A software system developed by Vanderbilt researchers to help improve operations of Nashville’s public transportation network won “Best Paper” at the 15th ACM/IEEE International Conference on Cyber-Physical Systems (ICCPS), held in Hong Kong May 13-16.
Currently, WeGo Nashville has 160 buses that cover 35 routes. However, its ridership is steadily increasing, leading to higher chances of disruptions like overcrowding, mechanical failures, and accidents.
To address the problems, Abhishek Dubey, associate professor of computer science and electrical and computer engineering, and his team designed a cloud-based tool called Vectura that incorporates artificial intelligence to provide insights into passenger flows, service delays, and the operational efficiency of WeGo service, which is testing the technology.
“Vectura leverages the data feeds already generated by modern transit networks, transforming raw figures into intuitive, interactive visualizations that highlight key performance metrics,” said Dubey, senior research scientist at the Institute for Software Integrated Systems (ISIS). “By doing so, it enables operators to quickly identify trends, pinpoint inefficiencies, and make informed decisions to enhance service quality.”
In the paper, the team also described using decision-making processes where the focus extends beyond immediate outcomes to consider long-term consequences of actions — an approach known as “non-myopic sequential decision procedures.”
Dubey explained that by doing so, the software can help anticipate problems and proactively position buses near areas with high likelihoods of disruptions and determine which vehicle to dispatch to solve a particular problem. Results showed 2% more passengers served and a reduction in deadhead miles by 40%.
“You don’t know what will happen in the future, but you plan for it by using some idea of what the environment is,” said Jose Talusan, lead author on the paper and a research scientist in ISIS. “Once a decision is performed, you wait and see how the environment reacts and then consider this for your next round of decisions.”
Dan Freudberg, WeGo’s deputy chief operating officer, is a co-author on the paper.
Funding for the research was provided through the Federal Transit Authority and National Science Foundation.
Contact: Lucas Johnson, lucas.l.johnson@vanderbilt.edu