Department of Mechanical Engineering, Faculty of Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
10.24200/sci.2025.66507.10094
Abstract
The automated guided vehicle path planning performance is highly dependent on the collected LiDAR data. This research centered on studying the LiDAR specifications for automated guided vehicle path planning problem using deep reinforcement learning. Three data collection approaches with differences in the data sample size and scanning range were considered to obtain 15 LiDAR specifications, and the training quality and the path planning performance were evaluated in order to find the optimal LiDAR specification. The most effective results were achieved using the approach of collecting the "distance and angle" pairs of the nearest obstacles. This method consistently reached all the defined targets in both seen and unseen static environments, with no collisions. Additionally, only limited collisions happened in an unseen dynamic environment containing moving obstacles. Notably, increasing the number of data samples can extend the training time, and the excessive data may also complicate decision-making with no significant benefits in terms of the path planning performance.
Sheikhbahaei, R. , Fallahi, M. H. and Yousefsani, S. A. (2025). Study of LiDAR specifications in automated guided vehicle path planning by deep reinforcement learning. Scientia Iranica, (), -. doi: 10.24200/sci.2025.66507.10094
MLA
Sheikhbahaei, R. , , Fallahi, M. H. , and Yousefsani, S. A. . "Study of LiDAR specifications in automated guided vehicle path planning by deep reinforcement learning", Scientia Iranica, , , 2025, -. doi: 10.24200/sci.2025.66507.10094
HARVARD
Sheikhbahaei, R., Fallahi, M. H., Yousefsani, S. A. (2025). 'Study of LiDAR specifications in automated guided vehicle path planning by deep reinforcement learning', Scientia Iranica, (), pp. -. doi: 10.24200/sci.2025.66507.10094
CHICAGO
R. Sheikhbahaei , M. H. Fallahi and S. A. Yousefsani, "Study of LiDAR specifications in automated guided vehicle path planning by deep reinforcement learning," Scientia Iranica, (2025): -, doi: 10.24200/sci.2025.66507.10094
VANCOUVER
Sheikhbahaei, R., Fallahi, M. H., Yousefsani, S. A. Study of LiDAR specifications in automated guided vehicle path planning by deep reinforcement learning. Scientia Iranica, 2025; (): -. doi: 10.24200/sci.2025.66507.10094
)