Collaborative partnerships among public transit agencies like the Chattanooga Area Regional Transportation Authority (CARTA), the Nashville Metropolitan Transit Authority, research universities and Google have sparked transit system improvements expected to yield cost savings and other benefits.
Using technology to optimize transit operations is not new. Companies like Optibus and Via have numerous solutions that rely on AI and other innovations. But these commercial solutions are not always carefully tailored for the underlying needs of smaller agencies, Philip Pugliese, CARTA’s general manager for planning and grants, told Government Technology.
“We’re hopeful that we can create a paradigm that’s helpful for small and midsize cities, and create solutions that not only make transit competitive, but then create the behavioral mode-shift that’s necessary to support those operations going forward,” he said.
One project, still largely in research and pilot stages, demonstrates the possibilities for how transit can grow in-house AI technology grounded in the realities of the place. It involves more than 50 million data points related to conventional and electric buses, as well as weather and traffic information, sought to better optimize paratransit operations at CARTA. The aim is to expand the technology to on-demand microtransit and, ultimately, across the fixed-route operations. Much of the research and real-world applications at the transit agency are still in the pilot phase, but have already shown the potential for $150,000 in projected annual fuel savings.
“We showed that we can actually improve the efficiency by 40 percent,” said Abhishek Dubey, associate professor of electrical engineering and computer science and director of the ScopeLab in the Institute for Software Integrated Systems at Vanderbilt University, which worked on the project. The work was partially funded by the National Science Foundation.
The ultimate goal, Pugliese said, is to apply the technology to fixed route, paratransit and microtransit “to be able to optimize the integration of all of those components and resources to make public transit more competitive, and cost-effective.”
“We have not operationalized some of the research elements just yet,” he said. “But we’ve demonstrated the efficacy in several areas.” These include optimizing for deploying the right type of vehicle — diesel, diesel hybrids or battery-electric buses — for the right type of service.
Transportation technology advancements that take into account real-time, on-the-ground realities like traffic, street networks and other variables used to plan dynamic routing or even forecast ridership and fuel consumption, have quickly catapulted transit into the high-tech space of data-driven decision-making and cloud computing. This is where Google enters the picture, to enable scaling beyond a small pilot across multiple transit operations.
“I think we are really at a place where we are seeing the potential for this, and the partnership with Google really allows us to use the cloud computing to innovate. We can’t do this without that,” Dubey said.
The technology produced by the project is not entirely deployed. The CARTA project has been in pilot mode, and another pilot involving on-demand microtransit operations is ready to move forward. It is now focused on field operations, where the same sorts of data points and analysis that led to operational improvements around route planning and travel times in paratransit can be applied to on-demand microtransit.
“Once that is successful, then we can really scale things up,” Dubey said.
“I think everyone will agree that we’re in a transformative phase for transportation,” Pugliese said. “And especially public transport. And with the rapid evolution of new technologies from machine learning, autonomy, electrification, connected vehicle technology, all coming online in a very rapid fashion, having a transit agency being able to leverage these partnerships with extensive academic researchers is a huge advantage, we believe, in informing our decision-making.”
