An ACO-MPC Framework for Energy-Efficient and Collision-Free Path Planning in Autonomous Maritime Navigation
Yaoze Liu, Zhen Tian, Qifan Zhou, Zixuan Huang, Hongyu Sun · 2025
This work presents a novel approach to optimizing energy dispatch in autonomous maritime systems by integrating metaheuristic search with model predictive control. Our proposed ACO-MPC framework combines Ant Colony Optimization (ACO with Model Predictive Control (MPC) to dynamically generate energy-efficient paths in a simulated sea surface environment, where renewable generation is influenced by wind speed and polar effects. A linear model, derived from real-world data, is embedded within the MPC formulation to accurately predict energy consumption, thereby enabling real-time optimization of renewable utilization, battery cycling, and backup power usage. Simulation results demonstrate that the ACO-MPC approach significantly outperforms conventional rule-based strategies and standard MPC methods, achieving both collision-free navigation and the lowest cumulative energy during the navigation to target points.