Calibrated RF-Fingerprinting Under Interference With Heterogeneous Transmission Protocols
Researchers propose a calibrated multi-label 1D-CNN RF-fingerprinting method that identifies overlapping Wi-Fi, LTE, and 5G transmitters despite co-channel interference.
The paper formulates RF-fingerprinting under co-channel interference as a multi-label classification problem solved with a 1D convolutional neural network, with calibration guarantees bounding average false negatives. It was validated on real data from the POWDER 5G testbed with 802.11a, 4G LTE, and 5G NR waveforms, achieving post-calibration accuracy between 73% and 97% depending on channel conditions. Calibration was robust to out-of-distribution interference, supporting spectrum policy enforcement in realistic high-contention wireless environments.
- Multi-label 1D CNN fingerprints simultaneous overlapping transmissions
- Calibration bounds average false negatives for spectrum policy violations
- Validated on POWDER 5G testbed across Wi-Fi, LTE, and 5G NR
- Accuracy 73-97% after calibration depending on channel conditions
From the original article
1 min read at arxiv.orgRadio Frequency(RF)-Fingerprinting is a spectrum monitoring technique that identifies specific transmitters based on hardware impairments imprinted within the emitted signal. Although widely researched, studies almost exclusively consider scenarios where only one transmitter is emitting at a time, limiting real world applicability.
By Tariq Abdul-Quddoos · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.