This paper examines the utilization of reinforcement learning, a form of intelligent technology, in the control of high-speed morphing flight vehicles (HMFVs). The proposed control scheme incorporates a hybrid drive control structure comprising ‘basic control’ and ‘intelligent compensation’. The open architecture allows for flexibility in selecting and designing the base controller, while the intelligent compensation controller employs the TD3 algorithm. The reward function incorporates dynamic constraints and penalties for unreasonable agent outputs, ensuring safe control. In state selection, the agent utilizes multiple control period data to capture differential and integral information, addressing traditional reinforcement learning’s steady-state error issues. After rigorous comparative simulation validation, the intelligent compensation control method presented in this study demonstrates a significant enhancement in the control performance of morphing aircraft during high-dynamic flight conditions, outperforming traditional methods. This advancement not only underscores the efficacy of the proposed approach but also contributes to the advancement of the field and potential real-world applications of morphing aircraft technology.
Reinforcement Learning Control Based TD3 for High-Speed Morphing Flight Vehicle
Lect. Notes Electrical Eng.
International Conference on Autonomous Unmanned Systems ; 2024 ; Shenyang, China September 19, 2024 - September 21, 2024
Proceedings of 4th 2024 International Conference on Autonomous Unmanned Systems (4th ICAUS 2024) ; Chapter : 11 ; 102-115
2025-04-09
14 pages
Article/Chapter (Book)
Electronic Resource
English
RESEARCH ON FLIGHT CONTROL METHOD OF MORPHING UCAV BASED ON REINFORCEMENT LEARNING
British Library Conference Proceedings | 2010
|