Experimental Evaluation of Rate Adaptation using Deep-Q-Network in IEEE 802.11 WLAN

Rate adaptation algorithm has an important role in the Wi-Fi network. It ensures that the nodes transmit at a suitable transmission rate with minimum packet errors on the receiving side. However, there are cases when the existing algorithms fail to adapt to the changes in the communication environment. In this paper, we propose a rate adaptation algorithm using a Deep Q-network (DQN), in which the DQN agent controls the transmission rate of a node in response to the communication environment. We also evaluate the proposed algorithm using the field-programmable gate array (FPGA) and software-defined radio (SDR). The experimental result shows that the proposed algorithm can adaptively select the suitable MCS and maintain the throughput.

I. Introduction

A high and stable data rate is one of the requirements required by new online services. However, any changes in the environment affect communication quality, including the data rate. In IEEE 802.11 standards, rate adaptation is a control mechanism that determines the throughput based on the communication environment. Among several transmission parameters, Modulation and Coding Scheme (MCS) plays a significant role in determining the data rate.