Aug 2026· International Journal of Communication Systems· Vol 39· 0 citations· 31 references
TL;DR
This study establishes a viable pathway toward integrating federated learning and green AI principles for the sustainable, intelligent operation of future 6G WSNs and introduces an adaptive client selection mechanism that prioritizes nodes with higher residual energy and better local model quality to participate in each training round, further enhancing sustainability.
Abstract
The advent of 6G wireless sensor networks (WSNs) demands highly efficient, sustainable, and intelligent energy management strategies to support massive device connectivity and data‐driven applications. Traditional centralized learning models are energy‐intensive and pose serious privacy risks, making them unsuitable for resource‐constrained WSN nodes. In this work, we propose a Federated Reinforcement Learning–based Green AI framework (Fed‐RL‐GAI) for sustainable energy optimization in 6G WSNs. Our approach combines federated averaging (FedAvg) with an energy‐aware deep reinforcement learning (DRL) agent at each node to minimize local energy consumption while optimizing global system performance. Specifically, nodes locally train lightweight DRL models to manage their sleep–wake cycles and transmission schedules based on dynamic environmental conditions, and periodically aggregate model updates without sharing raw data. We introduce an adaptive client selection mechanism that prioritizes nodes with higher residual energy and better local model quality to participate in each training round, further enhancing sustainability. Extensive simulations under realistic 6G WSN scenarios demonstrate that Fed‐RL‐GAI reduces energy consumption by up to 35%, extends network lifetime by 42%, and achieves faster convergence compared to traditional FL and centralized DRL baselines. This study establishes a viable pathway toward integrating federated learning and green AI principles for the sustainable, intelligent operation of future 6G WSNs.
An AI-driven framework integrates hybrid energy harvesting mechanisms with Deep Reinforcement Learning (DRL) to optimize energy efficiency in IoT systems and achieves up to 300% improvement in network lifetime under low-energy harvesting conditions.
Elkhatim Abuelysar Elmobarak Mohammed Ali· Islamic University Journal o...· 0 citations
Solar-powered sensor networks are reshaping agriculture by enabling continuous farm management and animal welfare monitoring through Internet-of-Things (IoT) devices, edge intelligence, and cloud analytics. Yet, their sustainability is threatened by two underexplored challenges: vulnerability to cyber-attacks and instability under fluctuating energy supplies. To address these issues, we propose a sustainable smart farm framework that ensures reliable monitoring under adversarial and resource-constrained conditions. Our approach employs deep reinforcement learning (DRL) to derive adaptive policies that jointly optimize monitoring fidelity and energy efficiency. To overcome DRL’s slow convergence and unstable adaptation in dynamic environments, we design a decision theory (DT)-guided transfer learning (TL) framework. Embedding DT principles accelerates policy learning while balancing trade-offs among monitoring quality, energy sustainability, and system resilience. Experimental evaluations on realistic farm scenarios show that DT-guided DRL reduces training runtime by 47.5% while outperforming TL-enhanced DRL in robustness and monitoring performance. This work makes three contributions: (1) unifying cyber resilience and energy adaptability in agricultural monitoring, (2) advancing methodological innovation with DT-guided TL for stable DRL convergence, and (3) providing design insights for sustainable agricultural cyber-physical systems. Collectively, these contributions extend sensor networks in agriculture by emphasizing resilience, sustainability, and intelligent adaptation.
Dian Chen, Zelin Wan, D. Ha et al.· ACM Transactions on Cyber-Ph...· 0 citations
Wireless Sensor Networks (WSNs) play a crucial role in the expanding landscape of the Internet of Things (IoT), yet they continue to face persistent challenges related to energy consumption, computational efficiency, and scalability. Although protocols like the Energy-Efficient Routing Protocol through Hybrid Algorithms (EERHA) have made progress in extending network lifespan using machine learning, they still encounter limitations particularly in managing processing overhead, adapting to changing conditions, and scaling to larger deployments. Unlike existing approaches that merely combine machine learning modules, DEERL-WSN (Distributed Energy-Efficient Reinforcement Learning for Wireless Sensor Networks), introduces a unified hierarchical learning framework where distributed reinforcement learning, lightweight graph neural networks, transfer learning, and federated optimization operate cooperatively. The novelty lies in (i) adaptive reward-driven routing using meta-learned objective weights, (ii) topology-aware clustering through lightweight GNN embeddings with significantly reduced computational complexity, (iii) transfer learning-assisted cluster-head prediction to eliminate repetitive optimization overhead, and (iv) federated deep reinforcement learning enabling scalable learning without centralized processing bottlenecks. These integrated innovations collectively address energy efficiency, scalability, and computational constraints simultaneously, which remain largely unresolved in existing WSN routing frameworks. DEERL-WSN addresses several bottlenecks found in earlier protocols and the simulation results show that DEERL-WSN significantly outperforms EERHA and other state-of-the-art methods.
Maheshkumar Patil, B. J, K. R et al.· 2026 7th International Confe...· 0 citations
A dynamic reward structuring framework within deep reinforcement learning to enable adaptive and balanced routing in IoT-WSNs and achieves significant performance gains, including approximately 30% improvement in energy efficiency, 25% reduction in latency, and 35% increase in network throughput compared with baseline methods.
Suresh Betam, S. Nagendram, Bathula Prasanna Kumar et al.· Scientific Reports· 0 citations
