Radio-frequency fingerprinting identification (RFFI), based on hardware inherent imperfections, is an effective technique for physical authentication in the Internet of Things (IoT). However, due to the nonstationary and open factors in realistic scenes, problems, such as inaccurate fingerprinting extractions and unreliable transmitter identifications, often occur. Specifically, in actual scenarios, there are challenges, such as varying channel status, variable transmission content, insufficient data set acquisition, and the unknown number of rogue devices. In this article, a robust and protocol-agnostic RFFI system is proposed for open-set recognition, which is achieved by a slicing-enhanced preprocessing with noise augmentation and a well-designed bone network with a modified open layer. Moreover, in order to verify the effectiveness and transferability of the designed RFFI system in applications of IoT, comparative experiments are conducted using a real-world data set of automatic-dependent surveillance-broadcast (ADS-B) signals and the open-source FIT/CorteXlab data set. The performance of the designed extractor and classifier, under varying channel and openness conditions, verifies the channel feasibility and openness generalization of our designed system. The results of the ablation experiment indicate that the designed RFFI system can achieve an accuracy of 96.28% in the classification of 50 registered IoT transmitters and 87.52% in open-set recognition of mixed 50 registered and 10 unregistered IoT transmitters.

Automatic modulation discrimination (AMC) is one of the critical technologies in spatial cognitive communication systems. Building a high-performance AMC model in intelligent receivers can help to realize adaptive signal synchronization and demodulation. However, tackling the intra-class diversity problem is challenging to AMC based on deep learning (DL), as 16QAM and 64QAM are not easily distinguished by DL networks. In order to overcome the problem, this paper proposes a joint AMC model that combines DL and expert features. In this model, the former builds a neural network that can extract the time series and phase features of in-phase and quadrature component (IQ) samples, which improves the feature extraction capability of the network in similar models; the latter achieves accurate classification of QAM signals by constructing effective feature parameters. Experimental results demonstrate that our proposed joint AMC model performs better than the benchmark networks. The classification accuracy is increased by 11.5% at a 10 dB signal-to-noise ratio (SNR). At the same time, it also improves the discrimination of QAM signals.

Specific Emitter Identification (SEI) distinguishes specific emitters among identical radar types by inspecting hardware fingerprints. Nevertheless, deep learning-based SEI poses challenges due to an excessive count of model parameters and high computational complexities, hindering its applicability in real-time and near-real-time situations. To address these challenges, we introduce a Multi-Resolution Convolutional Network (MRNet) capable of extracting signal features across both time and frequency domains. This model employs dilated convolution at varying expansion rates to glean multi-resolution features of signals, thereby enhancing accuracy without increasing parameters or computational efforts. An experiment conducted with ten real radar emitters examines aspects such as the depth and width of the convolutional network. The results reveal that MRNet outperforms conventional convolutional networks, achieving an approximate 8% boost in accuracy. When compared with multi-scale networks, MRNet exhibits superior performance with diminished complexity. This study emphasizes the advantages of employing MRNet in SEI, particularly in enhancing efficiency and interpretability.

Specific emitter identification (SEI) is a core technology for wireless device security that plays a crucial role in protecting wireless communication systems from various security threats. However, current deep learning-based SEI methods heavily rely on large amounts of labeled data for supervised training, facing challenges in non-cooperative communication scenarios. To address these issues, this paper proposes a novel contrastive asymmetric masked learning-based SEI (CAML-SEI) method, effectively solving the problem of SEI under scarce labeled samples. The proposed method constructs an asymmetric auto-encoder architecture, comprising an encoder network based on channel squeeze-and-excitation residual blocks to capture radio frequency fingerprint (RFF) features embedded in signals, while employing a lightweight single-layer convolutional decoder for masked signal reconstruction. This design promotes the learning of fine-grained local feature representations. To further enhance feature discriminability, a learnable non-linear mapping is introduced to compress high-dimensional encoded features into a compact low-dimensional space, accompanied by a contrastive loss function that simultaneously achieves feature aggregation of positive samples and feature separation of negative samples. Finally, the network is jointly optimized by combining signal reconstruction and feature contrast tasks. Experiments conducted on real-world ADS-B and Wi-Fi datasets demonstrate that the proposed method effectively learns generalized RFF features, and the results show superior performance compared with other SEI methods.