Survey of Dynamic Knowledge Graph for Urban Traffic:Construction,Representation and Application
In the field of intelligent transportation,various information collection devices have produced a mas-sive amount of multi-source heterogeneous data.These data encompass various types of information,including vehicle trajectories,road conditions,and traffic incidents,soured from devices such as traffic cameras,sensors,and GPS.However,the current challenge faced by researchers and practitioners is how to correlate and integrate the massive amount of heterogeneous data to facilitate decision support.To address this challenge,knowledge graph technology,with its powerful entity-to-entity modeling ability,has shown great potential in knowledge mining,representation,management,and reasoning,making it well-suited for intelligent transportation applica-tions.In this paper,we first review the construction techniques for geographic traffic graphs,multimodal knowl-edge graphs,and dynamic knowledge graphs,demonstrating the broad applicability of knowledge graphs in the field of intelligent transportation.Secondly,we summarize relevant algorithms of multi-modal knowledge graph representation learning and discuss dynamic knowledge graph representation learning in the field of intelligent transportation.Knowledge graph representation learning technology plays a crucial role in creating high-quality knowledge graphs by capturing and organizing the relationships between entities and their attributes within the transportation domain.This technology utilizes advanced machine learning algorithms to analyze and process the heterogeneous data from various sources to extract meaningful patterns and structures.We also introduce the completion technology and causal reasoning technology in dynamic transportation multi-modal knowledge graph,which is useful for improving the data of intelligent transportation systems.Comprehension ability and de-cision-making reasoning level have important theoretical significance and practical application prospects.Third-ly,we summarize the solutions of knowledge graph that provide important support for intelligent decision-mak-ing in several application scenarios.The utilization of knowledge graphs in intelligent transportation systems fa-cilitates real-time data integration and enables correlation analysis of diverse data sources to provide a holistic view of the traffic ecosystem.This comprehensive understanding empowers decision-makers to implement tar-geted interventions and proactive measures,ultimately mitigating traffic congestion and reducing the occurrence of accidents.Through the continuous refinement and enrichment of the traffic knowledge graph,the intelligent transportation system can adapt and evolve to address emerging challenges and optimize transport networks for enhanced efficiency and safety.Finally,we analyze and discuss the existing technical bottlenecks.The future of traffic knowledge graphs and their auxiliary applications are also prospected and discussed,highlighting the po-tential impact of this important technology on intelligent transportation systems.
NETL
NSTL
万方数据
