AI Tackles Traffic Congestion, a Global Issue
As cities across the globe continue to grow, so do the number of vehicles on the roads. Traffic congestion has become one of the world’s most widespread transport challenges, and existing traffic management infrastructure is increasingly overwhelmed by the growth in traffic volume.
The problems caused by traffic congestion go far beyond driver frustration. Prolonged sitting can have significant negative effects on mental and physical health. Traffic pollution and its by-products affect crop yields, ozone formation, and are toxic to aquatic life and environments. Further, congestion results in more travel time and fuel consumption causing economic strain on populations.
Traffic congestion in Toronto, Canada’s most populated city (nearly three million people), is a well-documented aspect of the city’s identity. Ranked globally in the top twenty most-congested cities, it is estimated that Toronto drivers spend nearly 250 hours waiting in rush hour traffic each year.
Consequently, finding solutions to improve traffic congestion safety and performance is a high priority for the City of Toronto. They are working with leading transportation experts across a variety of industries to not only mitigate Toronto’s congestion issues, but to set new standards for how we assess and respond to this growing issue.
Solutions from the UBC connected-vehicle testbed
City officials were first introduced to AURORA (the Automotive testbed for Reconfigurable and Optimized Radio Access) at the ITS Technology Innovation Forum held at UBC in March 2023, and quickly grew interested in investigating its use as a potential solution in downtown Toronto.
Led by UBC’s Dr. Dave Michelson, a professor in the Department of Electrical and Computer Engineering, AURORA conducts national, connected-vehicle research that is part of an initiative to promote safe, smart transportation. Michelson’s external roles as Vice-Chair of the IEEE’s Intelligent Transportation Systems Standards Committee (VT/ITSC) and Chair of the IEEE Digital Privacy Initiative’s Connected Vehicle Working Group provide his team with additional opportunities to contribute on the global stage.
AURORA’s impact has also been accelerated through UBC’s partnership with Rogers Communications—a collaboration that enhances research (across multiple sectors, from climate solutions to human health and social impact) by leveraging the capabilities of 5G.
“Urban transportation issues tend to be complex and involve many stakeholders” says Michelson. “When managed correctly, multi-partner collaborations of this sort foster the discussions that give us confidence that we are addressing the right problems and doing so correctly.”
THE TECHNOLOGY
The processes, developed through Michelson’s UBC research, evaluate and verify data relating to impacts on the intersections’ traffic configuration in order to optimize flow.
They were informed by results from a 2022 pilot project between Rogers and UBC that integrated AI sensors and traffic management systems into AURORA at five signalized intersections on the UBC Vancouver campus — which encompasses more than 400 hectares, and hosts a daily population of approximately 80,000 people. The pilot captured real-time data about traffic volumes, pedestrians, cyclists, speeds and congestion levels. The results estimate that in one year, the technology could reduce pedestrian wait times by a combined 2,500 days, vehicle wait times by 4,700 days, and carbon dioxide emissions by 75 tons.”
Based on these and other findings, Michelson aims to form an evaluation toolkit that could work across multiple technologies and solutions, with aims to establish these as the recognized standard for Canada’s Intelligent Transportation Systems network.
Insights and outcomes
In March 2024, Toronto City Council approved their pilot project to trial its first-ever use of AI technology to manage downtown Toronto traffic.
The NoTraffic Detection System, an AI-powered mobility platform that monitors, analyzes and optimizes traffic flow, was installed at five of the city’s busiest and consistently most-congested intersections. The system was set up to collect high-resolution traffic data on vehicle, cyclist and pedestrian traffic across four targeted time periods (morning, afternoon, evening, night).
Michelson’s team conducted assessments of the data and provided important insights regarding: techniques for assessing the performance of next generation traffic cameras that incorporate AI technology; development of teaching labs; development of digital privacy guidelines and a standard for assessing next-generation video detection systems; the suitability of the NoTraffic system for the planned trial; and how to deal with related issues such as digital privacy.
The initial data proves that the NoTraffic system is ready to be placed in optimization mode so that assessment of the NoTraffic system for traffic management improvement can be pursued, and that high-resolution traffic data can be used for a variety of applications across different time scales.
While connected vehicle technology shows strong potential to enhance safety and efficiency, North American adoption is stalled due to deployment concerns. Consequently, the team has proposed a three-stage deployment strategy that focuses on: prioritizing low-risk, high-readiness applications; minimizing risk; and building stakeholder confidence.
“Our analysis has shown that there’s a lot of room for innovation, and in areas that people didn’t realize was an issue” says Michelson. “For example, there is room to improve specification 660 and a need to develop industry standards for collecting and measuring traffic data.” The specification he refers to is the industry’s current video detection system standard, the Florida Department of Transportation’s Specification 660 (published in 1998), which is critically limiting in its parameters, particularly in that it doesn’t factor for non-vehicle traffic such as cyclists and pedestrians. Together, the teams recognized a need for a modern, consensus-based standard for verifying and validating video detection systems, which Michelson aims to pursue through the IEEE Standards process.
“What we’re doing in terms of strategic initiatives is trying to create opportunities and forge relationships to increase the application and breadth of the work,” says Roger Browne, Director of Traffic Management at City of Toronto. “You’ve got the city of Toronto as a bit of a playground testbed, working with a research team across the country at UBC (which is something we’ve never done before) to validate results in real time. I really appreciate the partnership between Rogers, UBC and ourselves on this project. We’re already hearing other groups say ‘we want to work with UBC, we want these kinds of relationships and results.’ You can see that there’s so much more to come in terms of our working relationship and I’m really looking forward to it.” The City of Toronto is now confidently moving toward the next phase of its AI-based traffic congestion management projects.
EXPLORE PARTNERSHIP OPPORTUNITIES WITH UBC