Flight maneuver intelligent recognition based on deep variational autoencoder network

Abstract The selection and training of aircraft pilots has high standards, long cycles, resource consumption, risk, elimination rate. It is the particularly urgent important requirement for current talent strategy national military to increase efficiency speed up all aspects pilot training, reduce cycle To this end, paper uses deep variational auto-encoder network adaptive dynamic time warping algorithms as support explore establishment an integrated evaluation system flight maneuver recognition quality evaluation, solve industry difficulty faced by data mining applications, achieve accurate reliable regimes under background mobility. will fully benefits existing airborne trainee pilots, personalized talents, rate elimination.