{"title":"Edge General Intelligence Through World Models, Large Language Models, and Agentic AI: Fundamentals, Solutions, and Challenges","authors":"Changyuan Zhao, Guangyuan Liu, Ruichen Zhang, Yinqiu Liu, Jiacheng Wang, Jiawen Kang, Dusit Niyato, Zan Li, Xuemin Shen, Zhu Han, Sumei Sun, Chau Yuen, Dong In Kim","doi":"10.1109/tccn.2026.3658762","DOIUrl":"https://doi.org/10.1109/tccn.2026.3658762","url":null,"abstract":"","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"42 1","pages":""},"PeriodicalIF":8.6,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146070222","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 : 2026-01-28DOI: 10.1109/TCCN.2026.3658760
Kai Zhong;Jinfeng Hu;Dongxu An;Huan Wan;Yiran Zhang;Xin Tai;Yongfeng Zuo;Ye Yuan;Cunhua Pan;Kah Chan Teh;Xianxiang Yu;Huiyong Li;Guolong Cui
Unimodular waveform design is a key technology in cognitive Multiple Input Multiple Output (MIMO) radar systems. Existing research mainly includes two categories: unimodular continuous/binary waveform design for detection, and unimodular continuous waveform design for Direction of Arrival (DOA) estimation without spectral constraints. Different from existing methods, our focus lies in investigating unimodular binary waveform design for DOA estimation within spectrally crowded environments. This problem is formulated as minimizing the mean square error (MSE) for DOA estimation, subject to the constraints of binary waveform and multiple spectral constraints. Due to the spectral constraint and nonconvex nature of the binary waveform constraint, the problem is NP-hard and challenging to solve directly. Fortunately, we observe that the problem can be decomposed into multiple more tractable subproblems by introducing auxiliary variables. Leveraging this characteristic, we propose an efficient Problem Decomposition-based Sequential Optimization (PDSO) method to tackle this problem. The method introduces two auxiliary variables to decompose the problem into two subproblems: one of which can be solved in closed-form, while the other is efficiently addressed by the Binary Alternating Directions Method of Multipliers (B-ADMM) algorithm. Compared to the existing methods, the proposed approach demonstrates superior performance in terms of computational cost, DOA resolution, and suppression of spectral interference.
{"title":"Binary Waveform Design for Spectrally-Compatible Cognitive MIMO Radar DOA Estimation","authors":"Kai Zhong;Jinfeng Hu;Dongxu An;Huan Wan;Yiran Zhang;Xin Tai;Yongfeng Zuo;Ye Yuan;Cunhua Pan;Kah Chan Teh;Xianxiang Yu;Huiyong Li;Guolong Cui","doi":"10.1109/TCCN.2026.3658760","DOIUrl":"10.1109/TCCN.2026.3658760","url":null,"abstract":"Unimodular waveform design is a key technology in cognitive Multiple Input Multiple Output (MIMO) radar systems. Existing research mainly includes two categories: unimodular continuous/binary waveform design for detection, and unimodular continuous waveform design for Direction of Arrival (DOA) estimation without spectral constraints. Different from existing methods, our focus lies in investigating unimodular binary waveform design for DOA estimation within spectrally crowded environments. This problem is formulated as minimizing the mean square error (MSE) for DOA estimation, subject to the constraints of binary waveform and multiple spectral constraints. Due to the spectral constraint and nonconvex nature of the binary waveform constraint, the problem is NP-hard and challenging to solve directly. Fortunately, we observe that the problem can be decomposed into multiple more tractable subproblems by introducing auxiliary variables. Leveraging this characteristic, we propose an efficient Problem Decomposition-based Sequential Optimization (PDSO) method to tackle this problem. The method introduces two auxiliary variables to decompose the problem into two subproblems: one of which can be solved in closed-form, while the other is efficiently addressed by the Binary Alternating Directions Method of Multipliers (B-ADMM) algorithm. Compared to the existing methods, the proposed approach demonstrates superior performance in terms of computational cost, DOA resolution, and suppression of spectral interference.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"5937-5952"},"PeriodicalIF":7.0,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146070256","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 : 2026-01-28DOI: 10.1109/TCCN.2026.3658765
Fan Wu;Yuxin Zhang;Gaolei Li;Jianhua Li;Jin Ma
In low-altitude networks, reliable and lightweight flight interconnection among AAVs mainly depends on protocols like Message Queuing Telemetry Transport (MQTT). However, existing protocol analysis methods often fail to adapt to unknown program functions caused by version updates due to a lack of protocol knowledge blueprints (PKB). Moreover, due to the lack of multiple rounds of verification in terms of grammar semantics and execution logic representation, the correctness, validity and credibility of the generated PKB in practical scenarios remain questionable. To address these challenges, this paper proposes a novel trustworthy protocol knowledge blueprints generation (TrustPKBG) framework for low-altitude AAV networks by large language model (LLM). In TrustPKBG, a Hierarchical Chain-of-Thought (HCoT) strategy is designated to automatically transform technical specifications into a structured, objective, and consistent PKB draft against the bias of human construction. Moreover, to enhance the abstract semantic similarity between PKB drafts and RFC documents, a memory-aware multi-agent collaboration mechanism is also presented, which enables closed-loop error detection and knowledge updating. To demonstrate the superiority of TrustPKBG, we also implement a Fuzzing-based PKB Verifier over MQTT/CoAP. Experimental results demonstrate that fuzzing with TrustPKBG can achieve an average code coverage rate of 33.71% for MQTT and 33.5% for CoAP, which is 50.83% higher than the baseline average coverage rate of 22.35% for MQTT and 27.6% for CoAP, with a valid test ratio of 74.57% for MQTT and 62.3% for CoAP. Furthermore, we explore PKB’s application in protocol code auditing by evaluating open-source MQTT and CoAP implementations for compliance verification, which demonstrate the effectiveness of the proposed methods.
{"title":"TrustPKBG: Trustworthy Protocol Knowledge Blueprints Generation via LLM for Low-Altitude AAV Networks","authors":"Fan Wu;Yuxin Zhang;Gaolei Li;Jianhua Li;Jin Ma","doi":"10.1109/TCCN.2026.3658765","DOIUrl":"10.1109/TCCN.2026.3658765","url":null,"abstract":"In low-altitude networks, reliable and lightweight flight interconnection among AAVs mainly depends on protocols like Message Queuing Telemetry Transport (MQTT). However, existing protocol analysis methods often fail to adapt to unknown program functions caused by version updates due to a lack of protocol knowledge blueprints (PKB). Moreover, due to the lack of multiple rounds of verification in terms of grammar semantics and execution logic representation, the correctness, validity and credibility of the generated PKB in practical scenarios remain questionable. To address these challenges, this paper proposes a novel trustworthy protocol knowledge blueprints generation (TrustPKBG) framework for low-altitude AAV networks by large language model (LLM). In TrustPKBG, a Hierarchical Chain-of-Thought (HCoT) strategy is designated to automatically transform technical specifications into a structured, objective, and consistent PKB draft against the bias of human construction. Moreover, to enhance the abstract semantic similarity between PKB drafts and RFC documents, a memory-aware multi-agent collaboration mechanism is also presented, which enables closed-loop error detection and knowledge updating. To demonstrate the superiority of TrustPKBG, we also implement a Fuzzing-based PKB Verifier over MQTT/CoAP. Experimental results demonstrate that fuzzing with TrustPKBG can achieve an average code coverage rate of 33.71% for MQTT and 33.5% for CoAP, which is 50.83% higher than the baseline average coverage rate of 22.35% for MQTT and 27.6% for CoAP, with a valid test ratio of 74.57% for MQTT and 62.3% for CoAP. Furthermore, we explore PKB’s application in protocol code auditing by evaluating open-source MQTT and CoAP implementations for compliance verification, which demonstrate the effectiveness of the proposed methods.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"6152-6174"},"PeriodicalIF":7.0,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146070257","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":"Multi-Objective Reinforcement Learning Based Dependent Task Scheduling with Service Caching in Mobile Edge Computing","authors":"Fuhong Song, Mingsen Deng, Huanlai Xing, Yanping Liu, Zhiwen Xiao, Lexi Xu, Xianfu Lei","doi":"10.1109/tccn.2026.3657056","DOIUrl":"https://doi.org/10.1109/tccn.2026.3657056","url":null,"abstract":"","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"58 1","pages":""},"PeriodicalIF":8.6,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146042780","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}
The increasing scarcity of spectrum resources, coupled with rising demand, has made effective spectrum management crucial. However, the complexity and spatio-temporal variability of spectral data present significant challenges for accurate spectrum prediction. This paper proposes a novel multi-band spectrum prediction model that integrates a hypergraph convolutional neural network (HGCN) with a simplified rectified linear unit-gated recurrent unit (ReLU-GRU) network which eliminate the reset gate. In this framework, the HGCN employs hypergraphs to represent spectral data, where nodes correspond to individual frequency bands and hyperedges capture multivariate relationships among them. The simplified ReLU-GRU is used to model the temporal dependencies between frequency bands, effectively fusing the extracted features for enhanced prediction performance. By replacing the traditional hyperbolic tangent (tanh) activation function with a linear rectification function (ReLU) in the state update process, the model mitigates the issue of gradient vanishing and accelerates the training process. To further improve convergence, an attention mechanism is incorporated to weight the output of hidden states. Experimental evaluation on a real-world spectral dataset from sensors in St. Gallen demonstrates that the proposed model achieves a 4.43% improvement in prediction accuracy compared to the traditional LSTM model and a 0.56% improvement over the GCN-GRU model, exhibiting superior stability. The results also show that the simplified ReLU-GRU is particularly effective in predicting highly variable data, outperforming the traditional tanh-GRU, especially in scenarios with significant fluctuations.
{"title":"Multi-Band Spectrum Prediction Algorithm Based on HGCN and Simplified ReLU-GRU","authors":"Lingzhao Zhang;Qin Wang;Haotian Chang;Haitao Zhao;Hongbo Zhu","doi":"10.1109/TCCN.2026.3657092","DOIUrl":"10.1109/TCCN.2026.3657092","url":null,"abstract":"The increasing scarcity of spectrum resources, coupled with rising demand, has made effective spectrum management crucial. However, the complexity and spatio-temporal variability of spectral data present significant challenges for accurate spectrum prediction. This paper proposes a novel multi-band spectrum prediction model that integrates a hypergraph convolutional neural network (HGCN) with a simplified rectified linear unit-gated recurrent unit (ReLU-GRU) network which eliminate the reset gate. In this framework, the HGCN employs hypergraphs to represent spectral data, where nodes correspond to individual frequency bands and hyperedges capture multivariate relationships among them. The simplified ReLU-GRU is used to model the temporal dependencies between frequency bands, effectively fusing the extracted features for enhanced prediction performance. By replacing the traditional hyperbolic tangent (tanh) activation function with a linear rectification function (ReLU) in the state update process, the model mitigates the issue of gradient vanishing and accelerates the training process. To further improve convergence, an attention mechanism is incorporated to weight the output of hidden states. Experimental evaluation on a real-world spectral dataset from sensors in St. Gallen demonstrates that the proposed model achieves a 4.43% improvement in prediction accuracy compared to the traditional LSTM model and a 0.56% improvement over the GCN-GRU model, exhibiting superior stability. The results also show that the simplified ReLU-GRU is particularly effective in predicting highly variable data, outperforming the traditional tanh-GRU, especially in scenarios with significant fluctuations.","PeriodicalId":13069,"journal":{"name":"IEEE Transactions on Cognitive Communications and Networking","volume":"12 ","pages":"5903-5912"},"PeriodicalIF":7.0,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146042796","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：