The invention discloses a hypersonic
flight vehicle reentry cooperative guidance method based on
reinforcement learning, and particularly relates to a hypersonic
flight vehicle reentry cooperative guidance method based on
reinforcement learning. Establishing a hypersonic velocity reentry
kinetic model and a multi-constraint reentry model; designing an
attack angle profile and a height energy profile, and obtaining analytic solutions of the
attack angle and the heeling angle;
intelligent decision making is carried out on heeling angle symbols according to a DQN
algorithm, the action space of the heeling angle symbols is expanded, and a step-shaped mixed reward function is designed by considering time
collaboration and falling angle
collaboration; the hypersonic aircraft reentry cooperative guidance method based on intelligent
reinforcement learning is obtained by training a heeling angle
intelligent decision-making model offline and giving cooperative time and a cooperative fall angle to obtain a guidance instruction online, the defect that in an aircraft guidance strategy, heeling angle symbols are turned frequently is effectively overcome, time cooperation and fall angle cooperation are met, and the guidance efficiency is improved. A
simulation experiment verifies that the method can well consider time and fall angle cooperation to carry out multi-hypersonic-velocity aircraft guidance.