acm-header
Sign In

Communications of the ACM

ACM TechNews

How AI Can Reduce Traffic Congestion, Fuel Consumption


View as: Print Mobile App Share:
Artist's rendering of an intelligent traffic control system.

Oak Ridge National Laboratory researchers used artificial intelligence and machine learning to design a computer-vision system to keep traffic moving efficiently through intersections.

Credit: cleantech.com

Researchers at Oak Ridge National Laboratory (ORNL) used artificial intelligence and machine learning to design a computer-vision system to keep traffic moving efficiently through intersections while minimizing fuel consumption.

The researchers used stoplight cameras from traffic management services firm Gridsmart to gather real-world data from images of vehicles passing through intersections, and trained the cameras to identify vehicle types and their estimated gas mileage.

Afterwards, they used the open source SUMO package to model traffic systems, and ran a simulation of a citywide traffic grid.

They enhanced the system with the ability to train a machine learning algorithm to control the traffic lights, and applied reinforcement learning to keep high-fuel-consumption vehicles moving, rather than idling at traffic signals.

Said ORNL's Thomas Karnowski, “What's interesting about it is you basically set up a system of rewards and penalties, and then you let the computer try different things until it learns to get the rewards and minimize the penalties."

From Government Computer News
View Full Article

 

Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account