Aircraft sensor Fault Diagnosis Method Based on Residual Antagonism Transfer Learning

With the rapid development of electronic information technology, aircraft has entered a completely electrified era, and the number of sensors has increased exponentially. Although key sensors have redundant designs, many aviation accidents in recent years are caused by sensor failures. Therefore, early detection of Aircraft sensor faults is of great significance for ensuring flight safety. Faced with a large number of unlabeled and uneven sample sensor data, a method for fault diagnosis of Aircraft sensors based on residual countermeasure migration learning is proposed. This method can help deep learning. The product neural network requires the limitation of a large number of labeled data, and uses the rich label data from different but related auxiliary fields to reuse and transfer the data of the target domain to achieve the purpose of transfer learning and realize the fault diagnosis of Aircraft sensors.