最新刊期
卷 52 ， 期 8 ， 2024
- 摘要：With the rapidly increase of the construction density of 5G micro base stations, the traditional cable power supply mode will greatly increase the construction cost and layout difficulty of 5G network. The co-transmission technique of signal and power over optical fiber as a new power supply mode, uses optical fiber as the transmission medium and co-transmits 1 064 nm energy light and 1 550 nm signal light, which can not only meet the high-speed 5G signal fronthaul, but also meet the centralized energy management of 5G micro base station. In this paper, a co-transmission scheme of signal and power over optical fiber based on space division multiplexing technology is proposed. 10 W energy light and 1.5 Gbit/s 5G new radio (NR) signal light are co-transmitted in the 10 km weakly-coupled seven-core fiber experiment. The 0.42 W electric power can be obtained by a photovoltaic converter with an optical-to-electrical conversion efficiency of 35%, which can drive a remote antenna unit (RAU). The fluctuation range of received optical power of energy light for 7 hours continuous monitoring is less than 0.4%. Meanwhile, the error vector amplitude value (EVM) of the received 5G NR signal is only 0.38%, achieving a record 6.3 W·（Gbit/s）·km electric power-capacity-distance product.关键词：space division multiplexing;5G fronthaul;high power laser;signal and power-over-fiber9|10|0更新时间：2025-12-08
- 摘要：With the increasing demand of optical communication data rate, the nonlinear effect in optical fiber has become the main factor limiting the growth of channel capacity. To solve this problem, this paper proposed a nonlinearity-tolerant geometrically-shaped (GS) four-dimensional (4D) modulation optimization method is proposed based on the 4D nonlinear interference (NLI) model. The transmission performance of four different modulation formats (including the 2D and 4D modulation formats designed for additive white Gaussian noise (AWGN) channel and optical fiber channel, respectively) are compared in detail via the simulation in the AWGN channel and single-span optical fiber transmission systems. The numerical results show that the constellation shaping based on the joint optimization of 4D NLI model and 4D space can not only achieve higher linear shaping gain but also provide better nonlinear tolerance. Taking the 5-channel wavelength division multiplexing transmission system as an example, the optimized NL-4D-1024 format can reduce the transmitted power by 0.8 dB and 0.4 dB compared with the traditional quadrature amplitude modulation (QAM) and 2D-GS modulation, respectively.关键词：multi-dimensional modulation formats;geometric shaping;nonlinear interference model;optical fiber channel;phase noise9|13|0更新时间：2025-12-08
- 摘要：Based on the four-branch CLEFIA-like dynamic cipher structure, this paper analyzes the 2m-branch CLEFIA-like dynamic cipher structure, proving the equivalence between the cyclically permuted transformation-based CLEFIA-like dynamic cipher structure and the CLEFIA cipher structure.Furthermore, the upper bound on the minimum number of differentially active rounds for the 2m-branch CLEFIA-like dynamic cipher structure is investigated. It is demonstrated if each round transformation
- 摘要：For the neural network-based speech Packet Loss Concealment (PLC), the input features are crucial factors that directly affect the final recovery performance. Additionally, the challenge of restoring high natural speech through PLC remains to be addressed. To effectively recover packet loss speech and improve its naturalness, this paper proposes a PLC method of speech signal based on the priori Mel-spectrum and neural vocoder. The proposed method adopts an asymmetric encoding and decoding network structure. At the encoding stage, this method utilizes two independent encoding networks to extract the latent time-frequency features from the waveform and Mel-spectrogram, respectively. At the decoding stage, the latent time-frequency features are jointly fed into a neural vocoder which is composed of several temporal adaptive denormalization layer to restore the lost speech signals and enhance the naturalness. Simulation experiments demonstrate that the proposed method outperforms two existing packet loss concealment algorithms in terms of perceptual evaluation of speech quality and short-time objective intelligibility.关键词：packet loss concealment;Mel-spectrum;neural vocoder;temporal adaptive de-normalization layer;time-frequency features11|15|0更新时间：2025-12-08
- 摘要：With the development of smart ocean, the transmission of multimedia data in underwater acoustic networks (UANs) has received much attention from scholars. The highly dynamic topology of UANs leads to incomplete data transmission between neighboring nodes, and the different portions of compressed data such as underwater images or videos have different effects on their reconstruction quality. Hence, UANs require coding mechanisms with unequal error protection (UEP) to encode and decode multimedia data. The recursive online fountain code with limited feedback (ROFC-LF) has the advantages of low overhead, less feedback and simple compiled codes, which is suitable for UANs. Combined with the characteristics of underwater acoustic channels, such as narrow bandwidth, long delay and energy limitation, this paper systematically analyzes the problem of the cycles existing in the build-up phase of ROFC-LF and proposes two optimization objectives to address the cycle problem as well as the UEP problem. In addition, a circle-avoidance ROFC-LF with UEP is presented for UANs. This coding mechanism reduces the number of useless encoded packets due to the presence of cycles in the largest component during the the build-up phase, which in turn decreases the energy consumption. To achieve the UEP property, a weighted-selection strategy is used in the build-up phase, whereas a priority strategy is employed in the completion phase. The proposed coding mechanism is analyzed based on the random graph theory, and the theoretical results are consistent with the simulation experimental results. The results show that the proposed coding mechanism can quickly recover important data while reducing the number of coded packets, and is suitable for transmitting multimedia data of varying importance in UANs with dynamically changing network topology.关键词：underwater acoustic networks;recursive online fountain code with limited feedback;unequal error protection;circle-avoidance;weighted-selection;data priority10|12|0更新时间：2025-12-08
- 摘要：Based on the decomposition of surface EMG (ElectroMyoGraphy) signals, the most primitive components of EMG signals are restored. By analyzing the features of the decomposed sEMG (Surface ElectroMyoGraphy) signals, the EEG (ElectroEncephaloGraph) information transmission law in neuromuscular system can be studied, and the bioelectric information transmission mechanism can be explored from the essence of human motion. The EEG and sEMG signals corresponding to 15% MVC and 30% MVC of 9 subjects are collected. Then, the motion unit action potential (MUAP) is obtained by STA template reconstruction decomposition of sEMG based on morphological decomposition. Then, the amplitude of MUAP, the number of MUAP and the firing rate are extracted as features, based on these features the trend of synchronous EEG signals and the transmission entropy value, the information transmission law of cortical muscles is explored. Result Under different strength levels, the three features under 30% MVC are larger than those under 15% MVC, but the change trends of the three features are same. When the EEG signal forms a peak or valley signal, MUAP Number, Amp and firing rate all show a trend of increasing. The increasing trend of MUAP amplitude is the most obvious, and the coupling effect of MUAP amplitude features and synchronous EEG signals is the best (transmission entropy value). Conclusion although different strength levels affect the strength of the extracted EEG signals, the overall information transfer law is consistent: when the intensity of muscle contraction EEG signals is enhanced to form peak or trough signals, the three features show an upward trend, and the MUAP amplitude feature is the best response among the three selected features, that is, this feature can better reflect the information transmission law of neuromuscular system in the process of human motion control.关键词：sEMG signal;EEG signal;sEMG decomposition;corticomuscular coupling;entropy of transfer4|12|0更新时间：2025-12-08
- 摘要：The problems of difficulty and low efficiency in decomposing multicomponent nonstationary nonlinear signal with complex time-frequency characteristics such as contiguity, overlap and intermittency in time-frequency domain are solved. Based on the time-frequency distribution of signal, a multicomponent nonstationary signal decomposition method called time-frequency filtering decomposition (TFFD) is proposed. TFFD gets the fitting IF curve which is consistent with the instantaneous frequency (IF) of the components by fitting the time-frequency datum points which can reflect the instantaneous characteristics and laws of the components in the signal. Based on the time-frequency coordinates of fitting IF curve, the distribution area of component is determined by setting the distance threshold condition. Thus, a time-frequency filter bank is constructed, which is based on fitting IF curve time-frequency coordinates as the central frequency and the bandwidth of distribution area as the passband width, to achieve time-frequency filtering decomposition for multicomponent signal. Through the analysis of the simulation and the actual signal with the representative time-frequency characteristics, and the comparison with the classical signal decomposition methods, it is proved that the TFFD method has good decomposition ability and efficiency.关键词：multicomponent signal;signal decomposition;time-frequency distribution;instantaneous frequency;time-frequency filtering8|18|0更新时间：2025-12-08
- 摘要：To address the challenge of weak target detection in nonhomogeneous sea clutter environments, this paper proposes an asymmetric encoder-decoder network (AED-Net) for sea clutter suppression based on complex-valued residual units and attention gates. The network takes the complex-valued signal generated by radar echoes passing through the matched filter as input. First, it replaces conventional convolutional units with complex-valued residual units to extract features of weak targets and sea clutter, enhancing the network’s capability of feature extraction while avoiding feature degradation. Then, attention gate modules are employed to selectively propagate the feature information extracted by each module in the encoding path to the corresponding modules in the decoding path. Finally, it yields the complex signal after sea clutter suppression as output. Due to the capability of independently selecting the input and output dimensions of each attention gate according to the network structure, the proposed network has an asymmetric encoder-decoder structure. Compared to typical symmetric encoder-decoder network, UNet, the introduction of complex-valued residual units and attention gates significantly reduces the redundancy of feature information, enhances feature extraction and transmission, and thus improves the sea clutter suppression performance. Meanwhile, complex-valued residual units have much smaller parameter size than convolutional units and the introduced attention gates greatly reduce the number of units in the decoding path, resulting in a significant reduction of total network parameters. Experimental results based on real sea clutter data demonstrate that compared to the complex value-UNet (CV-UNet) network, AED-Net achieves an average improvement of 9 dB in the output signal-to-clutter ratio (SCR) and can effectively operate at a minimum SCR reduction of 3 dB. Moreover, the number of parameters and the computational cost are reduced by 57.8% and 50%, respectively.关键词：sea clutter suppression;encoder-decoder network;residual unit;attention gate;complex-valued signal11|14|0更新时间：2025-12-08
- 摘要：This article proposes a dual-mode millimeter-wave on-chip transmission line (TL) based on the composite right-/left-handed (CRLH) theory, featuring both wide bandwidth and low loss. In addition, both the right-handed (RH) band supporting the forward waves and the left-handed (LH) band supporting backward waves are realized. To address the issues of high loss and high parasitic effect in millimeter-wave range and on semi-conductor substrate, a T-type CRLH on-chip unit cell is proposed. To further reduce the transmission loss and extend the bandwidth, the circuit topology and structural realization are investigated. In addition, dual modes including the RH and LH bands are realized by analyzing the dispersion curve of the CRLH transmission line. Based on Gallium Arsenide (GaAs)-based integrated passive device (IPD) technology, a 3rd-order periodic CRLH TL ranging from 18~44.5 GHz was fabricated and measured. The proposed transmission line features a compact size of 1.3 mm × 0.55 mm including GSG (Ground Signal Ground) pads, a minimum insertion loss of 1.18 dB, and a 3 dB fractional bandwidth exceeding 90%. These measurement results demonstrate that the proposed on-chip transmission line can be widely applied in on-chip functional circuits and systems for unique electromagnetic phase characteristics.关键词：radio-frequency circuits;millimeter-wave;on-chip circuits;transmission line;metamaterial4|11|0更新时间：2025-12-08
- 摘要：The flexible sensor based on photo plethysmo graphy (PPG) can detect heart rate (HR) and blood pressure (BP), but the calibration of their detection results is rarely reported. Therefore, this paper proposes a reflective PPG heart rate detection and blood pressure calibration system based on simulated blood circulation.The peristaltic pump is used to generate pulsating flow,and the frequency and pressure of simulated blood delivery are controlled by adjusting its rotational speed,thus causing the change of the volume of simulated blood in the elastic latex tube and changing the signal period and intensity of reflected light, which is closer to match the actual scenario of human pulse measurement process.The mean value of heart rate detection error of the system is 0.277 78, and the 95% consistency limit is (-2.595 62, 3.151 17). The goodness of fit of measured systolic blood pressure (SBP) and diastolic blood pressure (DBP) are 0.971 85 and 0.981 11, respectively. The mean value of mean deviation (MD) ± standard deviation (SD) of SBP and DBP detected by the calibrated flexible PPG sensor on four volunteers is (1.21±2.16) mmHg and (0.76±2.02) mmHg, respectively, which are in line with and far less than the standard index of 5±8 mmHg for measuring the accuracy of blood pressure monitors set by the association for the advancement of medical instrumentation (AAMI). The results show that this system can calibrate the flexible PPG sensor accurately and efficiently, which provides the calibration basis for the accurate blood pressure detection of portable wearable devices.关键词：flexible wearable sensor;photo plethysmo graphy;pulse wave transmission time;blood pressure calibration;heart rate detection4|13|0更新时间：2025-12-08
- 摘要：Optical fiber is the most stable medium for signal transmission in modern communication networks. Increasing the number of fiber modes is an effective way to meet the increasing demand for communication capacity. In this paper, a kind of uncoupled three-mode-group (LP01, LP11, LP21) graded-index fiber with low bending loss is designed. The strict cut-off conditions of the mode groups in the graded-index optical fiber are obtained by using the stratified method. In addition, the effective mode area and bending losses are calculated by a finite element simulation software, and the design region of three-mode-group transmission in the C+L (1 530~1 625 nm) band with low bending loss is obtained. It is revealed that the combined effects of the concave refractive index at the core center and the trench layer can promote the cutoff of radial higher-order modes in the fiber. The simulation results show that the designed fiber has a large effective index difference between mode groups that is ≥ 3×10-3, and the maximum bending loss is 1.3 dB/km when the bending radius is 30 mm, which meets the low bending loss condition recommended by ITU-T. In addition, at the wavelength of 1 550 nm, the degenerate modes in the same mode group have a relatively small differential mode group delay (≈176 ps/km), and the crosstalk between degenerate modes can be compensated by 4×4 multi-input multi-output technique. This fiber can be used in the C+L band to realize uncoupled space-division multiplexing signal transmission.关键词：few-mode fiber;graded-index;low bending loss;space-division multiplexing;optical fiber communications8|10|0更新时间：2025-12-08
- 摘要：In this study, a design method of multilayer structure composite absorber based on double layers of metasurface is proposed. The designed composite absorber consists of two layers of metasurface, top absorption enhanced skin and several support dielectric slabs. The unit cells of metasurface Ⅰ and Ⅱ are separately irregular-shaped metal patches connected by chip resistors and hexagonal metal rings loaded with chip resistors; the top absorption enhanced skin is a fiberglass enforced epoxy laminate; support dielectric slabs adopt PMI foam. The simulation results indicate that the absorption frequency bands with reflection coefficients below -10 dB and -20 dB are 2.80~23.64 GHz and 3.56~22.56 GHz, respectively. The measurement results show that, the absorption frequency bands with reflection coefficients below -10 dB and -20 dB are 2.36~23.87 GHz and 3.17~23.16 GHz, respectively; the reflection coefficient curves obtained by simulation and test have good consistency, which verifies the effectiveness of the design method. The simulation and test results show that the frequency band of -10 dB reflection coefficient at 50° oblique incidence is basically consistent with that at normal incidence, while the offset of start and stop frequencies is less than 0.8 GHz; further, the simulation and test results show that when the oblique incidence angle is 60°, the fractional bandwidth of the reflection coefficient below -10 dB is up to 141.8%, which indicates that the composite absorber designed in this study has the incidence stability in a wide angle range. In addition, the mechanism of ultrawideband and high absorption of the structure and the influence of the main structure parameters are analyzed; the results demonstrate that the top absorption enhanced skin can improve absorptivity of the whole structure by up to 0.2(1.0 represents 100% absorption), and the complementary enhancement design of the two layers of metasurface absorption frequency bands can improve oblique incidence stability obviously.关键词：absorber;metasurface;multi-layer structure;ultrawideband;oblique incidence11|15|0更新时间：2025-12-08
- 摘要：A new low-profile dual-band patch antenna for BeiDou system (BDS) applications is designed in this paper. The antenna is based on a single-layer patch antenna structure, and a new resonant frequency is introduced through loading “bridge-type” gaps, so that the antenna operates in dual frequency bands (resonant at 1 176 MHz and 1 575 MHz, respectively), realizing coplanar dual-frequency radiation. Meanwhile, four annular gaps are loaded and only one set of feeding pins is used for coupled feeding, which contributes to adjust impedance matching while reducing the number of feeding pins and simplifying the feeding network. In addition, “π type” metal branches are installed on the side of the antenna to further optimize impedance matching in high frequency band. The measured results show that the gain of the antenna is greater than 3.5 dBic in both the B1C band and B2a band, and the axial ratio are less than 3 dB in the above frequency bands. The spectrum shows that the developed antenna operates in the B1C and B2a frequency bands of Beidou system, and is expected to be applied to the terminal equipments of Beidou positioning system.关键词：patch antenna;dual-band;gaps loading;Beidou System;low profile13|14|0更新时间：2025-12-08
- 摘要：In this paper, a novel high-bandwidth density and low-power 7-order correlated NRZ (Non-Return-to-Zero) coding interface circuit for D2D (Die to Die) interconnection is proposed. In order to further improve the SNR (signal-to-noise ratio) and bandwidth density of 5-order correlated NRZ coding, this paper designs encoding and decoding circuits based on transmission and reception matrices. Based on the transmission matrix, a voltage-mode encoding circuit is designed at the transmitting end to effectively reduce power consumption. Based on the reception matrix, a decoding equalization circuit based on active adjustable inductor is designed at the receiving end to improve communication speed. In order to solve the problem of clock skew at the receiving end, this paper also designs an error correction circuit. The interface circuit is designed using 28 nm CMOS (Complementary Metal Oxide Semiconductor) technology, with a core area of 3 mm2, and can be applied to on-chip interconnects ranging from 10 to 50 mm. The backend simulation results indicate that, under the condition of a Nyquist frequency of 20 GHz and a channel loss of -8 dB, the receiver's narrowest eye width is 0.45 UI, with a bit error rate less than
- 摘要：The analog-to-digital converter (ADC), a key component bridging the analog and digital signal domains, is critical to the interface between these domains. However, existing research falls short in providing comprehensive models that take into account radiation effects on ADCs. To address the demands of simulating radiation effects in large-scale mixed-signal systems, this paper presents a method for developing behavioral-level models of ADCs that embody radiation effects. The paper begins by dividing the ADC into its constituent generic modules based on its working principles, and establishing behavioral models for each module using the very high speed integrated circuit hardware description language for analog and mixed signals (VHDL-AMS). These modules are then dynamically assembled according to fundamental principles to form the baseline ADC model under non-irradiated conditions. To address the radiation effects on ADCs, irradiation experiments were conducted to quantify the performance parameters of two ADC chips, HWD7710 and SAD9434, under total ionizing dose (TID) and neutron radiation (NR) exposure. The operating parameters affected by these radiation effects were meticulously measured. Least squares regression was then used to derive equations correlating the functional parameters of the ADC with radiation dose. Finally, TID and NR models for two different structured ADCs were developed by incorporating radiation parameter modules into the basic model based on the radiation relation equations. The universality and accuracy of the constructed ADC radiation effect model was verified by comparing the simulated results with experimental data. The relative deviation between the static parameter simulation results and the experimental results was within 5%. The results show that this method supports the modeling of radiation effects for various ADCs and multiple radiation effects.关键词：analog to digital converter;modeling method;total ionizing dose effect;neutron radiation effect;static parameters;Dynamic parameters6|15|0更新时间：2025-12-08
- 摘要：Field-circuit coupling simulation methods can achieve multi-level collaborative analysis across chips, packages, and systems. Multi-physics simulation methods can be used to design the signal integrity, power integrity, and reliability of chip packaging under multiple physical constraints. The combination of field-circuit coupling simulation and multi-physics simulation has become a crucial direction for advanced packaging simulation technology. By utilizing these two types of simulations together, early-stage optimization of chip packaging designs can be achieved. This article discusses the current research status of field-circuit coupling simulation and multi-physics coupling simulation. The article also proposes a frequency-domain method for non-linear field-circuit co-simulation and an electromagnetic-thermal coupling simulation method. The frequency-domain method avoids the long iterative process of field-circuit coupling simulation based on time-domain methods, solving the multi-physics electromagnetic-thermal coupling analysis problem of non-linear field-circuit under multi-physics constraints. The calculation results verify the accuracy and efficiency of the proposed method关键词：field-circuit coupling simulation method;multiphysics coupling simulation method;harmonic analysis method23|11|0更新时间：2025-12-08
- 摘要：In the variational quantum algorithm, the selection of parameterized quantum circuit topology is of great significance to the performance of the algorithm. There are some problems in the existing topology, such as the large number of quantum gates required by the fully connected topology, and the slight lack of the expressibility and entanglement capability of the ring topology. This paper proposes a new block-ring (BR) topology to solve the problems above, which can effectively reduce the size of parameters (i.e. the number of parameterized gates) while ensuring performance. n qubits are divided into several blocks equally, each block contains m qubits. Qubits in the block are connected in pairs, and the blocks are connected by ring structure. In order to construct parameterized quantum circuits with BR topology, we designed a algorithm of generating multiple-layer circuit, which can automatically generate quantum circuits composed of single-qubit gate
- 摘要：As the most commonly used sensor type in the field of magnetic sensing, Hall sensors have been widely used in industrial production, automotive electronics, aerospace, biomedicine and many other fields. In this work, the AlGaN/GaN heterojunction Hall sensor was developed by using the standard semiconductor chip manufacturing process, taking advantage of the high temperature tolerance and high mobility of wide-bandgap GaN semiconductor. The sensor in high temperature application environment requires the packaging material to resist high temperature, produce low stress and have little impact on the key indicators of the device. In this work, the influence mechanism of three different packaging materials on the performance of the sensor and the bonding wire is simulated, the changes of sensor sensitivity, offset voltage, temperature drift coefficient and other physical parameters before and after packaging with different packaging materials are verified by experiments. Finally, the packaging material with the best comprehensive performance and can be used in 550 K high temperature environment is selected. The offset voltage of the final packaged Hall sensor is less than 115 μV when the excitation current is kept at 1.0 mA, and the signal linearity is better than 0.3% in the range of 0~3.0 mA input current and 0~1.0 T magnetic field. The temperature drift coefficient is only -120.9 ppm/K in the temperature range of 300~550 K, which is better than the results reported in the literatures. The fabricated Hall sensor is expected to be applied in the magnetic field detection in extreme environments.关键词：AlGaN/GaN heterojunction;Hall sensor;package;high temperature stability;wire bonding6|13|0更新时间：2025-12-08
- 摘要：Person re-identification is an important research direction in the field of computer vision, aiming to identify and track the same person across different surveillance cameras. Compared with image-based re-identification methods, the video-based re-identification method has richer temporal and spatial information, making it more efficient in real-world applications. Due to the existence of various temporal relationships between video frames, valuable information such as motion patterns and fine-grained features can be obtained. Therefore, how to effectively extract these temporal and spatial clues has become a key issue in video-based re-identification. In this paper, a long and short time Transformer (LSTT) network based on a temporal relationship is proposed to address the video-based person re-identification problem. The module includes long and short term relationship modules to extract important temporal information and enhance feature representation. The long-term relationship module stores information for each frame using a memory cue and establishes global connections for each video frame. The short-term relationship module considers interaction between adjacent frames to learn fine-grained target information and improve feature representation. Additionally, to improve the model’s adaptability to different target features, a multi-scale module with convolution kernels of different sizes is designed. The module has multiple convolution receptive fields and can more comprehensively cover the target area, further improving the model’s generalization performance. Experimental results on three datasets, namely MARS, MARS_DL, and iLIDS-VID, demonstrate that the LSTT model achieves state-of-the-art performance.关键词：video-based person re-identification;Transformer;the long-term temporal relationship;the short-term temporal relationship;multi-scale module6|12|0更新时间：2025-12-08
- 摘要：There exist some inherent defects in the traditional Pareto dominance relation in many-objective optimization space, while some modified dominance relations have much room for improvement in balancing the convergence and diversity of many-objective evolutionary population. Based on this, a new dominance relation (dominance relation based on the Automatically reduced region Associated with the Reference vector, A2R) automatically shrinks the region (niche) associated with the reference vector. Specially, it reduces the size of the niche from generation to generation along the whole evolutionary process in order to achieve a balance of convergence and diversity adaptively. Furthermore, the A2R does not introduce any additional parameter. In addition, the crowding distance measure based on Lp-normal form (p equals 1/M, and M denotes the number of objectives) is used to improve the diversity of the solution set in many-objective space. Finally, the above two strategies are embedded into the framework of NSGA-II (Nondominated Sorting Genetic Algorithm II), and then a many-objective evolutionary algorithm based on A2R (MaOEA/A2R) is designed. The MaOEA/A2R is compared with other five representative many-objective evolutionary algorithms (MaOEAs) on the DTLZ and WFG benchmark functions with 5-, 10-, 15-, and 20-objective in terms of IGD (Inverted Generational Distance) and HV (HyperVolume) indicators. The empirical results overally show that MaOEA/A2R can obtain better convergence and diversity. Therefore, the proposed MaOEA/A2R is a promising many-objective evolutionary algorithm.关键词：evolutionary algorithm;many-objective optimization problem;modified dominance relation;many-objective evolutionary algorithm4|13|0更新时间：2025-12-08
- 摘要：The optimization problems in multiple areas can be modelled as many-objective optimization problems, which can be solved using many-objective evolutionary algorithms. However, it is difficult to balance convergence and diversity. To tackle this issue, this paper proposes a many-objective evolutionary algorithm based on dynamic decomposition and modified angle penalty distance referred to as DAEA (Duplication Analysis based Evolutionary Algorithm). DAEA decomposes the whole population into multiple clusters through dynamic decomposition, which is exempt from the predefined reference vectors and makes full use of the distribution information of the population itself to decompose. Then, DAEA selects solutions from each cluster based on modified angle penalty distance to balance convergence and diversity. Besides, DAEA operates mating selection according to Pareto dominance, knee points, and m-nearest angle binary tournament selection. Compared with nine many-objective evolutionary algorithms on 27 many-objective optimization problems, DAEA is effective on many-objective optimization problems with various shapes of Pareto front and stable on different numbers of objectives.关键词：multiobjective optimization;many-objective optimization;dynamic decomposition;angle penalty distance7|14|1更新时间：2025-12-08
- 摘要：Dynamic community discovery aims to detect community structure in dynamic complex networks, and has important research value for revealing the functions and evolution patterns of networks. Because the community structure of the adjacent snapshot networks is smooth, the community discovery result of the previous snapshot network can be used to supervise the community discovery process of the current snapshot network. However, existing methods are difficult to effectively extract these information to improve the performance of dynamic community discovery. In view of this, a dynamic community discovery method named NCI-SeNMF (Semi-supervised Nonnegative Matrix Factorization combining Node Change Information) is proposed, which can fuse node change information. NCI-SeNMF firstly uses k-core analysis method to extract the degeneracy core of every community network at the previous snapshot, and selects the nodes in the degeneracy core to construct the prior community membership information. Then, it quantifies the change degree of the local topology structure of the nodes in the adjacent snapshot networks, and applies it to further improve the prior community membership information. Finally, it integrates the prior community membership information through semi-supervised nonnegative matrix factorization model to discover dynamic communities. Extensive comparative experiments have been conducted on several synthetic and real-world dynamic networks, and the results show that NCI-SeNMF improves at least 4.8% in term of core evaluation metrics comparing with the existing dynamic community discovery methods.关键词：dynamic community discovery;semi-supervised nonnegative matrix factorization;k-core analysis;community network;complex networks6|15|0更新时间：2025-12-08
- 摘要：Currently, graph Transformers mainly add auxiliary modules in the traditional Transformer framework to model graph data. However, these methods have not improved the original Transformer architecture. Their data modeling accuracy needs to be further enhanced. Thus, this paper suggests a node classification method based on graph attention and improved Transformer. In the proposed framework, a topology enhancement based node embedding is constructed for graph structure reinforcement learning. Then, a secondary mask based multi-head attention is developed for aggregation and update. Finally, pre-Norm and skip connection are introduced to improve the interlayer structure of Transformer, which can avoid the over-smoothing problem caused by feature convergence. Experimental results demonstrate that compared to 6 typical baseline models, our method is able to achieve optimal evaluation results on all different indicators. Moreover, it can simultaneously handle the node classification task for both small and medium datasets and comprehensively improve the classification performance.关键词：node classification;graph attention network;Transformer;secondary mask;interlayer residual;multi-head attention6|13|1更新时间：2025-12-08
- 摘要：Network function virtualization simplifies the configuration and management of network services by migrating network functions from dedicated hardware devices to software middleboxes running on commercial servers. Under the environment of network function virtualization, the service function chain (SFC) composed of a series of ordered virtual network functions is becoming a mainstream alternative to host network services. The SFC deployment problem is to allocate the underlying physical network resources to the requirements of service function chains. It is challenging for infrastructure providers to obtain long-term high returns under limited resources. In this paper, we formally define the problem of SFC deployment and propose a novel method named graph attention network and imitation learning (GAT-IL) based on graph attention (GAT) network and imitation learning for SFC deployment. This method utilizes GAT to evaluate the potentials of each physical server, provides expert demonstrations through the Monte Carlo tree search algorithm, applies imitation learning to train the agent, and integrates the beam search strategy to optimize the solution space. Extensive experimental results show that the GAT-IL method proposed in this paper outperforms existing representative algorithms on performance metrics of average revenue-to-cost ratio and acceptance rate.关键词：network function virtualization;service function chain;graph attention network;imitation learning;beam search5|12|0更新时间：2025-12-08
- 摘要：Audio-visual bimodal emotion recognition is a research hotspot in the field of emotion computing. At present, emotion recognition methods cannot simultaneously extract local and global features of video, multi-modal data fusion is simple, loss function can not pay attention to misclassification of samples in model optimization, resulting in low accuracy of emotion recognition results. In this paper, an audio-visual emotion recognition method based on improved ConvMixer and focus loss function with dynamic weight is proposed. Spatial and temporal adjacent matrices were used instead of deep separation convolution in ConvMixer to extract global and local features in video spatial and temporal domain. A cross-modal temporal attention module is proposed to capture the temporal correlation between modals with a symmetrical structure to improve the feature fusion effect. The focus loss function with dynamic weight was calculated by the confusion matrix, and the proportion of error samples in the loss was increased differentially to optimize the model parameters. Experimental results on public data sets show that the proposed method can extract representative features, optimize the network structure effectively, and improve the accuracy of emotion recognition.关键词：emotion recognition;ConvMixer;attention mechanism;multi-modal feature fusion;focal loss function9|18|1更新时间：2025-12-08
- 摘要：Continuous-flow microfluidic biochip is a revolutionary technology for automation and miniaturization of biochemical experiments. As one of the key links in the automatic design of continuous microfluidic biochips, the control pattern assignment problem of multiplexers is an NP-hard combinatorial optimization problem. The existing particle swarm optimization algorithm for control pattern assignment problem has the disadvantages of falling into local optimal solution prematurely, slow convergence speed, and poor stability of the algorithm. In this paper, control pattern assignment algorithm based on hybrid discrete particle swarm optimization for continuous-flow microfluidic biochips is proposed. First, in order to accelerate the convergence speed of the proposed algorithm and avoid falling into a local optimum prematurely, a discrete adaptive region search strategy is proposed. Second, the stability of the proposed algorithm is improved by a sample-based social learning mechanism. Third, the optimal combination of the important parameters in the adaptive region search strategy is selected by equidistant sampling to further improve the results. The final experimental results show that the proposed algorithm optimizes the number of valves by an average of 19.01%, and improves the stability of the algorithm by 29.18%, and then performs well in practical biochemical applications.关键词：continuous-flow microfluidic biochips;control pattern assignment;discrete particle swarm optimization;sample learning;adaptive region search10|15|0更新时间：2025-12-08
- 摘要：The implementation of the cryptographic algorithm based on instruction extension is a lightweight scheme that balances both performance and area, which is especially suitable for the increasingly popular Internet of Things devices. The proposal of national cryptographic algorithms such as SM2, SM3, and SM4 is conducive to improving the security of self-controlled devices. However, the relevant research on instruction extensions for these algorithms is insufficient. RISC-V has become one of the most popular instruction set architectures due to its advantages of open source, simplicity, extensibility, etc. This paper mainly focuses on the instruction extensions and efficient implementation of the SM2, SM3, and SM4 algorithms based on a domestic open-source RISC-V processor. Specifically, this paper proposes an overall instruction expansion scheme based on the concept of hardware-software co-design; this paper conducts an in-depth analysis of the related cryptographic algorithms and comparison of the implementation schemes and then proposes efficient implementations of the hardware units, respectively. This paper designs and implements a coprocessor with a two-stage pipeline structure, sequential dispatching, out-of-order execution, and sequential write-back instruction execution modes, as well as an independent memory access unit and a large bit-wide register. The coprocessor takes over part of the control logic of the cryptographic algorithm, reducing hardware resource consumption. The experimental results show that the hardware structure of the cryptographic coprocessor designed in this paper is simplified, and the utilization of hardware resources is high. SM2, SM3, and SM4 algorithms occupy very few resources, but the execution rate decreases only to a certain extent compared with pure hardware implementation. The product of resource area and time spent has varying degrees of advantages compared to other relevant literature.关键词：RISC-V;co-processors;national cryptographic algorithm;instruction extensions;hummingbird E203;embedded system13|14|0更新时间：2025-12-08
- 摘要：The design of the loss function is crucial in deep face recognition. A common practice is to add a fixed margin term to all classes to modify the decision boundary between classes, compress the distance between intra-class features, and improve the ability of the model to separate features of different classes. However, adding the same margin term for all classes may ignore the inconsistency between classes in the face recognition dataset. In order to further improve the effectiveness of the model, we argue that the model should pay different attention to the samples of different classes according to the learning difficulty of the class. In this paper, we introduce a method for hard class mining based on the bias between the center of the class mean and the center of the class weight, called center bias estimation. The method proposed in this paper adaptively assigns margin terms of different sizes to different classes according to the value of center bias estimation. At the same time, to solve the problem of unstable calculation of center bias estimation in the early stage of training, we propose an adaptively changing convergence parameter to adjust the credibility of center bias estimation and design relevant experiments to prove the effectiveness of the convergence parameters. In the face verification baseline dataset, the proposed method in this paper is improved by 0.26% on average accuracy compared with the baseline method, reaching 96.62%. In two large face verification test datasets, when FPR is equal to 0.01%, the TPR scores of our method is improved by 0.58% and 0.22%, respectively, and the experimental results of 88.47% and 92.29% are obtained, and multiple experimental results show that our method is better than the general existing algorithms. The implementation code is published on https://github.com/TCCofWANG/FR-Centers-Bias.关键词：deep face recognition;hard class mining;class imbalance;center bias estimation;adaptive margin5|12|0更新时间：2025-12-08
- 摘要：Since the traffic flow is affected by multiple factors such as periodic characteristics and unexpected conditions, the prediction accuracy of existing models cannot satisfy the practical requirements. Under this background, this paper proposes a multimodal collaborative traffic flow prediction model based on error compensation (MCEC). To address the problem that traditional prediction models cannot take account of time series and covariates, this paper proposes a feature expansion method based on wavelet analysis, which introduces a clustering algorithm to obtain holiday labeling features, and uses congestion index, traffic accident map, and weather information as expanded features, and decomposes them on multiple scales. In the training phase, a multimodal collaborative model training was designed by adopting ARIMA (AutoregRessive Integrated Moving Average) model, LSTM (Long-Short-Term Memory network), a restricted dynamic time regularization technique, and a self-attentive mechanism to achieve the effect of fully learning each part of the data and optimally matching the model. In the error compensation stage, the obtained corresponding process values are input into the error compensation module based on SVR (Support Vector Regression) to learn and compensate the errors of each component, and reconstruct the prediction results. The MCEC is validated using a publicly available real highway data set. The results of a large number of comparison experiments at multiple time intervals show that the MAPE (Mean Absolute Percentage Error) of MCEC in traffic flow prediction reaches 17.02%,which has a higher prediction accuracy than other prediction models such as LSTM-SVR, ConvLSTM (Convolutional Long Short-Term Memory network), ST-GCN (Spatial Temporal Graph Convolutional Networks), MFFB (Multi-stream Feature Fusion Block), Transformer, indicating the validity and reasonableness of the MCEC model.关键词：traffic flow prediction;error compensation;multimodal cooperation;long-short-term memory neural networks;differential integrated moving average autoregressive models15|14|1更新时间：2025-12-08
- 摘要：At present, deep active learning (DAL) in the classification data labeling work has achieved outstanding success. How to select samples to improve the performance of models is still a difficult problem in deep active learning. We proposes a semi-automatic classification data labeling method based on weak label dispute (Dispute about Weak Label-based Deep Active Learning, DWLDAL). The method iteratively selects samples that is difficult for model to distinguish, and manually annotate these sample. This method contains pseudo label generator and weak label generator, pseudo label generator is trained on accurately annotated datasets to generate pseudo label for unlabeled data; weak label generator is trained on random data subset with pseudo labels. Weak label generator committee are used to determine which unlabeled data is the most controversial and should be manually annotated. We conducted experimental validation on the common datasets IMDB (Internet Movie Database), 20NEWS (20NEWSgroup), and chnsenticorp (chnsenticorp_htl_all) to address the issue of text classification. Three different voting decision-making methods are evaluated from the perspective of the accuracy of data annotation and classification tasks. The F1 score of data annotation in DWLDAL method is 30.22%, 14.07% and 2.57% higher than that in the existing method Snuba, respectively. The F1 score of classification task in DWLDAL method is 1.01%, 22.72% and 4.83% higher than that in Snuba method, respectively.关键词：deep active learning;text classification;pseudo label generator;weak label generator;balloting committee5|15|1更新时间：2025-12-08
- 摘要：In synthetic aperture radar (SAR) system, the three-dimensional rotation of ship targets in the presence of a medium and high sea state would lead to time-varying Doppler spectrum and image defocusing, which will adversely affect the subsequent information interpretation of ship targets in SAR images. Aiming at the refocusing problem of three-dimensional rotating ship targets, this paper proposes a SAR refocusing method for three-dimensional rotating ship target based on minimum entropy criterion and generative adversarial network, and designs the network structure of generator and discriminator. The generator transforms the defocused complex SAR ship image into range-Doppler domain, and estimates the phase error coefficient by range unit using phase error coefficient estimation network, and realizes the compensation of multi-order phase errors. The discriminator is composed of a complex-valued convolutional neural network, and all its elements, including convolution layer, activation function, feature mapping and parameters, are extended to the complex domain. The minimum entropy criterion and adversarial loss are introduced into the loss function to achieve unsupervised training and avoid the problem that it is difficult to obtain the target labeling samples of non-cooperative ships. Experiments on simulated data and Gaofen-3 data show that the proposed method achieves significant improvements in both refocusing accuracy and efficiency.关键词：synthetic aperture radar (SAR);generative adversarial network;ship targets;refocusing;minimum entropy criteria18|16|1更新时间：2025-12-08
PAPERS
- 摘要：Streaming communication is an essential communication scenario in optical fiber communication networks and mobile communication networks. Unlike traditional intermittent or block-oriented communication, the data transmission in streaming communication exhibits typical continuous streaming characteristics. Coupled codes, compared to traditional block codes, have shown significant performance improvement in streaming communication scenarios. Additionally, they inherit low encoding and decoding latency. These advantages make coupled codes an important candidate for channel coding in streaming communication scenarios. This paper first reviews existing coupled LDPC codes, including product-like coupled LDPC codes, partially reencoded coupled LDPC codes, spatially coupled LDPC (SC-LDPC) codes, and globally coupled LDPC (GC-LDPC) codes. Following that, this paper introduces a series of improved designs for coupled LDPC codes based on free-ride codes, and introduces a new class of coupled LDPC codes based on block Markov superposition transmission (BMST) techniques. Finally, this paper concludes with a discussion on future prospects and research directions of the coupled LDPC codes.关键词：channel coding;streaming communications;optical fiber communication networks;mobile communication networks;coupled codes;low-density parity-check codes;free-ride codes;block Markov superposition transmission13|15|0更新时间：2025-12-08
SURVEY AND REVIEW
- 摘要：In recent years, significant progress has been made in international fusion reaction research. The neutron detection technology which can meet requirements of nuclear energy systems such as controlled fusion reactors featuring high-energy and high-dose radiation has always been one of the core technologies in fusion research and application. Choosing wide bandgap semiconductor materials with strong radiation hardness to develop radiation detectors is an inevitable requirement for the development of radiation detectors in fusion installation. Diamond has excellent radiation hardness and ultra-fast time response, and is considered an ideal semiconductor material for fusion detection, especially for high-energy neutron diagnosis. Silicon carbide is also qualified to directly measure fusion neutrons, and its large-size epitaxial growth technology is quite mature. Large-area and highly sensitive neutron detectors prepared using silicon carbide can cover large-scale applications in fusion research installation. By studying these two types of wide bandgap semiconductor radiation detectors and achieving completely self-dependent preparation of high-performance devices, we can significantly improve the performance of China’s nuclear fusion reaction measurement systems, and support China standing at the leading edge in the future global energy revolution. Therefore, this article will elaborate on the key issues and research challenges in the development of these two wide bandgap semiconductor detectors in the application scenarios of fusion neutron detection, to assist in the development and application of fusion energy in China.关键词：neutron detector;diamond;silicon carbide;energy resolution;pulse response12|13|0更新时间：2025-12-08
PERSPECTIVE
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09064830号-19
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0