Smart Mobility Digital Twin for Automated Driving: Design and Proof-of-Concept
Kui Wang, Zongdian Li, Tao Yu, Kei Sakaguchi
TL;DR
This work tackles the challenge of coordinating automated driving in dynamic traffic by introducing a real-time mobility digital twin that couples edge and cloud computing with V2X sensing. The authors design a digital twin model that reflects static and dynamic elements and a platform that distributes perception, routing, and control tasks across RSU/OBU edges and cloud resources, enabling cooperative perception and route planning. A proof-of-concept with real traffic demonstrates feasibility, achieving latency well below a practical limit of $3.126$ s and enabling re-routing decisions based on road occupancy. The study provides a practical framework for improving traffic efficiency and safety in automated driving and sets the stage for real-world evaluations of system-wide benefits.
Abstract
During the past decade, smart mobility and intelligent vehicles have attracted increasing attention, because they promise to create a highly efficient and safe transportation system in the future. Meanwhile, digital twin, as an emerging technology, will play an important role in automated driving and intelligent transportation systems. This technology is applied in this paper to design a platform for smart mobility, providing large-scale route planning services. Utilizing sensing technologies and cloud/edge computing, we build a digital twin system model that reflects the static and dynamic objects from the real world in real time. With the smart mobility platform, we realize traffic monitoring and route planning through cooperative environment perception to help automated vehicles circumvent jams. A proof-of-concept test with a real vehicle in real traffic is conducted to validate the functions and the delay performance of the proposed platform.