Online FirstList of IssuesDocument Types

    Volume47,Issue10,2026
    Issue CoverIssue Contents

      Topics\: Theory and Technology of Power\-Computing Synergy

    • Grid carbon-intensity-aware power-computing coordinated task scheduling for data centers AI Introduction

      Wang Ziyi, Guo Shaoyong, Chen Jiewei, Xiao Yitao, Cai Tiantian, Zhang Zihao, Qi Feng
      Vol. 47, Issue 10, Pages: 1-18(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260382
      Abstract:With the rapid growth of AI, data-center energy consumption continues to increase. To balance computing service requirements and renewable-energy fluctuations, a grid carbon-intensity-aware power-computing coordinated scheduling method was proposed. First, a scheduling framework for geographically distributed heterogeneous data centers was developed to characterize how carbon emission attributes on the power supply side influence task allocation on the computing side. Then, considering workflow dependencies, heterogeneous resources, and spatiotemporal variations in grid carbon intensity, a GINEConv-PPO scheduling mechanism was proposed. A temporal planning algorithm extracted critical-chain positions and resource-contention risks to construct temporal-priority features, while GINEConv was integrated with PPO for low-carbon task allocation under quality of service (QoS) constraints. Simulations across workflow types and sizes showed that GINEConv-PPO met job-batch deadlines. Relative to the lowest emissions achieved by the compared algorithms in each scenario, its emission deviation remained near 0, versus approximately 6%~35% for the baselines.  
      Keywords:power-computing synergy;task scheduling;carbon emission;deep reinforcement learning  
      298
      |
      246
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Survey on access pattern protection techniques for searchable encryption protocols AI Introduction

      Gong Yunping, Li Xiong, Wen Bo, Wu Chunling, Tong Shihua
      Vol. 47, Issue 10, Pages: 19-37(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260460
      Abstract:To address the issue that searchable encryption protocols leak access patterns during query execution, which can lead to attacks such as query recovery, database reconstruction, and file injection, a systematic review of access pattern protection techniques was conducted. Firstly, a leakage lineage model centered on access patterns was constructed, and a four-dimensional technical classification framework covering ORAM/OMAP, obfuscation and differential privacy, distributed multi-server architectures, and hardware-based trusted execution environments was proposed. Representative schemes were analyzed and compared in depth, and an evolution graph of attacker-defender dynamics was mapped out. Secondly, a multidimensional evaluation framework encompassing security, efficiency, functionality, and deployability, along with a comprehensive assessment matrix, was established, and the industrial deployment status and open-source implementations were surveyed. Finally, future challenges and opportunities were discussed in areas such as post-quantum security, novel computing architectures, and integration with advanced search functionalities.  
      Keywords:searchable encryption protocol;access pattern leakage in lineage model;oblivious random access machine (ORAM);forward privacy;backward privacy  
      152
      |
      99
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Survey of the architecture and key technologies for computing-electricity collaboration AI Introduction

      Wang Xiaohui, Zheng Bihuang, Wu Chunpeng, Li Shuo, Wang Yiran, Xie Renchao, Tang Qinqin, Huang Tao
      Vol. 47, Issue 10, Pages: 38-64(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260315
      Abstract:The rapid development of artificial intelligence and emerging applications had driven sustained growth in computing demand, making the energy consumption of data centers an increasingly critical issue. Under the goals of high renewable-energy penetration and carbon reduction, the spatiotemporal mismatch and siloed operation of computing and power systems had constrained energy efficiency and grid regulation capabilities. Computing-electricity collaboration (CEC) integrates information and energy flows and converted migratable computing workloads into flexible resources for power-system regulation. Against this background, the state of the art in CEC was systematically reviewed. Firstly, the core concepts and fundamental connotations of CEC were clarified. Secondly, a six-layer reference architecture was constructed, and key technologies were examined from four dimensions: basic theory, perception and migration, analytical computation, and closed-loop control. The applicability, advantages, and limitations of major technical routes were also compared. Furthermore, two representative paradigms, namely “computing following power” and “power following computing”, were summarized. Finally, the major technical challenges were identified, and future directions, including intelligent scheduling, green microgrids, and standardization, were discussed.  
      Keywords:computing power network;computing-electricity collaboration;green collaborative computing;key technologies  
      712
      |
      831
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20

      Papers

    • Knowledge-driven autonomous orchestration method of service resources for the Internet of agents AI Introduction

      Qian Qijie, Song Yangzi, Qin Zhen, Zhong Xudong, Guo Yong’an, Ren Baoquan
      Vol. 47, Issue 10, Pages: 65-86(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260271
      Abstract:As a core resource allocation method, virtual network embedding (VNE) faces severe challenges in dynamic network environments. To address the poor adaptability and weak interpretability of existing deep reinforcement learning methods, which stem from their reliance on implicit knowledge encoding, a knowledge-driven resource orchestration framework was proposed. The framework pioneered the use of a temporal knowledge graph (TKG) as an explicit world model for the agent, enabling the structured representation and reasoning of network dynamics. Furthermore, a hybrid Actor-Critic agent was designed to integrate real-time states, historically curated experience, and domain heuristics through a dual-loop mechanism, achieving a synergy between rapid tactical adaptation and long-term strategic optimization. Experimental results demonstrate that the proposed framework achieves excellent performance in terms of the request acceptance ratio and revenue-to-cost ratio. In cross-topology transfer and online knowledge adaptation scenarios with frozen policy parameters, it maintains a consistent advantage; specifically, in exponential traffic scenarios, it improves upon the strongest baseline acceptance rate by 6.50% and achieves a relative improvement of over 80% compared to certain DRL baselines. This work provides key technical support for enabling efficient and reliable service provisioning in the Internet of agents.  
      Keywords:IoA;Knowledge-Driven;VNE;resource orchestration;TKG  
      219
      |
      93
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Wu Haiqin, Li Liang, Li Changfeng
      Vol. 47, Issue 10, Pages: 87-101(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260255
      Abstract:Task recommendation is a key service in crowdsourcing, which suffers from single points of failure and severe security and privacy issues. Existing privacy-preserving crowdsourcing task recommendation algorithms struggle to balance between data security and recommendation accuracy, and fail to accommodate workers’ personalized preferences for tasks. To address these, crowdsourcing was empowered with blockchain and federated learning, and a personalized and secure crowdsourcing task recommendation framework PBFRec was proposed. Specifically, the traditional centralized crowdsourcing platform was replaced with a consortium blockchain. Workers train local preference models using graph neural networks, and privacy-preserving global model aggregation and updates were achieved via differential privacy. Furthermore, a two-layer model aggregation method was designed based on data quality and a social relation-based model initialization method, which addressed the situation where malicious users submit low-quality model parameters and the cold-start problem. PBFRec satisfies Tε-local differential privacy and ensures the integrity of task recommendation model training. A prototype experiment on Hyperledger Fabric shows that PBFRec outperforms existing solutions in task recommendation quality and system performance. The recall rate is up to 2.9 times as much as that of existing approaches, and social relationships contribute positively to alleviating the cold-start problem.  
      Keywords:crowdsourcing;consortium blockchain;federated learning;task recommendation;graph neural network  
      171
      |
      143
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Yun Yanzhi, Meng Qingwei, Li Jiayue, Wang Keqing, Wang Dan
      Vol. 47, Issue 10, Pages: 102-114(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260361
      Abstract:To enhance the security and confidentiality of wireless communication signals, a novel two dimensional extended weighted fractional Fourier transform (2D-EWFRFT) communication method based on hyperchaos-driven dual quaternion scrambling for 3D constellation encryption was proposed. A hyperchaotic map with superior dynamic performance was constructed by employing the sine and exponential function. On the basis of dual quaternion theory, a 3D constellation encryption algorithm with joint rotation and translation was designed, effectively reducing computational complexity. The 2D-EWFRFT was constructed, where the interaction of multiple transform parameters further enriched signal forms. Furthermore, a cascade encryption strategy controlled by hyperchaos was presented, in which the 3D modulated signal was first encrypted via constellation scrambling and then processed by 2D-EWFRFT. The limitations of insufficient scrambling in single constellation encryption and weak parameter sensitivity of the conventional EWFRFT were overcome. Simulation results demonstrate that the proposed method can disrupt the distribution characteristics of modulated signals, scramble and diffuse the transmitted information, and enhance the secure transmission capability of wireless communications.  
      Keywords:physical layer security;3D constellation encryption;dual quaternion;2D-EWFRFT;hyperchaos  
      45
      |
      135
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Adversarial example detection method based on inter-layer activation evolution consistency AI Introduction

      Lin Yun, Wang Jiali, Zhang Sicheng, Wei Xiaolei
      Vol. 47, Issue 10, Pages: 115-129(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260186
      Abstract:Automatic modulation classification (AMC), a core technique for electromagnetic spectrum sensing, served as a key sensing task in applications such as spectrum monitoring, yet its adversarial vulnerability posed a severe security threat. To address the limitation that existing adversarial example detection methods only analyze static features while ignoring the dynamic cumulative effects of adversarial perturbations during forward propagation, an adversarial example detection method based on inter-layer activation evolution consistency was proposed. Activation representations from multiple key network layers were extracted, and autoencoders were trained to learn their activation manifolds. Distance matrices between samples and class-average activations were constructed in the reconstruction space, and adversarial examples were detected based on their inter-layer evolution consistency. Experimental results show that an average AUC exceeding 90% is achieved across two datasets, five typical attacks, and multiple signal-to-noise ratios (SNR), effectively enhancing the security of deep-learning-based spectrum sensing against adversarial attacks.  
      Keywords:electromagnetic spectrum sensing;intelligent sensing security;adversarial example detection;inter-layer activation evolution;class average activation  
      50
      |
      76
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • SSCE-YOLO: spectral-spatial collaborative enhancement for infrared remote-sensing object detection AI Introduction

      Cheng Deqiang, Yue Yang, Wang Yanchen, Kou Qiqi, Xu Feixiang, Sun Fengqian
      Vol. 47, Issue 10, Pages: 130-145(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260204
      Abstract:To address noise interference and complex background disturbance in infrared remote sensing object detection, a spectral-spatial collaborative enhancement algorithm termed SSCE-YOLO was proposed. The method constructed a hierarchical collaborative enhancement mechanism consisting of channel pre-screening, frequency-domain modulation, deep gated propagation, and global stability constraints. The frequency-spatial collaborative (FSC) filtering suppressed noise-related frequency bands while preserving target structures. The SpectralGNConv (SGNC) module injected FSC-enhanced features into multi-order channel interactions to continuously propagate denoising priors. Meanwhile, the channel-aware global context (CGC) employed mean-variance statistics to suppress high-fluctuation channels, while the efficient channel pre-selection attention (ECPA) module controlled computational overhead. Experiments on the VEDAI-IR and HIT-UAV datasets show that SSCE-YOLO improves mAP@0.5 from 59.2 to 63.7 and from 79.1 to 82.2, respectively, outperforming representative state-of-the-art methods while maintaining a favorable balance between detection accuracy and computational efficiency.  
      Keywords:infrared remote-sensing;object detection;YOLO;spectral-spatial collaboration  
      186
      |
      143
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Quality-aware contribution evaluation method for federated learning in data trading AI Introduction

      Tian Youliang, Zhi Rui, Wang Shuai, Jin Kunlong
      Vol. 47, Issue 10, Pages: 146-163(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260254
      Abstract:To address the problem that existing federated learning contribution evaluation methods suffered from insufficient accuracy, high computational overhead, or limited robustness, and thus could not effectively support contribution settlement and revenue allocation in data trading, a federated learning contribution evaluation method for data trading was proposed. First, a quality-aware client weight optimization method was designed to obtain foundational signals reflecting client quality differences by optimizing client aggregation weights. Then, weight variations, per-round marginal gains, and historical performance were incorporated, and multi-source signals were transformed into subjective opinions containing positive evidence, negative evidence, and uncertainty through subjective logic, thereby enabling dynamic contribution mapping. Experiments conducted on text, audio, and image datasets demonstrate that the proposed method achieves good performance in contribution evaluation accuracy, computational efficiency, and attack robustness, while also exhibiting dynamic adaptability and interpretability.  
      Keywords:federated learning;Contribution Evaluation;Quality-Aware Weight Optimization;subjective logic  
      165
      |
      159
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Fusion localization based on a dual-scale gated Mamba-CNN differentiable Kalman filter AI Introduction

      Jiang Rui, Zhang Teng, Zhang Tianchi, Shen Zihang, Wang Xiaoming, Xu Youyun
      Vol. 47, Issue 10, Pages: 164-186(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260238
      Abstract:To address the limitations of existing differentiable Kalman filter (DKF) methods, which commonly relied on recurrent neural-network architectures and therefore suffered from limited long-range memory, difficulty in simultaneously suppressing long-term sensor bias drift and transient measurement shocks, and degraded state-estimation stability under prolonged GPS outages in complex urban environments, a dual-scale gated Mamba-CNN differentiable Kalman filter (DSG-MC DKF) for fusion localization was proposed. The proposed algorithm removed the explicit control-input term from the prediction equation and feeds IMU sequences, measurement innovations, and outage masks into a dual-scale Mamba-CNN for data-driven gain prediction, thereby reducing the direct reliance on externally computed frame-rotation results and precisely specified kinematic models. Moreover, an explicit outage-mask gating mechanism was introduced, and a dual-scale parallel architecture consisting of Mamba and 1D-CNN branches was developed to decouple long-term bias compensation from high-frequency noise suppression. Experimental results show that, compared with the conventional GRU-based DKF, the proposed model improves localization robustness and accuracy while reducing single-step inference latency by approximately 80%. Furthermore, during a complete GPS outage lasting up to 10 s, the proposed model can still effectively preserve the kinematic and topological consistency of the vehicle’s trajectory.  
      Keywords:differentiable Kalman filter;fusion localization;structured state space model;GNSS outage  
      221
      |
      178
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • UAV RF signal detection based on time-frequency enhancement and lightweight detection network AI Introduction

      Wang Wei, Shi Zhongkai, Sun Minhong, Yang Shaofei, Shen Weiguo
      Vol. 47, Issue 10, Pages: 187-208(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260222
      Abstract:To address weak target energy, missed detection, and network redundancy in unmanned aerial vehicle (UAV) radio-frequency (RF) time-frequency map detection under low signal-to-noise ratio (SNR) conditions, a joint method combining time-frequency enhancement and lightweight detection was proposed. The proposed method selectively enhanced target-related regions using an adaptive foreground mask, multi-scale energy density, and connected-component interior weighting. A lightweight radio-frequency object detection network based on you only look once (RF-YOLO-Lite) was then constructed through P2/P5 scale reconstruction, a Ghost-based backbone, SlimNeck feature fusion, and a lightweight shared decoupled detection head. Experimental results show that the proposed method achieves 94.85% mAP@0.5 and 79.86% mAP@0.5:0.95 on the simulated dataset, with 1.39×106 parameters and 5.4 GFLOPs. Experiments on noise-corrupted measured data from DroneRFa (a large-scale dataset of drone radio frequency signals for detecting low-altitude drones) further demonstrate the robustness of the proposed method under low-SNR conditions. The method provides a favorable trade-off between detection accuracy and computational complexity.  
      Keywords:UAV RF signal;lightweight object detection;low SNR;time-frequency enhancement  
      250
      |
      121
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Xie Yaqin, Zhang Nian, Zhang Yu, Zhang Zhizhong, Li Jiamin, Wang Dongming
      Vol. 47, Issue 10, Pages: 209-226(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260290
      Abstract:To address the problem that cell-free networks relying on fixed access point (AP) was incapable of handling burst traffic demands and infrastructure-lacking scenarios, a rapid deployment method for dynamic unmanned aerial vehicle (UAV)-assisted cell-free networks based on deep reinforcement transfer learning and interference management (IM) was proposed. In the proposed method, UAV served as scalable nodes, and deep reinforcement learning (DRL) was employed to drive cooperative deployment, enabling both supplementary coverage and dynamic load balancing. A transfer learning mechanism was further introduced to adapt source-domain policies to target-domain dynamic scenarios, thereby substantially reducing the retraining overhead. Meanwhile, an IM-based distributed spectrum allocation strategy was incorporated to mitigate co-channel interference among densely deployed UAV. Simulation results demonstrate that, compared with conventional DRL-based strategies, the proposed method achieves faster convergence and better congestion relief, limits the user waiting probability to 2%, and improves the average user satisfaction by 44.3%. The proposed method overcomes the coverage limitations of fixed AP, providing an efficient and intelligent solution for bursty scenarios.  
      Keywords:UAV;deep reinforcement learning;transfer learning;interference management  
      231
      |
      172
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Hao Yaohui, Li Na, Chen Xin, Qin Yudong, Chang Heyu
      Vol. 47, Issue 10, Pages: 227-239(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260339
      Abstract:To solve the problems of insufficient interpretability and fragmented defense strategy generation in the reaso-ning process of existing cyber threat intelligence question-answering systems, a cyber threat intelligence question-answering framework integrating a knowledge graph and multi-agent collaboration was proposed. The framework was composed of three major functional modules: the ATT&CK knowledge graph, multi-agent collaboration, and model infe-rence enhancement. A bidirectionally reversible cyber threat knowledge graph was constructed from ATT&CK and STIX 2.1 data. Collaboration among multiple types of agents was implemented using the blackboard architecture and LangGraph technology. Model inference enhancement was achieved by integrating SP-CoT and graph retrieval, whereby attack attribution analysis and defense strategy generation were enabled. Experiments were carried out on the AthenaBench benchmark, and prominent performance improvements over comparison models were obtained. Experimental results demonstrate that the proposed framework delivers superior capabilities in multi-hop reasoning and defense strategy generation.  
      Keywords:Attack intent recognition;knowledge graph;ATT&CK;large language model;agent;defense strategy  
      53
      |
      22
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Privacy protection mechanism for cross-domain vehicle charging based on identity and anonymous credentials AI Introduction

      Yi Yuzi, Su Yiyue, Huang Jiwei, Liu Lin, Shi Jinqiao
      Vol. 47, Issue 10, Pages: 240-257(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260323
      Abstract:To address issues such as identity privacy leaks, lack of trust across management domains, and difficulty in efficiently auditing high-frequency billing data in electric vehicle cross-domain charging scenarios, a privacy protection mechanism for cross-domain vehicle charging was proposed. Firstly, a distributed registration and dynamic pseudonym mechanism was designed to achieve cross-domain identity anonymity and behavior unlinkability. Secondly, a hybrid cross-domain access control model was proposed, in which the consortium blockchain smart contract handled global access decisions and the charging service provider handled local dynamic authorization. In addition, a trusted billing and auditing mechanism was introduced to ensure data integrity and non-repudiation of charging records. Security analysis and experiment results demonstrate that the proposed scheme can provide comprehensive charging services for vehicles. It outperforms existing schemes in terms of authentication efficiency, privacy protection, and system scalability, effectively mitigating trust and privacy leakage risks in cross-domain charging scenarios.  
      Keywords:privacy protection;identity-based cryptography;anonymous credential;consortium blockchain  
      56
      |
      25
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20

      Comprehensive Reviews

    • Survey on deterministic endogenous security defense techniques for industrial Internet AI Introduction

      Zhang Zhiwei, Tang Guiyuan, Shen Yulong, Chen Yingyu, Xi Ning, Ren Baoquan, Jiao Chengwei, He Ji, Zhang Tao, Ma Jianfeng
      Vol. 47, Issue 10, Pages: 258-284(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260345
      Abstract:To address the structural mismatch between strict operational constraints of the industrial Internet and the conventional bolt-on, static-perimeter security paradigm, research on deterministic endogenous security defense was conducted. A deterministic endogenous security defense paradigm was proposed based on four categories of intrinsic determinism: entity, behavior, structure, and environment. Five key enabling technologies were systematically surveyed: endogenous identity management, task-behavior assurance, unknown-attack inference, active security defense, and security effectiveness evaluation. These were used to construct an endogenous security defense architecture covering prerequisites, system models, and four-dimensional technical implementation. The common bottlenecks of existing endogenous security practices in cross-path coordination and scenario adaptation were revealed through comparative analysis.  
      Keywords:industrial Internet;deterministic endogenous security;task-behavior assurance;unknown-attack inference  
      80
      |
      50
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Survey on website fingerprinting defense technologies for Tor network AI Introduction

      Wang Kangxu, Hu Hongchao, Cheng Guozhen, Ren Quan, Zhou Dacheng, Chang Huiyan, Liu Wenyan, Yang Xiaohan
      Vol. 47, Issue 10, Pages: 285-312(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260282
      Abstract:The security risks of Tor network architecture and website fingerprinting were clarified from the perspective of side-channel leaks such as packet timing and traffic bursts in encrypted traffic. A hierarchical classification framework for website fingerprinting defenses was proposed, and four defense categories were compared horizontally and vertically: traffic feature obfuscation, attack model misdirection, anonymity set enhancement, and cross-layer collaborative optimization. Public evaluation frameworks and datasets were reviewed, and an evaluation system integrating defense effectiveness, system overhead and deployment characteristics was constructed for defense selection. Core challenges including concept drift and adversarial robustness were summarized, and frontier research directions such as adaptive defense and native Tor integration were outlined.  
      Keywords:Tor network;cybersecurity;website fingerprint defense;comprehensive assessment  
      183
      |
      151
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20

      Correspondences

    • Ye Qing, Shao Zhenjie, Tang Yongli, Wang Yichuan, Hei Xinhong
      Vol. 47, Issue 10, Pages: 313-325(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260353
      Abstract:Aiming at the problems of high communication cost and large verification overhead in existing ring signature schemes under scenarios of large-scale ring members and multi-signature verification, a novel aggregate ring signature scheme based on homomorphic vector commitments and compressed ∑ protocols was designed. The scheme first generated individual ring signatures based on BLS signatures, (Zq,G1) vector commitments, and the compressed ∑ protocol for (Zq,G1) vector commitments. Furthermore, based on ElGamal-like vector commitments and the ∑ protocol for ElGamal-like vectors, it aggregated k individual ring signatures into one aggregated ring signature for transmission and verification. Theoretical analysis and experimental verification demonstrate that the proposed scheme does not require a trusted third party and satisfies security properties such as anonymity and unforgeability. Moreover, the total signature length and verification time for k individual signatures are both O(k), independent of the ring size, and the efficiency is significantly superior to similar schemes.  
      Keywords:aggregate ring signature;vector commitment;compressed Σ protocol;bilinear group  
      63
      |
      109
      |
      2
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    • Semantic communication with stable latent representation learning in wide-SNR multi-channel environments AI Introduction

      Yang Yanbo, Gao Xiaoting, Zhang Jiawei, Ma Jianfeng, Gao Lili
      Vol. 47, Issue 10, Pages: 326-346(2026) DOI: 10.11959/j.issn.1000-436x.TXXB260356
      Abstract:To address latent semantic distribution shifts across wide signal-to-noise ratio (SNR) ranges and diverse channels, a channel-aware stable semantic communication framework was proposed. It integrated channel-state-conditioned modulation, cross-channel consistency constraints, YUV-based decoupled reconstruction, and dynamic feature adaptation to enhance representation stability and reconstruction robustness. Stability was evaluated using sliced Wasserstein distance (SWD), Fréchet distance (FD), centered kernel alignment (CKA), mutual information (MI), and t-SNE. Experiments over AWGN, Rayleigh, and Rician channels demonstrate reduced distribution shifts, improved structural consistency, and enhanced cross-channel generalization while maintaining high reconstruction quality.  
      Keywords:semantic communication;channel awareness;wide SNR;YUV decoupled reconstruction  
      163
      |
      169
      |
      0
      <HTML>
      <L-PDF> <WORD> <XML> <Meta-XML>
      <Citation><Bulk Citation>
      Updated:2026-10-20
    Previous IssueNext Issue