A Path Planning Algorithm for Autonomous Ship Agents

The paper present a novel path planning algorithm for real-time navigation of autonomous ship agents in complex environments. The algorithm is a hybrid one of global and local methods, in which concept of convex hull is adopted for building a reverse roadmap as global information and with the guidance of reverse roadmap path is planned reactively in real-time with the motion of ship agents. The hybrid path planning algorithm not only can plan path for a single ship agent, but also can be integrated into a ship group behaviors model for marine traffic simulation. The simulation experiments of several challenging scenarios show that the algorithm is able to plan collision free path for autonomous ship agents.