Scholarlink

Research on employing knowledge graphs for Quadrotor Unmanned Aerial Vehicle (QUAV) fault diagnosis is relatively scarce. This study explores the construction of a fault knowledge graph for QUAV. The ontology framework is devised with a focus on structural principles and supplementary fault case data, reflecting the unique characteristics of QUAV-related faults. Leveraging the Qwen large language model of Alibaba Cloud, we streamline the information extraction process from text by integrating Named Entity Recognition (NER) and Relation Extraction (RE) tasks, resulting in improved efficiency and accuracy over specialized models. The use of Cypher serves as a pivotal tool for integrating knowledge graphs into the graph database, thereby facilitating their visualization. Through harnessing the powerful querying functionalities of Cypher, the experiments are executed that closely mimic the intricate fault analysis strategies adopted by maintenance professionals. The experimental results confirm that this knowledge graph provides a sound foundation for reasoning based on structural principles.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Fault Knowledge Graph for QUAV Construction Based on Qwen Large Model

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Zhuolun Li,
  • Shaojun Liang,
  • Yi Yang,
  • Chengjie Xu,
  • Ying Zheng

摘要

Research on employing knowledge graphs for Quadrotor Unmanned Aerial Vehicle (QUAV) fault diagnosis is relatively scarce. This study explores the construction of a fault knowledge graph for QUAV. The ontology framework is devised with a focus on structural principles and supplementary fault case data, reflecting the unique characteristics of QUAV-related faults. Leveraging the Qwen large language model of Alibaba Cloud, we streamline the information extraction process from text by integrating Named Entity Recognition (NER) and Relation Extraction (RE) tasks, resulting in improved efficiency and accuracy over specialized models. The use of Cypher serves as a pivotal tool for integrating knowledge graphs into the graph database, thereby facilitating their visualization. Through harnessing the powerful querying functionalities of Cypher, the experiments are executed that closely mimic the intricate fault analysis strategies adopted by maintenance professionals. The experimental results confirm that this knowledge graph provides a sound foundation for reasoning based on structural principles.