Decision Support from Genetic Algorithms for Ship Collision Avoidance Route Planning and Alerts

Ming-Cheng Tsou, Sheng-Long Kao, Chien-Min Su

Journal of Navigation · 2009 · 133 citations · 15 references

Concepts

TL;DR

When an officer of the watch faces complicated marine traffic, a decision‑support tool can reduce burden and improve safety, while route choices can also balance economy. The study applies a genetic algorithm to compute a safety‑critical, economically optimal collision‑avoidance route and equips VTS operators and OOWs with advance alerts. The method simulates the genetic algorithm within a GIS platform, integrating COLREGS and ship safety constraints to generate route recommendations. The system outputs an optimal safe avoidance turning angle, navigation restoration time, and restoration angle.

Abstract

When an officer of the watch (OOW) faces complicated marine traffic, a suitable decision support tool could be employed in support of collision avoidance decisions, to reduce the burden and greatly improve the safety of marine traffic. Decisions on routes to avoid collisions could also consider economy as well as safety. Through simulating the biological evolution model, this research adopts the genetic algorithm used in artificial intelligence to find a theoretically safety-critical recommendation for the shortest route of collision avoidance from an economic viewpoint, combining the international regulations for preventing collisions at sea (COLREGS) and the safety domain of a ship. Based on this recommendation, an optimal safe avoidance turning angle, navigation restoration time and navigational restoration angle will also be provided. A Geographic Information System (GIS) will be used as the platform for display and operation. In order to achieve advance notice of alerts and due preparation for collision avoidance, a Vessel Traffic Services (VTS) operator and the OOW can use this system as a reference to assess collision avoidance at present location.

References

15