Deep Reinforcement Learning for Scan Recognition and Path Planning of Rescue UAVs

Minchao Ma

doi:10.62051/ijcsit.v3n2.40

Authors

Minchao Ma

DOI:

https://doi.org/10.62051/ijcsit.v3n2.40

Keywords:

Deep Reinforcement Learning, Object Recognition, Markov Decision Process, Path Planning

Abstract

In recent years, rapid advances in communication and navigation technologies have led to the research of drones in disaster relief becoming a controversial topic of interest. These drones are capable of autonomously performing high-risk tasks under extreme and treacherous conditions, such as searching and locating disaster victims, effectively replacing human intervention. The purpose of this essay is to discuss the basic overview of rescue UAVs, to find how deep reinforcement learning techniques on the MATLAB platform can be used to improve the effectiveness of UAV recognition scanning and path planning. Subsequently discuss the future development path and potential applications of rescue UAV technology in a forward-looking manner.

Downloads

Download data is not yet available.

References

WEI Mingxin, HUANG Hao, HU Yongming, WANG Dezhi, & LI Yuebin. (2020). Deep learning-based monocular visual target localization tracking method for multi-rotor UAVs. Computer Measurement and Control, 28(4), 156–160. https://doi.org/10.16526/j.cnki.11-4762/tp.2020.04.032

VOLODYMYR M, KORAY K, DAVID S, et al. Human-level control through deep reinforcement learning [J]. Nature, 2019, 518(7540):29-33.

LI R, FU L, WANG L, et al. Improved Q-learning based route planning method for UAVs in unknown environment[C/CD]//Proceedings in the 15th International Conference on Control and Automation (ICCA). 2019, Edinburgh, UK.

ZHOU Bin, GUO Yan, LI Ning, et al. Path planning of UAV using guided enhancement Q-learning algorithm [J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(9):506-513.