Pub Date : 2026-01-22DOI: 10.1109/tccn.2026.3657147
Bo Chen, Jianping An, Bowen Gui, Liang Zeng
{"title":"High Quality and Secure Speech Transmission at Low Bitrate via Semantic-Acoustic Hybrid Coding for Low-Altitude Intelligent Systems","authors":"Bo Chen, Jianping An, Bowen Gui, Liang Zeng","doi":"10.1109/tccn.2026.3657147","DOIUrl":"https://doi.org/10.1109/tccn.2026.3657147","url":null,"abstract":"","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"50 1","pages":""},"PeriodicalIF":8.6,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146042786","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-20DOI: 10.1109/TCCN.2026.3656295
Song Han;Hongyun Zhang;Xinbin Li;Junzhi Yu;Zhixin Liu;Tongwei Zhang;Xin Zheng;Weigang Nie
This study investigates a joint resource allocation problem for the unknown time-varying underwater acoustic cooperative communication networks. A low-computational-cost and high-efficiency joint strategy decision-making approach, enhanced in three aspects: improving the quantity and quality of learning information, and optimizing the search method for the complex coupled strategy space, is proposed. First, we formulate a combinatorial multi-armed bandit learning model, which offers the separate learning spaces for each sub-strategy. Thus, the player can effectively learn the high-dimensional entire joint strategy by searching the low-dimensional sub-strategy space with low computational cost to avoid the complex direct search in the high-dimensional space. As a result, the solving difficulty caused by the complex coupling of joint sub-strategies can be significantly mitigated. Second, a hierarchical combinatorial bandit learning structure is formulated to improve the quantity and quality of learning information. The estimation layer is constructed to denoise the observed rewards, thereby providing higher quality learning information. The prediction layer is constructed to predict the rewards of all sub-strategies, thereby enriching learning information. Furthermore, in the decision-making layer, we propose an extended upper confidence bound 1 index function to achieve the effective integration of the higher quality denoised learning information and richer predicted learning information to improve the online decision. Finally, the simulation results verify that the proposed innovations can significantly improve the online decision-making efficiency for complex coupled joint strategies and effectively enhance the performance of underwater acoustic cooperative communication networks.
{"title":"Joint Resource Allocation for Underwater Acoustic Cooperative Communication Networks: A Hierarchical Combinatorial Bandit Approach","authors":"Song Han;Hongyun Zhang;Xinbin Li;Junzhi Yu;Zhixin Liu;Tongwei Zhang;Xin Zheng;Weigang Nie","doi":"10.1109/TCCN.2026.3656295","DOIUrl":"https://doi.org/10.1109/TCCN.2026.3656295","url":null,"abstract":"This study investigates a joint resource allocation problem for the unknown time-varying underwater acoustic cooperative communication networks. A low-computational-cost and high-efficiency joint strategy decision-making approach, enhanced in three aspects: improving the quantity and quality of learning information, and optimizing the search method for the complex coupled strategy space, is proposed. First, we formulate a combinatorial multi-armed bandit learning model, which offers the separate learning spaces for each sub-strategy. Thus, the player can effectively learn the high-dimensional entire joint strategy by searching the low-dimensional sub-strategy space with low computational cost to avoid the complex direct search in the high-dimensional space. As a result, the solving difficulty caused by the complex coupling of joint sub-strategies can be significantly mitigated. Second, a hierarchical combinatorial bandit learning structure is formulated to improve the quantity and quality of learning information. The estimation layer is constructed to denoise the observed rewards, thereby providing higher quality learning information. The prediction layer is constructed to predict the rewards of all sub-strategies, thereby enriching learning information. Furthermore, in the decision-making layer, we propose an extended upper confidence bound 1 index function to achieve the effective integration of the higher quality denoised learning information and richer predicted learning information to improve the online decision. Finally, the simulation results verify that the proposed innovations can significantly improve the online decision-making efficiency for complex coupled joint strategies and effectively enhance the performance of underwater acoustic cooperative communication networks.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"6104-6118"},"PeriodicalIF":7.0,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146223597","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-20DOI: 10.1109/TCCN.2026.3656389
Yuhan Ai;Qimei Chen;Dingzhu Wen;Mehdi Bennis
Powered by Artificial Intelligence (AI), AI-Generated Content (AIGC) has recently emerged as a promising approach for synthesizing multimodal content, including text, images, and videos. Diffusion model is one of the predominant frameworks within AIGC for generating diverse and photorealistic images with high fidelity. However, deploying diffusion models on practical resource-constrained edge devices confronts critical challenges for their limited computational abilities and severe communication overhead, which spurs research interests in distributed AIGC. Existing works on distributed AIGC primarily focus on either increasing training cost to pursue higher generation quality or sacrificing performance to achieve communication efficiency, merely considering their integration impacts. To address these issues, we propose a novel Cross-Modal Collaborative Diffusion Model (Co-Diff) framework that achieves computation-and-communication efficient image synthesis without compromising robust generalization. The core innovation lies in our split diffusion architecture, which strategically offloads computation-intensive reverse denoising to the server and edge devices perform lightweight forward diffusion with text-guided attention. In addition, we design a learnable quantization module that encodes high-dimensional noise into compact codes to minimize communication overhead. We establish a theoretical convergence analysis for Co-Diff, deriving a closed-form expression that reveal a quantifiable trade-off between denoising accuracy and computational latency. Based on the theoretical findings, we formulate a joint communication-and-computation optimization problem, which obtains an analytical solution among diffusion-step scheduling, processing-frequency allocation, and bandwidth distribution. Extensive experiments validate our theoretical analysis, and demonstrate the effectiveness of the proposed design for computation-and-communication efficient Co-Diff. Compared with conventional distributed learning frameworks, the proposed Co-Diff increases communication-and-computation efficiency by 72.8%, as well as boosts image synthesis quality by 12.9%.
{"title":"Cross-Modal Collaborative Diffusion Models for Distributed AI-Generated Content","authors":"Yuhan Ai;Qimei Chen;Dingzhu Wen;Mehdi Bennis","doi":"10.1109/TCCN.2026.3656389","DOIUrl":"https://doi.org/10.1109/TCCN.2026.3656389","url":null,"abstract":"Powered by Artificial Intelligence (AI), AI-Generated Content (AIGC) has recently emerged as a promising approach for synthesizing multimodal content, including text, images, and videos. Diffusion model is one of the predominant frameworks within AIGC for generating diverse and photorealistic images with high fidelity. However, deploying diffusion models on practical resource-constrained edge devices confronts critical challenges for their limited computational abilities and severe communication overhead, which spurs research interests in distributed AIGC. Existing works on distributed AIGC primarily focus on either increasing training cost to pursue higher generation quality or sacrificing performance to achieve communication efficiency, merely considering their integration impacts. To address these issues, we propose a novel Cross-Modal Collaborative Diffusion Model (Co-Diff) framework that achieves computation-and-communication efficient image synthesis without compromising robust generalization. The core innovation lies in our split diffusion architecture, which strategically offloads computation-intensive reverse denoising to the server and edge devices perform lightweight forward diffusion with text-guided attention. In addition, we design a learnable quantization module that encodes high-dimensional noise into compact codes to minimize communication overhead. We establish a theoretical convergence analysis for Co-Diff, deriving a closed-form expression that reveal a quantifiable trade-off between denoising accuracy and computational latency. Based on the theoretical findings, we formulate a joint communication-and-computation optimization problem, which obtains an analytical solution among diffusion-step scheduling, processing-frequency allocation, and bandwidth distribution. Extensive experiments validate our theoretical analysis, and demonstrate the effectiveness of the proposed design for computation-and-communication efficient Co-Diff. Compared with conventional distributed learning frameworks, the proposed Co-Diff increases communication-and-computation efficiency by 72.8%, as well as boosts image synthesis quality by 12.9%.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"5552-5565"},"PeriodicalIF":7.0,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146082000","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-24DOI: 10.1109/TCCN.2025.3647822
Odysseas G. Karagiannidis;Victoria E. Galanopoulou;Panagiotis D. Diamantoulakis;Zhiguo Ding;Octavia A. Dobre
The evolution of wireless communication systems requires flexible, energy-efficient, and cost-effective antenna technologies. Pinching antennas (PAs), which can dynamically control electromagnetic wave propagation through binary activation states, have recently emerged as a promising candidate. In this work, we investigate the problem of optimally selecting a subset of fixed-position PAs to activate in a waveguide, when the aim is to maximize the communication rate at a user terminal. Due to the complex interplay between antenna activation, waveguide-induced phase shifts, and power division, this problem is formulated as a combinatorial fractional 0-1 quadratic program. To efficiently solve this challenging problem, we use neural network architectures of varying complexity to learn activation policies directly from data, leveraging spatial features and signal structure. Furthermore, we incorporate user location uncertainty into our training and evaluation pipeline to simulate realistic deployment conditions. Simulation results demonstrate the effectiveness and robustness of the proposed models.
{"title":"Deep Learning Optimization of Two-State Pinching Antennas Systems","authors":"Odysseas G. Karagiannidis;Victoria E. Galanopoulou;Panagiotis D. Diamantoulakis;Zhiguo Ding;Octavia A. Dobre","doi":"10.1109/TCCN.2025.3647822","DOIUrl":"10.1109/TCCN.2025.3647822","url":null,"abstract":"The evolution of wireless communication systems requires flexible, energy-efficient, and cost-effective antenna technologies. Pinching antennas (PAs), which can dynamically control electromagnetic wave propagation through binary activation states, have recently emerged as a promising candidate. In this work, we investigate the problem of optimally selecting a subset of fixed-position PAs to activate in a waveguide, when the aim is to maximize the communication rate at a user terminal. Due to the complex interplay between antenna activation, waveguide-induced phase shifts, and power division, this problem is formulated as a combinatorial fractional 0-1 quadratic program. To efficiently solve this challenging problem, we use neural network architectures of varying complexity to learn activation policies directly from data, leveraging spatial features and signal structure. Furthermore, we incorporate user location uncertainty into our training and evaluation pipeline to simulate realistic deployment conditions. Simulation results demonstrate the effectiveness and robustness of the proposed models.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"4942-4956"},"PeriodicalIF":7.0,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145823389","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-19DOI: 10.1109/TCCN.2025.3646157
Irshad A. Meer;Karl-Ludwig Besser;Mustafa Ozger;Dominic A. Schupke;H. Vincent Poor;Cicek Cavdar
Multi-connectivity involves dynamic cluster formation among distributed access points (APs) and coordinated resource allocation from these APs, highlighting the need for efficient mobility management strategies for users with multi-connectivity. In this paper, we propose a novel mobility management scheme for unmanned aerial vehicles (UAVs) that uses dynamic cluster reconfiguration with energy-efficient power allocation in a wireless interference network. Our objective encompasses meeting stringent reliability demands, minimizing joint power consumption, and reducing the frequency of cluster reconfiguration. To achieve these objectives, we propose a hierarchical multi-agent deep reinforcement learning (H-MADRL) framework, specifically tailored for dynamic clustering and power allocation. The edge cloud connected with a set of APs through low latency optical back-haul links hosts the high-level agent responsible for the optimal clustering policy, while low-level agents reside in the APs and are responsible for the power allocation policy. To further improve the learning efficiency, we propose a novel action-observation transition-driven learning algorithm that allows the low-level agents to use the action space from the high-level agent as part of the local observation space. This allows the lower-level agents to share partial information about the clustering policy and allocate the power more efficiently. The simulation results demonstrate that our proposed distributed algorithm achieves comparable performance to the centralized algorithm. Additionally, it offers better scalability, as the decision time for clustering and power allocation increases by only 10% when doubling the number of APs, compared to a 90% increase observed with the centralized approach.
{"title":"Hierarchical Multi-Agent DRL-Based Dynamic Cluster Reconfiguration for UAV Mobility Management","authors":"Irshad A. Meer;Karl-Ludwig Besser;Mustafa Ozger;Dominic A. Schupke;H. Vincent Poor;Cicek Cavdar","doi":"10.1109/TCCN.2025.3646157","DOIUrl":"10.1109/TCCN.2025.3646157","url":null,"abstract":"Multi-connectivity involves dynamic cluster formation among distributed access points (APs) and coordinated resource allocation from these APs, highlighting the need for efficient mobility management strategies for users with multi-connectivity. In this paper, we propose a novel mobility management scheme for unmanned aerial vehicles (UAVs) that uses dynamic cluster reconfiguration with energy-efficient power allocation in a wireless interference network. Our objective encompasses meeting stringent reliability demands, minimizing joint power consumption, and reducing the frequency of cluster reconfiguration. To achieve these objectives, we propose a hierarchical multi-agent deep reinforcement learning (H-MADRL) framework, specifically tailored for dynamic clustering and power allocation. The edge cloud connected with a set of APs through low latency optical back-haul links hosts the high-level agent responsible for the optimal clustering policy, while low-level agents reside in the APs and are responsible for the power allocation policy. To further improve the learning efficiency, we propose a novel action-observation transition-driven learning algorithm that allows the low-level agents to use the action space from the high-level agent as part of the local observation space. This allows the lower-level agents to share partial information about the clustering policy and allocate the power more efficiently. The simulation results demonstrate that our proposed distributed algorithm achieves comparable performance to the centralized algorithm. Additionally, it offers better scalability, as the decision time for clustering and power allocation increases by only 10% when doubling the number of APs, compared to a 90% increase observed with the centralized approach.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"4957-4971"},"PeriodicalIF":7.0,"publicationDate":"2025-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11305124","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145785589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Satellite-aided low-altitude UAV networks utilize satellites to support low-flying drones with communication and navigation for tasks like sensing, surveillance, and delivery. A critical challenge is offloading UAV communication traffic and migrate services to satellites during network congestion. The decentralized structure, changing link conditions, and limited local visibility make it hard to coordinate service migration and request routing in such dynamic environments. To address these challenges, we propose a semantic graph-based multi-agent reinforcement learning (MARL) framework for satellite-aided UAV networks. We formulate service migration and routing as a semantic graph optimization problem, with the objectives of reducing communication delay and increasing network throughput. The framework incorporates two key components: a cyclic message-passing model that enables nodes to infer global network states from limited local observations, and a discrete denoising diffusion model that generates realistic, and dynamic topologies. Our framework leverages semantic feature extraction to further enhance decision-making in routing and service placement. Extensive simulations show that our approach achieves significant reductions in average transmission delay and improvements in the network throughput.
{"title":"Satellite-Aided Low-Altitude UAV Service Migration With Semantic Extraction and Generated Graphs","authors":"Qiongqiong Jia;Jie Zheng;Ling Gao;Jinping Niu;Rui Cao;Jie Ren","doi":"10.1109/TCCN.2025.3645425","DOIUrl":"10.1109/TCCN.2025.3645425","url":null,"abstract":"Satellite-aided low-altitude UAV networks utilize satellites to support low-flying drones with communication and navigation for tasks like sensing, surveillance, and delivery. A critical challenge is offloading UAV communication traffic and migrate services to satellites during network congestion. The decentralized structure, changing link conditions, and limited local visibility make it hard to coordinate service migration and request routing in such dynamic environments. To address these challenges, we propose a semantic graph-based multi-agent reinforcement learning (MARL) framework for satellite-aided UAV networks. We formulate service migration and routing as a semantic graph optimization problem, with the objectives of reducing communication delay and increasing network throughput. The framework incorporates two key components: a cyclic message-passing model that enables nodes to infer global network states from limited local observations, and a discrete denoising diffusion model that generates realistic, and dynamic topologies. Our framework leverages semantic feature extraction to further enhance decision-making in routing and service placement. Extensive simulations show that our approach achieves significant reductions in average transmission delay and improvements in the network throughput.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"5136-5147"},"PeriodicalIF":7.0,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145778020","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"BRFL: A Blockchain-based and Ring Signature-empowered Privacy-preserving Federated Learning Scheme for Low-altitude Networks","authors":"Linsheng Liu, Jiahui Chen, Haonan Fan, Keyu Xu, Peifeng Zhang, Xiaoguo Li, Tao Xiang","doi":"10.1109/tccn.2025.3645450","DOIUrl":"https://doi.org/10.1109/tccn.2025.3645450","url":null,"abstract":"","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"81 1","pages":""},"PeriodicalIF":8.6,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145777901","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-18DOI: 10.1109/TCCN.2025.3645473
Tan Li;Zhen Li;Hai Liu;Chao Yang;Tse-Tin Chan;Jun Cai
Proactive caching is essential for minimizing latency and improving Quality of Experience (QoE) in heterogeneous edge networks. While Federated Deep Reinforcement Learning (FDRL) shows promise for developing cache policies, it faces challenges such as an expanding action space and difficulty in balancing global knowledge sharing with local environmental adaptation. In this paper, we propose a Layer-wise Relevance Propagation-aided Personalized Federated (LRP-PFed) Deep Reinforcement Learning framework for edge caching to maximize system utility while satisfying caching constraints. To handle the expanding action space, we design a Multi-Head Double Deep Q-Network (MH-DDQN) that reshapes the action output layers into a multi-head structure, where each head generates a sub-dimensional action. Furthermore, we introduce an LRP-based adaptive personalization mechanism that dynamically determines the optimal number of personalized layers for each edge server during training. This approach enables automatic adaptation to heterogeneous environments while leveraging global information to accelerate learning convergence. Extensive experiments validate the effectiveness of our approach, showing that MH-DDQN achieves superior cache hit rates and reduced computational complexity compared to traditional DRL methods, while our LRP-guided personalization strategy achieves superior performance, scalability, and adaptivity compared to existing FDRL methods.
{"title":"Adaptive Layer-Wise Personalized Federated Deep Reinforcement Learning for Heterogeneous Edge Caching","authors":"Tan Li;Zhen Li;Hai Liu;Chao Yang;Tse-Tin Chan;Jun Cai","doi":"10.1109/TCCN.2025.3645473","DOIUrl":"10.1109/TCCN.2025.3645473","url":null,"abstract":"Proactive caching is essential for minimizing latency and improving Quality of Experience (QoE) in heterogeneous edge networks. While Federated Deep Reinforcement Learning (FDRL) shows promise for developing cache policies, it faces challenges such as an expanding action space and difficulty in balancing global knowledge sharing with local environmental adaptation. In this paper, we propose a Layer-wise Relevance Propagation-aided Personalized Federated (LRP-PFed) Deep Reinforcement Learning framework for edge caching to maximize system utility while satisfying caching constraints. To handle the expanding action space, we design a Multi-Head Double Deep Q-Network (MH-DDQN) that reshapes the action output layers into a multi-head structure, where each head generates a sub-dimensional action. Furthermore, we introduce an LRP-based adaptive personalization mechanism that dynamically determines the optimal number of personalized layers for each edge server during training. This approach enables automatic adaptation to heterogeneous environments while leveraging global information to accelerate learning convergence. Extensive experiments validate the effectiveness of our approach, showing that MH-DDQN achieves superior cache hit rates and reduced computational complexity compared to traditional DRL methods, while our LRP-guided personalization strategy achieves superior performance, scalability, and adaptivity compared to existing FDRL methods.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"4532-4546"},"PeriodicalIF":7.0,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145778018","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-17DOI: 10.1109/TCCN.2025.3645498
Huijie Yang;Tao Zhang;Jian Shen;Shan Jin
With the rapid development of Low-Altitude Intelligent Internet of Things (LAI-IoT), uncrewed aerial vehicles (UAVs) have become critical assistants for urban traffic management systems to acquire real-time data. They are widely deployed in key scenarios such as traffic accident scene image collection and dynamic traffic flow monitoring, providing essential data support for traffic scheduling and emergency response. Beyond the risk of raw data leakage, derivative information (including dataset cardinality, retrieval requests, and UAV communication trajectories) may also disclose sensitive situational information. If the urban planning brain repeatedly makes large requests for a specific intersection, the server may infer an upcoming major project, while disclosing the UAV provider’s total data volume reveals its operational scale and costs, disadvantaging it in future negotiations. Existing LAI-IoT communication schemes primarily focus on protecting users’ direct privacy, often overlooking the privacy of derivative data (e.g., dataset cardinality and communication trajectories), which can be exploited and misused by AGI models. Therefore, a privacy-preserving collaborative communication protocol for distributed UAVs based on a hierarchical key encapsulation mechanism is proposed. The protocol enables UAV datasets to accurately identify accident scenes while preserving multiple dimensions of privacy, including dataset cardinality at the UAV side, retrieval tag sets and their cardinalities at base stations, and UAV communication trajectories, without compromising communication efficiency. Theoretical analysis and experimental evaluation demonstrate that the proposed protocol maintains strong scalability and practicality while significantly reducing the success rate of cardinality and trajectory inference attacks, thereby enhancing the security of data communications in LAI-IoT scenarios.
{"title":"Privacy-Preserving Collaborative Communication for Distributed UAVs in Low-Altitude Intelligent IoT Networking","authors":"Huijie Yang;Tao Zhang;Jian Shen;Shan Jin","doi":"10.1109/TCCN.2025.3645498","DOIUrl":"https://doi.org/10.1109/TCCN.2025.3645498","url":null,"abstract":"With the rapid development of Low-Altitude Intelligent Internet of Things (LAI-IoT), uncrewed aerial vehicles (UAVs) have become critical assistants for urban traffic management systems to acquire real-time data. They are widely deployed in key scenarios such as traffic accident scene image collection and dynamic traffic flow monitoring, providing essential data support for traffic scheduling and emergency response. Beyond the risk of raw data leakage, derivative information (including dataset cardinality, retrieval requests, and UAV communication trajectories) may also disclose sensitive situational information. If the urban planning brain repeatedly makes large requests for a specific intersection, the server may infer an upcoming major project, while disclosing the UAV provider’s total data volume reveals its operational scale and costs, disadvantaging it in future negotiations. Existing LAI-IoT communication schemes primarily focus on protecting users’ direct privacy, often overlooking the privacy of derivative data (e.g., dataset cardinality and communication trajectories), which can be exploited and misused by AGI models. Therefore, a privacy-preserving collaborative communication protocol for distributed UAVs based on a hierarchical key encapsulation mechanism is proposed. The protocol enables UAV datasets to accurately identify accident scenes while preserving multiple dimensions of privacy, including dataset cardinality at the UAV side, retrieval tag sets and their cardinalities at base stations, and UAV communication trajectories, without compromising communication efficiency. Theoretical analysis and experimental evaluation demonstrate that the proposed protocol maintains strong scalability and practicality while significantly reducing the success rate of cardinality and trajectory inference attacks, thereby enhancing the security of data communications in LAI-IoT scenarios.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"5192-5205"},"PeriodicalIF":7.0,"publicationDate":"2025-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146026341","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-17DOI: 10.1109/TCCN.2025.3645467
Liang Dong;Jun Huang;Robert W. Heath
We present an attention-based transformer learning approach for dynamic resource allocation in multi-carrier non-orthogonal multiple access (NOMA) downlink systems. We propose transformer architectures for optimizing channel assignment under both time-invariant and time-varying channel conditions, with subsequent power allocation optimization for sum-rate maximization. For time-invariant channels, we employ an encoder-only transformer with multi-head attention mechanisms that processes channel-gain-to-noise ratio matrices to generate optimal channel-assignment matrices. For time-varying channels, we develop a hierarchical temporal correlation transformer that models user-specific temporal patterns before capturing inter-user dependencies. The custom loss functions address channel assignment constraints and temporal stability requirements. Numerical results demonstrate our approach’s superiority over baseline neural networks, achieving 95% accuracy in identifying optimal channel assignments while offering polynomial time complexity compared to the factorial complexity of exhaustive search methods.
{"title":"Transformer-Based Dynamic Resource Allocation for Multi-Carrier NOMA Systems","authors":"Liang Dong;Jun Huang;Robert W. Heath","doi":"10.1109/TCCN.2025.3645467","DOIUrl":"https://doi.org/10.1109/TCCN.2025.3645467","url":null,"abstract":"We present an attention-based transformer learning approach for dynamic resource allocation in multi-carrier non-orthogonal multiple access (NOMA) downlink systems. We propose transformer architectures for optimizing channel assignment under both time-invariant and time-varying channel conditions, with subsequent power allocation optimization for sum-rate maximization. For time-invariant channels, we employ an encoder-only transformer with multi-head attention mechanisms that processes channel-gain-to-noise ratio matrices to generate optimal channel-assignment matrices. For time-varying channels, we develop a hierarchical temporal correlation transformer that models user-specific temporal patterns before capturing inter-user dependencies. The custom loss functions address channel assignment constraints and temporal stability requirements. Numerical results demonstrate our approach’s superiority over baseline neural networks, achieving 95% accuracy in identifying optimal channel assignments while offering polynomial time complexity compared to the factorial complexity of exhaustive search methods.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"4926-4941"},"PeriodicalIF":7.0,"publicationDate":"2025-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145982236","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：