[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-2764":3},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":9,"source_name":10,"source_url":6,"published_at":11,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":16,"sources":23,"tags":25,"view_count":31,"doi":32,"paper":33,"created_at":52},2764,"An energy-aware federated intelligence framework for sustainable WSN-IoT ecosystems","https:\u002F\u002Fdoi.org\u002F10.1007\u002Fs12083-026-02291-x","With the rapid development of Internet of Things (IoT), the deployment of Wireless Sensor Networks (WSNs) as vital infrastructure for environmental monitoring, industrial automation, precision agriculture, and smart city applications has increased significantly. Despite this, WSN-IoT ecosystems are still hindered by the issues of limited energy, communication overhead, scalability, preserving privacy, and learning in an intelligent decentralized fashion. To overcome these challenges, this study suggests a federated energy-aware intelligence framework namely EcoSense-AI for sustainable WSN-IoT ecosystems. It combines Energy-Weighted Federated Learning, Adaptive Differential Privacy, Energy-Aware Graph Neural Network (EA-GNN)-based topology modeling, and multi-objective resource optimization, to facilitate secure, efficient, and adaptive edge-cloud collaborative intelligence. The effectiveness of EcoSense-AI was tested with a simulated WSN-IoT network of heterogeneous sensor nodes operating with different residual energy levels, communication ranges and moving network conditions. Distributed environmental sensing data in the device, edge and cloud layer were subjected to experimental analysis. The proposed framework was compared with FedAvg and FedProx for equal training rounds, communication setup and in energy constrained deployment settings. The accuracy, energy efficiency, privacy score, communication cost reduction and network lifetime extension were used as evaluation metrics to measure performance. The experimental results show that the proposed method, EcoSense-AI outperforms the baseline methods with 98.56% accuracy, 93.2% energy efficiency, and 0.97 privacy score, and increased network lifetime by 58.9% and decreased communication cost. The proposed framework for sustainable intelligent sensing and decentralized learning in next-generation WSN-IoT applications is proved to be effective through the results.","随着物联网（IoT）的快速发展，无线传感器网络（WSN）作为环境监测、工业自动化、精准农业和智慧城市应用的重要基础设施，其部署规模显著增长。尽管如此，WSN-IoT生态系统仍然受到能量受限、通信开销、可扩展性、隐私保护以及智能去中心化学习等问题的制约。为克服这些挑战，本研究提出了一种面向可持续WSN-IoT生态系统的联邦能量感知智能框架，即EcoSense-AI。该框架融合了能量加权联邦学习、自适应差分隐私、基于能量感知图神经网络（EA-GNN）的拓扑建模以及多目标资源优化，以实现安全、高效且自适应的边云协同智能。EcoSense-AI的有效性通过一个模拟WSN-IoT网络进行了验证，该网络由具有不同剩余能量水平、通信范围和动态网络条件的异构传感器节点组成。对设备层、边缘层和云层中的分布式环境感知数据进行了实验分析。在相同训练轮次、通信设置和能量受限部署条件下，将所提框架与FedAvg和FedProx进行了对比。采用准确率、能量效率、隐私评分、通信成本降低和网络寿命延长作为评价指标来衡量性能。实验结果表明，所提方法EcoSense-AI优于基线方法，准确率达到98.56%，能量效率为93.2%，隐私评分为0.97，网络寿命延长了58.9%，并降低了通信成本。实验结果证明了该框架在下一代WSN-IoT应用中实现可持续智能感知和去中心化学习的有效性。",null,"Peer-to-Peer Networking and Applications","2026-09-17T00:00:00Z","论文",10,false,78,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":13,"relevant":21,"comment":22},16,21,18,13,1,"提出面向WSN-IoT的联邦能效智能框架EcoSense-AI，在精度、能效、隐私与网络寿命上均有量化提升，对农业环境监测与精准农业的可持续感知具有参考价值。",[24],{"name":10,"url":6},[26,27,28,29,30],"智慧农业","农业物联网","边缘计算","联邦学习","数据隐私",0,"10.1007\u002Fs12083-026-02291-x",{"doi":32,"openalex_id":34,"authors":35,"venue":10,"cited_by_count":31,"oa_url":6,"card":45,"direction":49,"ingested_from":51},"W7169722039",[36,38,41,43],{"name":37,"orcid":9},"G Elumalai",{"name":39,"orcid":40},"Dhanalakshmi Gopal","https:\u002F\u002Forcid.org\u002F0009-0000-4123-176X",{"name":42,"orcid":9},"Jayaprakash Chinnadurai",{"name":44,"orcid":9},"V.V. Teresa",{"tldr":46,"method":47,"finding":48,"direction":49,"opportunity":50},"提出EcoSense-AI联邦智能框架，解决WSN-IoT能耗、隐私与通信瓶颈。","能量加权联邦学习、自适应差分隐私、能量感知GNN与多目标优化仿真。","准确率98.56%、能效93.2%、隐私0.97，网络寿命延长58.9%。","智慧农业 \u002F 农业物联网","可将该框架迁移至农田异构传感网，验证真实农业场景下的能效与隐私权衡。","openalex","2026-09-17T23:30:10.020578Z"]