EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Temporally-Aware Deep Reinforcement Learning for Dynamic Obstacle Avoidance in UAVs

Chang Liu, Shan Wang

Drones · 03.07.2026 · 10 (7) · 505

DOI: 10.3390/drones10070505 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Drones, online erschienen am 03.07.2026. Verfasst von Chang Liu und 1 weiteren Autor:innen. Die Arbeit nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Autonomous obstacle avoidance for UAVs in dynamic obstacle-dominated complex environments must address time-varying local collision risks from multiple directions under the constraints imposed by local sensing, environmental uncertainty, execution safety, and limited onboard computation. Planning-based methods often require frequent replanning or explicit obstacle prediction, whereas conventional reinforcement learning policies may produce myopic decisions under partial observability. To address these limitations, this study proposes a dynamic obstacle-avoidance framework that combines a temporal LiDAR representation with safety-aware action correction in recurrent reinforcement learning. Multi-layer LiDAR observations are constructed using sector-wise minimum pooling. Adjacent two-frame stacking and a CNN-LSTM architecture are then used to extract local geometric structures and short-term dynamic cues, and a velocity-control policy is optimized using Recurrent PPO. In addition, a smooth velocity-projection safety shield is introduced to modify policy outputs and reduce collision risk during both training and policy execution. Experiments conducted in mixed static–dynamic obstacle scenarios based on Gym-PyBullet-Drones show that the proposed method achieves an average success rate of 91.9% across four test configurations, with an average online computation time of 0.78 ms. Ablation studies further support the contributions of two-frame observations, LSTM-based temporal modeling, and the safety shield.

Zitieren

APA

Liu, C., & Wang, S. (2026). Temporally-Aware Deep Reinforcement Learning for Dynamic Obstacle Avoidance in UAVs. Drones, 10(7), 505. https://doi.org/10.3390/drones10070505

Vancouver

Liu C, Wang S. Temporally-Aware Deep Reinforcement Learning for Dynamic Obstacle Avoidance in UAVs. Drones. 2026;10(7):505. doi:10.3390/drones10070505

BibTeX
@article{Liu2026,
  author = {Liu, Chang and Wang, Shan},
  title = {Temporally-Aware Deep Reinforcement Learning for Dynamic Obstacle Avoidance in UAVs},
  journal = {Drones},
  year = {2026},
  volume = {10},
  number = {7},
  pages = {505},
  doi = {10.3390/drones10070505}
}
Kostenlos starten →