Rapid model-based predictive path planning algorithm for autonomous ship in constrained maritime environment

Autonomous navigation in constrained maritime environments requires path planning algorithms that are both dynamically consistent and computationally efficient. This paper proposes a rapid model-based predictive path planning method for underactuated autonomous surface vessels (ASVs), combining the Maneuvering Modeling Group (MMG) dynamics with an optimization strategy. Unlike geometry-based planners or computationally expensive metaheuristic methods such as Genetic Algorithms (GA) and Particle Swarm Optimization (PSO), the proposed approach ensures collision-free and dynamically feasible trajectories with logarithmic search complexity. The algorithm is validated through comprehensive simulations using the 1/80-scaled Duisburg Test Case ship model. Benchmarking results demonstrate that the method achieves up to two orders of magnitude faster runtimes compared to GA and PSO, while maintaining comparable accuracy. Scalability tests in complex strait-like environments confirm robustness under dense obstacle constraints, and PID-based tracking experiments verify that planned trajectories are executable under realistic actuation limits. The results highlight the suitability of the proposed planner for real-time autonomous ship navigation, offering a regulation-compliant, computationally efficient, and dynamically robust solution.