Short-term prediction of speed contour map and travel time by matching similar spatiotemporal traffic patterns

Abstract

Real time traffic condition prediction enables travelers to make smart decisions about departure time to avoid congestion on highway. In recent years, it has become possible to obtain a variety of observation data for traffic condition, so data-driven methods are continuously being developed. The Electronic Toll Collection System 2.0 (ETC2.0) is an example of a data collection system that gathers traffic conditions such as vehicle position, vehicle speed and time information in a wide area network, among others. While this ETC2.0 data can provide us useful information, it is relatively sparse. It is desirable to accurately predict traffic conditions from sparse real data. In this research, the short-term prediction of speed contour map and travel time are performed by the pattern matching of the spatiotemporal traffic state with the ETC2.0 data. Specifically, we use the ETC2.0 probe data and show that the method we used can predict speed contour map and travel time even with sparse data.