Enhancing Open Set RFF Recognition with cGAN: Generating Multiple Unknown Classes

Radio frequency (RF) fingerprint representing the inherent hardware characteristics of mobile devices has been employed to classify and identify wireless devices for the security of Internet of Things (IoT). Existing works on RF fingerprinting are usually based on the amplitude or phase of RF signal envelope, which leads to relatively coarse features. Moreover, the classification performance over small sample dataset is poor. To solve the problem, a novel device identification method based on RF fingerprinting with on deep learning is proposed. In particular, the RF signal are transformed into two dimensional representations by image preprocessing. Then the gray images representing the RF fingerprints are classified by employing classical CNN. To verify the performance of the proposed approach, a testbed is constructed by using MATLAB build framework of gray image preprocessing. Extensive experiment results show that the identification accuracy can reach at least 90%. Even with the sample rate of 20Gsps. Particularly, the accuracy of iPhone can reach 100%. It is verified that the proposed method can effectively classify mobile devices even with small sample RF fingerprints represented two dimensional gray images,

The security of ultra‐low‐power Internet of Things (IoT) systems is critical yet challenging due to significant energy constraints. These networks are vulnerable to impersonation and data poisoning attacks, where malicious entities can mimic legitimate devices to gain access or corrupt system integrity. While traditional cryptographic solutions are often too energy‐intensive for these environments, radio frequency (RF) fingerprinting offers a promising physical layer security alternative by using intrinsic hardware imperfections to uniquely identify devices. However, existing RF fingerprinting methods often overlook the severe energy budgets of battery‐powered IoT devices. To address this challenge, this paper introduces two complementary deep learning models for device identification in long range wide area network systems. The first, RFNet, is a full‐capacity convolutional neural network that achieves 97.48% identification accuracy. The second, TinyRFNet, is an ultra‐lightweight model designed for resource‐constrained hardware, maintaining 93.19% accuracy with over 34 times fewer parameters than RFNet. We further propose a dynamic, energy‐aware inference strategy that adaptively selects between these two models based on the device's remaining battery level, the model's prediction confidence, and the operational context. Extensive experimental evaluation on a dataset of 30 commercial LoRa devices demonstrates that this adaptive approach achieves an overall identification accuracy of 94.54% while reducing energy consumption by 17% compared to exclusively using the high‐accuracy model. This system provides robust protection against physical‐layer threats with minimal energy overhead, thereby extending the operational lifetime of devices in secure, ultra‐low‐power IoT deployments.

Most of the progresses achieved in radar high range resolution profile (HRRP) recognition rely on the closed set condition, where the test sample is from a known class. In realistic scenario, however, the test sample may be drawn from unknown classes, which is regarded as an open set recognition task. In such cases, conventional recognition algorithms will inevitably make a wrong prediction. In this paper, the open set problem is addressed by incorporating the extreme value theory (EVT) into the random forest (RF) classifier. The outputs of RF are analyzed to determine whether the test sample should be rejected as an unknown class. At the training phase, a Weibull-based extreme-value meta-recognition is introduced to describe the statistical characteristics of the known classes. At the testing phase, a probability estimation method is introduced to compute the probabilities of the test sample belonging to known and unknown classes based on trained Weibull distributions. The test sample is assigned to the class of highest probability. Experimental results demonstrate that the proposed method outperforms the state-of-art NN, 1-vs-set machine and W-SVM in rejecting unknown classes.

Wireless Physical Layer Identification (WPLI) aims at identifying or classifying authorized devices based on the unique Radio Frequency Fingerprints (RFFs) to guarantee physical-layer security between each individual device. WPLI is a promising technique in the security area of Internet of Things (IoT), which avoids the burden of conventional software-level security in IoT on the computation capability, on-chip storage and power consumption of devices. Moreover, WPLI directly extracts the RFFs from radio frequency signals at physical layer, which breaks the software-level security barriers of heterogeneous device networks in IoT. However, in the existing works in WPLI, a general work focusing on theoretical characterization of performance WPLI systems in IoT is still in absence. In this work, we evaluate the performance of WPLI in IoT by thoroughly characterizing user capacity of WPLI system as the metric. We utilize the concept of IoT by combining multiple devices to represent one user's secure identity to increase the user capacity of WPLI system. Based on the information-theoretic tool we proposed, the changes in user capacity of WPLI is characterized over typical real-world scenarios according to different number of devices assigned for one user. Various field experiments on classification performance of a practical multi- user-multi-device WPLI system with are conducted to validate the user capacity characterization for each scenario.

Smart grids incorporate diverse power equipment used for energy optimization in intelligent cities. This equipment may use Internet of Things (IoT) devices and services in the future. To ensure stable operation of smart grids, cybersecurity of IoT is paramount. To this end, use of cryptographic security methods is prevalent in existing IoT. Non-cryptographic methods such as radio frequency fingerprinting (RFF) have been on the horizon for a few decades but are limited to academic research or military interest. RFF is a physical layer security feature that leverages hardware impairments in radios of IoT devices for classification and rogue device detection. The article discusses the potential of RFF in wireless communication of IoT devices to augment the cybersecurity of smart grids. The characteristics of a deep learning (DL)-aided RFF system are presented. Subsequently, a deployment framework of RFF for smart grids is presented with implementation and regulatory aspects. The article culminates with a discussion of existing challenges and potential research directions for maturation of RFF.

Radio frequency fingerprint (RFF) identification is a promising technique for identifying Internet of Things (IoT) devices. This paper presents a comprehensive survey on RFF identification, which covers various aspects ranging from related definitions to details of each stage in the identification process, namely signal preprocessing, RFF feature extraction, further processing, and RFF identification. Specifically, three main steps of preprocessing are summarized, including carrier frequency offset estimation, noise elimination, and channel cancellation. Besides, three kinds of RFFs are categorized, comprising I/Q signal-based, parameter-based, and transformation-based features. Meanwhile, feature fusion and feature dimension reduction are elaborated as two main further processing methods. Furthermore, a novel framework is established from the perspective of closed set and open set problems, and the related state-of-the-art methodologies are investigated, including approaches based on traditional machine learning, deep learning, and generative models. Additionally, we highlight the challenges faced by RFF identification and point out future research trends in this field.

Spatial Location Constraint Prototype Loss for Open Set Recognition

Spatial Location Constraint Prototype Loss for Open Set Recognition

With the development of the Internet of Things, more and more devices need to be connected to the network. Wireless networks are open, and the network security of devices faces major challenges. The problem of device identification is particularly important. Some traditional signal source marking methods have greatly reduced reliability. The traditional authentication methods mainly rely on software addresses. The radio frequency(RF) fingerprint technology is an emerging device authentication technology, which identifies the device based on the inherent hardware defects of the device itself. The RF fingerprints reflect the inherent characteristics of the device and are difficult to tamper with. In this paper, we proposed an environment invariant feature extraction algorithm based on the cumulative distance. The experiments show that the algorithm proposed in this paper has good robustness. Within the allowable error range, the extracted RF fingerprint features of the same signal in different frequencies, different receivers, and different environments are invariant.KeywordsMarking methodsRF fingerprintI/Q data

Conditional feature generation for transductive open-set recognition via dual-space consistent sampling

Image classification methods are usually trained to perform predictions taking into account a predefined group of known classes. Real-world problems, however, may not allow for a full knowledge of the input and label spaces, making failures in recognition a hazard to deep visual learning. Open set recognition methods are characterized by the ability to correctly identify inputs of known and unknown classes. In this context, we propose GeMOS: simple and plug-and-play open set recognition modules that can be attached to pretrained Deep Neural Networks for visual recognition. The GeMOS framework pairs pre-trained Convolutional Neural Networks with generative models for open set recognition to extract open set scores for each sample, allowing for failure recognition in object recognition tasks. We conduct a thorough evaluation of the proposed method in comparison with state-of-the-art open set algorithms, finding that GeMOS either outperforms or is statistically indistinguishable from more complex and costly models.

In Synthetic Aperture Radar (SAR) image classification tasks, the performance depends on both the classifier and the dataset itself. However, in comparison with plenty of SAR classification methods, there is little work aimed at analyzing the distinguishability of the dataset. In the classification dataset, some classes are semantically different but their distinguishability is low, the classes are hard to be classified especially in some more practical cases that there are unknown classes without supervision exist. Referring to open set recognition (OSR), in this paper, we proposed the SAR Distinguishability Analysor (SAR-DA) to evaluate the distinguishability of the OpenSARUrban dataset. By modeling each class as a multivariate Gaussian distribution in latent space, SAR-DA can not only classify the classes having been seen in training phase, but also can recognize unknown samples if a test sample is out of each known distribution. Each class in OpenSARUr-ban is set unknown in turn, then we apply the SAR-DA on the split dataset in OSR and supervised setting. The distinguishability can be reflected by the unknown recognition recall rate. The experimental results show that the unknown recognition recall rate in OSR setting significantly decreased compared with those in supervised setting, indicating that even though the classes in OpenSARUrban are semantically different from each other, the latent distributions of some classes are quite similar and hard to be classified, thus these classes are of low distinguishability.

Few-shot open-set recognition (FSOSR) poses a significant challenge as it requires identifying unknown classes while maintaining the classification performance of known classes, despite having limited access to labeled training samples. Current methods often employ non-directional metric-based losses to encapsulate feature attributes within the embedding space, inadvertently disregarding the potential influence of spatial distribution deviations of feature representations on open-set recognition performance. To address this, we present a novel directional metric-based method termed Bilevel Direction Preserving (BiDirP). This method incorporates two direction-preserving regularizers operating at distinct levels, specifically at the instance and prototype levels. The combined application of these two direction-preserving regularizers effectively enhances the spatial separation between prototypes of different classes and refines the classification decision boundaries, which results in an improved discriminative ability to differentiate unknown classes within a broader open space. Comprehensive experiments on public benchmarks show that BiDirP can significantly improve the detection ability of unknown classes while correctly classifying known classes.

Understanding open-set recognition by Jacobian norm and inter-class separation

To develop access control mechanisms, particularly in terms of maintaining effective and secure access control within Internet of Things (IoT) networks. Whereas the sufficient use must be made of blockchain-based access control technology. This is because of the sheer volume of connected devices and the subsequent increase in transactions. This can negatively impact the performance and responsiveness of the networks. Thus, this article proposes a comprehensive approach that evaluates the requester’s reputation with respect to regulating access requests for IoT resources. The proposed approach combines fuzzy reputation with a decay algorithm. It then calculates the quantitative reputation value for each IoT user. This considers multiple variables, such as the Access Request Rate, frequency of requests, etc. This new reputation value serves as the basis for access-control decisions. Extensive simulations and experiments are conducted to evaluate the effectiveness of the proposed framework. For the simulation, we used a single-board Raspberry Pi. We also used a hybrid blockchain network environment comprising Geth and Hyperledger Fabric. We then analyzed and compared the performance of the proposed framework with that of the existing approaches. The results demonstrated that compared to the mathematical mechanism, the framework provides improved access control in IoT networks. The fuzzy-based reputation framework captures the dynamic nature of IoT environments, and effectively identifies trustworthy and malicious devices, whereas the decay algorithm reflects the most recent user behavior.