Quick Entry

  • Author login
  • Peer Review
  • Editorial
更多
  • Latest Articles
  • Online First
  • Archive
更多
Volume 42 期 8，2026 2026年第42卷第8期

    Xiong Kai, Liu Hanyang, Zhang Hanwen, Leng Supeng

    DOI：10.11959/j.issn.1000-0801.DXKX260060
    摘要：With the rapid development of the low-altitude economy, constructing a low-altitude intelligent network for aircraft has become a critical foundation for supporting the safe operation of low-altitude activities. However, existing deployment strategies for ground base stations are primarily designed for ground users, resulting in limited airspace coverage. Furthermore, the mobility of low-altitude aircraft is significantly higher than that of ground users, imposing stringent real-time requirements on beamforming and transmission power allocation of base stations. Focusing on structured air route scenarios, a collaborative optimization strategy for base station deployment and power allocation was proposed to maximize the overall air-ground transmission efficiency of the network. The optimization was divided into two stages. Firstly, a heuristic search algorithm was employed to globally optimize base station deployment positions, enhancing the geometric coverage performance of the network. Subsequently, a deep reinforcement learning framework based on the DeepSets architecture was constructed to address the issue of power allocation in highly dynamic scenarios, enabling three-dimensional dynamic tracking. Simulation results demonstrate that the proposed base station deployment scheme improves channel quality, achieving approximately 9.78% higher air-ground transmission rates compared to benchmark schemes. The reinforcement learning framework converges stably and quickly, providing effective three-dimensional tracking of aircraft users. This study offers theoretical support and practical guidance for building an efficient air-ground integrated low-altitude communication network.  
    关键词：Airspace corridor;base station deployment;power allocation;reinforcement learning;cooperative transmission   
    14661
    |
    109
    |
    0
    <HTML>
    <L-PDF><WORD><XML><Meta-XML>
    <引用本文> <批量引用> 156804145 false
    更新时间：2026-09-18

    Qin Pengtai, Meng Han, Wei Bin

    DOI：10.11959/j.issn.1000-0801.DXKX260283
    摘要：The large-scale development of the low-altitude economy puts forward new requirements for integration, high precision and high reliability in the communication, sensing and intelligent scheduling capabilities of the network. Traditional technologies are difficult to adapt to the complex characteristics of low-altitude scenarios, and the industry is facing common bottlenecks in architecture, networking technology, intelligent integration and resource coordination. From the perspective of core network research, the technical system of the low-altitude integrated communication, sensing and intelligence network was sorted out, the core network positioning of “one network for multiple purposes, integrated empowerment” was defined, and a four-layer overall architecture of “resource-enablement-operation-application” was constructed. Focusing on the three directions of networking, sensing and intelligence, core technical solutions such as “four-code-in-one” trusted access, integrated communication and sensing core network architecture, multi-source fusion sensing and intelligent service guarantee were proposed, technical difficulties and designs solutions were analyzed. It provides theoretical support and practical reference for the architecture design and technical research and development of the low-altitude integrated communication, sensing and intelligence network, and boosts the high-quality development of the low-altitude economy and the independent innovation of integrated communication and sensing technology.  
    关键词：low-altitude economy;integrated communication;intelligence;sensing and intelligence;networking   
    51
    |
    52
    |
    0
    <HTML>
    <L-PDF><WORD><XML><Meta-XML>
    <引用本文> <批量引用> 171931445 false
    更新时间：2026-09-18

    Wang Dong, Chen Wenjun, Yang Mengke, Li Hang

    DOI：10.11959/j.issn.1000-0801.DXKX260198
    摘要：With the low-altitude economy being designated as a national strategic emerging industry, the large-scale deployment of unmanned aerial vehicles has created an urgent demand for high-precision coverage planning in low-altitude networks. However, conventional planning methods relying on flight measurement data and empirical expertise face critical bottlenecks, including insufficient multi-source heterogeneous data fusion, accuracy degradation caused by decoupled modeling of environment, antenna, and channel, and the lack of interpretability in deep learning approaches. To address these challenges, a multi-source fusion digital twin architecture for wireless channels in low-altitude networks was proposed, comprising five core modules: data sensing and acquisition, data processing and fusion, digital twin model construction and updating, twin-driven network planning, and digital twin visualization. By hierarchically constructing four-layer digital twins—environmental terrain, physical sites, antenna beams, and wireless channels—high-fidelity mapping from geometric scenarios to electromagnetic propagation was achieved. Key technologies included a terrain-label-aware rapid decoupling algorithm for Grid-to-TIN digital elevation model conversion, a high-fidelity bounding box fitting method for automatic engineering parameter extraction, a dynamic beam twinning model based on weight files, and a data-knowledge dual-driven channel twinning paradigm integrating ray tracing with learnable antenna patterns. Validation based on 25 000 measured channel samples demonstrates that the proposed channel twinning model reduces the root mean square error (RMSE) of reference signal received power (RSRP) prediction by 24.3% and 42.2% compared to traditional two-ray models and pure ray tracing methods, respectively, with mean absolute error (MAE) reductions of 30.7% and 47.3%. These results verify the significant accuracy improvement achieved through multi-source fusion. This study provides theoretical foundations and technical support for precise site planning, three-dimensional beam optimization, and dynamic interference coordination in low-altitude intelligent networks, advancing channel modeling toward an intelligent fusion paradigm.  
    关键词：low-altitude intelligent network;wireless communication channel;Multi-Source Fusion Digital Twin;ray tracing;Coverage Prediction and Planning   
    70
    |
    62
    |
    0
    <HTML>
    <L-PDF><WORD><XML><Meta-XML>
    <引用本文> <批量引用> 158995947 false
    更新时间：2026-09-18

    Yang Mengqi, Wu Yiman, Zhou Renfei, Wu Peng, Pan Peng, Zheng Changliang

    DOI：10.11959/j.issn.1000-0801.DXKX260294
    摘要：Collaborative inference between low-earth-orbit (LEO) satellites and unmanned aerial vehicles (UAVs) provides a new solution for real-time intelligent sensing tasks. Existing collaborative computing methods based on task offloading or non-pipelined model partitioning tend to cause mutual waiting between communication and computation, resulting in low resource efficiency. To address this problem, a satellite-UAV collaborative inference scheme combining model splitting and pipeline parallelism was proposed. In this scheme, the deep neural network was partitioned, with the head and tail subnetworks deployed on the UAVs and the computation-intensive intermediate subnetwork deployed on the satellite. Complete inference was achieved through intermediate feature transmission. On this basis, the micro-batch pipeline parallel mechanism and non-orthogonal transmission were employed, allowing the computation and transmission stages of different micro-batches to be executed in parallel. Furthermore, with the objective of minimizing the end-to-end collaborative inference delay, the neural network deployment structure, micro-batch partitioning granularity, and uplink and downlink transmission power were jointly optimized. Simulation results show that the proposed scheme can effectively reduce the collaborative inference delay and improve system resource utilization.  
    关键词：Model splitting;distributed inference;pipeline parallelism;non-orthogonal multiple access   
    72
    |
    54
    |
    0
    <HTML>
    <L-PDF><WORD><XML><Meta-XML>
    <引用本文> <批量引用> 159231797 false
    更新时间：2026-09-18

    Liu Ya’nan, Shu Yun, Zhang Zheng, Qiu Shuo, Shen Jipeng, Wang Pengfei, Cang Jiyun

    DOI：10.11959/j.issn.1000-0801.DXKX260235
    摘要：To address the problem that selfish nodes in unmanned aerial vehicle (UAV) ad hoc networks could easily enter forwarding paths, thereby degrading transmission reliability, a comprehensive trust-constrained Q-learning secure routing protocol was proposed. This protocol was designed for multi-hop data aggregation scenarios, in which a comprehensive trust model was constructed by integrating direct trust, indirect trust, energy trust, and load trust, and dynamic weighting was achieved by employing the CRITIC weighting method. On this basis, a trusted candidate neighbor sets were formed according to the comprehensive trust value, and hop-by-hop routing decisions were made in combination with Q-learning. Simulation results showed that the interference of selfish nodes in the routing process was effectively suppressed, the packet delivery ratio was improved, and the end-to-end delay was reduced, thereby validating the effectiveness of the proposed protocol in supporting secure and reliable transmission in highly dynamic UAV ad hoc networks.  
    关键词：uav ad hoc network;selfish node;comprehensive trust;Q-learning;secure routing   
    37
    |
    49
    |
    0
    <HTML>
    <L-PDF><WORD><XML><Meta-XML>
    <引用本文> <批量引用> 158996481 false
    更新时间：2026-09-18

    Liu Chao, Tian Jiwei, Jia Nan, Zhang Huanle, Wang Buhong, Liu Zhouzhou

    DOI：10.11959/j.issn.1000-0801.DXKX260210
    摘要：With the continuous advancement of digital transformation in air traffic management (ATM) systems, their cybersecurity risks have been increasingly intensifying, and threat entity recognition has become a core component in constructing cybersecurity knowledge graphs. To address the deficiency that traditional sequence labeling models struggled to effectively capture textual syntactic structures and semantic dependencies, a hybrid method integrating graph convolutional network (GCN) and Transformer was proposed for ATM network threat entity recognition. Specifically, text graphs were constructed based on syntactic parsing and word co-occurrence relationships, and semantic subgraphs were generated and normalized through node importance ranking and breadth-first search. A hierarchical regularized deep graph convolutional neural network (HR-DGCNN) was designed, in which parent-child label similarity constraints were introduced to enhance the model's capability in identifying hierarchical threat entities. Meanwhile, global contextual semantic features were extracted by the Transformer encoder, and optimal entity sequence decoding was accomplished in combination with conditional random field (CRF). Comparative experiments were conducted on a self-constructed ATM threat corpus, and the results demonstrated that the proposed method outperformed baseline models such as BiLSTM-CRF and BERT-Tagging in terms of F1 score, precision, and recall. This method provides a novel solution for ATM system cyber threat entity recognition and possesses both theoretical and practical value.  
    关键词：air traffic management;cybersecurity;threat entity recognition;graph convolutional network;Transformer;knowledge graph   
    102
    |
    58
    |
    0
    <HTML>
    <L-PDF><WORD><XML><Meta-XML>
    <引用本文> <批量引用> 158995989 false
    更新时间：2026-09-18
查看更多

Copyright Information

Authorized by: China Association for Science and Technology
Sponsored by: China Institute of Communications
Posts and Telecom Press Co., Ltd.
Publisher: Beijing Xintong Media Co., Ltd.
Editor-in-Chief: Chen Shanzhi
Editorial Director: Li Caishan
Address: F2, Beiyang Chenguang Building, Shunbatiao No.1 Courtyard, Fengtai District, Beijing, China
Tel: 010-53879277/9279
E-mail: dxkx@ptpress.com.cn
Mailing Code: 2-397

ISSN (print): 1000-0801

lSSN (online): 2097-5589

CN: 11-2103/TN

Copyright Information Copyright Information
0