Table 1 Commonly employed methodologies for UAVs fault detection and diagnosis.
References | Methodology | Remarks |
|---|---|---|
Statistical feature extraction in UAV motors | Focuses on sound-based methods, lacks integration with deep learning | |
SVM classification with embedded feature extraction | Limited to specific motor faults, does not utilize complex feature sets | |
Convolutional neural networks with transfer learning for audio-based diagnosis | Employs CNNs but does not combine with other sophisticated methods | |
Compound fault labeling and diagnosis based on flight data and BIT record | Integrates multiple data sources but lacks advanced analytical techniques | |
Audio signal analysis for unbalanced blade detection | Utilizes traditional signal processing, limited in scope to specific fault types | |
Acoustic inspection system for wind turbines via UAVs | Focuses on structural health, not directly applicable to UAV internal faults | |
Frequency domain analysis for wind turbine inspection | Specific to wind turbines, does not translate directly to UAVs | |
Data-driven diagnosis under multiple operation conditions | Emphasizes data-driven methods without integrating novel feature extraction | |
Robust adaptive sliding mode for fault-tolerant control | Focuses on control solutions rather than diagnostic capabilities | |
Sensor fault diagnosis with Auto Sequential Random Forest | Utilizes advanced forestry methods but does not integrate diverse data features |
