Vehicle risk intervention method and system based on real-time multi-source data and knowledge graph

By combining real-time multi-source data and knowledge graphs with heterogeneous graph neural networks and temporal convolutional networks, a dynamic knowledge graph is constructed. This solves the problems of high false alarm rate and inaccurate intervention in traditional vehicle risk management, enabling rapid identification, graded early warning, and precise intervention of vehicle risks, thereby improving the level of traffic safety management.

CN121640711BActive Publication Date: 2026-07-24GUANGXIN INTELLIGENT CONSTR RES INST CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
GUANGXIN INTELLIGENT CONSTR RES INST CO LTD
Filing Date
2025-12-05
Publication Date
2026-07-24

AI Technical Summary

Technical Problem

Traditional vehicle risk management methods rely on a single data source, resulting in high false alarm and false negative rates. They lack the ability to jointly model multi-source heterogeneous data, making it impossible to achieve real-time and accurate risk behavior identification and intervention. Furthermore, existing intervention methods lack specificity and are prone to causing secondary accidents.

Method used

By employing a real-time multi-source data and knowledge graph approach, dynamic knowledge graphs are constructed by collecting data from vehicles, roadside facilities, and third parties. A joint model combining heterogeneous graph neural networks and temporal convolutional networks is used for risk identification, and intervention strategies are optimized through causal inference algorithms to achieve rapid identification, graded early warning, and adaptive intervention of vehicle risks.

Benefits of technology

It significantly reduced the false alarm rate and secondary accident rate, shortened the intervention response time, improved the accuracy of risk identification and the pertinence of intervention, and achieved the system's adaptive learning and continuous optimization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121640711B_ABST
    Figure CN121640711B_ABST
Patent Text Reader

Abstract

The application provides a vehicle risk intervention method and system based on real-time multi-source data and a knowledge graph, wherein the method comprises: collecting multi-source heterogeneous data and environmental data in real time, generating real situation features according to the environmental data; performing spatio-temporal alignment on the multi-source heterogeneous data to generate a standardized event stream. A dynamic knowledge graph is constructed based on ontology definition; risk behavior recognition is performed based on a joint model of a heterogeneous graph neural network and a time series convolution network, and a risk category and a risk probability are output. According to the risk category and the real situation features, a corresponding intervention action is matched in a pre-constructed intervention knowledge graph, and the intervention action is issued and executed. After the intervention, the data is collected, the intervention effect is evaluated through a causal inference algorithm, and the joint model and the intervention knowledge graph are updated and fed back. The application can quickly recognize vehicle risks, give a situational graded early warning, adaptively and accurately intervene, and reduce secondary accidents and false positive rates.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • CN120410211A

  • CN120823045A