Research on Fatigue Classification of Flight Simulation Training

Abstract Fatigue is an important factor affecting modern flight safety. It can easily lead to a decline in the pilot's operational capabilities, misjudgments and illusions. Moreover, it may even cause serious accidents. In this paper, wearable wireless physiological device used obtain pilot electrocardiogram data simulated experiments. Bioelectric signals have higher reliability than image information, are not affected by external environment (such as shooting angle light intensity). On other hand, neural networks been widely various classification regression tasks. study, EEG was collected driving simulator, after simple filtering preprocessing, time domain sent directly convolutional network, eliminating need for additional feature extraction operations. We found that network effectively amplify fluctuation details of train fatigue state recognition model. The results show accuracy model reaches 98%, which 26% 12% traditional k-nearest neighbor algorithm (KNN) support vector machine (SVM) model, respectively. based on established paper suitable status. This has practical significance reducing accidents caused fatigue, provides theoretical basis risk management development intelligent aircraft autopilot systems.