Region II University Transportation Research Center, NITTEC
Accurately predict border crossing delays.
Predictive Machine Learning Models and Smartphone App for Information Dissemination.
Machine Learning, Smart City, Transportation Management
Download PDFRead More
With over two million passenger vehicles crossing the Peace Bridge between the United States and Canada every year, there’s a large and growing need to more accurately predict delays caused by traffic patterns, security checks and tighter inspection procedures. Such delays cause an economic loss in excess of $4 billion annually for the US alone.
A large number of constantly changing variables—collected and applied to various models in real time—are necessary to accurately predict future waiting times based on current conditions. A two-step model first predicts the volume of traffic accumulating at a border crossing for a designated time period, then calculates predicted waiting times.
“Accurate border crossing waiting time predictions allow travelers to make better route choices and border authorities to manage checkpoints more efficiently.”
Complex data modeling based on real time inputs requires multiple state-of-the-art machine learning algorithms.
Data Science in Action
The two-step border crossing delay prediction model considers multiple factors like incoming traffic, staffing levels, the number of customer inspection booths open, and inspection times. Based on these variables, it can estimate future waiting times at the border more accurately, when compared with directly predicting the delays. In addition, an Android smartphone application is developed which enables travelers to view border crossing waiting times based on users’ preferences. The app also provides future waiting times and “next 15 minutes” updates frequently (every five minutes) and promotes user-reported and crowdsourcing data for means of informing travelers and border crossing authorities of forthcoming crossing delays.
The border crossing delay prediction studies minimize overall travel time, reduce negative environmental impacts, and enable border protection authorities to satisfy appropriate staffing levels and international security protocols.
The transportation industry is full of opportunities for data science applications. From traditional traffic management to emerging connected and automated vehicles, we are expecting to provide powerful data science tools to improve efficiency, safety, and sustainability of transportation systems.