High-speed gliding aircraft online trajectory optimization method based on deep neural network

The invention discloses a high-speed gliding aircraft online trajectory optimization method based on a deep neural network, and relates to the technical field of trajectory optimization, and the method comprises the steps: converting a high-speed gliding aircraft dynamics model into an optimal control problem; an adaptive pseudo-spectral method is adopted to convert an optimal control problem into a nonlinear programming problem; generating an optimal trajectory data set according to different initial state information of the aircraft; dividing the optimal trajectory data set into a training set and a test set, and training an offline deep neural network through the training set; inputting state variables in the test set into the trained deep neural network to obtain control variables of the aircraft, updating a heeling angle and attack angle control instruction, and enabling the aircraft to fly according to the newly generated control instruction; and inputting aircraft state information measured by a naviga