Paper Abstract and Keywords |
Presentation |
2020-08-04 10:00
[Invited Talk]
Reinforcement Learning Based Channel Selection Algorithm for IoT Devices and Its Application to Wireless Sensor Network for Building Monitoring System So Hasegawa (NICT), Ryoma Kitagawa, Takumi Ito, Takashi Nakajima (TUS), Song-Ju Kim (KU), Yoshito Watanabe, Yozo Shoji (NICT), Mikio Hasegawa (TUS) CCS2020-13 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
The IoT wave have spread and the number of IoT devices have rapidly increased.
In IoT system using numerous IoT devices which generate enormous traffic, it is considered effective to introduce a multi-channel selection function to avoid communication congestion.
We have proposed a channel selection algorithm based reinforcement learning for IoT devices with limited computational resource.
Furthermore, We have confirmed IoT devices implemented proposed method learn the surrounding communication environment and select optimal channel by experiments.
In this paper, we describe the requirements for dynamic channel selection technology in IoT networks constructed with fixed or mobile devices, and explain the demonstration experiment of building monitoring using the devices implemented our proposed scheme. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Machine Learning / Reinforcement Learning / Multi-Armed Bandit / IoT / Distributed Channel Selection / Building Monitoring System / / |
Reference Info. |
IEICE Tech. Rep., vol. 120, no. 124, CCS2020-13, pp. 5-10, Aug. 2020. |
Paper # |
CCS2020-13 |
Date of Issue |
2020-07-27 (CCS) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
CCS2020-13 |
|