Summary
International Workshop on Smart Info-Media Systems in Asia
2022
Session Number:SS1
Session:
Number:SS1-6
ns3-ai: Rate Control for Wireless LAN by Deep Q-Network
Tomoki Nakashima, Leonardo Lanante Jr, Muhammad Harry Bintang Pratama, Masayuki Kurosaki, Hiroshi Ochi,
pp.27-32
Publication Date:2022/9/15
Online ISSN:2188-5079
Summary:
Transmission rate control in wireless LANs is one of the factors that affect communication quality. Many transmission rate control algorithms have been proposed in previous studies. However, there are cases where existing algorithms cannot adaptively control the rate due to the dynamics of wireless communication. In this paper, we propose a transmission rate control method based on Deep Q-Network (DQN), in which a DQN agent learns information about the communication environment and adaptively controls the transmission rate in response to the communication environment. We evaluate the proposed DQN-based transmission rate control by using the ns3-ai framework and the ns-3 network simulator. Simulations show that the proposed method improves throughput by up to 95\% compared to the Minstrel existing method.