TY - GEN
T1 - The Trajectory Re-Planning Method of the Hyper-Sonic Vehicle Facing Sudden Threats
AU - Ren, Jie
AU - Yu, Jianglong
AU - Dong, Xiwang
AU - Ren, Zhang
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2024
Y1 - 2024
N2 - Aiming at the trajectory re-planning problem of multiple hyper-sonic vehicles with the time coordination under sudden threats, a time-coordinated fast trajectory re-planning method was proposed to avoid the no-fly zone. Firstly, the three-dimensional trajectory planning problem of the vehicle is decomposed into the longitudinal trajectory planning problem and the transverse trajectory planning problem. Secondly, a single sudden threat is regarded as a circular no-fly zone, and the circumvention of no-fly zone can only be considered in a transverse trajectory planning. Based on the in-depth study of the longitudinal trajectory planning with flight time constraints and range constraints, a reinforcement learning method is introduced into the transverse trajectory re-planning, and a fast tilting angle flipping self decision method based on the reinforcement learning is proposed. Finally, by training DQN to generate a tilting angle reversal decision maker, a fast three-dimensional trajectory re-planning method that can avoid the no-fly zone is realized. The proposed trajectory re-planning method not only has a fast trajectory re-planning speed, but also greatly improves the penetration probability against sudden threats. The simulation results show the feasibility and effectiveness of the proposed method.
AB - Aiming at the trajectory re-planning problem of multiple hyper-sonic vehicles with the time coordination under sudden threats, a time-coordinated fast trajectory re-planning method was proposed to avoid the no-fly zone. Firstly, the three-dimensional trajectory planning problem of the vehicle is decomposed into the longitudinal trajectory planning problem and the transverse trajectory planning problem. Secondly, a single sudden threat is regarded as a circular no-fly zone, and the circumvention of no-fly zone can only be considered in a transverse trajectory planning. Based on the in-depth study of the longitudinal trajectory planning with flight time constraints and range constraints, a reinforcement learning method is introduced into the transverse trajectory re-planning, and a fast tilting angle flipping self decision method based on the reinforcement learning is proposed. Finally, by training DQN to generate a tilting angle reversal decision maker, a fast three-dimensional trajectory re-planning method that can avoid the no-fly zone is realized. The proposed trajectory re-planning method not only has a fast trajectory re-planning speed, but also greatly improves the penetration probability against sudden threats. The simulation results show the feasibility and effectiveness of the proposed method.
UR - https://www.scopus.com/pages/publications/85200382020
U2 - 10.1109/ICCA62789.2024.10591861
DO - 10.1109/ICCA62789.2024.10591861
M3 - 会议稿件
AN - SCOPUS:85200382020
T3 - IEEE International Conference on Control and Automation, ICCA
SP - 449
EP - 454
BT - 2024 IEEE 18th International Conference on Control and Automation, ICCA 2024
PB - IEEE Computer Society
T2 - 18th IEEE International Conference on Control and Automation, ICCA 2024
Y2 - 18 June 2024 through 21 June 2024
ER -