Quality Evaluation Method for Variable Speed Limit Systems

Abstract

Advanced traffic management has used variable speed limit (VSL) systems to control traffic on freeways and urban arterial roads. A VSL system is expected to detect incidents and take measures to carry out any or all of the following: avoid accidents, postpone or alleviate congestion by providing warning messages, and harmonize traffic by reducing speed differences between vehicles. Methods for assessing two features of VSL systems, namely, incident detection and warning efficiency, are presented in this paper. The algorithm for warnings relies on incident detection. Therefore, the quality of the system's incident detection and warning effects need to be assessed separately. For the purpose of evaluating incident detection, the traffic state was reconstructed, discretized, and compared with the message signs of the VSL. The warning efficiency of the system was assessed by generating virtual vehicle trajectories and determining whether a driver experienced adequate warning. Scenarios for message signs and their influence on traffic safety were identified according to the distance of the warning from the tail of the congestion. Weightings for detections and missed detections based on the different scenarios were developed. A quality evaluation method was then used to grade the levels of the system in their incident detection and warning efficiency. The presented method was tested for the VSL system on Autobahn A99 near Munich, Germany.