An investigation of pilot emotion change detection based on multimodal physiological signals

Aviation safety assurance is the most important point in the course of civil aviation flight. During the flight, the control of the aircraft and the handling of emergencies are also closely related to the emotional changes of pilots. This paper collects pilot physiological signals, performs signal preprocessing, signal feature extraction, classification, etc., and finally establishes a model of emotion-physiological signal. The classification accuracy of this model for Valence, Arousal and Dominance is 0.9413, 0.9735 and 0.9804, respectively, which proves the feasibility of using biological signals to classify pilot emotions.