A methodology has been developed for predicting delays caused by congestion. The methodology is designed to give drivers a prediction of delays ahead of them. The methodology uses an equation based on the upstream and downstream traffic flows derived from loop-based traffic monitoring sites. It is essentially a form of queuing theory that estimates how long it will take to empty the road ahead of vehicles in front of the vehicle receiving the prediction information. The methodology has been tested using the VEDENS traffic flow simulation1.The equation provides a good prediction of the delays that will be experienced by drivers. The prediction works well for all cases except where an unforeseeable future event occurs, such as a blockage occurring or being removed after adriver receives the first prediction(s). The predictions of the methodology have been compared with those based on the latest journey times available at the time of the prediction.Becausejourney times are based entirely on historical events (how long it previously took vehicles to travel from one monitoring point to the next the predictions are much less accurate than those made using the methodology based on traffic flows. There is scope for developing the methodology into a practical system of driver information. (Author/publisher).
Abstract