The Federal Highway Administration has awarded $1.1 million for research to enhance detection of “vulnerable road users” within the Smart City Corridor overseen by the Center for Urban Informatics and Progress at the University of Tennessee at Chattanooga.
In addition to the funding awarded through the FHWA “Exploratory Advance Research” program, UTC and research partners will invest $300,000 to enable additional technology along the M.L. King Boulevard site to detect “vulnerable road users”—essentially, anyone not traveling inside an enclosed vehicle.
“It’s basically bicyclists, people on scooters, pedestrians, wheelchair users—everybody that’s not in a car or inside a vehicle,” said Dr. Mina Sartipi, founding director of CUIP, executive director of the UTC Research Institute and principal investigator on the funded research project. “The project title is ‘Advanced AI Research for Equitable Safety of Vulnerable Road Users,’ and we are working on fundamental research to develop advanced artificial intelligence algorithms” for traffic management systems that take the movement of “vulnerable users” into account.
The grant funds the addition of thermal cameras—which detect objects by infrared energy, or heat emitted—to a diverse array of existing sensors already installed at intersections along the Smart Corridor. Based on the combined data yielded by this technology, along with connected vehicles, Sartipi and her research team will develop algorithms for traffic management systems that incorporate prediction of movement by vulnerable road users, whose movements vary greatly depending whether on foot or on wheels.
“The major element that goes to the fundamental research part is sensor fusion,” Sartipi said. “That part, or the (artificial intelligence) part, is depending on the time of day, the natural light, the installed lighting, the weather and all of that, a camera or LiDAR (ultraviolet light detector), or radar or a thermal camera might be a better option. As part of this, we actually get data from more than one source and we infuse that data into the algorithms.
“This will ensure we have the most accurate observation of all road users at any given time and space and under any conditions.”
The algorithms developed will be tested and analyzed toward an accurate and real-time observation of intersections and prediction of traffic patterns that benefit urban mobility and improve safety.
Installation of thermal cameras funded by the federal grant begins this fall and will include some placed “mid-block,” Sartipi said, noting that movement by road users between intersections—such as crossing the street—also is a factor in traffic safety. Testing of sensor fusion-based algorithms is expected to be underway within 12 to 18 months.
The work is the latest in a continual sequence of research that began with the founding of CUIP in 2018.
“We are chipping away at it, working for four or five years on detection, and now working to make sure there is a system that can be applied under any weather condition, any road geometry condition and able to classify it. That’s what we’re going to be focusing on with this one,” Sartipi said.
“We have been working on detection and now we are saying, “OK, for this to be really applicable everywhere and generalizable, it should work under many, many different scenarios. This project is going to be working on that, and then we just keep enhancing.”
The University of Arizona and the University of San Francisco are collaboration partners and contributing funding along with UTC on this research project.