Professors Sharon Di (Civil Engineering), Qiang Du (Applied Mathematics), Gil Zussman, and Zoran Kostic (both Electrical Engineering) were recently awarded a $1.2 million grant from the National Science Foundation’s Cyber Physical Systems (CPS) program for their proposal "CPS: Medium: Hybrid Twins for Urban Transportation: From Intersections to Citywide Management."
With this funding, the team will create a virtual replica, or digital twin, of New York City that will continuously learn and dynamically update itself as the city traffic environment changes in real time. The twin will help traffic managers to monitor traffic patterns as they happen and quickly come up with adaptive management strategies. The researchers will use Columbia’s COSMOS, the only beyond-5G testbed in New York City, to get real-time traffic data, leveraging Cosmos’s rich sensor data and deep computational capabilities.
The digital twin is a hybrid of both machine learning and traffic modeling, reflecting the team’s multidisciplinary approach to how traffic congestion propagates in cities. They will train the system online by taking in real-time data collected from Cosmos sensors, including roadside infrastructure and in-vehicle sensors. With this data, the system can predict traffic conditions, accidents, and mitigate traffic congestion as well as optimize traffic flow so that people can travel across cities with fewer stops at intersections and reduced emission levels.