Wi-Fi 6 Cross-Technology Interference Detection and Mitigation by OFDMA: an Experimental Study

2025-03-07

AI Summary

To address performance degradation caused by cross-technology interference (CTI) in Wi-Fi 6, this paper proposes a lightweight, real-time CTI detection and scheduling co-design framework leveraging single-frame channel state information (CSI). Methodologically: (i) we empirically validate—on low-cost Wi-Fi 6 hardware—that CTI type identification is feasible using only a single CSI snapshot; (ii) we design a compact convolutional neural network (CNN) for low-overhead classification; and (iii) we develop a CTI-aware multi-user OFDMA scheduler, establishing a closed-loop “detect–feedback–avoid” mechanism. Our contributions include: (i) achieving high CTI classification accuracy in over-the-air experiments; (ii) reducing throughput loss by 35% compared to baseline schemes; and (iii) significantly improving multi-user spectral efficiency and system robustness under heterogeneous interference conditions.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Hardware-aware ML
• Application Domains: Internet of Things, Sensor Networks & Smart Cities

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Web Mining and Content Analysis: Web measurements
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Cross-Technology Interference (CTI) poses challenges for the performance and robustness of wireless networks. There are opportunities for better cooperation if the spectral occupation and technology of the interference can be detected. Namely, this information can help the Orthogonal Frequency Division Multiple Access (OFDMA) scheduler in IEEE 802.11ax (Wi-Fi 6) to efficiently allocate resources to multiple users inthe frequency domain. This work shows that a single Channel State Information (CSI) snapshot, which is used for packet demodulation in the receiver, is enough to detect and classify the type of CTI on low-cost Wi-Fi 6 hardware. We show the classification accuracy of a small Convolutional Neural Network (CNN) for different Signal-to-Noise Ratio (SNR) and Signal-to-Interference Ratio (SIR) with simulated data, as well as using a wired and over-the-air test with a professional wireless connectivity tester, while running the inference on the low-cost device. Furthermore, we use openwifi, a full-stack Wi-Fi transceiver running on software-defined radio (SDR) available in the w-iLab.t testbed, as Access Point (AP) to implement a CTI-aware multi-user OFDMA scheduler when the clients send CTI detection feedback to the AP. We show experimentally that it can fully mitigate the 35% throughput loss caused by CTI when the AP applies the appropriate scheduling.

Problem

• Detect and classify Cross-Technology Interference (CTI) using Wi-Fi 6 hardware.
• Mitigate throughput loss caused by CTI using OFDMA scheduling.
• Implement CTI-aware multi-user OFDMA scheduler with low-cost devices.

Innovation

• Uses CSI for CTI detection on Wi-Fi 6
• Implements CNN for interference classification
• CTI-aware OFDMA scheduler mitigates throughput loss