A Smarter, Safer Intersection: How AI and Industry Collaboration Are Improving Road Safety in Knoxville

A private-public partnership powered by AI is seeking to make Knoxville’s streets safer.

The AI TechX Consortium at the University of Tennessee, Cubic Corp., and the City of Knoxville have partnered to improve road safety in Knoxville and beyond. They are applying artificial intelligence (AI) research to make streets safer for drivers, pedestrians, and all other roadway users.

Project Beginnings

The foundation for this work was laid in 2023 when the University of Tennessee launched AI Tennessee, a statewide initiative focused on developing and deploying AI solutions that improve the health and prosperity of communities across the state.

From that initiative, the AI TechX Consortium was formed to connect academic researchers, industry partners, and seed funding to develop scalable, applied AI technologies. Dr. Airton G. Kohls, a researcher at the UT Center for Transportation Research, proposed using AI to improve vulnerable road users’ safety at intersections, and through the consortium, the right partners quickly came together.

“It was actually very serendipitous,” said Andrew Hall, director of software engineering at Cubic Corp. “Dr. Kohls has been a long-time collaborator with us. He mentioned this project through the AI TechX consortium, and it aligned perfectly with our computer vision expertise. Our president is a huge proponent of research collaboration with universities. It all just fell into place.”

For the City of Knoxville, the project supports its Vision Zero commitment, a promise to eliminate traffic deaths and injuries on Knoxville roadways by 2040. It also aligns with their larger commitment to leveraging data and technology to inform policy and service delivery.

Carter Hall, the City’s Deputy Chief Technology Officer, explains, “We want to utilize AI beyond just chatbots; we’re most excited to leverage it to be better in our essential functions as a city. Obviously, pedestrian and traffic safety is a huge part of that. We don’t just want to ‘do AI,’ we want to do projects that are enabled by AI.”

Phase I

The first phase of the project focuses on the intersection of Cumberland Avenue and Volunteer Boulevard in the heart of UT’s campus. It is one of Knoxville’s most active intersections, handling up to 25,000 vehicles and 18,000 pedestrians daily.

The research team is analyzing intersection trends and identifying risky maneuvers, including pedestrians crossing too soon, bike riders crossing multiple lanes, and cars running red lights.

The project uses computer vision and AI to interpret video in real time. Based on these observations, the objective is to dynamically adjust signal timing at the intersection, modifying pedestrian operations based on live conditions.

Caleb Knight, director of AI TechX, says, “What I love is that this is a hybrid approach that can change as needed. There are so many different scenarios on Cumberland Avenue — you have high student volumes during fall and spring, chaos on gamedays, and in the summer, everything slows down.”

Cubic’s Andrew Hall adds, “We are using actual observations, not just assumptions based on history or upcoming events. Being able to apply insights about vulnerable road users to adaptive signal timing is a big deal.”

Phase II

While Phase I is focused on this single intersection, the long-term vision is much more ambitious. “We want it to be scalable,” Dr. Kohls remarks. “We are not just focused on this intersection or even just on Tennessee. We want this to be used all over the country.”

Future phases will look at entire corridors or multiple high-risk intersections. The goal is to create solutions that can be replicated at other universities and metropolitan areas facing similar challenges.

Impact

For all collaborators, the motivation to be involved in this project is simple: safety.

“The stakes are really high,” said Kevin Heaslip, director of the UT Center for Transportation Research. “If a pedestrian or bicyclist gets hit by a car, there is a very real probability of injury or even death. We have thousands of students and visitors who could be vulnerable in this situation, and we need to use this technology to help keep them safe.”

“This isn’t AI for AI’s sake,” Andrew Hall elaborates. “This is applied AI that has an impact on every single one of us. We all use the roadway. Projects like this help us get where we are going more safely and more efficiently.”