Power domain Non-orthogonal multiple access (PDNOMA) has been widely considered to achieve the efficient short packet transmission (SPT). The most significant characteristic of SPT is that the finite blocklength coding is adopted, yielding a capacity backoff referred to as channel dispersion. However, the existing PD-NOMA schemes were widely investigated with the finite blocklength i.i.d. coding, which results in a large channel dispersion that dramatically decreases the capacity. To tackle this problem, we aim for investigating the PD-NOMA with the finite blocklength shell coding, which provides a smaller dispersion compared with the finite blocklength i.i.d. coding. On this basis, considering the demands of green communications, we conduct resource allocation to minimize the power consumption for the PD-NOMA scheme. Finally, simulation results verify the effectiveness of the proposed methods and compare the performance of the schemes with different coding strategies.

The convolutional neural network (CNN) based radio frequency fingerprint identification (RFFI), as an emerging device authentication technique, can identify wireless devices from their emitted radio-frequency (RF) transmissions. However, the variation of wireless channel may significantly impact accuracies of CNN-based RFFI systems, for instance, a CNN trained on signals collected on one day may fail to classify signals collected on other days. In this paper, we explore the time generalization of CNN-based RFFI systems by analyzing an open dataset that contains 6 WiFi devices operating on 4 days. Firstly, through visual analytics, it is found that equalizing signals can reduce the effect of the wireless channel variation, which is beneficial for the CNN to extract discriminative features and improve the classification accuracy. Secondly, convolutional autoencoder (CAE) based pre-training scheme is designed to obtain better generalization ability. Finally, as the input of the CNN, three signal representations are investigated in time, frequency, and time-frequency domains, namely in-phase and quadrature (IQ) samples, discrete Fourier transform (DFT) results and spectrogram, respectively. Experimental results show that the IQ-based CNN can reach the best performance, and the classification accuracy exceeds 95% for WiFi devices operating on different days.

A user equipment (UE)-assisted hybrid coherent and noncoherent transmission scheme is designed for cell-free massive multiple-input–multiple-output (MIMO), operating in the presence of phase offsets caused by both imperfect radio frequency (RF) chains and asynchronous reception. First, considering the effects of both factors, we derive closed-form spectral efficiency (SE) expressions for hybrid transmission under conjugate beamforming and zero-forcing (ZF) precoders. Based on these expressions, we introduce the concept of superposition gain to clarify the rationality of hybrid transmission. To leverage its advantages, we propose an access point (AP) grouping algorithm and its enhanced version, which groups the serving APs based on downlink (DL) equivalent channel at the UE side. Since the DL equivalent channel incorporates information on both RF and delay phase offsets, hybrid transmission based on this algorithm can simultaneously address both. Additionally, to implement this UE-assisted hybrid transmission with low overhead, we propose a signaling interaction strategy that extends traditional processes by introducing ZF-based DL beamforming pilot transmission for obtaining the DL equivalent channel, along with indication method for reporting the grouping results. Moreover, to further improve performance, a sequential convex approximation power allocation algorithm is proposed for hybrid transmission. Finally, simulations show that UE-assisted hybrid transmission achieves nearly fivefold improvement in 95%-likely SE over coherent transmission in the presence of both RF and delay phase offsets.

Multisource localization (MSL) for overlapped signals has attracted much attention, and the existing methods rely on the combination of multitype measurements, which puts higher requirements on the receiver. Besides, these methods can not avoid the problem of measurement-source association. In view of this, on the basis of broadband signal time-frequency spectrogram detection (TFSD), we propose an MSL scheme for overlapped signals based on index ambiguity elimination, which can avoid the above-mentioned measurement-source association problem by extracting the pure part of each signal component. Specifically, we first conduct the time-frequency transformation of the received signal, and analyze the overlapping types of the time-frequency blocks (TFBs) in two aspects: 1) inter-TFB, i.e., the overlapping types between the TFBs and 2) intra-TFB, i.e., the overlapping types between signal components contained in the TFB. On this basis, the TFB nonoverlapping part extraction algorithm is designed to eliminate the overlap between TFBs. Afterward, the signal segmentation algorithm based on the signal characteristic change mechanism is designed to obtain the pure part of each signal component, that is, eliminating the index ambiguity of each signal component contained in the extracted nonoverlapping TFB. Finally, the angle-of-arrival (AOA) information of multiple receivers for a certain signal can be obtained through the AOA estimation method, as well as the location of source device corresponding to the signal can be estimated by the triangulation method. Simulation and experiment results verify the effectiveness of the designed scheme.

In light of the demands of green communications and latency constraints in 5G, this paper investigates the resource allocation to minimize the power consumption for 2-user uplink non-orthogonal multiple access (NOMA) schemes with finite blocklength codes. In contrast to the existing research, we consider the achievable bound developed with the finite blocklength Gaussian shell codes (FBGSC) as it is greater than the widely used i.i.d. Gaussian achievable bound. To sufficiently explore the potential of NOMA, three typical 2-user uplink NOMA schemes are investigated in this paper, in terms of the classical power domain NOMA, rate splitting multiple access (RSMA) and NOMA with joint decoding (NJD). Simulation results indicate that with FBGSC, NJD always outperforms the other two schemes. In addition, it is found that, with FBGSC, RSMA outperforms NOMA only if the channel quality difference is small. To the best of our knowledge, we are the first to investigate uplink NOMA and RSMA with the shell bound developed by Scarlett (2017), which is the state of the art for their corresponding channels in the finite blocklength regime. Besides, we are also the first to investigate the power/energy-efficient transmission for all these three NOMA schemes with the shell codes.

Time-frequency localization (TFL) based intelligent wideband spectrum sensing is capable of achieving precise dynamic spectrum management. Recent studies demonstrate that object detectors can achieve excellent TFL performance in simple electromagnetic scenarios when trained with massive and labeled datasets. However, in real-world concurrent cognitive scenarios that allow users to reuse the same frequency band under a certain interference constraint, the phenomenon of signal overlapping in the time-frequency domain will seriously degrade the performance of object detector. To the best of our knowledge, no comprehensive analysis has been conducted to assess the impact of overlapping in TFL. To fill this research gap, we analyze the impact of overlapping and identify three challenges: variety of overlapping, hard to label, and feature destruction. To enhance the robustness of the detector, we first adopt a self-supervised learning (SSL) framework based on a masked autoencoder. This framework aims to pre-train a backbone with excellent feature extraction ability using unlabeled dataset to overcome variety of overlapping and labeling difficulties. Subsequently, we develop a transformer based robust TFL (TRTFL) detector. This detector is designed to leverage both time-frequency correlation and fine-grained features, effectively addressing issues related to feature destruction. Finally, simulation results demonstrate the superiority of the proposed method and the effectiveness of SSL framework and TRTFL. Compared to existing detectors, the TRTFL achieves superior feature extraction, yielding a mean average precision (mAP) of 90.70% in overlapping signal scenarios. Moreover, the TRTFL with SSL can achieve an mAP of up to 95.08% outperforming the state-of-the-art.

With a great potential of improving the service fairness and quality for user equipments (UEs), cell-free massive multiple-input multiple-output (mMIMO) has been regarded as an emerging candidate for 6G network architectures. Under ideal assumptions, the coherent joint transmission (CJT) serving mode has been considered as an optimal option for cell-free mMIMO systems, since it can achieve coherent cooperation gain among the access points. However, when considering the limited fronthaul constraint in practice, the non-coherent joint transmission (NCJT) serving mode is likely to outperform CJT, since the former requires much lower fronthaul resources. In other words, the performance excellence and worseness of single serving mode (CJT or NCJT) depends on the fronthaul capacity, and any single transmission mode cannot perfectly adapt the capacity limited fronthaul. To explore the performance potential of the cell-free mMIMO system with limited fronthauls by harnessing the merits of CJT and NCJT, we propose a CJT-NCJT hybrid serving mode framework, in which UEs are allocated to operate on CJT or NCJT serving mode. To improve the sum-rate of the system with low complexity, we first propose a probability-based random serving mode allocation scheme. With a given serving mode, a successive convex approximation-based power allocation algorithm is proposed to maximize the system’s sum-rate. Simulation results demonstrate the superiority of the proposed scheme.

The convolutional neural network (CNN) is effective to classify radio frequency (RF) signals of unmanned aerial vehicles (UAVs), although the variation of UAV operating channels can degrade the performance of the CNN. As the CNN is data hungry, an intuitive solution is to capture UAV signals of all channels to train the CNN. However, the signal collection is time-consuming and expensive. Hence, this paper proposes a data augmentation scheme based on the frequency characteristics of UAV signals, which approximately simulates UAVs operating on different channels. With this scheme in the training pipeline, we use signals of a single UAV channel to train CNNs that can classify UAVs operating on arbitrary channels. Extensive indoor and outdoor experiments are conducted, and the collected signals are also released as part of the technical contributions of our work. The experimental results show that the proposed data augmentation scheme can improve the classification accuracy of CNNs by 60%.

Cell-Free massive multiple-input multiple-output (MIMO), which provides high spectral efficiency (SE) through the coherent joint transmission, has drawn much academic research interest. However, the implementation of coherent joint transmission is restricted by the phase offset caused by hardware defect and asynchronous reception. In this paper, by combining the advantages of the high performance provided by coherent joint transmission and the phase offset robustness provided by non-coherent joint transmission, we propose a hybrid transmission scheme based on access point (AP) grouping, which ensures decent coherent joint transmission by minimizing the phase offset between collaborative APs within the group, and avoids large phase offset between groups by executing non-coherent joint transmission. Moreover, to facilitate simulation verification of the performance of the proposed scheme under various preprocessing methods, we derive a generalized version of SE expression of the hybrid transmission scheme using successive interference cancellation. Furthermore, simulations are provided to verify the effectiveness of our proposed scheme.

Time-frequency localization based intelligent wideband spectrum sensing is essential for achieving precise dynamic spectrum access and management. Currently, object detection based detectors can achieve excellent time-frequency localization performance in a simple electromagnetic environment. However, the overlapping of signals in the time-frequency domain can corrupt signal features and degrade detector performance. In this paper, we propose a transformer based robust time-frequency localization (TRTFL) detector that fully extracts the features of the time-frequency domain correlation to improve its robustness in solving the aforementioned problem. Furthermore, to exploit the convolution operation for mining fine-grained features, we embed a convolutional layer with a small kernel in the transformer block. Finally, simulation results validate the advantages of TRTFL compared to existing detectors and demonstrate its robustness for overlapping signals in spectrogram.

Compared with the traditional spectrum sensing methods that can only detect the presence or absence of signals, deep learning-based time-frequency localization (TFL) methods can obtain two-dimensional time-frequency information (TFI). However, for time-frequency domain overlapped signals, TFL methods can only obtain the contour TFI of the whole time-frequency block (TF-Block), but cannot obtain the TFI and corresponding relationship of each component signal contained in the TF-Block, which is named index ambiguity here. To use spectrum resources more efficiently, it is need to mine the usage of time-frequency resources in multidimensional space as much as possible, such as the time-frequency resources occupation and direction-of-arrival (DoA) of each component signal, which can help the users to avoid interference in spectrum reuse. In this paper, a processing framework is designed to eliminate the index ambiguity. Based on the result of TFL, rank features are extracted using the sliding window method to characterize the signal property, and then a signal segmentation algorithm is designed to obtain the concrete composition of overlapped signals. Finally, the DoA of each component signal is obtained based on the signal segmentation information. Simulation results demonstrate that the proposed method can efficiently and accurately extract multidimensional information to eliminate the index ambiguity of overlapped signals.

The advancements in unmanned aerial vehicle (UAV) technology have brought immense convenience to society. However, unauthorized UAVs pose a serious threat to personal privacy, public safety, and aviation security. Therefore, accurate UAV detection and classification are crucial. Moreover, with the increased popularity of UAVs, the likelihood of multiple UAVs appearing in the same area simultaneously has also dramatically increased. Recent studies demonstrate that object detectors, such as FasterRCNN and YOLO, can be used to detect and classify multiple UAVs based on spectrograms. To our best knowledge, the object detectors are directly used to classify UAV without considering the characteristics of the UAV signal spectrogram, which results in a decrease in recognition performance. In this article, we analyze the characteristics of the UAV signal spectrogram in detail and conclude two problems, i.e., prior anchor mismatch and cross-domain detection, hindering the implementation of object detector for UAV recognition. To solve prior anchor mismatch, we propose an anchor-free detector based on keypoint and design a novel keypoints matching algorithm to improve recognition performance. To solve cross-domain detection, we propose an adversarial learning-based data adaptation method, which can generate domain-independent and domain-aligned features. Finally, the experiments adopt practical spectrogram and synthetic spectrogram to verify the superiority of the proposed anchor-free detector and the effectiveness of the proposed data adaptation method.

With the increasing security concerns of the satellite network due to the broadcasting nature and the inherent openness of satellite-terrestrial communications, the satellite spectrum and terrestrial node resource trading based cooperation in cognitive satellite terrestrial networks (CSTNs) has gained a lot attention. However, the existing literature has not well considered the fairness issue in resource trading, which may cause cooperation failure between the satellite and terrestrial networks when their own benefits are impaired. To tackle this issue, in this paper we propose a two-layer hierarchical game framework for a multi-terrestrial base stations (BSs) CSTN scenario to guarantee the fairness of resource trading between the satellite and terrestrial networks and thus achieve a win-win situation for both networks. Specifically, a coalition formation game is adopted to study the cooperative behaviors among the terrestrial BSs. Herein, we propose a distributed merge-and-split based coalition formation algorithm to determine the coalition structure, of which the stability, convergence, and complexity are theoretically investigated. Moreover, a Stackelberg game is introduced to model the competition between the satellite and terrestrial BSs, where the satellite acts as the leader and the terrestrial BSs act as the followers. The Stackelberg equilibrium (SE) for the Stackelberg game is derived based on the backward induction method. We then design a distributed algorithm to obtain the coalition structure and SE for the proposed two-layer hierarchical game framework. Finally, simulations are presented to validate our theoretical results.

Over the past few years, commercial drones have grown in popularity. However, the pervasive use of drones may pose a range of secure risks to sensitive areas such as airports and military bases. Hence, drone detection and identification are critical and necessary for governments and security agencies. This paper proposes a radio frequency identification (RFI) system for drones based on spectrogram and convolutional neural network (CNN). Specifically, spectrogram is used to represent fine-grained time-frequency characteristics of drone signals. Then CNN is designed to infer drone types by identifying their spectrograms. In practice, drones have different operating channels, and any one of them can be selected for signal transmission. It means that the carrier frequencies of their signals are unknown, which may result in misclassifications. To address this problem, we collect drone signals from all potential frequency bands, and demonstrate that carrier frequency offset (CFO) compensation can significantly improve the system performance. Experimental evaluation is performed in real wireless environments involving 6 drones and a Universal Software Radio Peripheral (USRP) X310 platform. Moreover, the proposed spectrogram-based CNN can reach the best performance compared with the IQ-based and FFT-based CNNs. The classification accuracy is beyond 98% for drones operating on arbitrary channels.

In physical layer security, the geographical secure region as a practical security metric has received considerable interest. This paper derives the distribution functions of signal-to-interference-plus-noise ratios at Bob and Eve with Rician fading channels, which is used to calculate the secrecy outage probability for different Eve’s locations. Then, the secure region can be determined in geographical two-dimensional plane by comparing the secrecy outage probability with a threshold. Based on this, two optimization problems of the beamforming vector at Bob used for transmitting the artificial noise are constructed to expand the secure region, with and without the knowledge of Eve’s location, respectively. Simulation results are provided to depict the secure region with different antenna numbers, and illustrate that the secure region can be effectively expanded by using the optimized beamforming vector at Bob.

In this paper, we study the automatic modulation classification in a non-orthogonal multiple access system. To mitigate the effect of interference, a likelihood-based algorithm and a fourth-order cumulant-based algorithm are proposed. Different from the maximum likelihood classifier for a single signal without interference, a likelihood function of the far and near users' signals is derived. Then, a marginal probability for the far user is obtained by using the Bayesian formula. Hence, the modulation type can be determined by maximizing the marginal probability. The high computational complexity of the likelihood-based algorithm renders it impractical; accordingly, it serves as a theoretical performance bound. On the other hand, we construct a feature vector through the estimated fourth-order cumulants of the received signal including the superposed signal and noise. For each modulation pair, using the mean and covariance matrix of the estimated feature vector, its probability density function can be obtained. Then, the key is to calculate the mean and covariance matrix of the estimated feature vector. To solve this problem, the moments of the superposed signal are derived. Therefore, modulation classification can be performed by maximizing the probability density function. Extensive simulations verify that the two proposed algorithms perform well under a wide range of signal-to-noise ratios and observation lengths.

Automatic modulation classification (AMC) plays a pivotal role for spectrum monitoring in cognitive radio (CR). The objective of this work is to investigate the AMC problem for multiple-input multiple-output (MIMO) systems employing orthogonal frequency division multiplexing (OFDM) under imperfect timing synchronization scenarios. Two major contributions are made in this work: i) specific higher-order cumulants (HOC) are proved to be robust to symbol timing offset (STO). ii) with the feature of multiple HOCs, a random forest based AMC algorithm is proposed for MIMO-OFDM systems, which is robust to timing synchronization error. Numerical simulations are conducted to verify the validation of the algorithm.

This paper presents a novel approach for blind identification of space-frequency block codes. Based on the random matrix theory, we propose principal component sequence as a discriminating feature, which is detected by sliding window in the frequency-domain. With this feature, Euclidean distance is employed for decision making. The proposed algorithm does not need priori knowledge of the signal parameters such as channel coefficients, the modulation mode or noise power. Meanwhile, this algorithm is compatible to identify single-antenna systems and spatial multiplexing of different orders without another module. Moreover, the simulations show the proposed algorithm adapts to multipath fading effectively with a short observation period.

In this study, the authors propose a non‐parametric algorithm to implement the estimation of transmit‐antenna number, which is a prerequisite for blind interception process of multiple‐input multiple‐output orthogonal frequency division multiplexing signals in frequency selective fading. Specifically, a series of test statistics are constructed by exploiting the eigenvalues of the sample covariance matrices from each subcarrier, followed by a combination of these test statistics. As a consequence, the number of transmit antennas can be determined after a serial binary hypothesis testing. The theoretical analysis and simulation results verify the rapid convergence and high reliability of the proposed algorithm at a relatively low signal‐to‐noise ratio.