Research on multi-UUV mission planning for mine countermeasures based on preferred multi-objective optimization

Abstract

Facing rising maritime security threats, this study presents T-MOEA/D, a sophisticated evolutionary algorithm enhancing UUVs’ capabilities in mine detection and neutralization. The algorithm tackles the multi-objective challenge by balancing time and energy, integrating user preferences into its optimization process. It leverages genetic operators such as dual chromosome encoding and partially mapped crossover to evolve efficient solutions, outperforming T-NSGA-II and T-NSGA-III in hypervolume and operational time. The UUV, directed by T-MOEA/D, navigates to operational areas and employs StyleGAN and YOLOv9 for accurate mine perception, crucial for executing mine countermeasure tasks. The system’s effectiveness is confirmed through Unity3D simulations and real-world tests, demonstrating its practicality and reliability. The study’s findings offer strategic guidance for planning large-scale mine countermeasure missions with multiple UUVs, ensuring operational efficiency and safety in complex underwater environments. The Pareto optimal solutions align with user preferences, reflecting a tailored approach to mine countermeasure missions.

Keywords

• UUV
• Anti-mine technology
• YOLOv9
• Multi agent collaboration
• Unity3D
• T-MOEA/D