A study of EEG-based flight phase recognition

During different flight phase, pilots are confronted with different situations, which lead to obvious difference of the electroencephalography (EEG) signals. Through EEG data the flight phase can be recognized, with the help of some equipments such as Synthetic Vision System (SVS), we can optimize cockpit information interaction, reduce pilots' operation mistakes, improve flight safety. For flight training, only well trained pilots can fly safely. However, the universal training programme mayn't be appropriate to every trainees. With the help of EEG-based flight recognition we can deduce the effectiveness of the simulation training. Combined with training index, we can also estimate whether the "overtraining" has happened, in which the trainees just repeat operating process mechanized without any high level cognition. This method will make great contributions to improve the training efficiency and help to create personalized training plans. In this paper, the main work is to verify and realize a novel method used to recognize flight phase. 20 subjects were selected to collect EEG signals. Different feature extraction methods and classifiers were combined to find a most effective combination. The classification accuracy reaches 0.980, which means this method is capable for flight phase recognition.