At a busy downtown intersection filled with cars, scooters, bikes and pedestrians, Raleigh traffic engineers are testing new approaches to managing congestion as the city continues to grow.
The city is combining existing traffic cameras with artificial intelligence and mapping software to analyze how people and vehicles move through selected intersections in near real time. City officials say the goal is not to eliminate traffic, but to make traffic lights operate more efficiently and safely, especially in areas with high pedestrian activity.
“One of the core challenges of transportation engineering is moving vehicles efficiently,” said Jed Niffenegger, City of Raleigh Transportation Manager. “We can't change the fact that intersections are busy. What we can do is make sure they operate as efficiently as possible.”
The pilot program relies on computer vision technology to convert live video into data. The software tracks turning movements, traffic volumes and different travel modes, replacing manual processes that once required staff to stand at intersections and manually count vehicles.
“The cameras allow us to make changes more quickly,” Niffenegger said. “Thanks to analytics, the amount of work involved has been reduced to a fraction of what it used to be.”
Raleigh has approximately 250 closed-circuit traffic cameras installed throughout the city. Currently, due to computing limitations, AI systems can analyze data from about 10 to 12 cameras at a time. Transportation staff is rotating intersections into the program based on locations where signal timing studies and corridor reviews are underway, such as along Glenwood Avenue and parts of downtown.
This technology allows engineers to fine-tune peak signal timing based on real-world conditions rather than assumptions, sometimes in just seconds.
“We're going to look at the evening rush and see if the traffic lights are running too long or if we need to start a little later,” Niffenegger said. “It’s about fine-tuning what we’re already doing.”
City officials say it's important to focus on intersections. Road widening has been done for years to ease congestion, but engineers say intersections are often the real bottlenecks.
“If we can maximize efficiency at intersections, we can delay costly road widening projects,” Niffenegger said. “This saves us money and allows us to invest elsewhere where we can have a bigger impact.”
James Alberk, the City of Raleigh's emerging technology manager, said the project is part of a broader effort to better understand how cities behave as their populations grow. Raleigh added more than 500,000 residents last year, putting new pressure on roads not designed for today's traffic.
“The camera records video, and we convert that video into data,” Alverque said. “That data helps traffic engineers understand what’s happening across the city at any given moment, not just with one camera.”
Data is visualized using Raleigh's high-resolution three-dimensional model, also known as a digital twin. City staff can compare traffic conditions before and after a signal change and assess whether the adjustment is working as intended.
The pilot initially focused on vehicular traffic. After validating these results, the city expanded its analysis to include pedestrians and cyclists, especially in downtown areas where pedestrian and scooter use is high.
“We wanted to understand all modes of transportation,” Alberque said. “GIS allows you to integrate traffic data with other information and analyze it in one place.”
City officials emphasized that privacy protections are built into the system. The City does not store videos or collect personally identifiable information. Data is limited to anonymous counts aggregated over 15-minute intervals, including whether the object is a vehicle, pedestrian, or bicycle.
“There is no identifying information whatsoever,” Alverque said. “We are not recording video.”
This technique is not used to automatically change signal timing. Instead, data is reviewed by engineers and used to inform decision-making, and humans maintain control throughout the process.
During the pilot, the city verified the AI results by running manual counts in parallel with the software. Alverque said more than a dozen test cases have been reviewed so far, giving staff confidence in the accuracy of the data.
Beyond traffic management, Raleigh's camera network already supports other city departments. Police, 911 dispatchers, and stormwater crews use cameras to better understand on-the-ground conditions during emergencies such as flooding.
Expanding the system beyond the pilot stage will require additional investment. Alverque said the city has been experimenting with technology while relying on existing infrastructure and limited resources.
“Technology is advancing very rapidly,” he said. “We are trying to be thoughtful about how we invest and make the right decisions as circumstances change.”
For now, city leaders say the pilot provides a way to improve safety and efficiency at busy intersections without widening the roadway, an approach they believe will become increasingly important as Raleigh continues to grow.
