Estimating Freeway Level-of-Service Using Crowdsourced Data

In traffic operations, the aim of transportation agencies and researchers is typically to reduce congestion improve safety. To attain these goals, need continuous accurate information about situation. Level-of-Service (LOS) a beneficial index operations used monitor freeways. The Highway Capacity Manual (HCM) provides analytical methods assess LOS based on density highway characteristics. Generally, obtaining reliable data every road in large networks using traditional fixed location sensors cameras expensive otherwise unrealistic. Traditional intelligent system facilities are limited major urban areas different states. Crowdsourced an emerging, low-cost solution that can potentially safety operations. This study incorporates crowdsourced provided by Waze propose algorithm for assessment hourly basis. proposed exploits various features from big (crowdsourced user alerts speed/travel time variation) perform classification machine learning models. Three categories model inputs introduced: Basic statistical measures speed; travel reliability measures; number alerts. Data collected were calculate ground truth LOS. results reveal estimation accuracy 10% (accuracy = 0.93, Kappa 0.83). method was also tested confirmed after coronavirus disease 2019 (COVID-19) with severe breakdown due stay-at-home policy. extendible freeways other locations. this research provide freeway segments, regardless availability sensors.