A Generalizable Model-and-Data Driven Approach for Open-Set RFF Authentication

Xie, Renjie; Xu, Wei; Chen, Yanzhi; Yu, Jongmin; Hu, Aiqun; Ng, Derrick Wing Kwan; Swindlehurst, A. Lee

doi:10.1109/tifs.2021.3106166

Cited by 73 publications

(15 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The subtle fixed frequency in the baseband signal is proved unreliable, which needs to be removed before feature extraction [29] [35]. And in [36], the authors utilize a model-and-data driven method for emitter identification, where synchronization is performed by a novel neural operator. In our research, the discrete Fourier transform (DFT) and the chirp z-transform (CZT) are utilized to estimate the carrier frequency offset (CFO).…”

Section: B Generation Of Paired Samplementioning

confidence: 99%

Specific Emitter Identification Based on Paired Sample and Complex Fourier Neural Operator

Zha¹,

Li²,

Qiu³

et al. 2022

Preprint

View full text Add to dashboard Cite

<p>Specific emitter identification (SEI) techniques utilize the physical-layer fingerprints embedded in their received signals to identify the unique emitter. Fingerprints originate from hardware impairments during transmitter manufacturing. Considering the different manifestations of fingerprints in different domain perspectives, a novel neural operator with the time and frequency domain attention mechanism is proposed, which is named the complex Fourier neural operator (CFNO). And to further enhance the individual discriminability, the demodulation reconstruction waveform and the synchronized waveform are converted into a new sample format, named the paired sample. With the paired sample, the proposed network acquires prior communication knowledge and a baseline of the ideal transmitter. In this work, different network structures are investigated. And performances of the paired sample are fully evaluated. Compared with existing SEI methods based on deep learning models, results show the effectiveness of the CFNO structure, as well as the proposed paired sample scheme. The significance of this paper is that the proposed framework fully utilizes the domain knowledge in the communication field, resulting in excellent performance improvement. </p>

show abstract

Section: B Generation Of Paired Samplementioning

confidence: 99%

Specific Emitter Identification Based on Paired Sample and Complex Fourier Neural Operator

Zha¹,

Li²,

Qiu³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…In order to retain device-relevant information and to obtain discriminative RFFs, a maximum likelihood (ML) RFF extractor was previously proposed in [13]. Given a training set T = {(x i , y i )} N i=1 with N samples, the ML RFF extractor F (•) in [13] is obtained by solving the optimization problem:…”

Section: B ML Rff Extractormentioning

confidence: 99%

“…where 0 ≤ λ < 1 is a hyper-parameter that balances the learning effects for the raw and augmented signals. The first term in the objective function of ( 9), measuring the amount of device-relevant information extracted from the raw signal, is the same RFF learning objective as the one in our previous work [13]. The second term is the objective corresponding to the proposed augmented training.…”

Section: B Learning Dr-rff Extractor F (•)mentioning

confidence: 99%

“…For instance, in [11], [12], the authors extracted the RFFs using a convolutional neural network (CNN), a popular type of deep neural network (DNN). In our previous work [13], a softmax-based metric learning approach was adopted for enabling device authentication by strengthening the generalizability of the DL-based RFFs, such that unknown devices can be recognized even with the existence of unknown device aging. Simliarly, in [14], a metric learning approach based on triplet loss was proposed to enable unknown device authentication.…”

Section: Introductionmentioning

confidence: 99%

“…Thanks to the ability of DNNs to learn the RFF features themselves, the classification performance of these DL-based RFFs is noticeably better than that obtained with handcrafted RFFs. Despite these promising improvements, the above mentioned DL-based methods, e.g., [11]- [13], are typically trained as a maximum-likelihood (ML) estimator. In particular, their performance heavily relies on the quality of the collected training data.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Disentangled Representation Learning for RF Fingerprint Extraction under Unknown Channel Statistics

Xie¹,

Wang²,

Yu³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Deep learning (DL) applied to a device's radio-frequency fingerprint (RFF) has attracted significant attention in physical-layer authentications due to its extraordinary classification performance. Conventional DL-RFF techniques, trained by adopting maximum likelihood estimation (MLE), tend to overfit the channel statistics embedded in the training dataset. This restricts their practical applications as it is challenging to collect sufficient training data capturing the characteristics of all possible wireless channel environments. To address this challenge, we propose a DL framework of disentangled representation learning (DRL) that first learns to factor the input signals into a device-relevant component and a device-irrelevant component via adversarial learning. Then, it synthesizes a set of augmented signals by shuffling these two parts within a given training dataset for training of subsequent RFF extractor. The implicit data augmentation in the proposed framework imposes a regularization on the RFF extractor to avoid the possible overfitting of device-irrelevant channel statistics, without collecting additional data from unknown channels. Experiments validate that the proposed approach, referred to as DR-RFF, outperforms conventional methods in terms of generalizability to unknown complicated propagation environments, e.g., dispersive multipath fading channels, even though all the training data are collected in a simple environment with dominated direct line-of-sight (LoS) propagation paths.

show abstract

Robust Anomaly Detection via Radio Fingerprinting in LoRa-Enabled IIoT

Halder

Newe²

2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

A Generalizable Model-and-Data Driven Approach for Open-Set RFF Authentication

Cited by 73 publications

References 32 publications

Specific Emitter Identification Based on Paired Sample and Complex Fourier Neural Operator

Specific Emitter Identification Based on Paired Sample and Complex Fourier Neural Operator

Disentangled Representation Learning for RF Fingerprint Extraction under Unknown Channel Statistics

Robust Anomaly Detection via Radio Fingerprinting in LoRa-Enabled IIoT

Contact Info

Product

Resources

About