A Study on Travel Time Estimation in Diverging Road Sections of Uninterrupted Traffic Flow based on Timestamp Data Analysis

Various methods using GPS or vehicle detectors have been developed in estimating the travel time of individual road sections, but several problems still exist in terms of estimation accuracy. In particular, at the upstream of diverging road sections such as highway junctions and urban road intersections, traffic flow starts to diverge according to the traveling direction of vehicles. In such a case, the diverged flow behaviors can differ from each other, and differences in travel time towards each direction can be observed even in the same road section. Accordingly, it is necessary to estimate the travel time of a diverging road section by the traveling directions. For this purpose, this study provides a method for estimating the travel time of different traveling direction at the upstream of diverging highway sections using the GPS-based time stamp data. Three sequential steps with a few statistical approaches are provided in this stage: divergence detection in data, classification of diverged data, outlier filtering. In ‘data divergence detection’, a new statistical analysis method that can detect the occurrence of data divergence is provided, and it is analyzed that the method is useful in finding the threshold of determining data divergence. In ‘classification of diverged data’, a statistical approach is presented to classify the data by travel directions, and it is found that our modified method shows superior performance compared to others. In ‘outlier filtering’, a simple moving average is used to remove the data showing abnormal behaviors, but it is analyzed that this approach requires further improvement. The performance of the proposed method is tested using a microscopic simulation program. Through the tests, it is shown that the proposed method has reasonable performance in estimating the travel in the same road section by the travel directions.