Development of UAV Failure Diagnosis Technology using EPs-Based Classification Model and EPs Networks

With the advancement of the UAV (Unmanned Aerial Vehicle) industry, the occurrence of unexplained UAV accidents has been continuously increasing. Traditional malfunction detection in UAV systems relies solely on sensor measurement monitoring and operator judgment, making accurate failure detection challenging. In this paper, an automatic UAV Failure Diagnosis technology was developed based on emerging patterns to detect the presence of imminent failures while monitoring UAV condition. Our studies showed that the EPs-based classifier is more accurate than other diagnosis method such as GBT. Additionally, proposed a failure diagnosis visualization method using a network structure, which allows for the analysis of failure causes and relationships.