The Development of Regional Vessel Traffic Congestion Forecasts Using Hybrid Data from an Automatic Identification System and a Port Management Information System

Abstract

The present study proposes a new method that forecasts congestion in the area near a port by combining the automatic identification systems of ships and port management information data. The proposed method achieves 85% accuracy for one-day-long ship congestion forecasts. This accuracy level is high enough to act as a reference value for both manned and unmanned operation situations for autonomous vessels in port areas. The proposed forecast algorithm achieves 95% accuracy when used for a one-hour ship congestion forecast. However, the accuracy of the algorithm is degraded to almost half when the automatic identification system or the port management system is used independently.