Collision Risk Prediction for Small Ships in South Korea Via Optimization of Wireless Communication Period

Abstract

Since the emergence of COVID-19, there has been a global surge in demand for marine leisure activities. In Korea, the population using marine leisure has risen approximately 192% to 20,406 people, compared to 6,966 people in the year 2000, indicating a continuous growth over the past two decades. Maritime transportation has become increasingly intricate worldwide due to the development of increasingly autonomous, larger, and faster ships. To effectively address potential hazards in such complex traffic environments, it is imperative to anticipate future scenarios and respond rapidly. However, small vessels account for the highest proportion of marine accidents, exhibit movements that exceed the communication period, complicating their behavior prediction. This study aims to identify the appropriate communication interval and prediction methodology for estimating the navigational risk associated with small ships. To achieve this, prediction data were generated for Korean fishing boats using point-based and motion-based prediction methods and communication periods. The accuracy of these predictions was assessed by employing the root mean square error metric and a maritime traffic risk model based on existing data. The findings demonstrate that the point based prediction method is more accurate in predicting the future risk of small ships by approximately three times compared to the motion-based prediction method. Among the communication intervals analyzed in this study, 5-s interval is recommended to ensure accurate navigational predictions. The significance of this study lies in its determination of the optimal prediction method and communication period for predicting the navigational risk of small ships, which has practical implications for enhancing maritime safety.