UAV flight control method based on deep reinforcement learning
Conference paper
Bai, S., Li, B., Gan, Z. and Chen, D. (2021). UAV flight control method based on deep reinforcement learning. 2021 International Conference on Cyber-Physical Social Intelligence (ICCSI). Beijing, China 18 Dec 2021 - 20 Mar 2022 Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/ICCSI53130.2021.9736242
Authors | Bai, S., Li, B., Gan, Z. and Chen, D. |
---|---|
Type | Conference paper |
Abstract | Aiming at the intelligent perception and obstacle avoidance of UAV for the environment, an obstacle-avoidance flight decision method of UAV based on image information is proposed in this paper. Add Gate Recurrent Unit (GRU) to the neural network, and use the deep reinforcement learning algorithm DDPG to train the model. The special gates structure of GRU is utilized to memorize historical information, and acquire the variation law of the environment of UAV from the time sequential data including image information and UAV position and speed information to realize the dynamic perception of obstacles. Moreover, the basic framework and training method of the model are introduced, and the generalization ability of the model is tested. The experimental results show that the proposed method has better generalization ability and better adaptability to the environment. |
Keywords | GRU; DDPG; Image information |
Year | 2021 |
Publisher | Institute of Electrical and Electronics Engineers (IEEE) |
Digital Object Identifier (DOI) | https://doi.org/10.1109/ICCSI53130.2021.9736242 |
Accepted author manuscript | License File Access Level Open |
Publication dates | |
23 Mar 2022 | |
Publication process dates | |
Accepted | 18 Nov 2021 |
Deposited | 30 Mar 2022 |
Web address (URL) of conference proceedings | https://ieeexplore.ieee.org/xpl/conhome/9736193/proceeding |
https://openresearch.lsbu.ac.uk/item/8z98q
Download files
211
total views90
total downloads3
views this month1
downloads this month