IEEE Transactions on Network Science and Engineering最新文献

{"title":"Multi Layer Jamming Detection and Classification in 6G-Enabled Satellite-Air-Ground Integrated Networks Using Machine Learning","authors":"Shikhar Verma;Tiago Koketsu Rodrigues;Nei Kato;Masayuki Ariyoshi;Yohei Hasegawa","doi":"10.1109/TNSE.2026.3676399","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3676399","url":null,"abstract":"Future 6G networks aim to ensure ultra-reliable, high-throughput communication not only in dense urban environments but also across remote, rural, and disaster-prone regions.To support such ubiquitous connectivity, the Satellite-Air-Ground Integrated Network (SAGIN) has emerged as a promising multi layer architecture, integrating Low Earth Orbit (LEO) satellites, High Altitude Platform Stations (HAPS), Unmanned Aerial Vehicles (UAV), and terrestrial networks.However, the heterogeneous and dynamic nature of SAGIN significantly expands its attack surface, with multi layer jamming attacks posing a severe threat to communication reliability and network security. These jamming attacks can originate from space, air, or ground layers and may simultaneously target multiple communication domains, disrupting both data and control links. Despite their critical impact, research on the characteristics and correlations of multi layer jamming remains scarce, largely due to the lack of integrated platforms and publicly available datasets. Additionally, varying conditions across layers—such as mobility, power constraints, and spectrum usage—make real-time jamming classification extremely challenging. Traditional threshold-based approaches often result in high false alarm rates and poor generalization. To address these issues, we proposed a novel random forest based supervised machine learning classification model capable of real-time detection, source and target identification, and jamming type classification in SAGIN. To train the proposed random forest algorithm, we propose a novel synthetic dataset generation framework that models coordinated jamming scenarios under diverse SAGIN conditions. Moreover, this paper presents the first in-depth analysis of multi layer jamming characteristics and inter-layer correlations. Comprehensive evaluation results—comprising statistical analysis, correlation studies, confusion matrices, performance metrics, robustness assessment, and benchmarking against state-of-the-art machine learning models—validate the effectiveness, robustness, and generalization capability of the proposed framework.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8224-8240"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147736945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Topology Linearization for Multi-Agent Systems Security: Mitigating Malicious Propagation via Path Decomposition","authors":"Yan Liu;Sujie Shao;Shaoyong Guo;Chenyu Wang;Chao Yang;Zhibin Zang","doi":"10.1109/TNSE.2026.3680460","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3680460","url":null,"abstract":"The rapid adoption of Large Language Model (LLM)-based Multi-Agent Systems (MAS) is hindered by severe security risks, where high-connectivity topologies facilitate the explosive propagation of prompt injections and hallucinations. To address this, we propose the Chain-Type Secure Path Reconstruction Method (C-SPRM), a topological defense framework that fundamentally constrains threat lateral movement by transforming arbitrary interaction graphs into linearized logical execution chains based on path decomposition theory. We formulate the topology reconstruction as a Spatio-Temporal Attack Surface (STAS) minimization problem and develop a Graph-to-Sequence Path Decomposition via Reinforcement Learning (G2S-PDR) model to generate optimized execution sequences that balance concurrency risks against latency. Additionally, an adaptive topology virtualization mechanism is introduced to resolve topological bottlenecks in high-degree networks like star or fully connected graphs. Experimental results demonstrate that C-SPRM significantly reduces the effective blast radius of malicious agents and exponentially increases the difficulty of attack propagation, achieving superior system robustness with controllable computational overhead compared to traditional interaction architectures.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8481-8498"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147736961","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ARRANGE: A Secure Decentralized Federated Learning Framework for Heterogeneous LEO Satellite Constellation","authors":"Liang Zhao;Shenglin Geng;Ammar Hawbani;Yunhe Sun;Tianyu Li;Ammar Muthanna;Zhi Liu","doi":"10.1109/TNSE.2026.3677423","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3677423","url":null,"abstract":"Low Earth Orbit (LEO) satellite constellations play an increasingly vital role in Earth observation and data-driven applications. However, traditional centralized processing frameworks are hindered by limited downlink bandwidth and high satellite mobility, making real-time and large-scale data transmission inefficient. To tackle these challenges, we propose ARRANGE, <underline>A</u> secu<underline>R</u>e and decent<underline>R</u>alized Feder<underline>A</u>ted Lear<underline>N</u>ing framework specifically designed for hetero<underline>G</u>eneous LEO satellite Const<underline>E</u>llation. ARRANGE enables inter-satellite task offloading and decentralized model aggregation, while incorporating symmetric and asymmetric encryption to ensure data security and integrity during transmission. To address the straggler effect and data heterogeneity, we introduce an improved Genetic Algorithm (GA) that optimizes task scheduling across satellites with varying resources and institutional affiliations. Simulation results on real-world benchmark datasets demonstrate that ARRANGE significantly reduces system delay and improves model accuracy compared to existing FL strategies. This work provides a scalable and secure solution for enabling on-orbit collaborative intelligence in dynamic satellite environments.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8379-8396"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147737026","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Robust Beamforming for RSMA-Aided Bistatic Integrated Sensing and Communication in Low-Altitude Wireless Networks","authors":"Zhijie Lyu;Xuantao Lyu;Ye Wang;Rui Wang;Yi Gong","doi":"10.1109/TNSE.2026.3683486","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3683486","url":null,"abstract":"The integrated sensing and communication (ISAC) technology enhances the next generation of wireless networks by introducing a new approach to secure communication. In this work, a novel rate-splitting multiple access (RSMA)-aided bistatic ISAC system is proposed to serve multiple users and to estimate the parameters of a moving target. The design of the beamforming matrix aims to both maximize weighted sum-rate (WSR) and minimize the Cramér-Rao bound (CRB) for the target under a power budget constraint. In particular, we consider a scenario where a dual-functional radar communication (DFRC) waveform is emitted with a trade-off between WSR for multiple users and CRB for a target. Inspired by the robust downlink communication capabilities of RSMA, users exhibit improved performance compared to the space division multiple access based ISAC, as it effectively addresses multi-user interference by decoding it partially and treating the remaining interference as noise. In order to strike a balance between communication and sensing requirements, the system combines two metrics of DFRC through a coefficient. Additionally, methods such as weighted minimum mean square error, semidefinite relaxation, and successive convex approximation are employed to tackle the issue of non-convexity. The simulation results demonstrate the performance gain and effectiveness of the proposed architecture over the baseline scheme.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8702-8719"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147796125","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Cross-Modal Generative Semantic Communications Powered by Semantic Knowledge Base","authors":"Zechuan Fang;Mengying Sun;Sen Wang;Xiaodong Xu;Haixiao Gao;Jinghong Huang;Shujun Han;Ping Zhang","doi":"10.1109/TNSE.2025.3650269","DOIUrl":"https://doi.org/10.1109/TNSE.2025.3650269","url":null,"abstract":"The rapid advancement of generative artificial intelligence (GAI) has opened new avenues for semantic communication (SemCom). In this paper, we propose CG-SemCom, a unified cross-modal generative semantic communication framework powered by shared semantic knowledge bases (SKBs). CG-SemCom leverages GAI technologies as semantic extractor at the transmitter and generative reconstructor at the receiver, enabling flexible and interpretable cross-modal transmission. We further develop VCG-SemCom, a visual transmission-oriented implementation of CG-SemCom. Specifically, the transmitter employs a vision-language large model (vLLM) to extract concise semantic information in the form of visual difference description, which is transmitted via the joint source-channel coding (JSCC). At the receiver, the diffusion model (DM)-based generative reconstructor synthesizes the target image. The knowledge retrieval mechanism tailored for shared SKBs is introduced to guide semantic extraction and ensure consistency. Additionally, a deep reinforcement learning (DRL)-driven inference agent is proposed to dynamically optimize the generation process at the receiver. To address semantic noise caused by knowledge misalignment and module mismatch, a dual-level error detection and retransmission mechanism is introduced. Moreover, we propose a novel generation similarity metric to evaluate reconstruction quality without requiring access to the original image. Extensive experiments demonstrate that the proposed VCG-SemCom achieves superior information transfer efficiency compared to SemCom benchmark schemes with up to 6% improvement in semantic fidelity and over 4 times reduction in bandwidth consumption.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"5568-5585"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146026487","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DynNet: A Lightweight Distributed Framework for Topology Obfuscation in Dynamic Networks","authors":"Xuanbo Huang;Kaiping Xue;Lutong Chen;Zixu Huang;Jian Li;Hyundong Shin","doi":"10.1109/TNSE.2026.3673521","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3673521","url":null,"abstract":"Network Topology Obfuscation (NTO) has emerged as a promising scheme to conceal the physical layout of networks, thereby preventing adversaries from targeting critical nodes or links. By constructing deceptive virtual topologies, NTO schemes can obscure essential components and deploy honeypots to mislead and detect malicious behavior. However, existing NTO methods suffer from two major limitations. First, their high computational overhead makes them impractical for dynamic environments such as vehicular networks, where rapid topology adaptation is essential. Second, most current solutions depend on centralized controllers, typically within Software-defined Networking (SDN) architectures, which limits their applicability in decentralized or ad hoc settings. To address these challenges, we propose <bold>DynNet</b>, a distributed and probabilistic NTO framework that adapts to dynamic topologies. Each node in DynNet independently handles probing packets (e.g., packets with low time-to-live (TTL) values) by probabilistically choosing to respond truthfully, return deceptive data, or remain silent. This probabilistic framework allows nodes to rapidly update their behavior in response to topology changes with minimal overhead. To ensure consistency, we propose a bucket-based probabilistic mapping to ensure that packets from the same flow consistently trigger the same response. Our evaluation shows that DynNet maintains effective topology concealment in highly dynamic scenarios while significantly reducing computational and communication overhead. The results demonstrate its suitability for deployment in distributed, rapidly changing network environments.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"7684-7701"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147606106","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Resource Allocation for IRS-Assisted Secure NOMA-ISAC Network","authors":"Zhengqiang Wang;Miao Chen;Yongjun Xu;Haibo Zhang","doi":"10.1109/TNSE.2026.3677130","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3677130","url":null,"abstract":"In the design of dual-function waveforms for an integrated sensing and communication (ISAC) system, confidential information transmitted by the base station (BS) to legitimate users is designed to be embedded in the radar sensing signal. This generates a critical security concern, as the information becomes vulnerable to eavesdropping by the sensing target, threatening the physical layer security of the system. To address this problem, this paper introduces a non-orthogonal multiple access (NOMA) protocol to mitigate the interference problem in the multiple-input single-output ISAC network. Additionally, the sensing component within the dual-function waveforms is employed not only for target detection but also as artificial noise to interfere with the potential eavesdropper. Furthermore, deploying an intelligent reconfigurable surface (IRS) with optimized phase design in the ISAC network can further enhance the communication security by reshaping the wireless environment. For the IRS-assisted multi-user NOMA-ISAC network, we jointly optimize the resource allocation problem for the precoding matrix of the BS and the phase shift matrix of the IRS, considering constraints including decoding order, quality of service for legitimate users, security, sensing performance, and upper bound on the transmit power, to maximize system throughput. Given the non-convex nature of the optimization problem, an efficient joint optimization algorithm based on semidefinite relaxation and alternating optimization algorithm is proposed to find the solution for the problem. Simulation results validate the convergence of the proposed algorithm and demonstrate its superiority over other benchmark algorithms.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8139-8157"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147665290","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"FedAIM: Graph-Driven Scheduling for Concurrent Asynchronous Federated Learning in Computing Power Networks","authors":"Guoming Yang;Shaoyong Guo;Yitao Xiao;Xinran Mao;Sai Huang;Feng Qi;Jiakai Hao","doi":"10.1109/TNSE.2026.3675692","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3675692","url":null,"abstract":"Asynchronous federated learning (AFL) faces multi-job scheduling challenges in computational power networks (CPNs). In large-scale CPN environments, the decision space is prohibitively large, and it is challenging to extract critical scheduling information. This leads to asynchronous jobs contending for computing and communication resources, which affects the staleness of local models and results in degraded training efficiency. To address these issues, we propose FedAIM, a graph-driven scheduling framework for multi-job AFL in CPNs. FedAIM encodes the system state into computation, communication, and staleness views to capture the interdependence between jobs and computing nodes in resource scheduling and model optimization, and leverages reinforcement learning for job scheduling. It combines two key components: a Mixture-of-Views Router (MoV-R), which applies context-aware gating across multiple graphs to generate lightweight representations that filter redundancy while preserving interference-critical signals; and an asymmetric Actor–Critic scheduler, where the Actor exploits MoV-R features for low-latency job–device allocation and the Critic integrates a global view for long-term evaluation. Experiments on large-scale simulations and a 50-node Raspberry Pi testbed show that FedAIM improves accuracy by 13–19%, reduces communication by 43%, and lowers computation cost by 45%, achieving a favorable trade-off between learning quality and system efficiency.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8158-8177"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147665310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Edge-Aided Collaborative Inference for Visual Language Navigation-Based Intelligent Vehicles","authors":"Jinkai Zheng;Tom H. Luan;Guanjie Li;Yuan Chang;Yalun Wu;Yuan Wu;Haixia Peng","doi":"10.1109/TNSE.2026.3679193","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3679193","url":null,"abstract":"In Visual Language Navigation (VLN)–based autonomous driving, the limited field of view of an individual vehicle hinders the acquisition of comprehensive environmental information, thereby affecting navigation reliability. Collaborative perception alleviates this issue by enabling information-sensing vehicles (ISVs) to collect visual observations along the trip, which information-requesting vehicles (IRVs) can then exploit to enhance scene understanding and navigation decisions. To further reduce the substantial reasoning burden that exceeds the computational capacity of individual vehicles, edge devices (EDs) equipped with large language models (LLMs) can provide efficient VLN inference services. While such collaboration significantly enriches scene awareness, it also introduces several challenges. First, ISVs and EDs possess heterogeneous resources and behave as self-interested entities that incur energy, communication, and computation costs, necessitating appropriate incentive mechanisms to sustain their participation. Second, an ISV may be simultaneously covered by multiple EDs, and selecting the most suitable ED depends on the dynamic characteristics of wireless links, vehicle states, and edge resources. Furthermore, the computation cost of large-model inference at the edge is difficult to model accurately due to the complexity of VLN pipelines. To address these challenges, this paper proposes an edge-aided collaborative inference framework that coordinates ISVs, IRVs, and EDs within a hierarchical incentive structure to enable stable and efficient cooperation. A vehicle–edge matching algorithm is developed to determine optimal associations, and an empirical validation framework is introduced to accurately estimate the costs of large-model–based VLN inference at the edge. Simulation results demonstrate that the proposed framework converges rapidly and substantially improves the utilities of ISVs, EDs, and IRVs, as well as the social welfare.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8445-8461"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147736928","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Heterogeneous Hypergraph Multi-Agent Learning for UAV Collaboration in Disaster Scenarios","authors":"Ziqi Rong;Zhishu Shen;Wanwei Zhan;Qiushi Zheng;Tiehua Zhang;Jiong Jin","doi":"10.1109/TNSE.2026.3680588","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3680588","url":null,"abstract":"Uncrewed Aerial Vehicles (UAVs) play a critical role in disaster scenarios by providing emergency communication and data collection services where terrestrial infrastructure is unavailable. However, effective multi-UAV collaboration remains challenging due to dynamic ground node distributions and the presence of high-order correlations that conventional methods struggle to capture. To this end, we propose a heterogeneous hypergraph multi-agent reinforcement learning framework that jointly optimizes throughput, fairness, and energy consumption in UAV collaboration. By leveraging heterogeneous hypergraphs, the framework captures complex many-to-many interactions between UAVs and ground nodes, embedding these high-order correlations into actor–critic networks with entropy regularization. This design enables the transformation of local observations into globally optimized UAV collaboration. Extensive simulations demonstrate that the proposed method outperforms state-of-the-art baselines, achieving up to a 10% increase in throughput, a 5% improvement in fairness, and a 4% reduction in energy consumption.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8362-8378"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147736929","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}