Cross-Domain Lifelong Reinforcement Learning for Wireless Sensor Networks

Wireless sensor networks (WSNs) with energy harvesting (EH) are expected to play a vital role in intelligent 6G systems, especially in industrial sensing and control, where continuous operation and sustainable energy use are critical. Given limited energy resources, WSNs must operate efficiently to ensure long-term performance. Their deployment, however, is challenged by dynamic environments where EH conditions, network scale, and traffic rates change over time. In this work, we address system dynamics that yield different learning tasks, where decision variables remain fixed but strategies vary, as well as learning domains, where both decision space and strategies evolve. To handle such scenarios, we propose a cross-domain lifelong reinforcement learning (CD-L2RL) framework for energy-efficient WSN design. Our CD-L2RL algorithm leverages prior experience to accelerate adaptation across tasks and domains. Unlike conventional approaches based on Markov decision processes or Lyapunov optimization, which assume relatively stable environments, our solution achieves rapid policy adaptation by reusing knowledge from past tasks and domains to ensure continuous operations. We validate the approach through extensive simulations under diverse conditions. Results show that our method improves adaptation speed by up to 35% over standard reinforcement learning and up to 70% over Lyapunov-based optimization, while also increasing total harvested energy. These findings highlight the strong potential of CD-L2RL for deployment in dynamic 6G WSNs.

UAV-Enabled IoT Networks: A SWIPT Energy Harvesting Architecture with Relay Support for Disaster Response

Due to the wide application of unmanned aerial vehicles (UAVs) as relays to establish Disaster Response Networks (DRNs), an effective model of energy harvesting (EH) and energy consumption for the UAV-aided Disaster Response Network (DRN) is rising to be a challenging issue. This is mainly manifest in Internet of Things (IoT) scenarios where multiple users are looking to communicate with the UAV. In this paper, the possibility of connecting an UAV with several users is investigated where the UAV as a relay receives data from a DRN and delivers to another network considering two IoT scenarios. The first scenario represents a conventional method with limited UAV energy where low communication rates and inadequate service coverage for all users are challenges. But in the second scenario, a Simultaneous Wireless Information and Power Transmission (SWIPT) technique is used to serve users. Considering potential limitations in transmission energy of users within disaster networks, the SWIPT technique is applied to maximize energy acquisition by the UAV, leading to improve the efficiency of the investigated scenario. Finally, the required energy of the UAV to serve the largest number of users in the shortest possible time is clarified. Furthermore, by Considering the relationship between energy and UAV flight time and defining the UAV flight time optimization problem, optimal network parameters are obtained. Simulation results show the effectiveness of the proposed scenario.

A Novel Energy-aware Design for a D2D Underlaid UAV-Aided Cognitive Radio Network

This paper investigates the effectiveness of a free disaster response network (free-DRN) and an energy harvesting enabled DRN (EH-enabled DRN) enhanced by cognitive radio (CR) technology and unmanned aerial vehicles (UAVs). In the EH-enabled DRN scenario, where device-to-device (D2D) communication harvests energy from cellular users, significant improvements are demonstrated in terms of energy efficiency (EE) and communication rate compared to the free-DRN approach. The impact of user density and UAV specifications on network performance is analyzed, addressing the challenge of optimizing the duration of energy harvesting for both cellular users and D2D devices in the EH-enabled DRN scenario. Additionally, simulation results reveal an optimal UAV height that ensures efficient network operation for varying densities of D2D devices. Overall, numerical and simulation findings highlight the superior performance of the EH-enabled DRN approach, showcasing the positive effects of enabling D2D links and improved EE. Notably, reducing energy harvesting duration and cellular user density can further enhance the EE of the DRN by up to 3dB.

Delay Optimization of a Federated Learning-based UAV-aided IoT network

This paper explores the integration of power splitting(PS) simultaneous wireless information and power transfer (SWIPT) architecture and federated learning (FL) in Internet of Things (IoT) networks. The use of SWIPT allows power-constrained devices to simultaneously harvest energy and transmit data, addressing the energy limitations faced by IoT devices. The proposed scenario involves an Unmanned Arial Vehicle (UAV) serving as the base station (BS) and edge server, aggregating weight updates from IoT devices and unicasting aggregated updates to each device. The results demonstrate the feasibility of FL in IoT scenarios, ensuring communication efficiency without depleting device batteries.