A high-definition map architecture for transportation digital twin system construction

Abstract

Digital twin systems for transportation are widely regarded as a core technology for enabling full life-cycle management of traffic information and providing intelligent decision support, with the goals of improving traffic efficiency and reducing accident risks. However, existing research primarily focuses on simulation and prediction of traffic flow, lacking a unified framework that integrates static infrastructure, dynamic states, and microscopic behaviors. To address this gap, this paper proposes a lightweight behavior-cognitive architecture for high-definition (HD) maps to support multiscale information representation in transportation digital twin. It consists of three layers: (1) a global road network layer that models transportation infrastructure and static geographic features; (2) a dynamic target layer organizing the real time status and trajectory of traffic participants; (3) a behavioral cognition layer for behavior interpretation and understanding. Based on this architecture, a construction method for transportation digital twin systems is developed and validated through experiments conducted in real-world traffic scenarios and simulation environments. The results demonstrate that the proposed approach achieves high adaptability and accuracy, offering effective support for building digital twin systems in complex traffic environments.