Research on Aircraft Brake Sensor Fault Diagnosis Based on Joint Learning of Hybrid Features

Takeoff and landing is a stage of frequent accidents for civil aircrafts, and braking is an important system that affects flight safety during this stage. One of the main issues that constrain the safe and reliable operation of modern brake systems is the difficulty in diagnosing multi-source heterogeneous coupling sensor signal faults. In this paper, a brake sensor signal fault diagnosis method based on joint learning of hybrid features is proposed. This method combines the probability distribution model and the SAE-CNN deep learning model. By the joint learning of the hybrid features composed of statistical features and deep semantic features, the abnormal signal feature submerged in multi-source, heterogeneous and coupling data are identified. The experimental results show that the method is of high efficiency, and improves the accuracy and robustness of the results, which enhances the safety and reliability of the brake system operation.