AVKG: An interlinked Cross-domain Knowledge Graph for Smart Autonomous Vehicle Decision Making
Huihai Wang, Junfeng Jiao, Rui Zhu, Gengchen Mai
Figshare
Problems Identified (4)
Urban Data Integration for AVs: Autonomous vehicle deployments require timely, integrated urban data to support routing and higher-level mobility applications.
Heterogeneous GIS Information Silos: Dynamic urban knowledge reasoning and discovery over conventional GIS databases is difficult because heterogeneous sources, formats, and standards create information silos.
Urban Data Integration for AVs: Autonomous vehicle deployments require timely, integrated urban data to support routing and higher-level mobility applications.
Heterogeneous GIS Information Silos: Dynamic urban knowledge reasoning and discovery over conventional GIS databases is difficult because heterogeneous sources, formats, and standards create information silos.
Proposed Solutions (5)
Autonomous Vehicle Knowledge Graph: The paper develops AVKG, a machine-readable semantically structured graph repository integrating diverse urban contextual data for AV operation.
AV Ontology Schema: The paper develops an AV knowledge graph schema/ontology to represent semantic relations among concepts used in AVKG.
Cross-domain Urban Subgraphs: AVKG is organized into five urban subgraphs covering road networks, incidents, activities, points of interest, and AV operators.
GeoSPARQL Triple Store: AVKG is stored in a GraphDB triple store supporting SPARQL and GeoSPARQL queries for spatiotemporal semantic reasoning.
Autonomous Vehicle Knowledge Graph: The paper develops AVKG, a machine-readable semantically structured graph repository integrating diverse urban contextual data for AV operation.
Results (3)
Spatiotemporal Semantic Reasoning Support:
Competency Question Demonstration:
Urban Awareness Platform:
Research Domain
Autonomous vehicle knowledge graphs and urban GIS data integration