最新刊期
卷 40 ， 期 12 ， 2024
- 摘要：With the rapid development of intelligent transportation systems, sidelink communication technology for cellular vehicle-to-everything (C-V2X) has become the key technology to enable vehicle-to-vehicle, vehicle-to-infrastructure, and vehicle-to-pedestrian communications. 3rd Generation Partnership Project (3GPP) has finished research on sidelink communication technology enhancement, aiming to meet the diverse needs of V2X services, and is actively advancing research and standardization on 6G evolution technologies such as integrated sensing and communication (ISAC), artificial intelligence and machine learning (AI/ML). The development trends of sidelink communication technology in C-V2X were firstly introduced. Then, the key evolution points of 3GPP Release 18 sidelink communication technology were discussed. Finally, considering the research on 6G technology in 3GPP Release 19, a perspective on the standard future development direction of C-V2X was provided.关键词：C-V2X;sidelink unlicensed;sidelink positioning;ISAC;AI/ML830|3512|0更新时间：2024-12-30
- 摘要：With the rapid development of vehicle-road-cloud integrated system (VRCIS), physical space and cyber space were deeply intertwined, resulting in a brand-new transformation of automobile safety. New challenges arise in the integrated security domain, where safety, security and information security were deeply overlapped. Based on a comprehensive analysis the integrated security problem of intelligent connected vehicle (ICV) and limitations of existing security defense methods, guided by the endogenous security existence theorem, a dynamics heterogeneous redundancy (DHR) based endogenous security construction technology was proposed to solve the triple security overlap problem with integrated construction effects, achieving known and unknown threat defense without relying on prior knowledge. The results of a large number of endogenous safety white-box pile injection tests show that the endogenous safety DHR architecture had 100% differential mode suppression capability. In order to promote the overall goal of achieving endogenous safe boarding, a close range planning scheme for intelligent connected vehicles with the core of “pre blocking, mid defense, and post traceability” had also been explored.关键词：intelligent connected vehicle;dynamic heterogeneous redundant;security & safety;endogenous security884|1229|0更新时间：2024-12-30
- 摘要：Vehicular-road collaborative roadside sensing systems are a crucial part of developing “integrated vehicular-road-cloud systems” and implementing the digital transformation and upgrading of transportation infrastructure. Based on current application needs and industrial progress, key technologies and the standardization status of the system were analyzed, and the testing tools and verification results developed by the research team were presented.The results of testing and verification show that some deployed roadside sensing systems still have a significant technological gap to be fulfilled, and with the aid of testing tools, system performance can be greatly improved. The necessity of key technologies, the usability of existing standards, and the high value of testing tools have all been confirmed.关键词：vehicular-road collaboration;roadside sensing system;sensing technology;fusion algorithm;standardization;testing verification692|1128|0更新时间：2024-12-30
- 摘要：Due to openness of wireless communication, Internet of vehicles (IoV) is vulnerable to many cyber-attacks such as denial of service, spoofing and fuzzy attacks. Therefore, random forest (RF) and gradient boosting decision tree-based stacking intrusion detection (RF-IDS) model was proposed. Firstly, the adaptive synthetic sampling (ADASYN) algorithm was adopted to generate more similar samples through the nearest neighbor sampling strategy in order to balance the training samples of different categories, and form a relatively symmetric dataset. Secondly, GBDT was used to evaluate the importance of features and select sample data with important features to build a lightweight classifier. Finally, the k-fold cross-validation stacking method was used to reduce the probability of overfitting. RF, GBDT and LightGBM classifiers serve were used as base-learner. The RG-IDS model was tested by CICIDS 2017 and NSL-KDD datasets. The experimental results demonstrate that RG-IDS model can achieve a higher F1-score.关键词：Internet of vehicles;intrusion detection;ADASYN;GBDT;stacking413|558|0更新时间：2024-12-30
- 摘要：In order to improve the accuracy of detecting the cyber-attacks in Internet of vehicles, hyper-parameter optimization convolution neural network-based ensemble Intrusion detection system (CNES) was proposed. In CNES, the convolution neural network (CNN) was adopted to serve as based learner in ensemble learning. Moreover, the particle swarm optimization was utilized to optimize the hyber-parameters of the CNN, and then CNN model was optimized. Confidence averaging and concatenation techniques were constructed to improve the accuracy. The performance of the proposed CNES was measured based on Car-Hacking and CICIDS2017 datasets. This shows the effectiveness of the proposed CNES for cyber-attack detection. The CNES achieves F1 score of 100% on Car-Hacking dataset.关键词：Internet of vehicles;intrusion detection;convolution neural network;particle swarm optimization algorithm;ensemble learning381|914|0更新时间：2024-12-30
Topic：
- 摘要：For frequency division duplexing (FDD) cell-free massive multiple input multiple output (MIMO) systems, maximizing downlink capacity relies on designing a precoding matrix that achieves high beamforming gain and effectively suppresses multi-user interference. The accurate acquisition of downlink channel state information (CSI) is essential for calculating precoding matrix. However, due to the imperfect reciprocity between uplink and downlink in FDD systems, securing low-overhead and high-precision downlink CSI remains an industry challenge. Therefore, a CSI measurement and feedback scheme on the basis of co-located spatial user clusters was proposed, along with a two-stage precoding optimization scheme that integrated cluster-level CSI feedback. Firstly, equivalent downlink spatial correlation was obtained based on the calculation of uplink and downlink statistical reciprocity, and co-located space user clusters were jointly constructed using the downlink channel quality and equivalent spatial correlation. Secondly, a dynamic CSI coherent measurement scheme was designed based on large-scale fading characteristics and CSI pilot port capacity constraints, and a statistical CSI feedback scheme for co-located space clusters was developed, significantly reducing feedback overhead. Finally, during the CSI measurement phase, large-scale slowly varying interference elimination precoding was designed using the statistical covariance matrix between cluster users. During the downlink scheduling phase, real-time multi-user interference elimination precoding within clusters was designed on the basis of the signal-to-leakage-and-noise ratio (SLNR) algorithm. The two were cascaded to form the optimized downlink precoding weight values for each user. Simulation results show that the proposed CSI feedback optimization scheme reduces the feedback overhead by 34% compared with the path gain information feedback strategy in the literature, and the two-stage precoding optimization scheme improves the FDD non-cellular massive MIMO spectral efficiency by 10.8% compared with the SLNR precoding.关键词：FDD cell-free massive MIMO;co-located spatial user cluster;CSI measurement and feedback;two-stage precoding1099|896|0更新时间：2024-12-30
- 摘要：Addressing the challenges of high pilot overhead and computational complexity in traditional channel estimation methods for millimeter-wave communication systems aided by reconfigurable intelligent surfaces (RIS), a novel channel estimation algorithm based on the alternating direction method of multipliers (ADMM) for multi-user joint scenarios was proposed. Firstly, leveraging the shared base station (BS)-RIS component within the cascaded channels of different users, the common subspace of user signals was estimated and projected to effectively mitigate noise interference. Secondly, considering the correlation and sparsity of multi-user cascaded channels, an optimization problem for the sparse cascaded channel matrix and common scaling factor matrix was formulated. To solve this non-convex joint optimization problem, the ADMM algorithm was employed to alternately optimize the sparse matrix and the common scaling factor matrix. Simulation results demonstrate that, compared to existing methods, the proposed ADMM-based channel estimation algorithm for multi-user joint scenarios reduces pilot overhead by approximately 50%.关键词：massive MIMO;reconfigurable intelligent surface;channel estimation;alternating direction method of multipliers377|768|0更新时间：2024-12-30
- 摘要：In the space environment, visual sensors capture images with single colors, weak textures, and few feature points, making image stitching difficult. The Hessian matrix and scale space were constructed to obtain matching points through non maximum suppression, using Harris corner detection algorithm to construct feature point descriptors. RANSAC algorithm was used to filter the matching results and obtain the homography matrix for concatenation. In addition, satellite images were used in the experimental section to compare the processing performance of Harris corner detection algorithm with other algorithms. The results show that the Harris corner detection algorithm has better image stitching performance, and can be used to quickly stitch satellite images using the limited computing power of onboard processing units based on the Harris corner detection algorithm.关键词：Harris corner detection;SURF algorithm;RANSAC;image stitching178|545|0更新时间：2024-12-30
- 摘要：Network traffic classification is crucial for network security maintenance and management, and it has been widely applied in tasks, such as quality of service (QoS) assurance and intrusion detection. To address the issues of traditional traffic classification models, such as insufficient feature extraction and low classification accuracy, a dual-modal network traffic classification method based on group mix attention (GMA) with a transformer and a bi-directional long short-term memory (Bi-LSTM) network, named group mix transformer and Bi-LSTM for traffic classification (GMTBLC), was proposed. In the data preprocessing phase, packet-level images within sessions were generated from the payloads of data packets to reduce information interference. In the classification phase, the images were firstly processed by the packet group mix transformer (PCMT) module, which utilized the transformer and GMA to capture global features. Simultaneously, session images were processed by the spatio-temporal feature extraction (SFE) module, of which the spatial features of packets were extracted by a convolutional neural network with residual connections, and temporal features of packets were extracted by a bi-directional long short-term memory network. In the fusion classification layer, the above global and spatiotemporal features were integrated using a dynamic weighting mechanism to complete network traffic classification. Experimental results on ISCX and USTC-TFC2016 datasets demonstrate that the proposed model achieves a classification accuracy of 99.31%, with precision, recall, and F1-score all above 98%, and outperforms the other models in classification effectiveness.关键词：traffic classification;deep learning;attention mechanism;Transformer;LSTM433|1033|0更新时间：2024-12-30
- 摘要：In recent years, mobile communication technology has developed rapidly, with 5G networks being widely deployed globally, and research on 6G technology is also underway. With the rapid economic development and increased investment in satellite-based facilities, China's satellite communication, satellite navigation and other satellite application industries are expected to enter a period of rapid development, and the trend of building a satellite-terrestrial integrated network will become the future direction of communication development. The demand for frequency resources in satellite communication overlaps with that of ground mobile communication systems, and the problems of co-frequency interference and strong power suppression interference between mobile communication and satellite communication are becoming increasingly prominent. The frequency sharing technology for 5G and 6G mobile communication satellite-terrestrial integration was analyzed from the perspectives of technology research, policy formulation, and station layout, providing reference for building a satellite-terrestrial integrated global seamless communication system.关键词：satellite-terrestrial integrated;co-channel interference;high power pressing;frequency sharing275|1218|0更新时间：2024-12-30
- 摘要：The network topology structure based influence maximization algorithms are greatly influenced by the network structure, which leads to unstable performance of social networks of different scales and different topology structures. In view of this problem, a improved Transformer model based social network influence maximization algorithm was proposed. Firstly, the high influential nodes of the society network were selected based on the k-shell decomposition method. Seconcly, the topology structure information and connection framework information of the candidate nodes were discovered by use of the random walk strategy. Finally, the Transformer model was improved, in order to support scalable node feature sequences, and the improved Transformer model was taken advantage to predict the seed nodes of the social network. Validation experiments were carried on six real social networks of different scales. The results show that the proposed algorithm realizes a good influence maximization performance on social networks of different scales and topology structures, and the time efficiency of the seed node recognition has been increased significantly.关键词：social network;influence node;influence maximization;information propagation;neural network272|600|0更新时间：2024-12-30
Research and Development
- 摘要：To address issues such as low availability of computing power in ultra-large scale computing clusters of intelligent computing centers, low maturity of domestically produced technologies, bottlenecks in large-scale networking efficiency, and complex operations and maintenance, a system based on cloud computing technology for constructing a ten-thousand card cluster in an intelligent computing center was proposed. A ten-thousand card cluster was constructed using 18 432 NPU units and an optimized RDMA network. A multi-plane network architecture was adopted, in conjunction with SDN technology to achieve RDMA network tenant isolation. The network load balancing strategy was optimized, resulting in a link load balancing error of less than 10% and an All-Reduce bandwidth of over 35 GB/s. By employing the optimized distributed storage protocol, the model’s breakpoint recovery time was reduced to half of its original duration. The validation results demonstrate that the domestic NPU ten-thousand card cluster, with the collaborative optimization of software and hardware, can not only meet the training needs of large models with hundreds of billions of parameters but also support the training tasks of large models with trillions of parameters.关键词：supercomputer cluster;intelligent computing center;ten-thousand card cluster;artificial intelligence1335|1408|0更新时间：2024-12-30
- 摘要：Along with the country’s vigorous promotion of new infrastructure construction and the rapid development of industrial Internet, industrial passive optical network (PON) technology is increasingly widely used in the industrial field. The policy background of industrial PON technology development was discussed in depth, its key technologies in broadband enhancement, network protection, deterministic delay, open intelligence, and security protection were elaborated, and the application practice of industrial PON in typical business scenarios was analyzed.关键词：industrial PON;gigabit optical network;deterministic delay;programmable;edge arithmetic;data security265|831|0更新时间：2024-12-30
- 摘要：5G serves as the pivotal infrastructure for the emerging era of the Internet of everything, with intelligentization being its crucial direction. In the 5G-Advanced (5G-A) phase, AI enabled core network running can substantially enhance network efficiency, facilitate real-time perception, and precisely cater to diversified service demands through the deep integration of AI technology and network running mechanisms. For the 5G-A network, the overall intelligent architecture, key technologies, and applications of AI enabled core network operations were highlighted. Firstly, the development trend of AI enabled 5G-A core network running was analyzed and key intelligent applications were elaborated. Secondly, AI enabled core network running architecture combining four layers and four dimensions was proposed, encompassing a central intelligent analysis layer, a control plane intelligent endogenous layer, an edge real-time inference layer, and an on-device lightweight inference layer. This architecture enables customized, real-time data collection across four dimensions: slices, network elements, users, and service flows, thereby offering intelligent analysis services such as data perception, training, and inference. Then, five key technologies and capabilities of AI enabled core network running were introduced. Finally, focusing on the typical service requirements of AI enabled core network running, the relevant solutions and their results were analyzed and summarized, thus providing new ideas for the application and implementation of 5G-A core network intelligence.关键词：5G-Advanced;artificial intelligence;core network;network data analysis404|2349|0更新时间：2024-12-30
- 摘要：At present, the conventional heating methods in rural areas of northern China present challenges to carbon emission reduction efforts.Consequently, selecting appropriate clean energy heating methods in rural areas has emerged as a significant concern. The TRNSYS was used to simulation software to construct three decentralized clean energy heating systems-natural gas boiler heating, air source heat pump heating, and solar-coupled air source heat pump heating. Then, a typical rural household in Shanxi Province was taken as an example, simulations were conducted for the three systems. The research findings indicate that the solar-coupled air source heat pump heating system performs excellently in heating stability and operational efficiency. It exhibits the lowest operating costs over its lifespan, Moreover, the solar-coupled air source heat pump heating system is more conducive to carbon emissions reduction. These research results provide an important reference for choosing clean energy heating methods in rural areas of northern China, which can help guide policymaking and resource allocation and promote the popularization and application of clean energy heating in rural areas.关键词：clean energy heating;solar energy;air source heat pump;TRNSYS;economics455|1255|0更新时间：2024-12-30
Engineering and Application
- 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.
- Total Visits：1035510 Visits Today：3559
0