Time management at red lights would improve considerably thanks to this AI
An AI can make it unnecessary to wait at traffic lights when there is no pedestrian nearby or crossing.
In general, we already know that most traffic lights work according to a programming that has been made to them. Then there are others that have buttons for pedestrians to request for it to turn green, although these are not always reliable. But in the future maybe an AI will take over this function.
Now the German Federal Ministry of Transport and Digital Infrastructure has funded a project called KI4LSA which it is researching with other partners improve traffic flow in cities. To achieve this, they began installing high-resolution cameras and radar sensors.
As explained in New Atlas, they experimented with an intersection controlled by traffic lights in the town of Lemgo. This population of approximately 40,000 was used to an AI studies a particularly busy area.
Thanks to the installed technology, a first recording of the number of vehicles, waiting times at red lights and speed at which they cross the intersection.
With this information, an algorithm was trained using machine learning to cross-reference the obtained data with stimuli from real-time traffic via cameras and sensors. All in an effort to find a way to minimize wait times.
In the simulations performed, it is estimated that AI improves traffic flow by up to 15%, although it should be noted that it left it at that: simulations, it has not been made for practical purposes at the moment. The next step will be activate it at this same intersection in the coming monthsbut still under surveillance.
Granted, it’s not amazing technology, but so far it’s barely made the leap into reality and it seems more than logical that progress is being made in this direction that would help decongest traffic in cities and therefore make them more pleasant.