Advancing UAV Sensor Fault Diagnosis Based on Prior Knowledge and Graph Convolutional Network DOI Creative Commons
Hui Li, Chaoyin Chen, Tiancai Wan

et al.

Machines, Journal Year: 2024, Volume and Issue: 12(10), P. 716 - 716

Published: Oct. 10, 2024

Unmanned aerial vehicles (UAVs) are equipped with various sensors to facilitate control and navigation. However, UAV highly susceptible damage under complex flight environments, leading severe accidents economic losses. Although fault diagnosis methods based on deep neural networks have been widely applied in the mechanical field, these often fail integrate multi-source information overlook importance of system prior knowledge. As a result, this study employs spatial-temporal difference graph convolutional network (STDGCN) for sensors, where structure naturally organizes diverse sensors. Specifically, layer enhances feature extraction capability nodes, modules designed extract dependencies from sensor data. Moreover, ensure accuracy association graph, research introduces UAV’s dynamic model as knowledge constructing graph. Finally, diagnostic accuracies 94.93%, 98.71%, 92.97% were achieved three self-constructed datasets. In addition, compared commonly used data-driven approaches, proposed method demonstrates superior capabilities achieves highest accuracy.

Language: Английский

Large Language Models for UAVs: Current State and Pathways to the Future DOI Creative Commons

Shumaila Javaid,

Hamza Fahim, Bin He

et al.

IEEE Open Journal of Vehicular Technology, Journal Year: 2024, Volume and Issue: 5, P. 1166 - 1192

Published: Jan. 1, 2024

Language: Английский

Citations

6
Advancing UAV Sensor Fault Diagnosis Based on Prior Knowledge and Graph Convolutional Network DOI Creative Commons
Hui Li, Chaoyin Chen, Tiancai Wan

et al.

Machines, Journal Year: 2024, Volume and Issue: 12(10), P. 716 - 716

Published: Oct. 10, 2024

Unmanned aerial vehicles (UAVs) are equipped with various sensors to facilitate control and navigation. However, UAV highly susceptible damage under complex flight environments, leading severe accidents economic losses. Although fault diagnosis methods based on deep neural networks have been widely applied in the mechanical field, these often fail integrate multi-source information overlook importance of system prior knowledge. As a result, this study employs spatial-temporal difference graph convolutional network (STDGCN) for sensors, where structure naturally organizes diverse sensors. Specifically, layer enhances feature extraction capability nodes, modules designed extract dependencies from sensor data. Moreover, ensure accuracy association graph, research introduces UAV’s dynamic model as knowledge constructing graph. Finally, diagnostic accuracies 94.93%, 98.71%, 92.97% were achieved three self-constructed datasets. In addition, compared commonly used data-driven approaches, proposed method demonstrates superior capabilities achieves highest accuracy.

Language: Английский

Citations

0