Research on UAV autonomous deformation strategy based on deep learning

To improve the autonomous deformation capability of morphing aircraft, a new method was proposed, which established an autonomous deformation strategy knowledge base based on deep learning. A longitudinal nonlinear dynamic model of variable-sweep-wing unmanned aerial vehicle (UAV) was obtained by using multi-rigid-body modeling. DATCOM was used to perform aerodynamic analysis, and then aerodynamic coefficient fitting formula was calculated. Based on prior knowledge, a deformation strategy was designed, which could generate training data for deep learning. A deep neural network (DNN), so-called autonomous deformation strategy knowledge base, was trained by using these data, which could output optimal sweep angle according to current flight states and flight mission. Simulation results show that this DNN gives morphing U A V a better speed maneuverability in deceleration task and a better directional maneuverability in turning task, which can enhance UA V's combat capability.