Reinforcement Learning For Wifi Channel Access

Abstract

WiFi’s success is largely a testament to its cost-effectiveness, convenience and ease of integration with other networks. There is little more than a single access point needed for the initial setup of a wireless network. The lack of wires allows a suddenly-increased number of users to access network services within their networking environment from any convenient location. However, wireless networks have a frequent issue of losing packets that’s usually caused by poor WiFi signal, network interference and long distance connection. Packet loss is data lost during a transmission, leading to a slower connection than usual and reduced reliability of communication between Wireless Local Area Networks and devices. As a result, this study not only analyzes the problem, but also demonstrates the solution to maintain a reliable network connection within an area with multiple access points and devices, by using machine learning paradigm, particularly, Reinforcement Learning, and the program will recommend the appropriate channel for the access point from this process.