Texas A&M researchers working on technology to improve AI used in traffic light switchers
Their ultimate goal is to reduce commuter wait times at intersections
COLLEGE STATION, Texas (KBTX) - Researchers at Texas A&M are working on technology to make traffic light switchers more efficient with the goal of reducing wait times at intersections.
Dr. Guni Sharon, a Texas A&M computer science and engineering professor who is leading the project, says current switchers use a very basic form of artificial intelligence that’s based mostly on the number of cars waiting at any particular intersection. When the switcher notices a certain number of cars waiting at a red light, it turns green. When all those cars have passed through the intersection, it shifts to yellow and then red again.
This new technology Sharon and his team are developing is capable of teaching those switchers how to take other factors into account when causing the traffic light it controls to change colors. Some of those factors include approaching cars from other directions, speed limits, the number of lanes on the road, and even traffic flows and trends based on the time of day.
”Many other outlets suggested using what’s called deep reinforcement learning for each intersection to use a deep neuro-network to learn and optimize the actuation of the intersection,” Sharon said.
Based on computer simulations, Sharon says the technology has reduced intersection wait times by as much as 20%.
“What we’re trying to do is to train a much simpler control function that can be understood and regulated by traffic engineers,” Sharon said.
Sharon says this technology is still in its somewhat early phases and a few years away from seeing real-world applications.
Copyright 2021 KBTX. All rights reserved.