An Edge-Driven Digital Twin Structure for Connected and Autonomous Vehicles

Connected and Autonomous Vehicles (CAVs) generate continuous streams of data that must be processed efficiently to support real-time monitoring, safety, and intelligent decision-making. However, handling such high-volume data using centralized systems introduces latency and scalability challenges. This paper presents an edge-based Digital Twin framework designed for real-time vehicular data processing using a time-series architecture. The proposed system integrates a real-world GPS dataset to simulate multiple vehicles and employs MQTT-based communication for lightweight data transmission. A Digital Twin engine deployed at the edge processes incoming telemetry data, performs anomaly detection based on speed variations and behavioral patterns, and stores results in a time-series database for efficient querying. Additionally, a RESTful API layer and interactive map visualization enable real-time monitoring and analysis of vehicle trajectories and detected anomalies. Experimental results demonstrate that the system supports scalable multi-vehicle simulation, efficient data handling, and effective anomaly identification. The proposed framework provides a practical and lightweight solution for intelligent transportation systems and future smart mobility applications.