Prediction Method of Aircraft Maneuvering Range Based on Deep Learning Neural Network

In terms of orbit change docking, space rendezvous, and other issues, how to quickly and accurately evaluate the subsequent reachable range of the aircraft is the key to solving such application problems. The traditional method for this task is to use models and numerical integration, which is computationally complex. In this paper, a four-layer deep neural network is established to predict the subsequent reachable range of the aircraft under different maneuvering capabilities, and a relative state-based data preprocessing method is designed to provide training data for the network. The network is trained and tested using the data generated by the traditional model, and the error of the network prediction is analyzed. Using the trained network to predict the maneuverable reach of an aircraft takes about half the time compared to using traditional model integral calculations.