACT In this study, we implemented a method for maritime collision avoidance using a DRL (Deep Reinforcement Learning) algorithm. The DRL-based method utilizes neural networks to recognize target ships (TSs) identified from AIS data and perform appropriate collision avoidance maneuvers...

... following control of intelligent ships based on real-time wave compensation was also proposed (Mu, 2024a). The elastic force contraction algorithm was used to solve the unmanned container ship path planning (Wang, 2024). Some novel technologies were utilized to make a more robust controlling. The wave...

... their chances of survival. In response to this challenge, this paper proposes a method for identifying man-overboard events based on attitude information. By employing deep learning classification algorithms, this method can timely detect man-overboard situations, ensuring the safety of individuals at sea...

... ABSTRACT With the increasing demand of ship energy saving, ship type optimization is becoming more and more important to ship industry. This paper proposes an intelligent optimization algorithm that optimizes the ship's bow and stern separately. The bow is optimized based on total resistance...

.... neural network optimization problem air emission liquified natural gas (lng) algorithm artificial intelligence consumption path optimization weather forecast duration optimization algorithm optimization path point path point density effectiveness draught machine learning meteorological...

.... (2020) proposed a real-time prediction algorithm for ship vertical acceleration based on a Gated Recurrent Unit (GRU) model within a recurrent neural network. Yuan et al. (2023) proposed a hybrid prediction model for coupling a semi-submersible platform's hull motion and mooring system. Through...

... is transformed into heading angle control problem by Line-Of-Sight (LOS) algorithm. Finally, the MPC algorithm is applied to make the ship follow the target path based on the ship response model, coupled with a Disturbance Observer (DO) compensating for errors caused by restricted-water effects. Simulation...

... safe navigation in mixed environments where conventional and autonomous vessels coexist. decision support system risk management bearing artificial intelligence th navigational scenario rule 17 rule 15 navigation simulation algorithm colreg classification rule 7 appropriate dss module...

...) in the PIV experiment covered the entire tank area, capturing particle motion induced by sloshing flow at 1000 FPS using a high-speed monochrome camera to minimize noise and enhance particle contrast. The U-Net algorithm was employed to segment the water region from images containing both water and air...

... measurement before assembly. An image-based measurement system utilizes stereo camera technology, which estimates distances by analyzing the disparity between images captured by different cameras. In this study, an algorithm was developed to reconstruct high precision 3D coordinates from 2D images of huge...

... and time-varying characteristics of equipment operation, resulting in insufficient energy consumption prediction accuracy. Therefore, this study combines the actual energy consumption conditions of the shipyard and uses the Lasso artificial intelligence algorithm to establish an accurate air compressor...

... algorithm dataset rmse depositional environment accuracy sway data ann deviation prediction lstm network effectiveness lstm model Proceedings of the Thirty-fifth (2025) International Ocean and Polar Engineering Conference Seoul/Goyang, Korea, June 1-6, 2025 Copyright © 2025...

... algorithm wang machine learning ksmote fault diagnosis fault diagnosis performance data imbalance ratio control system fault diagnosis model failure mode diagnosis ksfa subsea bop Proceedings of the Thirty-fifth (2025) International Ocean and Polar Engineering Conference Seoul/Goyang, Korea...