Protocol for UAV fault diagnosis using signal processing and machine learning
Luttfi A Al-Haddad1, Alaa Abdulhady Jaber2, Nibras M Mahdi2
1Training and Workshops Center, University of Technology- Iraq, Baghdad 10066, Iraq.
STAR Protocols
|October 2, 2024
Summary
This study introduces a fault diagnosis protocol for unmanned aerial vehicles (UAVs) using signal processing and artificial intelligence. The method enables accurate fault detection in various UAV models through vibration analysis and machine learning.
Area of Science:
- Aerospace Engineering
- Artificial Intelligence
- Signal Processing
Background:
- Safe operation of unmanned aerial vehicles (UAVs) necessitates robust fault diagnosis systems.
- Existing methods may lack comprehensive approaches for real-time fault identification.
Purpose of the Study:
- To present a standardized protocol for UAV fault diagnosis.
- To leverage signal processing and artificial intelligence for enhanced fault detection accuracy.
Main Methods:
- Collecting vibration-based signal data using 3-axis accelerometers.
- Preprocessing data and extracting relevant features.
- Applying machine learning algorithms (deep neural networks, SVM, k-NN) for fault classification.
Main Results:
- The developed protocol demonstrates accurate fault detection capabilities.
- The methodology is adaptable to diverse UAV platforms.
Conclusions:
- The proposed protocol offers a reliable framework for UAV fault diagnosis.
- Integration of signal processing and AI enhances UAV safety and operational integrity.
