Pipeline for Fast Digital Twin Development and Integration in Driving Simulation

The use of digital mapping with precise GPS coordinates has allowed intelligent navigation, which is now ubiquitous in vehicles. The flat 2D imagery provided by Google maps was recently enhanced by the introduction of dynamic 3D representations of the Earth. Unity3D and Unreal Games Engines now offer Application Programming Interfaces that can be leveraged for geospatial applications. This technology, which offers visually compelling results for flight simulation and drone applications, opens new opportunities for driving simulation. This paper presents an innovative approach for developing a drivable Digital Twin of Philadelphia’s Roosevelt Boulevard to enhance urban planning and traffic management using advanced simulation technologies. We introduce a complete pipeline that integrates geospatial imagery with data from Google Maps and OpenStreetMap (OSM) through tools like CityEngine and RoadRunner, enabling the creation of highly detailed, editable 3D urban scenes. This methodology facilitates rapid modifications to urban landscapes, exemplified by the integration of a bus lane into existing road infrastructure, demonstrating significant advancements over traditional methods. The core of our approach is a dynamic traffic flow model developed within the Unity driving simulator, utilizing probabilistic distributions and real-world data from the NGSIM dataset to mirror actual traffic conditions accurately. This model supports the simulation of realistic, varied, and dynamic traffic patterns, crucial for testing and evaluating urban traffic scenarios and infrastructure changes. The implementation showcases the potential of digital twins for transforming urban planning and traffic systems by providing a reliable platform for scenario testing and decision-making.