Machine Learning Models Could Improve Transit in Chattanooga – Government Technology

Integrating new technologies related to on-demand transit and introducing electric vehicles into public transit operations requires a deep knowledge of those systems and how best to deploy them. In Chattanooga, artificial intelligence could soon help meet that end.

The Chattanooga Area Regional Transit Authority (CARTA) was recently awarded $3.9 million in grants from the National Science Foundation (NSF) and the U.S. Department of Energy (DoE) to continue research projects with Vanderbilt University and other academic partners to develop machine learning models for insights into how to best deploy electric buses and on-demand transit. 

“We’re facing many of the same challenges that many other transit agencies are facing, in terms of funding, operational efficiency and providing useful transit to the public to increase ridership. So we’re looking at ways in which we can create a multimodal ecosystem and to do that as efficiently as possible,” said Philip Pugliese, transportation system planner at CARTA. 

CARTA has had an electric shuttle fleet going back to 1992 and, in 2016, introduced three large 35-foot battery-electric buses. That transition prompted the agency to begin working with researchers at Vanderbilt and other academic partners to better understand how to expand and deploy the electrification process more efficiently, said Pugliese.

“We have funding to expand our electric fleet, and understanding how best to deploy those vehicles, and how best to justify the capital expense of battery electric technology, versus diesel and hybrid-diesel is very important to us,” he added. 

The research will offer insight into how to best partner the operational realities of electric buses with various transit needs. Data across a number of metrics will be explored across variables such as climate, route details such as distance and grade, as well as charging costs, which can differ greatly depending on the time of day. The initial project will look at what routes are most ideal for EVs. 

The NSF awarded the project $2.1 million, while the DoE awarded $1.8 million. 

“And then with our new Department of Energy project we’re looking at ways in which we can harness any opportunities to integrate dynamic micro-transit type service, in connection with our fixed-route fleet,” said Pugliese. “So, really looking to the future, and looking at access to neighborhoods and with connection to our high-frequency corridor type routes.” 

Electric vehicles, transit officials and researchers assert, operate much differently than conventional fossil fuel burning vehicles or hybrid electric buses. For example, EVs tend to perform better in high volume traffic than a diesel bus.

“At the end of the day we want to know how to allocate buses to specific trips, specific routes. And which buses should be allocated to which routes,” said Abhishek Dubey, assistant professor of computer science and computer engineering at Vanderbilt University.

“Ideally, if you had the whole fleet as electric vehicles, which is very expensive and very, very difficult to have … you would have electric vehicles everywhere. But you cannot. So the question becomes, ‘With the limited electric and hybrid fleet that you have, where will you get the most impact?’” said Dubey. “Which routes should you really be using electric vehicles, and which are the rest of the routes you can actually put your diesel vehicles on?”

The project will develop machine-learning models which take in all of these various metrics. CARTA and the researchers plan to develop the project to share these analysis tools with other agencies which are also embarking on electric fleet transitions. 

“We think that with this type of modeling technology, it will give agencies the opportunity to really expand the deployment of electric vehicle technology,” said Pugliese.

Looking for the latest gov tech news as it happens? Subscribe to GT newsletters.

Source

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More