An artificial potential field-based simulation approach for maritime traffic flow

The collision avoidance of intelligent vessels is an important issue to promote traffic fluency and safety in the waterway. This paper proposes an approach to simulate maritime traffic flow based on historical data and Agent-based models. In the first one, the combination of collision potential fields of different obstacles is described using the Artificial Potential Field method, i.e. piers, and buoys. The different static models are established according to the sharps and measures of the obstacles. Furthermore, the agent model is designed for simulating vessels' behaviors where simulated vessels always choose their routes based on the consideration of the trade-off between the collision potential distribution and operation convenience. The simulation result indicated that the collision avoidance action of vessels of waterway is described exactly using the Artificial Potential Field method indeed. Subsequently, by creating a large amount of Agent-based vessels with this method, arbitrary maritime traffic flow can be simulated in the software platform. Particularly, not only the simulated traffic flow might conform to the statistical characteristics of the real one, but also any individual simulated vessel should behave like a real sailing vessel. Eventually, the efficiency of the proposed approach is validated in the case in The Yangtze River in Wuhan section.