Integration of road context information into knowledge graph for intelligent analysis of road accidents

Given the increasing complexity of road infrastructure, analyzing road accidents and developing effective prevention strategies has become critical. By examining historical accident data, researchers can identify patterns and trends in accident occurrences. The utility of this data is significantly enhanced by incorporating contextual information from diverse data sources. This study proposes a knowledge graph (KG) based on a labeled property graph model, which integrates temporal, spatial, and semantic accident knowledge with road context for intelligent accident analysis and data management. The integrated road-accident context graph (IRACG) presented here incorporates road context, including road topology and hazard information. Historical traffic accident data were used to extract comprehensive semantic accident details, which were then combined with road topology from the geospatial road network and hazard information from remote sensing sources. After constructing the IRACG using the Neo4j graph database, scenario-based experiments were performed to validate its applicability, focusing on accident pattern analysis and avoidance routing. The results demonstrate that the IRACG effectively integrates various types of accident knowledge, facilitating complex queries and in-depth analyses not achievable with traditional databases. This study highlights the importance of integrating detailed road-level context into KGs for more effective accident analysis and prevention.