AI, a tool against traffic jams? – Octopus.ca

Long queues at traffic lights could be a thing of the past, thanks to a new artificial intelligence system developed by researchers at Aston University.

The system in question, the first of its kind, consults images from traffic cameras in real time and adapts the frequency of traffic lights to compensate, which, according to the researchers, allows vehicles to continue to circulate, and reduces thus congestion.

Said system relies on a learning mode where a program is able to understand when it is doing wrong, and then try a new way of doing things, or continue to improve until it makes progress.

In tests, the system proved to be significantly more efficient than all other methods tried, which generally rely on hand-designed transitions.

In 2019, congestion in urban areas across the UK was estimated to lead to the average Briton wasting around 115 hours of their life in traffic jams, as well as wasting £894 in petrol and lost wages. One of the main causes of this congestion is the lack of synchronization of traffic lights.

The researchers designed a traffic simulator, Traffic 3D, to train their program, teaching it to handle various traffic and weather conditions. When the system was tested at a real intersection, it subsequently adapted to these real conditions, despite the fact that it had only been confronted with simulated environments. Specialists therefore believe that Traffic 3D could be effective in several real places.

“Like a game”

Dr. Maria Chli, explained that the team “designed the whole thing as a traffic control game. The program gets a “reward” when it takes a car through the intersection. Each time a car has to wait or there is a traffic jam, the reward is negative. We do not send signals or data; we only control the reward system”.

Currently, the main method of automating traffic lights depends on magnetic induction loops; a cable is laid on the road and detects cars driving over it. The program counts everything and then reacts to the data. Since the AI ​​created by the Aston University team “sees” heavy traffic before cars cross the traffic light and quickly makes a decision, the system responds more quickly, and with more flexibility, say the researchers. researchers.

For his part, Dr. George Vogiatzis, indicates that “the reason why we based our program on learned behaviors is that it allows him to understand situations never experienced before. We tested this with a physical obstacle causing congestion, rather than a traffic light timing issue, and the system still performed well. As long as there is a causal link, the computer will understand ultimately what is this link. It is a very powerful system”.

The researchers hope to be able to test their system on real roads this year.

Don’t miss any of our content

Encourage Octopus.ca

We wish to give thanks to the author of this write-up for this outstanding web content

AI, a tool against traffic jams? – Octopus.ca


Find here our social media profiles and other pages that are related to them.https://www.ai-magazine.com/related-pages/