More Is Better: Data Augmentation for Channel-Resilient RF Fingerprinting

Soltani, Nasim; Sankhe, Kunal; Dy, Jennifer G.; Ioannidis, Stratis; Chowdhury, Kaushik R.

doi:10.1109/mcom.001.2000180

Cited by 84 publications

(58 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, we propose techniques that improve the model accuracy for all the devices in the dataset. Closer to our paper, Xie et al [52] and Soltani et al [53] presented DAG schemes for RFP. The latter introduced a scheme that involves DAG at the transmitter's side, which is not applicable to RFID tags.…”

Section: Background and Related Workmentioning

confidence: 72%

“…Moreover, as far as we know, no work has proposed the usage of federated machine learning (FML) to improve the performance of CNN-based RFP. Although this problem is extremely relevant, only recently it has received attention from the community [39,52,53]. Although very similar in target, the solution proposed in [39] assumes that the transmitter is able to somehow modify the transmitted waveform, which is not the case for RFID tags since they are passive devices.…”

Section: Background and Related Workmentioning

confidence: 99%

“…The latter introduced a scheme that involves DAG at the transmitter's side, which is not applicable to RFID tags. Moreover, the schemes in [52,53] are validated on simulated datasets with a population of only 10 devices. Neither of the datasets are shared, including the large-scale dataset in [53], which makes the results not replicable.…”

Section: Background and Related Workmentioning

confidence: 99%

“…Moreover, the schemes in [52,53] are validated on simulated datasets with a population of only 10 devices. Neither of the datasets are shared, including the large-scale dataset in [53], which makes the results not replicable.…”

Section: Background and Related Workmentioning

confidence: 99%

“…To perform RFP, it is possible to exploit DAG on both the devices involved in a communication, i.e. tag and reader [53]. However, to apply data augmentation on the tag side we would have to modify it, reducing the real applicability of this work.…”

Section: Data-augmented Radio Fingerprintingmentioning

confidence: 99%

See 4 more Smart Citations

The Tags Are Alright

Piva

Maselli

Restuccia

2021

Proceedings of the Twenty-Second International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Net

View full text Add to dashboard Cite

Millions of RFID tags are pervasively used all around the globe to inexpensively identify a wide variety of everyday-use objects. One of the key issues of RFID is that tags cannot use energy-hungry cryptography, and thus can be easily cloned. For this reason, radio fingerprinting (RFP) is a compelling approach that leverages the unique imperfections in the tag's wireless circuitry to achieve largescale RFID clone detection. Recent work, however, has unveiled that time-varying channel conditions can significantly decrease the accuracy of the RFP process. Prior art in RFID identification does not consider this critical aspect, and instead focuses on customtailored feature extraction techniques and data collection with static channel conditions. For this reason, we propose the first large-scale investigation into RFP of RFID tags with dynamic channel conditions. Specifically, we perform a massive data collection campaign on a testbed composed by 200 off-the-shelf identical RFID tags and a software-defined radio (SDR) tag reader. We collect data with different tag-reader distances in an over-the-air configuration. To emulate implanted RFID tags, we also collect data with two different kinds of porcine meat inserted between the tag and the reader. We use this rich dataset to train and test several convolutional neural network (CNN)-based classifiers in a variety of channel conditions. Our investigation reveals that training and testing on different channel conditions drastically degrades the classifier's accuracy. For this reason, we propose a novel training framework based on federated machine learning (FML) and data augmentation (DAG) to boost the accuracy. Extensive experimental results indicate that (i) our FML approach improves accuracy by up to 48%; (ii) our DAG approach improves the FML performance by up to 19% and the single-dataset performance by 31%. To the best of our knowledge, this is the first paper experimentally demonstrating the efficacy of FML and DAG on a large device population. To allow full replicability, we are sharing with the research community our fully-labeled 200-GB RFID waveform dataset, as well as the entirety of our code and trained models, concurrently with our submission. CCS CONCEPTS• Security and privacy → Network security.

show abstract

Section: Background and Related Workmentioning

confidence: 72%

Section: Background and Related Workmentioning

confidence: 99%