Deep Learning-based RF fingerprinting approaches struggle to perform well in cross-domain scenarios, particularly during hardware warm-up. This often-overlooked vulnerability has been jeopardizing their reliability and their adoption in practical settings. To address this critical gap, in this work, we first dive deep into the anatomy of RF fingerprints, revealing insights into the temporal fingerprinting variations during and post hardware stabilization. Introducing HEEDFUL, a novel framework harnessing sequential transfer learning and targeted impairment estimation, we then address these challenges with remarkable consistency, eliminating blind spots even during challenging warm-up phases. Our evaluation showcases HEEDFUL's efficacy, achieving remarkable classification accuracies of up to 96% during the initial device operation intervals-far surpassing traditional models. Furthermore, cross-day and cross-protocol assessments confirm HEEDFUL's superiority, achieving and maintaining high accuracy during both the stable and initial warm-up phases when tested on WiFi signals. Additionally, we release WiFi type B and N RF fingerprint datasets that, for the first time, incorporate both the time-domain representation and real hardware impairments of the frames. This underscores the importance of leveraging hardware impairment data, enabling a deeper understanding of fingerprints and facilitating the development of more robust RF fingerprinting solutions.

We leverage realistic datasets and evaluate online learning strategies for downlink beam-pair selection in mmWave networks, including classical contextual-bandit methods, deep reinforcement learning approaches, and our proposed two-stage approach, termed warm-start Neural-Linear Thompson Sampling (NLTS). Unlike standard methods that learn solely online, our approach pre-trains a neural network offline using past channel measurements and signal strengths, then performs lightweight Bayesian updates in the learned feature space during deployment. Experiments show that our proposed warm-start NLTS significantly accelerates learning, improves beam-pair selection accuracy, and enhances spectral efficiency outperforming all baselines with only a modest increase in the decision latency. These findings highlight the potential of combining supervised pre-training with online bandits for faster, more reliable beam-forming in next-generation wireless networks.

Radio Frequency Fingerprinting (RFFP) has emerged as a promising approach for device authentication by exploiting hardware-specific impairments embedded in transmitted signals. Yet existing methods largely overlook a major drawback: RFFP sensitivity to temperature–a critical factor influenced by both internal and environmental conditions–which can significantly alter device signatures and degrade classification performance. In this paper, we propose a novel temperature-aware RFFP framework that explicitly incorporates device temperature information into the learning process to improve robustness and generalization. We evaluate the proposed method on a real-world Bluetooth Low Energy (BLE) dataset collected across multiple devices and environmental conditions. Experimental results demonstrate that temperature-aware modeling consistently outperforms other temperature mitigation baselines, achieving significant improvements in classification accuracy, particularly under unseen temperature and environmental conditions.

The U.S. Federal Public Key Infrastructures (FPKI) relies on digital certificates and a network of Certificate Authorities (CAs) to build trust between federal agencies and their commercial partners. However, the current certificate generation process of FPKI faces two critical issues: The emerging quantum threats that jeopardize the security of conventionalsecure digital signatures and the systemic vulnerabilities introduced by centralized signing operations. While threshold PostQuantum (PQ) signatures offer a promising foundation for compromise-resilient trust, their adoption within the FPKI remains constrained by strict standard compliance requirements. ML-DSA is a prominent construction for thresholding within the NIST-PQ standards. However, existing threshold constructions of ML-DSA diverge from the NIST standard, limiting algorithmic compliance and hindering adoption in regulated settings such as FPKI, which require thorough cryptanalysis and revalidation. Hence, no fully standard-compliant threshold PQ solution currently meets FPKI's need for a distributed and quantum-resistant trust infrastructure. We introduce SHIELD, an efficient threshold NIST-PQcompliant signature framework that replaces its classical centralized counterpart in certificate generation within the FPKI. SHIELD provides several key properties that overcome the limitations of prior approaches. First, it achieves algorithmic FIPS204 compliance and seamless functional interchangeability with standard ML-DSA, without altering the signing algorithm. Second, it conforms to FPKI's structural and operational requirements, provides high security against adversarial environments, and prevents fraudulent certificates by mitigating single-key compromise risks. Finally, we provide the opensource implementation of SHIELD. We empirically evaluated the performance of SHIELD under diverse network latency conditions to validate its practical conformance with FPKI operational needs.

This paper presents a novel impersonation attack framework that aims to fool RF Fingerprinting (RFFP) identification systems by synthesizing signals that replicate the hardware-specific impairments of a target device. Our framework leverages unsupervised learning to enable accurate impairment estimation, combined with signal processing-based generation to synthesize high-fidelity adversarial signals. Unlike prior works that assume full access to the legitimate (victim) RFFP classifier, we consider a more realistic attack strategy where the adversary performs the attack from a completely different transceiver hardware. We further evaluate our proposed attack under realistic and challenging deployment settings, including over-the-air transmission in both Line-of-Sight (LoS) and Non-Line-of-Sight (NLoS) scenarios. Extensive experiments conducted on a Bluetooth Low Energy (BLE) device testbed demonstrate that our attacks remain highly effective even under severe access constraints, significantly outperforming existing baselines in terms of targeted attack success rates by over 80

The reciprocity of channel state information (CSI) collected by two devices communicating over a wireless channel has been leveraged to provide security solutions to resource-limited IoT devices. Despite the extensive research that has been done on this topic, much of the focus has been on theoretical and simulation analysis. However, these security solutions face key implementation challenges, mostly pertaining to limitations of IoT hardware and variations of channel conditions, limiting their practical adoption. To address this research gap, we revisit the channel reciprocity assumption from an experimental standpoint using resource-constrained devices. Our experimental study reveals a significant degradation in channel reciprocity for low-cost devices due to the varying channel conditions. Through experimental investigations, we first identify key practical causes for the degraded channel reciprocity. We then propose a new wavelet-based CSI reconstruction technique using wavelet coherence and time-lagged cross-correlation to construct CSI data that are consistent between the two participating devices, resulting in significant improvement in channel reciprocity. Additionally, we propose a secret-key generation scheme that exploits the wavelet-based CSI reconstruction, yielding significant increase in the key generation rates. Finally, we propose a technique that exploits CSI temporal variations to enhance device authentication resiliency through effective detection of replay attacks.

Machine learning (ML)-based network anomaly detection methods are proven to provide automated network protection from traffic misbehavior and authorized system access through data monitoring and analysis. However, conventional centralized methods present risks for data privacy and breaches. By facilitating distributed model training over a number of network nodes, Federated Learning (FL) emerges as a key enabler for effective anomaly detection yet while preserving the privacy of the data. This paper studies FL-based detection approaches under two different Deep Learning models, CNN and MLP. We use XGBoost for feature selection and the two UNSWNB15 and CICDDoS2019 datasets for assessing the effectiveness of each model through the evaluation of standard performance metric criteria, namely the recall, precision, accuracy, and F1score metrics. Our experimental findings indicate that integrating XGBoost-based feature selection with the CNN model yields superior performance on the UNSW-NB15 dataset, whereas the MLP model benefits more from the same integration when applied to the CICDDoS2019 dataset.

Barrett modular reduction and multiplication are essential primitives for efficient modular computation in cryptographic schemes, including post-quantum standards such as machine learning (ML) key encapsulation mechanism (KEM) and ML-digital signature algorithm (DSA). To protect against faults that compromise correctness and security, we introduce the first efficient fault-detection mechanisms tailored to these operations. For modular reduction, we leverage word-based representations with compact word-sum checks that exploit algebraic input-output relations to ensure computational integrity. For modular multiplication, we adopt a tunable hybrid strategy: early stages apply word-sum checks, while later stages use partial recomputation with encoded inputs, providing robust protection against injected faults. Formal analysis, fault-injection simulations, and hardware/software implementations show that our methods detect a wide range of faults with minimal performance and area overhead. Evaluation results demonstrate overheads of 3.43% and 7.15% for 512-bit and 1024-bit inputs in modular reduction, and 26.47% and 27.22% for 2048-bit inputs in modular multiplication in the number of clock cycles in software. Moreover, in hardware, we observed reasonable overheads: less than 27.5% in area and 2.1% in delay for modular reduction, and less than 23.5% in area and 16.2% in delay for modular multiplication. These results confirm the practicality of our methods for secure yet efficient integration.

RF data-driven device fingerprinting through the use of deep learning has recently surfaced as a potential solution for automated network access authentication. Traditional approaches are commonly susceptible to the domain adaptation problem where a model trained on data from one domain performs badly when tested on data from a different domain. Some examples of a domain change include varying the device location or environment and varying the time or day of data collection. In this work, we propose using multifractal analysis and the variance fractal dimension trajectory (VFDT) as a data representation input to the deep neural network to extract device fingerprints that are domain generalizable. We analyze the effectiveness of the proposed VFDT representation in detecting device-specific signatures from hardware-impaired IQ signals, and evaluate its robustness in real-world settings, using an experimental testbed of 30 WiFi-enabled Pycom devices under different locations and at different scales. Our results show that the VFDT representation improves the scalability, robustness and generalizability of the deep learning models significantly compared to when using raw IQ data.

Secret-key generation and agreement based on wireless channel reciprocity offers a promising avenue for securing IoT networks. However, existing approaches predominantly rely on the similarity of instantaneous channel measurement samples between communicating devices. This narrow view of reciprocity is often impractical, as it is highly susceptible to noise, asynchronous sampling, channel fading, and other system-level imperfections – all of which significantly impair key generation performance. Furthermore, the quantization step common in traditional schemes introduces irreversible errors, further limiting efficiency. In this work, we propose a novel approach for secret-key generation by using wavelet scattering networks to extract robust and reciprocal CSI features. Dimensionality reduction is applied to uncover hidden cluster structures, which are then used to build hidden Markov models for efficient key agreement. Our approach eliminates the need for quantization and effectively captures channel randomness. It achieves a 5x improvement in key generation rate compared to traditional benchmarks, providing a secure and efficient solution for key generation in resource-constrained IoT environments.

Deep learning-based radio frequency fingerprinting (RFFP) has become an enabling physical-layer security technology, allowing device identification and authentication through received RF signals. This technology, however, faces significant challenges when it comes to adapting to domain variations, such as time, location, environment, receiver and channel. For Bluetooth Low Energy (BLE) devices, addressing these challenges is particularly crucial due to the BLE protocol's frequency-hopping nature. In this work, and for the first time, we investigated the frequency hopping effect on RFFP of BLE devices, and proposed a novel, low-cost, domain-adaptive feature extraction method. Our approach improves the classification accuracy by up to 58% across environments and up to 80% across receivers compared to existing benchmarks.

Detecting irregular patterns that suggest potential threats or system flaws plays a vital role in anomaly detection, which is essential for maintaining the security and integrity of network systems, especially in IoT environments where devices are often vulnerable and widely distributed. This study uses the UNSW-NB15 dataset, a vast collection of network traffic data, to examine several Machine Learning (ML) and Deep Learning (DL) approaches for network anomaly detection. In order to improve model performances, the study uses strategies such the Synthetic Minority Over-Sampling Technique (SMOTE) to address the issue of class imbalance. The capabilities of a number of widely used machine learning (ML) algorithms, such as Decision Trees, Random Forests, KNN, XGB, and widely used deep learning (DL) models, such as CNN, ANN and LSTM, to identify unusual patterns in various attack and typical behavior scenarios are assessed. Our findings provide important information for future advancements in anomaly detection techniques by highlighting the significance of feature selection, class balancing, and model resilience in successfully differentiating unusual behaviors in network traffic. XGB emerged as the most successful approach in this study, with its enhanced performance largely attributed to the synergy between its robust ensemble framework and the class balancing achieved through SMOTE.

Power Line Communication (PLC) systems are facing increasing security threats as adversaries leverage low-cost Software-Defined Radios (SDRs) to launch physical-layer attacks, e.g., jamming and Radio Frequency Fingerprinting (RFF), for communication disruption and unauthorized device tracking, respectively. This paper investigates the dual role of Radio Frequency (RF) wireless jamming for PLC environments, through two distinct scenarios: (i) friendly RF jamming for privacy preservation of (cabled) PLC devices against unauthorized RFF, and (ii) adversarial RF jamming to degrade the performance of legitimate RFF-based authentication systems. We conducted various systematic experiments using nine USRP X310 SDRs connected to actual PLC couplers exchanging signals modulated according to the Binary-Phase Shift Keying modulation scheme to analyze the behavior of RFF in PLC scenarios under different RF jamming levels. Our results demonstrate, for the first time, that strategic RF jamming effectively obscures device fingerprints in cabled PLC communications while maintaining communication quality, with bit error rates remaining acceptable across most configurations. We also demonstrate that device identification accuracy degrades significantly as the jamming intensity increases. Our findings establish fundamental trade-offs between privacy protection and authentication reliability, providing insights for the design of robust PLC systems.

In recent years, there has been a rising trend towards emerging applications (e.g., brain-computer interaction and haptics-based autonomous cars) with diverse requirements. To effectively enable these applications via autonomous operation and intelligent analytics, one can use a metaverse. In a metaverse, we have two spaces: (a) a meta space based on a virtual model that performs analysis and resource management and (b) a physical space comprised of real world entities. A metaverse effectively enables emerging applications by performing three main tasks: (a) distributed learning of metaverse models; (b) instantly serving the end-users; and (c) sensing of the physical environment and sharing it with the meta space for synchronized operation. To perform these tasks, efficient wireless resource management is needed. Therefore, a novel resource scheduling framework for the wireless metaverse to enable various applications is proposed. Our aim is to minimize the cost of learning and sensing in metaverse. Subsequently, we formulate a problem that fulfills the reliability as well as latency constraints of the service-requesting users. We assign multiple resource blocks to learning and sensing devices/units, whereas we use a concept of puncturing for service-requesting devices/users upon arrival. We use a scheme that is based on block successive upper-bound minimization and convex optimization for solving our formulated problem. Finally, we use an empirical cumulative distribution function vs. cost and cost vs. metaverse entities for numerical evaluations.

Deep learning (DL)-based RF fingerprinting (RFFP) technology has emerged as a powerful physical-layer security mechanism, enabling device identification and authentication based on unique device-specific signatures that can be extracted from the received RF signals. However, DL-based RFFP methods face major challenges concerning their ability to adapt to domain (e.g., day/time, location, channel, etc.) changes and variability. This work proposes a novel IQ data representation and feature design, termed Double-Sided Envelope Power Spectrum or EPS, that is proven to significantly overcome the domain adaptation challenges associated with WiFi transmitter fingerprinting. By accurately capturing device hardware impairments while suppressing irrelevant domain information, EPS offers improved feature selection for DL models in RFFP. Our experimental evaluation demonstrates the effectiveness of the integration of EPS representation with a Convolution Neural Network (CNN) model, termed EPS-CNN, achieving over 99% testing accuracy in same-day/channel/location evaluations and 93% accuracy in cross-day evaluations, outperforming the traditional IQ representation. Additionally, EPS-CNN excels in cross-location evaluations, achieving a 95% accuracy. The proposed representation significantly enhances the robustness and generalizability of DL-based RFFP methods, thereby presenting a transformative solution to IQ data-based device fingerprinting.

New capabilities in wireless network security have been enabled by deep learning that leverages and exploits signal patterns and characteristics in Radio Frequency (RF) data captured by radio receivers to identify and authenticate radio transmitters. Open-set detection is an area of deep learning that aims to identify RF data samples captured from new devices during deployment (aka inference) that were not part of the training set; i.e. devices that were unseen during training. Past work in open-set detection has mostly been applied to independent and identically distributed data such as images. In contrast, RF signal data present a unique set of challenges as the data forms a time series with non-linear time dependencies among the samples. In this paper, we introduce a novel open-set detection approach for RF data-driven device identification that extracts its neural network features from patterns of the hidden state values within a Convolutional Neural Network Long Short-Term Memory (CNN+LSTM) model. Experimental results obtained using real datasets collected from 15 IoT devices, each enabled with LoRa, wireless-Wi-Fi, and wired-Wi-Fi communication protocols, show that our new approach greatly improves the area under the precision-recall curve, and hence, can be used successfully to monitor and control unauthorized network access of wireless devices.

Deep Learning-based RF fingerprinting has emerged as a game-changer for offering robust network device authentication and identification solutions. However, it struggles in cross-time scenarios, particularly during hardware warm-up phases. This often-overlooked vulnerability jeopardizes the reliability of these solutions. In response to this critical gap, we dive deep into the anatomy of RF fingerprints, revealing insights into temporal variations in DL-based RF fingerprinting during and post hardware stabilization. Introducing HEEDFUL, a novel framework harnessing sequential transfer learning and targeted impairment estimation, we address these challenges with remarkable consistency, eliminating blind spots even during challenging warm-up phases. Our extensive evaluation showcases HEEDFUL's efficacy, achieving remarkable classification accuracies of up to 96% during the initial intervals of device operation-far surpassing traditional models. Cross-domain assessments confirm HEEDFUL's superiority, achieving a steady 87% classification accuracy across warm-up intervals on the Day 2 dataset. Additionally, we release a WiFi RF fingerprinting dataset that, for the first time, incorporates both the time-domain representation and real hardware impairments of the frames. This inclusion underscores the importance of leveraging actual hardware impairment data, enabling a deeper understanding of fingerprints and facilitating the development of more resilient solutions.

Software-Defined Networking (SDN) provides the flexibility to dynamically manage network paths, facilitating efficient traffic flow and mutipath routing, thereby improving network resiliency to congestion. This study investigates three distinct SDN-enabled routing strategies: shortest path routing, Equal-Cost Multi-Path (ECMP) routing, and bandwidth-based multipath routing. The comparative result analysis across these three routing strategies revealed that bandwidth-based routing consistently outperforms Shortest Path and ECMP routing across key performance metrics. The study found that Bandwidth-based routing maintains lower delay and jitter, indicating its superior ability to manage network traffic efficiently even as data flow increases. Furthermore, it demonstrates a more modest increase in packet loss, underscoring its effective congestion management. These findings suggest that Bandwidth-based routing provides a more reliable and efficient network performance, particularly in high-traffic conditions, making it a preferable solution for SDN implementations seeking to optimize data flow and network stability.