Philadelphia digital twin simulates road redesigns before construction

Much of Roosevelt Boulevard in Philadelphia consists of 12 lanes of traffic, and tens of thousands of people travel along it each day. Now, a team of engineers and urban planners has built a drivable digital twin of the road, allowing alternative layouts to be tested in simulation before construction.

The need is stark. Roosevelt Boulevard has 56 signalized intersections, 94 unsignalized intersections with 129 access points, and 54 crossovers between its inner and outer lanes. Between 2013 and 2017, it was the site of 2,946 crashes, resulting in 81 serious injuries and 62 deaths—14% of all fatal crashes in Philadelphia.

The project was led by Rahul Mangharam of the School of Engineering and Applied Science and Erick Guerra of the Weitzman School of Design, with Helen Loeb of the virtual automotive startup Jitsik. Their StreetSavii platform combines precise GPS coordinates, geospatial imagery and 3D Earth models, using urban-planning tools including CityEngine and RoadRunner to recreate actual driving conditions.

The digital twin also contains a traffic-flow model designed to reflect changing conditions. That lets planners rapidly prototype options such as integrating a bus lane into the existing roadway, then examine predicted effects on safety, cost and traffic before work begins. The team’s research appears in Transportation Research Record.

So what changes in practice? Philadelphia planners can compare more road designs earlier, when changes are cheaper and do not yet require construction. The source describes the platform as a way to rapidly prototype road layouts and calculate predicted safety, cost and traffic outcomes before construction. The platform remains a simulation: the reported performance comes from its creators, and the material here does not include an independent real-world validation of its predictions.