Development of a Road Condition Recovery Time Estimation System for Winter Snow Events

Abstract

This research develops a Normal Condition Regain Time (NCRT) estimation system, which automatically determines the NCRT at detector stations on the metro-freeway network for given snow events. The NCRT process is based on the findings that the speed level during the recovery process reaches a stable free-flow-speed (FFS), whose value is generally lower than the pre-snow FFS at a same location. Further, the speed-density (U-K) relationship of the traffic flow after snow is cleared exhibits a similar but shifted-down pattern of the normal-day U-K relationship at a given location. In this study, the after-snow traffic condition with a stable but shifted-sown pattern of the normal-day U-K relationship is defined as the ‘wet-normal’ condition, and the NCRT is defined as the time when the U-K data during a snow event starts to follow the wet-normal U-K pattern at a given station. The NCRT estimation system first collects the traffic and weather data for the metro-freeway network and determines the normal-day U-K relationships for the detector stations whose traffic data include both uncongested and congested regions. The normal-day U-K relationships are then applied to calibrate the wet-normal U-K patterns at given locations using the traffic data collected during snow events. Finally, the NCRTs are determined for each station by comparing the U-K data trajectory during a given event with the wet-normal U-K pattern at given locations. The NCRT estimation system has been applied to a set of the sample snow events.