skip to main content

Title: Radio Frequency Fingerprinting on the Edge
Deep learning methods have been very successful at radio frequency fingerprinting tasks, predicting the identity of transmitting devices with high accuracy. We study radio frequency fingerprinting deployments at resource-constrained edge devices. We use structured pruning to jointly train and sparsify neural networks tailored to edge hardware implementations. We compress convolutional layers by a 27.2× factor while incurring a negligible prediction accuracy decrease (less than 1%). We demonstrate the efficacy of our approach over multiple edge hardware platforms, including a Samsung Galaxy S10 phone and a Xilinx-ZCU104 FPGA. Our method yields significant inference speedups, 11.5× on the FPGA and 3× on the smartphone, as well as high efficiency: the FPGA processing time is 17× smaller than in a V100 GPU. To the best of our knowledge, we are the first to explore the possibility of compressing networks for radio frequency fingerprinting; as such, our experiments can be seen as a means of characterizing the informational capacity associated with this specific learning task.
; ; ; ; ; ; ; ; ;
Award ID(s):
Publication Date:
Journal Name:
IEEE Transactions on Mobile Computing
Page Range or eLocation-ID:
1 to 1
Sponsoring Org:
National Science Foundation
More Like this
  1. 5G and open radio access networks (Open RANs) will result in vendor-neutral hardware deployment that will require additional diligence towards managing security risks. This new paradigm will allow the same network infrastructure to support virtual network slices for transmit different waveforms, such as 5G New Radio, LTE, WiFi, at different times. In this multi- vendor, multi-protocol/waveform setting, we propose an additional physical layer authentication method that detects a specific emitter through a technique called as RF fingerprinting. Our deep learning approach uses convolutional neural networks augmented with triplet loss, where examples of similar/dissimilar signal samples are shown to the classifiermore »over the training duration. We demonstrate the feasibility of RF fingerprinting base stations over the large-scale over-the-air experimental POWDER platform in Salt Lake City, Utah, USA. Using real world datasets, we show how our approach overcomes the challenges posed by changing channel conditions and protocol choices with 99.86% detection accuracy for different training and testing days.« less
  2. Various hardware accelerators have been developed for energy-efficient and real-time inference of neural networks on edge devices. However, most training is done on high-performance GPUs or servers, and the huge memory and computing costs prevent training neural networks on edge devices. This paper proposes a novel tensor-based training framework, which offers orders-of-magnitude memory reduction in the training process. We propose a novel rank-adaptive tensorized neural network model, and design a hardware-friendly low-precision algorithm to train this model. We present an FPGA accelerator to demonstrate the benefits of this training method on edge devices. Our preliminary FPGA implementation achieves 59× speedupmore »and 123× energy reduction compared to embedded CPU, and 292× memory reduction over a standard full-size training.« less
  3. In this paper, we introduce a deep spiking delayed feedback reservoir (DFR) model to combine DFR with spiking neuros: DFRs are a new type of recurrent neural networks (RNNs) that are able to capture the temporal correlations in time series while spiking neurons are energy-efficient and biologically plausible neurons models. The introduced deep spiking DFR model is energy-efficient and has the capability of analyzing time series signals. The corresponding field programmable gate arrays (FPGA)-based hardware implementation of such deep spiking DFR model is introduced and the underlying energy-efficiency and recourse utilization are evaluated. Various spike encoding schemes are explored andmore »the optimal spike encoding scheme to analyze the time series has been identified. To be specific, we evaluate the performance of the introduced model using the spectrum occupancy time series data in MIMO-OFDM based cognitive radio (CR) in dynamic spectrum sharing (DSS) networks. In a MIMO-OFDM DSS system, available spectrum is very scarce and efficient utilization of spectrum is very essential. To improve the spectrum efficiency, the first step is to identify the frequency bands that are not utilized by the existing users so that a secondary user (SU) can use them for transmission. Due to the channel correlation as well as users' activities, there is a significant temporal correlation in the spectrum occupancy behavior of the frequency bands in different time slots. The introduced deep spiking DFR model is used to capture the temporal correlation of the spectrum occupancy time series and predict the idle/busy subcarriers in future time slots for potential spectrum access. Evaluation results suggest that our introduced model achieves higher area under curve (AUC) in the receiver operating characteristic (ROC) curve compared with the traditional energy detection-based strategies and the learning-based support vector machines (SVMs).« less
  4. The accurate identification of wireless devices is critical for enabling automated network access monitoring and authenticated data communication in large-scale networks; e.g., IoT networks. RF fingerprinting has emerged as a potential solution for device identification by leveraging the transmitter unique manufacturing impairments of the RF components. Although deep learning is proven efficient in classifying devices based on the hardware impairments, trained models perform poorly due to channel variations. That is, although training and testing neural networks using data generated during the same period achieve reliable classification, testing them on data generated at different times degrades the accuracy substantially. To themore »best of our knowledge, we are the first to propose to leverage MIMO capabilities to mitigate the channel effect and provide a channelresilient device classification. For the proposed technique we show that, for Rayleigh channels, blind partial channel estimation enabled by MIMO increases the testing accuracy by up to 40% when the models are trained and tested over the same channel, and by up to 60% when the models are tested on a channel that is different from that used for training.« less
  5. Cloud deployments now increasingly provision FPGA accelerators as part of virtual instances. While FPGAs are still essentially single-tenant, the growing demand for hardware acceleration will inevitably lead to the need for methods and architectures supporting FPGA multi-tenancy. In this paper, we propose an architecture supporting space-sharing of FPGA devices among multiple tenants in the cloud. The proposed architecture implements a network-on-chip (NoC) designed for fast data movement and low hardware footprint. Prototyping the proposed architecture on a Xilinx Virtex Ultrascale + demonstrated near specification maximum frequency for on-chip data movement and high throughput in virtual instance access to hardware accelerators.more »We demonstrate similar performance compared to single-tenant deployment while increasing FPGA utilization (we achieved 6× higher FPGA utilization with our case study), which is one of the major goals of virtualization. Overall, our NoC interconnect achieved about 2× higher maximum frequency than the state-of-the-art and a bandwidth of 25.6 Gbps« less