Fault Prediction of Firefighting Unmanned Aerial Vehicles Based on VMD-KAN-LSTM

This study introduces an integrated model combining VMD, KAN, and LSTM for fault prediction in firefighting drones during critical operations. VMD extracts multi-scale features, KAN captures nonlinear characteristics, and LSTM analyzes time series data to predict faults. The model shows higher accuracy and generalization than traditional methods, providing a novel health management solution for firefighting drones.