The Control Method of Hypersonic Vehicle Based on Machine Learning

Hypersonic aircraft has nowadays become the research hot button in various countries. However, since the control system of hypersonic aircraft is complicated, many effects such as nonlinear factors, state coupling, aerodynamic parameter changes, etc., and also the flight environment and aerodynamic characteristics change drastically, traditional flight control design methods are not capable anymore. For avoiding the application limitations of existing hypersonic aircraft attitude control methods, this paper proposes a hypersonic aircraft attitude control method based on deep reinforcement learning methods, which provides stably control the aircraft in flight scenarios with uncertain environment and drastic changes in aerodynamic characteristics attitude.