Research Article
A Knowledge Graph for UAV Mission Planning Systems
@INPROCEEDINGS{10.1007/978-3-031-67162-3_7, author={Xiaofang Duan and Rong Chai and Siya Zhang and Chengchao Liang}, title={A Knowledge Graph for UAV Mission Planning Systems}, proceedings={Communications and Networking. 18th EAI International Conference, ChinaCom 2023, Sanya, China, November 18--19, 2023, Proceedings}, proceedings_a={CHINACOM}, year={2024}, month={8}, keywords={UAV Mission planning Knowledge graph Knowledge reasoning GNN}, doi={10.1007/978-3-031-67162-3_7} }- Xiaofang Duan
Rong Chai
Siya Zhang
Chengchao Liang
Year: 2024
A Knowledge Graph for UAV Mission Planning Systems
CHINACOM
Springer
DOI: 10.1007/978-3-031-67162-3_7
Abstract
Unmanned aerial vehicles (UAVs) are increasingly applied in various mission scenarios due to their versatility, scalability and cost-effectiveness. In practical UAV application scenarios, UAV mission planning systems (UMPSs) highly rely on efficient mission planning and resource scheduling strategies in order to meet the requirement of UAV missions. However, the dynamic and complex mission environment poses challenges to mission planning in UMPSs. To tackle this problem, knowledge graph technology can be utilized which plays a critical role in managing intricate relationships and constraints among UAVs, missions and environments, ensuring the efficient information update of the mission environment as well as the intelligent perception of the environment knowledge. In this paper, we first summarize the process of constructing knowledge graphs for UMPS, and discuss the concepts and procedures of knowledge modeling, extraction, fusion and representation in detail. Then, we investigate the knowledge reasoning problem in UMPSs which is of particular importance for achieving information update in the mission execution environment, and propose a graph neural network (GNN)-based knowledge reasoning method that leverages the advantages of both graph structure and the path inferring technique to predict the missing entities or relations in the knowledge graph. Finally, the effectiveness and applicability of the proposed knowledge reasoning method are verified via simulations.
