MPC Researchers Exploit Big Data and Machine Learning to Improve Traffic Management
Posted: Mar 25, 2021
The Utah Department of Transportation deploys more than 100 continuous count stations to collect traffic volumes at different locations in order to achieve balanced geographic coverage. Even so, it is still difficult to capture traffic characteristics of the entire Utah road network within a short time period to predict traffic disruptions from variables such as traffic incidents, work zones, adverse weather, and others. Researchers at the University of Utah found that machine learning techniques can fill this gap by analyzing existing traffic data. The technique may be employed to partially or fully supplant expensive and labor-intensive short-duration traffic count programs for predicting traffic volume and reliability changes. The technique will be useful in transportation operation analysis, congestion management, and accident prevention and other traffic management needs.
The Utah Department of Transportation deploys more than 100 continuous count stations to collect traffic volumes at different locations in order to achieve balanced geographic coverage. Even so, it is still difficult to capture traffic characteristics of the entire Utah road network within a short time period to predict traffic disruptions from variables such as traffic incidents, work zones, adverse weather, and others. Researchers at the University of Utah found that machine learning techniques can fill this gap by analyzing existing traffic data. The technique may be employed to partially or fully supplant expensive and labor-intensive short-duration traffic count programs for predicting traffic volume and reliability changes. The technique will be useful in transportation operation analysis, congestion management, and accident prevention and other traffic management needs.