TY - GEN
T1 - Flight Control of a Canard Rotor/Wing VTOL Aircraft Based on Reinforcement Learning Under Transition Flight Mode
AU - Zhang, Yinong
AU - Zhao, Huimin
AU - Zhu, Dehai
AU - Wang, Ban
AU - Huang, Jiangtao
AU - Gao, Zhenghong
N1 - Publisher Copyright:
© 2023, Beijing HIWING Sci. and Tech. Info Inst.
PY - 2023
Y1 - 2023
N2 - The vertical takeoff and landing (VTOL) aircraft has developed rapidly in recent years due to its excellent performance in terms of takeoff, landing and cruising. As a relatively complete concept of VTOL aircraft, the canard rotor/wing (CRW) aircraft has the advantages of high hovering efficiency and fast cruising speed. In this paper, the concept of reinforcement learning is introduced, and a neural network controller which does not distinguish the flight mode and does not depend on the model is designed. Firstly, the dynamic model of the studied CRW aircraft in transition flight mode is established. Then, a pitch angle neural network control strategy is designed for transition flight mode based on the deep deterministic policy gradient (DDPG) algorithm. Finally, the effectiveness of the proposed pitch angle control scheme based on reinforcement learning is verified by simulation tests.
AB - The vertical takeoff and landing (VTOL) aircraft has developed rapidly in recent years due to its excellent performance in terms of takeoff, landing and cruising. As a relatively complete concept of VTOL aircraft, the canard rotor/wing (CRW) aircraft has the advantages of high hovering efficiency and fast cruising speed. In this paper, the concept of reinforcement learning is introduced, and a neural network controller which does not distinguish the flight mode and does not depend on the model is designed. Firstly, the dynamic model of the studied CRW aircraft in transition flight mode is established. Then, a pitch angle neural network control strategy is designed for transition flight mode based on the deep deterministic policy gradient (DDPG) algorithm. Finally, the effectiveness of the proposed pitch angle control scheme based on reinforcement learning is verified by simulation tests.
KW - Canard rotor/wing aircraft
KW - Reinforcement learning control
KW - Transition flight mode
KW - VTOL
UR - http://www.scopus.com/inward/record.url?scp=85151048645&partnerID=8YFLogxK
U2 - 10.1007/978-981-99-0479-2_184
DO - 10.1007/978-981-99-0479-2_184
M3 - 会议稿件
AN - SCOPUS:85151048645
SN - 9789819904785
T3 - Lecture Notes in Electrical Engineering
SP - 1985
EP - 1994
BT - Proceedings of 2022 International Conference on Autonomous Unmanned Systems, ICAUS 2022
A2 - Fu, Wenxing
A2 - Gu, Mancang
A2 - Niu, Yifeng
PB - Springer Science and Business Media Deutschland GmbH
T2 - International Conference on Autonomous Unmanned Systems, ICAUS 2022
Y2 - 23 September 2022 through 25 September 2022
ER -