A Survey on Prediction of PQoS Using Machine Learning on Wi-Fi Networks

Abstract

Deployment of Wi-Fi networks increases considerably as a last hop for Internet access in homes and buildings. Wi-Fi as a best-effort network cannot satisfy user expectations for carrier-grade quality and today, users consider Internet service providers responsible for Wi-Fi quality degradation. Thus, notion of perceived quality of service (PQoS) and quality of experience (QoE) have been extensively used as quality indicators of user satisfaction. This paper provides a review of state-of-the-art literature to investigate machine learning techniques for estimating PQoS in Wi-Fi networks. In this paper, only literature that attempted to estimate QoE based on network performance parameters on Wi-Fi networks have been investigated and categorized based on their problem area and machine learning techniques. The review indicated promising achievements using machine learning techniques to identify wireless problems and accordingly improve the quality. However, estimation of PQoS requires further research with recent IEEE 802.11 technology advancements and machine learning approaches.