Pilot fatigue detection method based on artificial intelligence

The invention discloses a pilot fatigue detection method based on artificial intelligence. The pilot fatigue detection method comprises the steps that heart rate data and pulse signal data of a pilot are acquired and denoised; performing feature extraction on the de-noised data to obtain heart beat interval data; performing root mean square and standard deviation calculation to obtain RMSSD, SDNN and PNN50; performing feature extraction on the RMSSD, the SDNN and the PNN50 to obtain feature data; taking the feature data as input of a data set, selecting 80% of the feature data from the data set as a training set of a K-means algorithm, randomly selecting a plurality of feature data as a data set centroid, and training and iterating the data set to obtain different clusters arranged in sequence; taking the remaining 20% of feature data in the data set as a test set, testing, verifying a test result, and if the test result is judged to be accurate, taking the test result as a prediction result; and outputting a