초록

본 연구의 목표는 미래에 발생할 이벤트가 교통흐름에 미치는 영향을 예측하여, 지자체가 보유하고 있는 교통 정보 제공 시스템에 교통상황 예측 정보를 제공할 수 있도록 예측 교통정보를 산출하고 제공할 수 있는 기술을 개발하는 것이다.
미래에 발생할 이벤트(교통수요변화, 기상변화, 공사, 행사, 사고, 돌발 상황)가 교통 흐름에 미치는 영향을 사전에 예측하여 도로운영자가 선제적 교통제어전략(교통정보제공 등)을 수립할수 있도록 미래 교통상황에 대한 예측정보를 산출하고 제공하는 교통상황예측 정보 제공 시스템을 개발하였다.

Abstract ▼

A predicted traffic information provision system is developed considering the future events such as construction, accidents, abrupt change in traffic demands, adverse weather. The traffic prediction algorithm is divided into long-term prediction and short-term prediction. The long-term prediction is

A predicted traffic information provision system is developed considering the future events such as construction, accidents, abrupt change in traffic demands, adverse weather. The traffic prediction algorithm is divided into long-term prediction and short-term prediction. The long-term prediction is made based on reproducible travel patterns using travel speeds collected on days with similar attributes. Meanwhile, the short-term prediction is included to reflect the variation of real-time traffic speeds by matching the currently bserved traffic speed with the most similar historical data using K-nearest neighbor algorithm. The long-term prediction and long-term prediction is combined using a weight function. The weather information like forecast information and precipitation data is used for the traffic prediction for the rainy weather. The road construction data like construction period and the number of lanes closed are used to consider the impact by road construction on the future traffic conditions. The performance of the algorithm is evaluated using real data collected from the urban expressway network in Seoul Metropolitan City. The evaluation shows promising results for the applicability of this method in practice.