• Journal of Korean Society of Transportation
• /
• v.24 no.5 s.91
• /
• pp.7-18
• /
• 2006

Data for travel information Provision are regularly aggregated to Provide travel time information in a reliable and convenient manner and to manage traffic data and information efficiently. In most of practices in Korea, the GPS based travel time data are mainly aggregated every 5 minutes As a result, some probes can't pass by a link within aggregation interval and thereby create uncompleted link passing data. But these data are mostly generated during the congested times and therefore a method that uses such uncompleted link passing data are required. This study estimated queue dissipation length, green time and cycle that use GPS spot speed and developed a link travel time estimation method using such uncompleted link passing data. It also presents method and the overall process of using such data to estimate link travel time in a more accurate manner. As a result, MAPE 1.98% and MAE 4.75 sec of link travel time accuracy improvement has been reported, which is not much different from the real link travel time. The method Proposed here would be an alternative to increase the amount of GPS probe data, especially in congested urban arterial case.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.2
• /
• pp.39-57
• /
• 2018

The purpose of this study was to construct an artificial intelligence model that learns and estimates the relationship between vehicle queue length and link travel time in urban areas. The vehicle queue length estimation model is modeled by three models. First of all, classify whether vehicle queue is a link overflow and estimate the vehicle queue length in the link overflow and non-overflow situations. Deep learning model is implemented as Tensorflow. All models are based DNN structure, and network structure which shows minimum error after learning and testing is selected by diversifying hidden layer and node number. The accuracy of the vehicle queue link overflow classification model was 98%, and the error of the vehicle queue estimation model in case of non-overflow and overflow situation was less than 15% and less than 5%, respectively. The average error per link was about 12%. Compared with the detecting data-based method, the error was reduced by about 39%.