AI could make traffic signals less chaotic, roads more orderly for motorists in the future, MIT researchers say

Most of us do not like to wait at a traffic signal until the light turns green. Sometimes, we try to bypass the signal and end up getting stuck in the process, after which we have to pay a hefty fine. Sometimes this also leads to accidents. It is true that waiting at the signal is very tiring and costs us in terms of fuel wastage and time, harming the environment. Motorists and policy planners want a practical alternative, albeit minor, to put an end to this nuisance. Researchers at the Massachusetts Institute of Technology (MIT) think they have uncovered one.

The team tried to find ways to ensure that motorists did not have to wait at a traffic intersection for a signal to change colour. Instead, what if they reach the signal exactly when it is green. Although this is difficult to achieve for human drivers, it can be done with consistency by an autonomous vehicle that uses artificial intelligence,

Using AI, the vehicle speed can be adjusted in such a way that it reaches the next signal prematurely and passes it without waiting for the color to turn green.

In his study, published on pre-print server arXivthe researchers demonstrated machine learning Approach that can learn to control a fleet of self-propelled vehicles to keep the flow of traffic smooth. Led by graduate student Vindula Jayawardena, the team of researchers say their approach reduces fuel consumption and emissions while improving average vehicle speed.

“It’s a really interesting place to intervene. Nobody’s life is better because they were stuck at a crossroads,” was senior author Kathy Wu. Cited as saying.

But there is another complication. The researchers want the system to learn technology that saves fuel while reducing travel times. These objectives may be inconsistent. Wu says that although he wants the automobile to go faster to save travel time, he wants it to slow down or not run at all to reduce emissions. These competitive landscapes can be extremely troubling for the learning agent.

As a result, the researchers devised a workaround known as reward shaping. He provided the systems domain with information that it could not learn on its own using reward shaping. He penalized the system in this scenario, whenever the vehicle came to a complete halt, it would learn to avoid doing so in the future.

Once they had built it, they tested their control algorithms using a traffic simulation platform with an intersection. As the automobiles approached the intersection, their system did not cause any stop-and-go traffic. When cars are forced to a complete stop due to traffic stop ahead, it is known as stop-and-go traffic.

In simulations more cars pass through the same green phase, which outperforms models simulating human drivers. Compared to previous optimization strategies aimed at avoiding stop-and-go traffic, their approach resulted in higher fuel savings and lower emissions.

If all vehicles on the road are autonomous and connected to their systems, they could reduce fuel consumption by 18 percent and CO2 emissions by 25 percent, while improving travel speed by up to 20 percent, he says. . Even if only two percent of vehicles are autonomous, they can offer at least 50 percent of the total fuel and emissions reduction benefits.


Source link

Sharing Is Caring:

Leave a Comment