An Edge-Driven Digital Twin Structure for Connected and Autonomous Vehicles
B. Aishwarya
Dept. of CSE (Artificial Intelligence & Machine Learning)
ACE Engineering College
(JNTUH)
Hyderabad, India
aishwaryabandi147@gmail.com
Mrs. J. Bhargavi
Assistant Professor, Dept. of CSE (Artificial Intelligence & Machine Learning)
ACE Engineering College
(JNTUH)
Hyderabad, India
bhargavi.jangam@aceec.ac.in
T. Anirudh Singh
Dept. of CSE (Artificial Intelligence & Machine Learning)
ACE Engineering College
(JNTUH)
Hyderabad, India
thakuranirudh1920@gmail.com
B.Pradeep
Dept. of CSE (Artificial Intelligence & Machine Learning)
ACE Engineering College
(JNTUH)
Hyderabad, India
mail2pradeepbandari@gmail.com
Abstract—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.
Keywords—Digital Twin, Edge Computing, MQTT, Time-Series Database, Anomaly Detection, Connected Vehicles, GeoLife Dataset, Real-Time Processing
