Please wait a minute...

Journal of Beijing University of Posts and Telecommunications

  • EI核心期刊

Current Issue

  • Review

    • A Review of New-Modal MIMO Technologies in 6G
    • Ying WANG Ye YUAN Yuanbin Chen
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 1-13.
    • Abstract ( 71 )       
    • The advancement of multiple-input multiple-output (MIMO) technologies results in larger array apertures and more densely packed array elements. This progress has spurred the development of novel technologies like extremely large-scale MIMO, holographic MIMO, and extremely large-scale reconfigurable intelligent surfaces, collectively known as the new-modal MIMO. As the sixth generation of the mobile communi-cation system (6G) moves towards a higher frequency spectrum, including millimeter waves and terahertz, the double evolutions of new-modal MIMO substantially increase the Rayleigh distance, resulting in nonnegligible considerations of near-field communications. This article investigates far-field and near-field channel modeling, estimation, and beamforming techniques tailored for different paradigms of new-modal MIMO, highlighting their technical details and new characteristics. Finally, prospective research directions are discussed for new-modal MIMO, demonstrating its beneficial potential to enhance 6G mobile com-munication systems.
    • Supplementary Material | Related Articles

    Paper

    • A Quantization Bit Allocation Method Based on Semantic Importance in Digital Semantic Communication
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 14-21.
    • Abstract ( 52 )       
    • Digital semantic communication can be compatible with existing communication systems while retaining the advantages of semantic communication, and quantization is the key to realize digital semantic communication. Quantization in digital semantic communication requires multiple quantizers to quantify multi-dimensional semantic features. Due to the limited hardware and the limited number of quantization bits, a bit allocation scheme for quantizers is necessary. To solve this problem, a bit allocation algorithm based on semantic importance is proposed. Firstly, a quantized bit allocation problem based on semantic importance is constructed. Under the limit of the maximum number of bits, the importance of different semantic information is considered to minimize the distortion caused by quantization and transmission. Then, a quantization bit allocation algorithm based on reinforcement learning is proposed with the bit allocation range as the action space and the semantic feature as the state space. Finally, the proposed algorithm is trained and the optimal bit allocation strategy is obtained. The simulation results show that the proposed algorithm converges quickly. In the task scenario of image classification, the cross entropy of the proposed algorithm decreases by 48.16% compared with the benchmark algorithm, and the classification accuracy increases by 12.65%.
    • Supplementary Material | Related Articles
    • Combination Auction-based Online Scheduling for the Deterministic IP Networks
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 22-28.
    • Abstract ( 20 )       
    • Deterministic IP (DIP) networking uses frequency synchronization to improve the network scalability. However, frequency synchronization introduces a more complex forwarding model, involving non-convex and non-differentiable operations, which increases the complexity of scheduling problems. Moreover, most scheduling schemes still rely on online heuristic algorithms and offline mechanisms, leading to either inefficient scheduling performance or high computational complexity. To address these issues, an online flow scheduling algorithm is proposed based on the combinatorial auction, where the resources are items, the flows are bidders, and the network is the auctioneer. Non-differential operations are carried out during the bundle enumeration phase before the flows arrive. When they arrive, TS flows only need to select a resource bundle from the feasible set. Additionally, we carefully design the resource’s pricing function which is a linear increasing function. The important time-critical applications could still access to the network even if they come latter. Finally, we conduct simulations that demonstrate the effectiveness of the proposed algorithm.
    • Supplementary Material | Related Articles
    • Traffic Classification Using Domain-Based Graph Matching
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 29-34.
    • Abstract ( 26 )       
    • This paper proposes a domain-based graph matching approach to address the current challenges in network traffic classification, including data encryption, uneven distribution, and user privacy concerns. The method relies solely on non-content features to characterize network flow characteristics and employs graph matching algorithms to reduce inter-class imbalances, enabling coarse-grained clustering and reliable graph matching. Firstly, an unsupervised clustering framework is designed, which studies the diverse distributions and category similarities of traffic data based on a limited set of features. This unsupervised clustering helps mitigate network disparities by aggregating network sessions into a few clusters with extracted primary features. Next, the correlation between clusters from the same network is used to construct a similarity graph. Finally, a graph matching algorithm is proposed, which combines graph neural networks and graph matching networks to reveal reliable correspondences between different network relationships. This allows for associating clusters in the test network with clusters in the initial network, enabling the labeling of test clusters based on associated clusters in the training set. Simulation results demonstrate that the proposed method achieves an accuracy rate of 96.8%, which is significantly superior to existing approaches.
    • Supplementary Material | Related Articles
    • Modeling Static and Dynamic Joint Relationship for 3D Pose Estimation
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 35-43.
    • Abstract ( 32 )       
    • There are enormous challenges for 3D pose estimation due to non-rigid characteristics and occlusion. And modeling joint relationship is a key technology for that. It is difficult to deal with joint missing or skewing caused by occlusion or singularity if only modeling static joint relationships from human physical construction. It is beneficial to improve 3D pose estimation when extracting pose semantics and modeling dynamic joint relationship but the wholeness and hierarchy on human joint relationship can’t be neglected. This work proposes modeling static and dynamic joint relationships together for 3D pose estimation. Mutual information algorithm is used for obtaining a joint relationship map which used to group human joints. Then accumulate the human joint groups based on three level human joint freedoms level by level. A cascading estimation and group joint feature sharing networks are designed to model static joint relationship. Multi-group attention mechanisms are present for each level for extrating pose semantic feature to model dynamic joint relationships. A data enhancement policy by category balancing and pose reorganizing is presented further for improving model robustness. The experiments are carried out based on Human3.6M, MPI-INF-3DHP and MPII datasets extensively. The results show that the average error of our model is reduced 0.2 mm and the average accuracy is increased 0.2% at least when comparing with other advanced models. And the model performance is improved significantly when our data enhancement policy is carried out.
    • Supplementary Material | Related Articles
    • Research on Offloading Strategy of Mobile Edge Computing by Data Compression in Blockchain
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 44-50.
    • Abstract ( 32 )       
    • The combination of Mobile Edge Computing (MEC) and blockchain can enhance the computing power of “mining” nodes. Data compression can reduce the size of uninstalled data, which reduces the uninstallation delay. Using this advantage, this paper introduces data compression into the mobile edge computing network that supports blockchain, and designs a computing delay minimization offloading strategy to meet the requirements of node mining for two scenarios: shared MEC server computing resources and dedicated MEC server computing resources. For the shared MEC server computing resource scheme, a multi-dimensional resource allocation problem with minimum system computing delay was established by jointly optimizing the user offloading rate and compression rate. Due to the existence of variable coupling and max-max function, the established problem was not convex. By introducing relaxation variables and auxiliary variables, the problem was transformed into a convex problem and the optimal solution was obtained. For the dedicated MEC server computing resource scheme, jointly optimize the unloading rate, compression rate and computing resources to establish a resource allocation problem that minimizes computing delay. Due to the existence of variable coupling and max-max function, the optimization problem is non-convex, and the problem is transformed by relaxation variables and auxiliary variables. And with the aid of Block Coordinate Descent (BCD) algorithm to the problem is decomposed into two convex problems, then puts forward an iterative algorithm based on BCD to get the optimal solution. Finally, the correctness of the proposed algorithm and the superiority of the proposed strategy in calculating the delay are verified by simulation.
    • Supplementary Material | Related Articles
    • User Selection Method for 5G IoT System with Spatially Correlated Channel under the Condition of Limited Feedback
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 51-58.
    • Abstract ( 28 )       
    • Most of the existing user scheduling algorithms are based on the incoherence of Multiple Input Multiple Output (MIMO) channels, while there may be coherence among channels in practical systems. Meanwhile, in the Multi-user MIMO (MU-MIMO) system, in order to reduce the overhead of the uplink channel, the user only needs to feedback part of Channel State Information (CSI) to the base station, so the multi-user interference is inevitable. This paper analyzes the influence of the coherence conditions of MIMO channels on the upper limit of user capacity and transmission rates, and then deduces the low-complexity transmission rate based on the residual interference caused by limited feedback. For MU-MIMO limited feedback systems with channel coherence, a user selection method based on reinforcement learning is proposed. The proposed selection method can avoid recalculating the achievable rate in each cycle, thus greatly reducing the computational complexity. When the system is in the coherent channel environment, the proposed algorithm improves the throughput.
    • Supplementary Material | Related Articles
    • Deep learning based algorithm for low light target detection
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 59-65.
    • Abstract ( 36 )       
    • The images acquired in complex low-light environments are prone to problems such as low contrast and loss of detail information, and the results of using existing target detection methods are not satisfactory. In this paper, the YOLOv8 algorithm is improved for the special characteristics of low-light environment detection to improve the reliability of detection in low-light environment. Firstly, the backbone feature extraction part of this network adopts C3_ResBlock to improve the multi-scale and weak feature extraction ability in low-light detection; subsequently, the hollow space pyramid structure SE_ASPP is introduced to extract the information of the complex scene by using different expansion rates to maintain the computational volume while improving the training effect; finally, the adaptive fusion of SKNet and GAM attention mechanism. SKNet contains adaptive induction field mechanism, which can select more important and effective space for multi-scale feature extraction and fusion, and GAM can adjust the importance degree of each channel to improve the feature extraction and characterization ability of the network model. Numerical experiments show that compared with YOLOv8, the proposed YOLO-RSG algorithm in this paper improves the mAP by 3.60% in the ExDark dataset, which can effectively improve the performance of target detection in low illumination scenes and has good stability and applicability, and can better meet the needs of target detection in low-light scenes.
    • Supplementary Material | Related Articles
    • AoI based multi-UAV iot task assignment and trajectory planning
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 66-73.
    • Abstract ( 30 )       
    • Age of Information (AoI) describes the time from the time when information is sent to the time when it is received, which can accurately measure the value of data. In order to minimize the average AoI in emergency communication, unmanned aerial vehicles (UAVs) are introduced as information relay and a task assignment and trajectory optimization algorithm based on deep reinforcement learning framework is proposed. Firstly, the relationship between the original AoI minimization problem and UAV-assisted emergency communication is analyzed and proved, and the AoI minimization problem is divided into two stages to solve. Secondly, in order to reduce the ineffective flight time of drones, new planning strategies have been developed to make the trajectory smoother by incorporating disaster nodes along the way into the trajectory. Finally, in order to avoid the UAV repeatedly visiting the same rescue group, a centralized information sharing mechanism is designed to save energy consumption and information distribution time. Experimental results show that compared with traditional optimization algorithms, the proposed optimization algorithm can achieve a smaller information age.
    • Supplementary Material | Related Articles
    • Mural inpainting algorithm based on semantic reasoning and joint learning
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 74-81.
    • Abstract ( 20 )       
    • We propose a mural restoration algorithm based on semantic inference and dynamic joint learning to address the issues of semantic inconsistency and disordered structural textures caused by the lack of semantic constraints and isolated restoration of texture structure in existing deep learning methods for mural restoration. Firstly, a mural restoration framework based on joint learning is constructed, and a joint hierarchical network is designed to divide the mural into high-level semantics and low-level semantics, enabling hierarchical restoration of different semantics. Then, a joint global generation module is designed to model the global semantics of the mural through autoregressive modeling and infer the repaired global semantic information. Next, a joint local generation module is constructed, which introduces a context aggregation block to learn the contextual information of the mural and generate local information for the mural. Finally, a joint attention mechanism is introduced to enable collaborative training between the global semantic restoration module and the local restoration module, overcoming the issues of error accumulation and semantic inconsistency caused by isolated restoration. Experimental results on real Dunhuang murals demonstrate that the proposed method achieves better structural and texture consistency in the restoration results compared to the baseline methods, and objective evaluation metrics outperform the comparison algorithms.
    • Supplementary Material | Related Articles
    • Pattern-reconfigurable Satellite Receiving Antenna Based on Liquid Metal
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 82-86.
    • Abstract ( 24 )       
    • Based on the Yagi antenna principle, a reconfigurable four-arm helical antenna with a liquid metal reflector as its directional diagram has been proposed in this paper. The antenna is composed of a four-arm helical antenna, which serves as the main radiating unit, and a hollow column acting as a reflector for directional diagram switching. The directional diagram can be reconstructed by injecting liquid metal into different hollow columns. The mechanism of the reflection process has been theoretically studied and experimentally verified. The final designed antenna achieves beam switching at 5 radiation angles within a 360 ° coverage range, with an increased antenna directivity. The radiation frequency range covers the B1 frequency band of the BeiDou satellite and the L1 frequency band of GPS, ranging from 1.5 to 1.64 GHz.
    • Supplementary Material | Related Articles
    • A DGA Domain Name Detection Method Based on Multi-level Feature Extraction
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 87-92.
    • Abstract ( 27 )       
    • To tackle the problems that the existing domain detection methods of domain generation algorithm (DGA) cannot fully extract and utilize the domain features and the detection methods based on word embedding are prone to cause the loss of important information, a DGA domain name detection method based on multi-level feature extraction (DDMFE) is proposed. Firstly, the vector representations of domains are obtained by word embedding, and the domain character features are extracted to obtain preprocessing samples. Secondly, the domain vectors are processed by a multi-level feature extraction network to capture the contextual and semantic information of the domains and fuse different domain information to generate a text-level feature representation of the domains. Finally, to calculate the domain classification probability, a feed-forward neural network is used to process the domain character features, an improved capsule network is used to process the domain text features, and a fusion operation is used to generate the domain classification probability for domain detection. After experimental validation, the proposed method improves the accuracy of DGA domain name detection and DGA algorithm recognition by 1.1%~8.6% and 1.8%~3.1%, respectively, which provides a good detection performance.
    • Supplementary Material | Related Articles
    • A Lightweight Model for Identifying Tilted Mixed Color License Plates
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 93-99.
    • Abstract ( 22 )       
    • At present, the license plate recognition model based on neural network is large, which is not conducive to deployment in edge computing equipment. In addition, the recognition rate of oblique license plates and mixed color license plates is low. To this end, a lightweight neural network combination model for recognizing tilted mixed color license plates is proposed. The model first uses the improved P-YOLO algorithm based on YOLO algorithm to achieve license plate detection, classification, and localization; Then, an improved N-G-LPRNet algorithm based on LPRNet algorithm is used to recognize license plate characters. Finally, compared with the improved algorithm model, the test results on the CCPD dataset show that the P-YOLO algorithm significantly improves the mAP index for detecting tilted license plates during the license plate detection stage. Combined with character recognition networks, the recognition accuracy for conventional license plates, slightly tilted license plates, and strongly tilted license plates is improved by about 1.3%, 70.7%, and 63.8%, respectively; In the character recognition stage, under the premise of mixed color license plate training, the N-G-LPRNet algorithm improves the recognition rates of blue and green license plates by about 40.65% and 32.26%, respectively; The final P-YOLO-N-G-LPRNet combination model has a comprehensive recognition rate of 98.16%, occupying approximately 5MB of space. It has obvious advantages in high recognition rate and lightweight.
    • Supplementary Material | Related Articles
    • Research and Implementation of an Active Self-Reconfiguration Mechanism for Array Processors
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 100-106.
    • Abstract ( 21 )       
    • With the development of various image and video recognition applications, the demand for visual data processing and understanding has exploded. As a result, the demand for a real-time, intelligent user experience is increasing, and the diversity and volatility of applications make it necessary for reconfigurable computing architectures to be both flexible and efficient. However, the method of realizing dynamic reconfiguration using the traditional upper computer suffers from the problems of long reconfiguration times and low reconfiguration efficiency, which greatly limit the performance improvement of reconfigurable structures. In order to reduce the reconfiguration time and accelerate the hardware reconfiguration process, an instruction flow-driven active self-reconfiguration method is proposed. The method makes full use of the rich PE resources of the reconfigurable array, monitors the execution state of the array autonomously through the PE, and autonomously realizes functional reconfiguration according to the monitoring results. The experimental results show that compared with the traditional centrally controlled reconfigurable array processor, the computational speed is improved by 25% and the configuration time is reduced by 27% in achieving the dynamic reconfiguration process of de-block filtering.
    • Supplementary Material | Related Articles

    Report

    • Energy-efficient multi-user edge computing for streaming tasks
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 107-114.
    • Abstract ( 34 )       
    • In a multi-user mobile edge computing (MEC) system, mobile users can upload their own tasks to the edge server on the access network, thereby effectively reducing the processing overhead of their tasks. In a MEC system, to ensure the real-time execution of tasks with long data collecting duration, a streaming task processing scheme is proposed, where the data collection and local computing, the offloading transmission and edge computation, are carried out in different time slots. Under this scheme, specific size of the task, more importantly energy consumption for executing the task, is related to the time length of data collection. To find the most energy-efficient way for completing the streaming tasks, the problem of minimizing the overall power consumption is formulated to jointly optimize the duration of each stage for completing the task, together with the multi-user offloading ratio and bandwidth allocation. In order to solve the intractable non-convex problem, block coordinate descent method is utilized to separate the optimization variables into two parts. Exploiting the analytical structure of the problem, optimal solution of these two parts of variables is obtained with bisection search and golden section search. Simulation results show that the proposed method has extremely low computational complexity and can significantly reduce the overall system power consumption.
    • Supplementary Material | Related Articles
    • Research on the Deployment Strategy of UAV Location for Forest Fire Monitoring
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 115-121.
    • Abstract ( 24 )       
    • The application of Unmanned Aerial Vehicles (UAVs) to forest fire monitoring can improve the efficiency of firefighting and rescue. However, in the complex environment of forest fires, the deployment of UAVs faces problems such as high energy consumption, low offloading efficiency, and dynamic changes in the environment. Therefore, an air-ground-assisted edge computing framework is investigated, in which the UAV collects fire scene data at the fire scene and provides edge computing services, and the command center provides edge computing services with high computational power. In order to provide efficient computing services, a UAV location deployment scheme based on multi-agent reinforcement learning is designed, which first determines the area that needs UAV to provide computing services based on the fire spreading speed and distance, and then designs an autonomous deployment strategy based on multi-agent reinforcement learning that minimizes the system cost to obtain the optimal location of the UAV in the designated task area. The final simulation results demonstrate that the proposed scheme can effectively reduce the total cost of UAV deployment.
    • Supplementary Material | Related Articles
    • Optimal Control-Based Load Balancing Strategy for Vehicular Edge Computing First Network
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 122-127.
    • Abstract ( 20 )       
    • With the deployment of the Internet of vehicles and the constant emergence of intelligent vehicular applications, the computing power of vehicular terminals is no longer sufficient to meet the demand of intelligent vehicular applications. A large number of vehicular computing tasks are offloaded to the edge computing network. It has become a challenge for research on the vehicle edge computing network to ensure fast and reasonable allocation of computing tasks, reduce latency, and improve user service quality. This paper focuses on the resource allocation problem of vehicular task offloading and comprehensively considers factors, such as the local computing latency of vehicular tasks, the computing latency of vehicular tasks in edge servers and network latency, and the network latency generated by the migration of vehicular tasks in edge clouds. In this paper, we propose an optimal control-based load balancing strategy for vehicular edge computing first network. The strategy solves the load balancing, resource allocation, and task migration by considering the load status of the edge server and the disturbance of the computing load. Simulation experiments show that regardless of whether there is disturbance or not, the load balancing effect of the strategy proposed in this paper is better than the state-of-the-art methods.
    • Supplementary Material | Related Articles
    • Multi-Channel Residual Hybrid Dilated Convolution with Attention for Word Sense Disambiguation
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 128-134.
    • Abstract ( 22 )       
    • Aiming at insufficient generalization ability of current WSD (word sense disambiguation) model, Multi-Channel Residual Hybrid Dilated Convolution with Attention (MRHA) WSD model is proposed. Linguistic knowledge is used to construct disambiguation features, 3 vectorization methods are used to vectorize disambiguation features to form 3-channel word embedding matrix, and positional coding is deeply fused with 3-channel word embedding matrix. A complex convolutional encoder is designed to increase expressive ability of WSD model. Experiments are conducted on SemEval-2007: Task#5 and SemEval-2021: Task#2. Experimental results show that compared with the newest WSD model using Clustered Sense Labels (CSL) and Multi-Channel Convolutional Neural Networks with Multi-Head Attention (MCNN-MA), average bias of the proposed method is respectively reduced to 1.345% and 2.157%.
    • Supplementary Material | Related Articles
    • Balance Control of an Unmanned Bicycle Based on Linear Extended Observer and Non-singular Terminal sliding Mode Control
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 135-143.
    • Abstract ( 24 )       
    • In order to realize the robust control of unmanned bicycle under different terrain conditions, different load conditions and different speed scenarios, a robust controller was designed by combining Linear Expanding State Observer (LESO) and Non-singular Terminal Sliding Mode Control (NSTSMC) method, and a physical prototype experiment platform was built to verify its performance. Considering the linear variable parameter (LPV) model of the unmanned bicycle, The coupling between the steering angle dynamics and the lean angle is eliminated by feed-forward compensation; the internal uncertainties and external disturbances of the system are combined into lumped disturbances, and a linear extended state observer is introduced to construct an improved LPV model that only includes the lean angle. A non-singular terminal sliding surface is chosen, and the equivalent control and nonlinear control of the controller are designed by using the improved LPV model. The results of numerical simulation and physical prototype experiments show that the unmanned bicycle can realize self-balancing motion on four different grounds: granite road, asphalt road, cement road and lawn. Its lean angle can be stabilized at [-0.006,0.006](rad) range; it also has strong robustness for the mass load within 16.5kg, the changing vehicle speed of 1.2m/s-2.4m/s, and the pulse interference generated when crossing the speed bump.
    • Supplementary Material | Related Articles
    • A Multi-Party Authentication Model for High-Frequency Cross-Origin Access
    • Journal of Beijing University of Posts and Telecommunications. 2024, 47(5): 144-150.
    • Abstract ( 22 )       
    • Secure cross-origin authentication is essential to providing effective protection for sensitive data and services in high-frequency cross-origin access. The existing cross-origin authentication work mostly focuses on improving the reliability of authentication credentials, and lacks effective protection for authentication service providers. Therefore, a multi-party authentication model COMPA for high-frequency cross-origin access is proposed. Firstly, by analyzing the functional differences between classical consensus and consensus for multi-party authentication, a safe Practical Byzantine fault-tolerant algorithm SPBFT is proposed to realize multi-party security authentication between nodes. Secondly, a network reconfiguration algorithm is designed to locate and replace the risky authentication participants based on the authentication results, and the authentication network is reorganized to make it more flexible. Design simulation experiments to verify the authentication effect of the model and its resistance to malicious attacks. The results show that the model can achieve safe and reliable authentication, fault location and network elastic reorganization within 20 seconds, with good effectiveness and robustness, and controllable time cost.
    • Supplementary Material | Related Articles