A Researcher is Using Machine Learning to Reduce Vehicle Pollution

Tuesday - November 14 2023, 07:13 UTC - 1 year ago

Vehicle pollution is a global and local problem that significantly contributes to air pollution. A researcher has developed a three-pronged strategy using machine learning to create traffic light management systems that will lessen emissions from vehicles, as well as account for social and environmental needs. These systems will improve air quality by controlling traffic signals and providing location-specific air quality levels to city dwellers.

Vehicle pollution is a significant contributor to air pollution worldwide making it both a global and local problem. It is responsible for the release of pollutants such as carbon monoxide, nitrogen oxides, particulate matter, and hydrocarbons into the atmosphere.

The United States Environmental Protection Agency estimates that greenhouse gas emissions from transportation account for about 29 percent of total US greenhouse gas emissions rising more from 1990 to 2021 than any other sector.

Decreasing emissions .

Now, in order to lessen this dangerous form of air pollution, a researcher is using machine learning to create traffic light management systems that are socially and environmentally conscious making them ideal at lessening emissions from vehicles.

Yu Yang, an assistant professor of computer science and engineering in Lehigh University’s P.C. Rossin College of Engineering and Applied Science, and his team are working on a three-pronged strategy that would enable more regular traffic flow with fewer or shorter stops. Cars sitting at stoplights in crowded cities significantly contribute to localized air pollution. Individuals who have asthma or other medical disorders that increase their sensitivity to airborne particulate matter are particularly affected by this activity.

The researchers will create a low-cost, mobile air-quality sensor device to locate high-pollution areas and understand the environmental needs of various locations. For example, a hospital location may be home to a higher number of vulnerable people who require more protection.

"We’ll use those data to then develop a spatial-temporal graph diffusion learning model to determine the traffic situation in our test-bed city of Newark, New Jersey," said Yang. "In other words, what is both the traffic and the air pollution like at different points of time in different locations?" .

Using a reinforcement learning technique, the researchers will explore how traffic signal regulation enhances air quality by taking into account traffic signals around the city and investigating their functioning.

The first of its kind .

"This is the first project of its kind to incorporate a social component into a traffic control system," said Yang. "We’re taking both a technical and a social perspective to solve a real-world problem." .

The researchers will also take into account human-system interactions, which include how individuals actually choose and use automobiles in urban settings. Yang argues that previous research in this field makes the assumption that people choose their rides at random. However, that is untrue indicating that the team’s algorithm can be improved by using this kind of data to make it more accurate at predicting how users really interact with the system.

Yang’s main objective is to create a web-based system that shows city dwellers location-specific air quality levels so that they can make educated decisions about what they do, when they do it and where they do it. This system would be complemented by a traffic management model that would allow city transportation officials to control signals in real time adapting them to change their settings to improve air quality.