Making Traffic Jams History - AI Traffic Light System Could Significantly Reduce Congestion

Making Traffic Jams History - AI Traffic Light System Could Significantly Reduce Congestion

Aston University researchers devised a new artificial intelligence system that surpasses all existing systems.

Making Traffic Jams History - AI Traffic Light System Could Significantly Reduce Congestion

A new artificial intelligence system analyzes live video footage and adjusts the lighting accordingly.

Americans spent 6.9 billion hours stuck in traffic in 2014. The average commuter burned an additional 19 gallons of petrol during traffic delays. Each year, this equates to $160 billion in lost time and fuel.

In many major US cities, traffic can waste more than 100 hours per year for the average motorist. That's enough time to take two and a half weeks off in a regular company. Researchers, thankfully, are attempting to minimize traffic congestion, whether through the development of autonomous automobiles or the application of artificial intelligence in traffic signals.

Long lineups at traffic lights, for example, might be a thing of the past according to Aston University researchers' new artificial intelligence technology (AI). The groundbreaking technology monitors live camera footage and changes traffic signals to compensate, keeping traffic moving and reducing congestion.

The technology employs deep reinforcement learning, in which software identifies when it is not doing well and tries a different strategy - or continues to improve when it is performing well. In testing, the system outperformed all existing techniques, which frequently rely on manually-designed phase transitions. Traffic signal timing issues are a key source of congestion.

The researchers created Traffic 3D, a cutting-edge photo-realistic traffic simulator, to train their computer, training it to manage various traffic and weather circumstances. Despite being taught purely on simulations, the system adapted to real traffic junctions when tested on a real junction. As a result, it might be useful in a variety of real-world situations.

"We created this up as a traffic control game," Dr. Maria Chli, a reader in Computer Science at Aston University, stated. When the computer successfully navigates an automobile through a crossroads, it is rewarded. There is a negative reward every time a car needs to wait when there is a traffic bottleneck. We essentially have no input; we merely control the reward system."

At the moment, the most common type of traffic signal automation used at intersections is based on magnetic induction loops; a wire rests on the road and detects vehicles passing over it. That is counted by the software, which then reacts to the data. Because the Aston University team's AI'sees' high traffic volume before the cars pass through the lights and makes its choice then, it is more responsive and can respond faster.

"The reason we based our software on learned behaviors is so that it can grasp circumstances it hasn't directly experienced before," stated Dr. George Vogiatzis, senior lecturer in Computer Science at Aston University. We tried this with a physical obstruction that causes congestion rather than traffic signal phasing, and the system performed admirably. The computer will eventually figure out what the causal relationship is as long as there is one. It's a really powerful mechanism."

This year, the researchers want to begin testing their system on actual roadways.

Post a Comment

Previous Post Next Post