Physical-layer key generation exploits the randomness and unpredictability of wireless channels to enhance the security of wireless communications. However, factors such as asynchronous measurements, channel noise, and hardware impairments undermine the ideal reciprocity of channel state information (CSI) observed by legitimate users, thereby limiting key generation performance. To address this issue, this paper proposes a jointly optimized contrastive learning network (JOCLNet), which consists of a joint learning model and a contrastive learning model. The joint learning model, built upon a convolutional neural network (CNN) and a feedforward neural network (FNN), separates non-reciprocal components such as noise and measurement errors from the raw CSI of legitimate users, while extracting reciprocal components for key generation, thus improving robustness in noisy environments. To overcome the inability of existing deep-learning-based SKG schemes to resist passive eavesdropping attacks, the contrastive learning model introduces a contrastive loss on top of the joint learning model. This enables a dual objective: enhancing reciprocity between the processed CSI of legitimate users while ensuring that the processed data of an eavesdropper remains highly uncorrelated with that of legitimate parties. Furthermore, a complete key generation scheme is designed based on the proposed JOCLNet. Experimental results demonstrate that the proposed scheme achieves strong resistance to eavesdropping and robust adaptability to noise.

Industrial Internet of Things (IIoT) increasingly relies on Gigabit Ethernet (1000BASE-T) as the physical back-bone for interconnecting industrial devices, while the rapid growth of IIoT nodes has intensified concerns about physical-layer identity spoofing and unauthorized access. Recently, device fingerprinting has emerged as a promising approach to achieving secure authentication at the physical layer. However, existing 1000BASE-T fingerprint extraction methods rely on randomly scrambled signals, leading to degraded authentication reliability. In addition, the absence of radio-frequency (RF) frontend modules—commonly defined in wireless systems—within 1000BASE-T transmitters prevents the direct application of conventional hardware-imperfection models. To overcome these challenges, this paper first introduces test mode (TM) signals as highly consistent and controllable reference inputs, and on this basis, proposes an interpretable high-pass filter (HPF) fingerprint model. The model characterizes the high-pass response of the transmission link using a single-pole system, establishes a monotonic relationship between filter parameters and waveform morphology, and extracts stable fingerprint features accordingly. Furthermore, a closed-loop physical-layer authentication framework is developed, integrating signal acquisition, preprocessing, feature extraction, and device identification. Experimental results demonstrate that the proposed method achieves 100% identification accuracy under standard sampling conditions and preserves perfect recognition over the entire tested sampling-rate range. Moreover, the method exhibits substantially enhanced noise robustness compared with baseline methods, and retains 95.59% accuracy after a 30-day interval in temporal stability evaluations.

Channel state information (CSI) serves as a critical indicator of wireless signal conditions and is widely regarded by researchers for its sensitivity in detecting changes within the channel. However, traditional sensing technologies often require substantial data and intricate learning algorithms, highlighting an urgent need for advancements in lightweight sensing technologies. These technologies should leverage simpler terminal devices, reduced data volumes, and more straightforward classification algorithms to achieve sensing capability that are comparable to those offered by more complex and established methods. This article concentrates on the lightweight application of wireless sensing and encompasses the following key contributions: 1) the development of a lightweight sensing model utilizing a single-antenna multicarrier system, which introduces a CSI ratio model that adapts multiantenna techniques for single-antenna settings and 2) the enhancement of feature stability through the introduction of a complex-plane fitting method using artificial vector, alongside a dual receiver-based method for cross-scene feature generation aimed at producing stable and high-quality auxiliary features. Experimental results show that the feature extraction capability of the single-antenna multicarrier CSI ratio model is close to that of traditional multiantenna scheme. On the gait dataset, when the enhanced CSI ratio is used as a feature, the accuracy is nearly 100%, surpassing the 93% accuracy of the original amplitude feature. On the gesture dataset, the combination of the enhanced CSI ratio and position and environment independent features achieves an accuracy of 96%, which is superior to using the original CSI amplitude feature alone. An analysis of resource consumption shows that the lightweight SVM model incurs very low computational overhead during decision-making, validating the potential of this scheme in terms of efficiency and practical application.

Wireless channel key generation technology is an important mechanism to guarantee the security of wireless network, but influenced by the key length and the actual electromagnetic environment, wireless channel key generation technology is faced with the challenge of high-key generation rate (KGR) and low-key disagreement rate (KDR). The existing key generation methods also lack the full use of the channel state information (CSI). We propose a key generation method based on multisubcarrier phase difference to expand the randomness source dimension, eliminate the phase bias, offset part of the noise influence, and set the threshold screening data to reduce the influence of measurement error. We further propose a key generation method based on resampling of kernel density estimation (KDE), which yields highly reciprocal randomness sources by resampling the results of KDE of phase difference values. To fill the metric gap of whether a method keeps low KDR while increasing the KGR, the evaluation metric of effective improvement ratio (EIR) is proposed. The two methods we proposed have a higher EIR than the method of using multiple-input and multiple-output (MIMO) and increasing the quantization level, achieving the goal of increasing the KGR while maintaining the low KDR. The KGR can reach about 12146 bits/s, and the KDR is 1.83%. The keys obtained by both methods can effectively prevent passive eavesdropping and meet the randomness requirements.

The idea of physical layer security is to use the characteristics and damage of the propagation medium to ensure secure communication in the physical layer. Channel state information is a fine-grained value from the physical layer, which describes the amplitude and phase of each sub-carrier in the frequency domain. It estimates the channel information by representing the channel properties of the communication link. Due to the uniqueness and short-term reciprocity of the channel, both the sender and receiver can obtain almost the same and random CSI in a short time, and can generate the same and random key to realize secure communication between the two parties. At present, most of the traditional methods of obtaining CSI are costly, bulky, and limited in regions. In order to verify the validity, feasibility and stability of extracting CSI with Nexmon firmware and using it to generate keys, this paper verifies its performance.

The access control mechanism in LoRa has been proven to have high security risks. In order to improve the secure access ability of LoRa terminals, this paper presents a physical layer-based authentication system for security enhancement. Different from the security access technology of cryptography, a lightweight gateway architecture called LW-LoRaWAN is proposed to realize a data frame-based authentication with radio frequency fingerprint (RFF). A novel RFF feature of Cross Power Spectral Density (CPSD) is used to achieve a fast authentication with one single frame. Theoretical analysis and experimental results show that the proposed system not only reinforces the authentication security of LoRa network but also protects the LoRa terminals against the Sybil attacks. The LW-LoRaWAN provides new security approach from physical layer for LoRa network.

Radio frequency (RF) fingerprint is the inherent hardware characteristics and has been employed to classify and identify wireless devices in many Internet of Things applications. This paper extracts novel RF fingerprint features, designs a hybrid and adaptive classification scheme adjusting to the environment conditions, and carries out extensive experiments to evaluate the performance. In particular, four modulation features, namely differential constellation trace figure, carrier frequency offset, modulation offset and I/Q offset extracted from constellation trace figure, are employed. The feature weights under different channel conditions are calculated at the training stage. These features are combined smartly with the weights selected according to the estimated signal to noise ratio at the classification stage. We construct a testbed using universal software radio peripheral platform as the receiver and 54 ZigBee nodes as the candidate devices to be classified, which are the most ZigBee devices ever tested. Extensive experiments are carried out to evaluate the classification performance under different channel conditions, namely line-of-sight (LOS) and nonline-of-sight scenarios. We then validate the robustness by carrying out the classification process 18 months after the training, which is the longest time gap. We also use a different receiver platform for classification for the first time. The classification error rate is as low as 0.048 in LOS scenario, and 0.1105 even when a different receiver is used for classification 18 months after the training. Our hybrid classification scheme has thus been demonstrated effective in classifying a large amount of ZigBee devices.