Search

In this paper, we investigate the grant-free communications with adaptive period for industrial Internet of Things, where only a fraction of devices are active at a time. To the best of our knowledge, this is the first work to exploit the non-continuous temporal correlation of the received signal for joint user activity detection (UAD), channel estimation and signal detection, while all the previous work requires continuous transmission. Two schemes are proposed toward this purpose, namely periodic block orthogonal matching pursuit (PBOMP) and periodic block sparse Bayesian learning (PBSBL), which outperform the previous schemes in terms of the success rate of UAD, bit error rate and accuracy in period estimation and channel estimation. The CramC)r-Rao lower bounds (CRLBs) of channel estimation by PBOMP and PBSBL are derived. It is shown that the two proposed approaches have close CRLBs and normalized mean square error at high SNR.

In this article, a novel 1.05-1.25 GHz filtering antenna feeding network with reconfigurable frequency and polarization is proposed. It simply consists of a reconfigurable in-/out-of-phase filtering power divider (RFPD) and a 90° hybrid coupler. By controlling the varactors in the RFPD without any switch component, the proposed 3-port feeding network offers continuous reconfigurable filtering responses with four switchable output patterns including: equal power dividing with phase difference of negative or positive 90°, single-ended output at either output port. With this proposed feeding network connecting to a regular dual-polarized antenna, the resulted antenna system can easily realize reconfigurable polarization of vertical polarization (V-LP), horizontal polarization(H-LP), left-handed circularly polarization (LHCP) and right-handed circularly polarization (RHCP), in addition to its reconfigurable filtering responses. A prototype using microstrip technology is designed, fabricated and measured to demonstrate the proposed technique. Good agreement between simulated and measured results is obtained. The measured results show that the feeding network covers a frequency range of 1.05-1.25 GHz, and the error of phase difference and amplitude imbalance are less than 5° and 0.5 dB, respectively. The resulted antenna beams can be freely switched in four polarization states with good performance.

In this paper, we propose a deep neural network (DNN) aided codebook beamforming scheme for frequency-division-duplex (FDD) millimeter-Wave (mmWave) massive multiple-input multiple-output (MIMO) systems under multipath. Different from the time-division-duplex (TDD) systems, for FDD systems the reciprocity between the down-link (DL) and up-link (UL) channel state information (CSI) does not hold in general, requiring an extra CSI feedback stage. Based on the previous theoretical analysis and measurements, however, partial reciprocities, including the spatial directional angles and number of propagation paths, do exist for FDD mmWave systems. Given such partial reciprocities, we propose a beamforming scheme based on a predefined codebook. Furthermore, we consider the channels with multipath where multiple codewords can be selected for some mobile stations (MSs) to utilize more than one propagation paths, which improves the link reliability. We then propose the corresponding beamforming strategy that not only identifies the number of paths, but also reconstructs the individual single-path channels which are easier to deal with at the beamforming stage. To select the optimal codewords for these single-path channels with lower complexity, we then present a DNN-aided approach that predicts the optimal codewords. The integration of DNN resolves not only the issue of power leakage, but also angular ambiguity in the codebook. Simulation results demonstrate that the performance of the proposed DNN-aided approach is very close to that of an exhaustive search.

Denoising is an indispensable step in seismic data processing. Deep-learning-based seismic data denoising has been recently attracting attentions due to its outstanding performance. In this letter, we investigate the architecture of deep Convolutional Networks (ConvNets) for seismic data denoising. The untrained ConvNets are served as a generative network to a single seismic data profile with Gaussian noise. Starting with random initialized parameters, the generative networks with various handcrafted architectures have different ability to map the seismic data at iterations and can separate the Gaussian noise as residuals. For the purpose of exploring the ability of Gaussian noise separation, the depth, width, and skip connection as the main components of generative network are assembled as various architectures to fit Gaussian noise, clean, and noisy seismic data, respectively. Then, the favorable network architecture with high and low impedance (an ability to hinder data reconstruction) to noise and seismic data is adopted as prior model to seismic data denoising task. Furthermore, a stopping criterion is designed for the data fitting process to obtain the latent clean seismic data automatically. The proposed method does not need data sets for training and it makes use of network architecture as prior. Extensive experiments both on synthetic and field data demonstrate the effectiveness of the selected ConvNet and the advantages are evaluated by comparing the denoising results with f-x multi-channel singular spectrum analysis (MSSA) and state-of-the-art unsupervised neural network (NN)-based method.

Due to the constraints of natural environments, acquired prestack seismic data is usually not complete, which seriously affects subsequent seismic data processing. With the progress of deep learning, many neural networks with different structures have been applied to missing seismic data interpolation. Among them, U-net can efficiently interpolate the regularly and irregularly missing seismic traces with small gap. While for consecutively missing seismic traces with big gap, the interpolation results for low amplitude missing components need to be further improved. In this letter, we analyze the variation of interpolation results for consecutively missing seismic traces during the traditional U-net training process, and find that U-net tends to only interpolate the high amplitude missing components. Meanwhile, due to the distribution difference between low and high amplitude seismic data, one U-net model is insufficient to interpolate both high and low amplitude missing components with a wide amplitude range. To improve the interpolation results of single U-net, we propose a multistage training process to train multiple U-net models. Each U-net model focuses on interpolating different missing components with a small amplitude range. In this way, more accurate interpolation results for low amplitude missing components can be obtained. Comparison experiments conducted on synthetic and field seismic data show that, under the same number of training epochs, the proposed training process can produce more accurate interpolation results comparing with traditional single U-net.

Missing traces interpolation is a basic step in the seismic data processing workflow. Recently, many seismic data interpolation methods based on different neural networks have been proposed. The existing research shows that when the seismic data is consecutively missing, the larger gap for missing traces, the more difficulty of interpolation, due to the convolution operation in the neural network can only capture local relations. In this letter, we incorporate the coordinate attention block to the Unet for 2D successively missing traces interpolation. The hybrid loss function combined with structural similarity (SSIM) and L1 norm is used as the loss function to further improve the interpolation performance of the designed network. Comparison experiments on 2D synthetic and field seismic data show that the interpolation results obtained by the proposed method are more accurate and reasonable compared with Unet and Unets equiped state-of-the-art similar modules.

Fault detection is a crucial task in seismic structure interpretation. Convolutional neural network (CNN)-based methods, in general, require large amount of labeled data for network training. One way to build the labeled data is to create synthetic seismic images with corresponding fault labels. However, it is hard to ensure that the synthetic data have the same fault feature distributions as the field data, which may lead to inaccurate and unreliable prediction results. Another way is to manually label the faults, which is time-consuming and subjective. In this letter, we propose that using knowledge distillation (KD) to improve the performance of fault detection by integrating the features from large number of synthetic samples and a small number of field samples. We distill knowledge from an ensemble of two teacher CNNs to train a student CNN (applied to final target) for seismic fault detection. In our work, one segmentation teacher CNN is trained on synthetic samples with known ground truth fault labels and another classification teacher CNN is trained on field samples with manually picked labels. Then, a classification student network is trained on samples generated by voting the results from two teacher models. The student CNN learns not only the general fault characteristics in the synthetic data but also the specific fault features of the target field data. Test on the field data shows that the student CNN highlights seismic fault more accurately with higher resolution than the teacher CNNs.

The hybrid beamforming architecture where a combination of analog and digital precoders is employed has attracted much attention recently for its flexible trade-off between power consumption and performance. A codebook-based beamforming scheme is generally considered as a more practical solution rather than the adaptive beamforming schemes due to the large number of antennas. However, the constant amplitude constraint brought by the phase shifters (PSs) in the analog precoder always makes the codebook design problem a tricky and complicated one. In this work, we consider a novel combined PS (CPS) architecture where one high-resolution PS is connected to each RF chain in parallel with low-resolution PSs to each antenna. The corresponding codebook design aiming for ideal flat beam gains is then proposed for the CPS architecture. Simulation results reveal that the proposed scheme achieves a good trade-off between performance and energy efficiency compared with conventional single PS (SPS) and the recently proposed double PS (DPS) architecture.

This paper proposes a 7-dimensional ( 7D ) Cellular Neural Network (CNN) based offset quadrature amplitude modulation filter bank multicarrier (OQAM-FBMC) encryption scheme for seven core fiber. The chaotic sequences generated by 7D CNN are applied to produce the masking vectors to encrypt the phase, carrier frequency, and time. In order to verify the performance of the encryption scheme, 70 Gb/s (7×10 Gb/s) encrypted OQAM-FBMC signal transmission over 2 km 7 core fiber is experimentally demonstrated. The key space of 7D CNN can reach 101575 and the scrambling degree can be maintained at 100% regardless of the number of symbols. The experimental results also show that when some keys are compromised, the system's bit error rate (BER) can still reach above 0.46, which effectively ensures the security of the system. Due to its good performance in security, the proposed scheme has important application prospects in future optical access network.

Both information age and delay are critical performance metrics of emerging time-sensitive applications. However, their inherent correlation in the finite block length (FBL) regime has remained uninvestigated as it is affected by block length and update rate in a complex manner. In this paper, closed-form expressions for average age of information (AoI), peak AoI (PAoI) and delay are derived for an FBL Last-Come First-Served system with retransmission and non-preemption policies, based on which a comprehensive analysis of the relationship among the three metrics in the FBL regime is presented. It is proved that there exists a strong tradeoff between delay and AoI/PAoI given a block length, and that AoI, PAoI and delay have positive correlation given an update rate, regardless of the weight. With the goal of minimizing delay and AoI simultaneously, the weighted sum of delay and PAoI (upper bound on AoI) is formulated and proved to be convex with respect to block length and update rate. A low-complexity optimization algorithm is developed with a closed-form expression of the optimal update rate, whose performance approaches the Pareto boundaries of the PAoI-delay and the AoI-delay regions, at much lower complexity than exhaustive search.

In this article, we focus on a wireless-powered body area network in which the simultaneous wireless information and power transfer (SWIPT) technique is adopted. We consider two scenarios based on whether sensor nodes (SNs) are equipped with battery. For the first time, energy consumption minimization with throughput heterogeneity (ECM-TH) problem is addressed for both scenarios. For the battery-free scenario, a low-complexity time allocation scheme is proposed. This scheme solves the ECM-TH problem based on a hybrid method of gradient descent and bisection search algorithms. Consequently, compared with the interior-point method, our scheme has a lower computational complexity for the same energy consumption performance of the network. For the battery-assisted scenario, the nonconvex ECM-TH problem is first transformed into a convex optimization problem by introducing auxiliary variables. Then, a joint time and power allocation scheme based on the Lagrange dual subgradient method is proposed to solve it. Compared with the battery-free scenario, energy consumption and outage probability are both decreased in the battery-assisted scenario. Moreover, we address a special case wherein the feasible set of the above-mentioned ECM-TH problems may be empty owing to poor channel conditions or high throughput requirements of SNs.

A fast beam training scheme for the massive antenna arrays is a key to frequency division duplex (FDD) mmWave systems, as channel reciprocity between the down-link (DL) and up-link (UL) channels does not hold in general, requiring feedback mechanism for DL beam selection. Due to the large number of antennas in a massive array, it is not very practical to conduct an exhaustive search, especially when considering the small angular coverage of one directional narrow beam (hence the number of candidate beams). To address such a challenge, we consider the 3D beam matching problem for a multi-user massive multiple-input multiple-output (MIMO) system, and propose a two-stage hierachical codebook along with the corresponding fast beam training scheme. The proposed codebook contains a primary codebook for hierarchical beams and an auxiliary codebook for narrow beams following the limited resolution of the phase shifters (PSs). The fast beam training scheme based on the codebook not only reduces beam training overhead, but also is applicable to the scenario where there exist multiple propagation paths for one mobile station (MS). Numerical results show that the proposed scheme not only enjoys a lower beam matching complexity (i.e. the training overhead), but also achieves a comparable performance to the exhaustive search scheme.

An interference alignment (IA) and independent component analysis (ICA)-based semiblind scheme, referred to as IA-ICA, is proposed for downlink dual-user power-domain nonorthogonal multiple access (NOMA) systems in high-reliability low-latency (HRLL) Internet of Things (IoT). At the base station (BS), one user is converted to constructive interference to the other user via phase alignment of each symbol. At both user ends, ICA is used for semiblind signal detection. The phase rotation via nonredundant precoding at the BS does not introduce any spectral overhead, while only 1–2 pilot symbols are required for elimination of ICA incurred ambiguity. Closed-form expressions are derived for the users’ symbol error rate (SER) performance in Rayleigh fading with 4-quadrature amplitude modulation (4-QAM), which matches the simulation results very well. Based on the analytical results, we propose an efficient power allocation algorithm that is based on statistical channel state information (CSI) only, and therefore, the signaling overhead involved is negligible. In particular, a near-optimal SER performance can be achieved with equal power allocation between the two users. The proposed IA-ICA-based semiblind NOMA system demonstrates a much better SER performance than the existing approaches even though they are under perfect CSI. Hence, it is a feasible solution for HRLL IoT, with high reliability and very low spectral and signaling overheads.

A novel tunable vector-sum filtering power divider (FPD) with arbitrary output phase difference is proposed. It is constructed by a tunable FPD [which has reconfigurable in-/out-of-phase output and arbitrary power division ratios (PDRs)] on the left and a wideband branch-line coupler (BLC) on the right. The output wave from the left FPD forms two orthogonal vectors at each output port of the BLC. The orthogonal vectors are summed up and result in a new vector with phase that is purely determined by the amplitude ratio of the orthogonal vectors. As a result, the phase difference between the new vectors at the outputs of the BLC can be continuously tuned by adjusting the PDR of the left FPD. Detailed theoretical analysis has been given. It shows that the phase difference between the output ports can be flexibly tuned from 0 °C to 360 °C within the full frequency tuning range.

The vertical spatial distribution of dust particles in radio frequency (RF) discharge plasma sheath is studied. The dust particles are levitated above the powered electrode in the discharge. From the distribution of the dust particles of different sizes generated by mixing of suitable gases, the effect of changing the RF power on the dust cloud behavior is studied. It is found that the distribution of the different sized dust particles in the sheath with variable bias differs significantly from that of single sized dust particles with fixed bias. As the RF power increases, the distribution range as well as the levitation height of the dust cloud above the powered electrode decrease. A theory for the observed results is also given. Based on the experimental and numerical results, we attribute the observed dependence to a combination of the enhanced sheath electric field and the ion drag force, but the decrease of the sheath thickness with increase of the discharge power has only minor contribution.

In this paper, the design methodology of a microstrip angular displacement sensor based on an impedance transformer with a quasi-Chebyshev frequency response is proposed. This sensor with impedance transforming function can realize the angular displacement detection by bandwidth sensing, and matching sensing simultaneously. The insertion loss in stopband achieves 24.9 dB to 32.6 dB with an average bandwidth selectivity of 2.25 MHz/degree of the sensor centered at 2.4 GHz with two poles of 1.71 GHz and 2.81 GHz. Meanwhile, the average matching level of return loss for displacement sensitivity are 0.18 dB/degree and 0.15 dB/degree, respectively. It is the first time to apply an impedance transformer into the angular displacement sensor design for both bandwidth and matching level sensing among reviewed literatures. In this design, the load-to-source impedance transformation is designed for 250-to- $50~\Omega $ , and by adding a rotated stub on the proposed microstrip transformer, the sensing function is developed with angular displacement.

In this article, for the first time, a novel four-port coupler with reconfigurable multifunctions of filtering rat-race coupler (FRC), filtering branch-line coupler (FBC), and conventional single-ended two-port multipole bandpass filter (BPF) is proposed. It simply consists of four varactor-loaded stepped-impedance resonators that are arranged symmetrically. Through appropriate coupling-type (electrical coupling or magnetic coupling) conversion, 90° or −90° phase shift between the adjacent resonators can easily be obtained and works as a key factor in the transformation between different filtering coupler modes. With this simple configuration, the proposed coupler achieves reconfigurable FRC and FBC with not only operating frequency tuning over a wide tuning range but also bandwidth and power division ratios (PDRs) control for each filtering mode. In addition, by imposing open circuit at any two ports, the four-port coupler can be converted to the conventional single-ended two-port BPFs with reconfigurable filter orders of second, third, and fourth. The measured results show that the FRC and FBC can continuously cover the frequency ranges of 1.21–1.67 and 1.2–1.61 GHz with −1-dB bandwidth tuning of 60–120 and 55–110 MHz, respectively. The 3- and 6-dB PDRs have been demonstrated for both filtering coupler modes. In single-ended two-port BPF mode, the frequency tuning ranges are measured of 1.12–1.6, 1.16–1.6, and 1.2–1.61 GHz, respectively, for the proposed two-port second-, third-, and fourth-order BPFs. Measured and simulated results are in good agreement, validating the proposed design method.

In this article, we propose a semi-blind full-duplex (FD) amplify-and-forward (AF) relay system with adaptive self-interference (SI) processing assisted by independent component analysis (ICA) for low-latency and high-reliability (LLHR) Internet of Things (IoT). The SI at FD relay is not necessarily canceled as much as possible like the conventional approaches, but is canceled or utilized based on a signal-to-residual-SI ratio (SRSIR) threshold at relay. According to the selected SI processing mode at relay, an ICA-based adaptive semi-blind scheme is proposed for signal separation and detection at destination. The proposed FD relay system not only features reduced signal processing cost of SI cancelation but also achieves a much higher degree of freedom in signal detection. The resulting bit error rate (BER) performance is robust against a wide range of SRSIR, much better than that of conventional FD systems, and close to the ideal case with perfect channel state information (CSI) and perfect SI cancelation. The proposed system also requires negligible spectral overhead as only a nonredundant precoding is needed for ambiguity elimination in ICA. In addition, the proposed system enables full resource utilization with consecutive data transmission at all time and same frequency, leading to much higher throughput and energy efficiency than the time-splitting and power-splitting-based self-energy recycling approaches that utilize only partial resources. Furthermore, an intensive analysis is provided, where the SRSIR thresholds for the adaptive SI processing mode selection and the BER expressions with ICA incurred ambiguities are derived.

Convex edges are located on convex surfaces of hidden geometry spaces, resulting in shorter geometrical distance than observed. Convex edges reveal the close relationships between two connected nodes and have a key role in many research fields and applications, including social, trust, and protein–protein interaction networks. We first quantitatively define convex edges using hidden geometry space defined by Ricci curvature and show these to be capable of predicting highly interacted edges in social networks. From empirical studies of social networks, we find that convex edges often occur on two connected nodes with similar popularity and allow for frenemy (or versatile node) identification, and that almost all nodes maintain an upper bound of convex edges. These findings present further evidence of the limited attention in an attention economy from a different perspective. Using three well-known synthetic networks, we show that convex edges can reflect the formation of a network from scatter (concave) to aggregation (convex). Taken together, our results show that convex edges are a novel mechanism to investigate social networks from different disciplines and offer insights into many network phenomena.

In this article, a novel tunable isolation network and a novel ring-type reconfigurable single-to-balanced (SE-BAL), power-dividing (PD), and single-to-single-ended (SE) bandpass filter is proposed. The tunable isolation network consists of a simple $\pi $ -type resistive network with tunable varactors, while the proposed filter consists of three ring-type stepped impedance resonators (SIRs) connected by varactors in between. By simply combining the proposed tunable isolation network and the proposed ring-type reconfigurable filter, it can achieve multiple reconfigurable functions among SE-BAL filtering, PD filtering, and conventional SE filtering with fully matched three ports, arbitrary power-division ratios (PDRs), high output isolations, and excellent amplitude and phase balance, in addition to continuous frequency and independent bandwidth tuning. The proposed filter is demonstrated using the microstrip line technology. The measured results show that the SE-BAL, PD, and two SE filtering functions can continuously cover the frequency ranges of 1.03–1.61, 1.01–1.53, 1.03–1.56, and 1.06–1.48 GHz with −1-dB bandwidth tuning of 35–80, 50–140, 60–130, and 70–140 MHz, respectively; the 1-, 3-, and 5-dB PDRs have been demonstrated for both SE-BAL and PD filtering modes. The measured output return losses and isolation are better than 19 and 18 dB, respectively, for SE-BAL filtering mode and 20 and 16 dB, respectively, for PD filtering mode. Good agreement is obtained between the simulated and measured results.

A dual-band dual-polarized base station antenna for the fifth-generation (5G) mobile system is presented in this paper. The proposed antenna covers the frequency bands from 3.3 to 3.8 GHz (the lower band) and from 4.8 to 5.0 GHz (the upper band) with good isolation between its ports (≥ 20 dB). It consists of two double-oval-shaped dipoles, two double-oval-shaped feeding lines and a cavity reflector. In this design, parts of the dipole antenna structure are used as the feeding lines and it is found that using one arm of the dipole to feed the whole antenna can improve impedance matching. The dual-band performance is achieved by integrating a small oval-shaped loop within the large oval-shaped loop without increasing the size of the radiating patch. The size of the radiating patch is only 0.26λ0 × 0.26λ0 (λ0 is the free-space wavelength at 3.3 GHz). The cavity reflector improves the gain performance and reduces the overall size of the antenna, which is only 0.66 λ0 × 0.66λ0 × 0.2λ0. The antenna has an average realized gain of 7.56 dBi in the lower band and 7.42 dBi in the higher band. Meanwhile, for both bands, the radiation pattern is stable, and the half power beamwidth is within 65° ± 5°. Both simulated and measured results demonstrate that the antenna is a very good candidate for 5G mobile base stations.

Broadband power line communication (BPLC) and television white space (TVWS) are regarded as promising candidates for indoor broadband applications of the Internet of Things. However, they share the access to the very high frequency (VHF) band, which could cause harmful interference and performance degradation to each other. In this paper, a TVWS regulated BPLC system is proposed for point-to-multipoint downlink communication, which integrates the requirement of primary user sensing and the permissible transmission power spectral density (PSD) for TVWS users into the BPLC standard, regarding VHF band access. This integration guarantees minimum interference level between TVWS and BPLC and allows higher transmission PSD for BPLC users in VHF, and hence higher capacity and coverage for BPLC.