Best Papers
 Online Curriculum Reinforcement Learning Based UAV Training for Disaster Network Recovery 


Vol. 49,  No. 1, pp. 12-22, Jan.  2024
10.7840/kics.2024.49.1.12


PDF Full-Text
  Abstract

Extensive research has been conducted on the development of a flying ad-hoc network (FANET) utilizing unmanned aerial vehicles (UAVs) for the purpose of establishing a wireless network. The advantage of FANET lies in its ability to construct a network in a three-dimensional space, thereby offering a valuable solution even in situations where obstacles are confined to a two-dimensional plane. This unique characteristic of FANET enables it to be a promising solution for reconstructing networks in disaster or emergency scenarios. This paper presents a novel algorithm for reconstructing an ad-hoc network using UAVs, specifically addressing scenarios involving malfunctioning relay nodes that hinder the delivery of source data to the intended destination. To overcome the challenges associated with decentralized control, an autonomous decision-making solution is proposed for UAVs based on deep reinforcement learning. Furthermore, curriculum learning and recurrent neural networks are employed to tackle issues related to sparse and delayed rewards. To enhance practicality, the algorithm enables UAVs to make decisions based on partial and incomplete information about the surrounding environment. Experimental results demonstrate the effectiveness of the proposed algorithm in successfully reconstructing ad-hoc networks across diverse scenarios.

  Statistics
Cumulative Counts from November, 2022
Multiple requests among the same browser session are counted as one view. If you mouse over a chart, the values of data points will be shown.


  Related Articles
  Cite this article

[IEEE Style]

D. Lee, J. Eo, M. Kwon, "Online Curriculum Reinforcement Learning Based UAV Training for Disaster Network Recovery," The Journal of Korean Institute of Communications and Information Sciences, vol. 49, no. 1, pp. 12-22, 2024. DOI: 10.7840/kics.2024.49.1.12.

[ACM Style]

Dongsu Lee, Jeyeon Eo, and Minhae Kwon. 2024. Online Curriculum Reinforcement Learning Based UAV Training for Disaster Network Recovery. The Journal of Korean Institute of Communications and Information Sciences, 49, 1, (2024), 12-22. DOI: 10.7840/kics.2024.49.1.12.

[KICS Style]

Dongsu Lee, Jeyeon Eo, Minhae Kwon, "Online Curriculum Reinforcement Learning Based UAV Training for Disaster Network Recovery," The Journal of Korean Institute of Communications and Information Sciences, vol. 49, no. 1, pp. 12-22, 1. 2024. (https://doi.org/10.7840/kics.2024.49.1.12)
Vol. 49, No. 1 Index

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)






pISSN : 1226 - 4717
eISSN : 2287 - 3880