“I imagine a tool that will serve more as a personal assistant to the passenger,” Michael Hutchins, data analyst at GoDurham, the North Carolina city’s transit provider, said of a potential future for AI in transit where riders are not only kept abreast of delays, but offered information related to other routes or travel options. He examined the technology during a panel Tuesday on its use and possibilities in public transit, organized and hosted by Optibus, a transit tech provider.
GoDurham is now experimenting with AI-powered tools on a project known as “traffic service priority,” where onboard technology captures information like on-time performance, GPS positioning and scheduling to aid in outcomes that improve transit speeds in heavy traffic.
The endeavor leverages existing data, processes this with AI, “in order to signal traffic signals and decrease the travel time in those areas,” Hutchins said, noting the technology currently has a limited deployment but officials hope to scale it wider.
“What we want it to do is improve the customer experience by taking a vehicle and moving it through high-traffic areas and improve time,” he said.
Hutchins admits his pragmatism toward the early stages of using AI-enabled technology, but acknowledges it has shown promise for data-intensive tasks.
“We’re very green. We’re simply trying to just discover it,” he said. “Figure out how we can create efficiencies in our processes. And do things in less time.”
The ability to take large amounts of data from numerous sources and synthesize it for decisions related to the complex job of routes and scheduling is where AI can shine, said Sharad Agarwal, Optibus managing director for North America.
“The ability to aggregate just giant, giant sums of data is really going to be the biggest driver. And we’ve seen that a lot when it comes to on-time performance,” Agarwal said during the panel, indicating AI is also being used for “scenario testing,” where a single scenario can be tested in a number of various fashions, aiding in the scheduling of routes and other concerns. “I think AI is really going to help make a lot of efficiency gains, especially when it comes to running a transit operation.”
Considering numerous factors and data sources when setting routes and schedules is essential, Hutchins agreed, highlighting the need to balance various priorities and challenges like on-time performance, ridership, driver shortages and more.
“And you try to bring all of these different things into the forefront, and figure out, ‘How do I take this [siloed] data and put it into a space where I can [act on] all these things at the same time,’” he said. “My hope is that in the future AI will be able to basically grab all these different things and put it in the planner, for example, and can actually make a determination on how to construct the routes.”
