An Entropy-Based WLAN Channel Allocation using Channel State Information
Israel Elujide, Yonghe Liu · 2020
Low-cost access points have proliferated wireless local area networks (WLAN) providing main wireless access in many unmanaged networks. The majority of these APs rely on received signal strength indication (RSSI), known to be unreliable, as a measure of the wireless link quality. However, an accurate link measurement is a precursor to channel selection which in turn allows more efficient use of the wireless resources, especially in a crowded and dense wireless environment. In this paper, we present CSI-EWCA, an entropy-based WLAN channel allocation model using channel state information to combat the unreliability in RSSI. To develop a self-reliant system that is independent of CSI data for devices with low computational power, we develop a machine learning model to predict channel spectral entropy from physical layer network information extracted from the Linux kernel. Our experimental results show that CSI-EWCA can consistently select a channel with high throughput and low jitters and fewer retries.