항만 도시 교통물류 안전 증진을 위한 실시간 기상 변화 및 항만 영향권 특성 별 사고 영향요인 분석에 관한 연구

초록

Port safety management should consider a variety of cargo shifting within trucks and containers, occurring at and near port areas. In particular, it is crucial for port safety management to consider not only incidents directly 'at-port' but also those in the surrounding 'near-port' areas, including the port influence area. This is significant because of the potential for high crash severity at near port areas, given the substantial truck traffic that could lead to large-scale crashes. Therefore, developing management strategies for port city safety requires identifying key risk factors that influence crash severity in each port area. During this process, because the key factors influencing crash severity may vary as one gets closer to the port center, it is essential to take into account the size of the port influence area. This study collected and matched both crash and weather data to consider various variables. Additionally, this study developed four machine learning-based crash severity models, including Naive Bayes Classification, Support Vector Machine, Extreme Gradient Boosting, and Light Gradient-Boosting Machine. Furthermore, the identification of key factors influencing high crash severity is determined through the application of an eXplainable Artificial Intelligence technique. It is expected that findings derived from this study can contribute to policy-making efforts aimed at enhancing traffic safety in the port area.

키워드