A Dynamic Self-Adapting Mechanism for ZigBee Performance Assurance Under Wi-Fi Interference

As the increasing number of the wireless devices, the unlicensed industrial scientific medical band becomes more and more crowded. The highly uncertain cross technology interference (CTI) has shown to adversely affect the performance of low-power wireless sensor networks (WSN). Because of the incompatible physical and medium access control layer, the low-power WSN suffers serious packet loss during the transmission. It raises the need for agile methods that can assess the channel conditions and assure the performance and robustness of wireless sensor networks. In this paper, we present a light-weight and self-adapting CTI mitigation strategy that quantifies the channel idle state and improves the packet reception rate (PRR) of the communication. Based on the background received signal strength indicator, we propose a new metric to quantify the constantly changing channel state in the temporal domain. We employ logistic regression as the machine learning classifier to decide whether the communication is viable over the interfered link. Through the collected Wi-Fi traffic, we find that the channel is idle in most of the time, which motivated the design of our scheme. Extensive experiments show that our scheme can achieve over 91% of the PRR, which gain the improvement of 40%. Carrying on further, our scheme consumes less energy via degrading packet loss rate in the energy consumption part.