КОНФЕРЕНЦІЇ ВНТУ електронні наукові видання, 
Перспективи розвитку машинобудування та транспорту-2025

Розмір шрифта: 
REINFORCEMENT LEARNING FOR AUTONOMOUS DRONE NAVIGATION IN INDOOR ENVIRONMENTS
Gulzhan Bakytyvna Kashaganova, Birzhan Önerbayuly Kosherbayev, Byrzhan Dabayevna Sharipova

Остання редакція: 2025-06-11

Анотація


This paper presents a reinforcement learning (RL) approach for autonomous drone navigation in complex indoor environments. Traditional navigation systems struggle with dynamic layouts and GPS-denied conditions. By leveraging RL algorithms such as Deep Q-Networks (DQN), a drone can learn optimal policies for obstacle avoidance, path planning, and goal-reaching behavior through trial and error. Simulations conducted in a virtual 3D environment demonstrate the system’s ability to generalize navigation strategies across varied layouts, achieving efficient and collision-free flight without prior maps or human intervention.


Ключові слова


reinforcement learning, autonomous drone, indoor navigation, deep Q-network, obstacle avoidance, simulation, intelligent control

Посилання


  1. Mnih, V., Kavukcuoglu, K., Silver, D., et al. (2015). Human-level control through deep reinforcement learning. Nature, 518(7540), 529–533.
  2. Lillicrap, T. P., Hunt, J. J., Pritzel, A., et al. (2016). Continuous control with deep reinforcement learning. arXiv preprint arXiv:1509.02971.
  3. Kendall, A., Hawke, J., Janz, D., et al. (2019). Learning to drive in a day. 2019 International Conference on Robotics and Automation (ICRA), 8248–8254.
  4. Chen, Y. F., Liu, M., Everett, M., & How, J. P. (2017). Decentralized non-communicating multiagent collision avoidance with deep reinforcement learning. 2017 IEEE International Conference on Robotics and Automation (ICRA), 285–292.
  5. Garcia, J., & Fernández, F. (2015). A comprehensive survey on safe reinforcement learning. Journal of Machine Learning Research, 16(1), 1437–1480.
  6. Koenig, N., & Howard, A. (2004). Design and use paradigms for Gazebo, an open-source multi-robot simulator. 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2149–2154.
  7. Zhang, Q., et al. (2020). Autonomous drone navigation using deep reinforcement learning. Sensors, 20(12), 3400.

Повний текст: PDF