Enabling ZigBee Link Performance Robust Under Cross-Technology Interference

Abstract

The unlicensed ISM spectrum is becoming increasingly populated by ubiquitous wireless networks, such coexistence can inevitably incur the cross-technology interference (CTI) for the scarcity of spectrum resources. Existing approaches for defending against such interferences often modify hardware condition, which can be ineffective when encountering with the massive deployment of legacy ZigBee devices and uncertain WiFi APs. In this paper, we build an Adaptive Transmission Estimation (ATE) model to mitigate the negative effect on ZigBee side based on a channel quality quantification metric called Channel Idle Degree (CID) and logistic regression, which can indicate ZigBee devices to adjust packet rate efficiently. Particularly, our mechanism that is a lightweight, and software-level approach without any modification on the hardware of devices, can be easily implemented on the off-the-shelf ZigBee devices. Extensive experimental results show that, under the coverage of different WiFi traffic, our lightweight mechanism can achieve 4$\times$ and 1.5$\times$ performance gains over WISE and CII. Particularly, when WiFi traffic is 4 Mbps, we can still obtain over 90%. Furthermore, the energy consumption efficiency can be reduced markedly.