Deep Reinforcement Learning for UAV Path Planning in Low-Altitude Environments
Advances in Engineering Technology Research · 20.03.2026 · 16 (1) · 146
DOI: 10.56028/aetr.16.1.146.2026 ↗Kurz gefasst
Fachartikel in der Zeitschrift Advances in Engineering Technology Research, online erschienen am 20.03.2026. Verfasst von Jing Guo und 5 weiteren Autor:innen.
Abstract (Originalsprache)
This paper addresses the UAV path planning problem in complex 3D environments and proposes a reinforcement learning method based on D3RQN (Dueling Double Deep Recurrent Q-Network). The approach extends the DQN framework by incorporating value–advantage decomposition, double Q-network estimation, and recurrent neural network modeling. A decomposed reward function is designed to balance efficiency, safety, and reachability by considering progress, energy consumption, altitude deviation, and obstacle proximity. To validate the proposed method, a 3D simulation environment with obstacles and disturbances was constructed, and comparative experiments were conducted against DDQN, D3QN, and DRQN. Experimental results demonstrate that D3RQN significantly outperforms the baselines in terms of convergence speed, success rate, and average return, achieving more efficient and stable path planning in complex scenarios. This study provides a practical solution for UAV autonomous navigation in low-altitude economy applications.
Zitieren
APA
Guo, J., Zhang, S., Yu, T., Xu, S., Xu, Y., & Wang, E. (2026). Deep Reinforcement Learning for UAV Path Planning in Low-Altitude Environments. Advances in Engineering Technology Research, 16(1), 146. https://doi.org/10.56028/aetr.16.1.146.2026
Vancouver
Guo J, Zhang S, Yu T, Xu S, Xu Y, Wang E. Deep Reinforcement Learning for UAV Path Planning in Low-Altitude Environments. Advances in Engineering Technology Research. 2026;16(1):146. doi:10.56028/aetr.16.1.146.2026
BibTeX
@article{Guo2026,
author = {Guo, Jing and Zhang, Shenglong and Yu, Tengli and Xu, Song and Xu, Yubin and Wang, Ershen},
title = {Deep Reinforcement Learning for UAV Path Planning in Low-Altitude Environments},
journal = {Advances in Engineering Technology Research},
year = {2026},
volume = {16},
number = {1},
pages = {146},
doi = {10.56028/aetr.16.1.146.2026}
}
EUVIMED