Lower production costs have inspired studies on unmanned aerial vehicles (UAV) for wireless communication. However, limited transmission power and size of the UAV make it challenging to use advanced communication models while meeting the growing need for high data rates and energy efficiency (EE). In this paper, we study an energy-efficient UAV network enhanced by an intelligent reflecting surface (IRS) with simultaneous wireless information and power transfer (SWIPT), where the IRS is employed to improve the EE of ground user equipment (GUE). The goal is to maximize the average EE by jointly controlling the UAV's flying route, IRS phase steer, UAV transmission power, and power splitting (PS) ratio of the energy transfer technology. The formulated problem of maximizing the average EE is non-convex and thus challenging to be solved. To address this problem, we propose a deep reinforcement learning (DRL) approach. The modified reward function is implemented to enhance the efficiency of the DRL agent, which is formulated based on the expected signal-to-interference-plus-noise ratio (SINR) map. Simulation results demonstrate that the proposed DRL algorithm achieves lower energy consumption, higher data rate, and improved EE compared to the comparison algorithm.
Wireless body area networks (WBANs) exhibits immense potential for application in ubiquitous healthcare systems. However, in dense environments where multiple WBANs are deployed such as hospitals, the performance of WBANs is deteriorated by inter-network interference. In this study, we apply Kuramoto model to the transmission power control of users in WBAN environments with the purpose of addressing the issue of inter-network interference, and thereby improving the Quality of Service (QoS) fairness in WBAN environment. For this purpose, we define the utility function which expresses the QoS of a user and synchronize this utility function value of each user for fair resource allocation. The synchronization of the utility function is obtained by Kuramoto model and coupling coefficient defined in Kuramoto model is dynamically adjusted for faster convergence. The range of convergence of the Kuramoto model is derived analytically. From the simulation results, we can verify that the performance of the proposed algorithm is better than those of the existing algorithms with regard to QoS fairness, power, and data rate. In addition, we show that the convergence speed of the proposed algorithm with adaptive coupling strength parameter outperforms the Kuramoto model with fixed coupling strength parameter.
With the increasing computational demands of Internet of Things (IoT) applications, air-ground integrated networks (AGIN), leveraging the capabilities of Unmanned Aerial Vehicles (UAVs) and High-Altitude Platform (HAP), provides an essential solution to these challenges. In this paper, we propose a framework that facilitates local computing at IoT devices and offers the flexibility to offload tasks to aerial platforms when necessary. Specifically, we formulate a multi-objective optimization model aiming at simultaneously minimizing energy consumption and reducing task latency by adjusting control variables such as transmit power, offloading decisions, and UAV placement in a distributed network of IoT devices. Our proposed framework employs Deep Deterministic Policy Gradient (DDPG) techniques to dynamically optimize network operations, allowing for efficient real-time adjustments to network conditions and task demands. The performance of the proposed algorithm is compared to traditional algorithms, including the Whale Optimization Algorithm (WOA), Gradient Search with Barrier, and Bayesian Optimization (BO). Simulation results show that this approach significantly minimizes energy consumption and latency, outperforming conventional optimization methods. Additionally, scalability tests confirm that our framework can efficiently integrate an increasing number of IoT devices and UAVs.
Considering the exponential growth of wireless devices with data-starving applications fused with artificial intelligence, the significance of wireless network scalability using distributed behavior and fairness among users is a crucial feature in guaranteeing reliable service to numerous users in the network environment. The Kuramoto model is described as nonlinear self-sustained phase oscillators spinning at varying intrinsic frequencies connected through the sine of their phase differences and displays a phase transition at a specific coupling strength, in which a mutual behavior is accomplished. In this work, we apply the Kuramoto model to achieve a weighted fair resource allocation in a wireless network, where each user has different quality of service (QoS) requirements. Because the original Kuramoto model is the synchronization model, we propose a new weighting parameter for representing requirement of each node resource and modify the Kuramoto model to achieve weighted fair resource allocation for users with different QoS requirements. The proposed modified Kuramoto model allocates all users the resource based on their weight among contending nodes in a distributed manner. We analyze the convergence condition for the proposed model, and the results reveal that the proposed algorithm achieves a weighted fair resource allocation and with potentially high convergence speed compared to previous algorithm.
13 August 2025 in IEEE Transactions on Vehicular Technology
Space-air-ground integrated networks (SAGIN) have gained significant attention as a key solution to address the challenges of limited spectrum resources and unstable user accessibility caused by the growing number of connected devices. However, the extensive coverage and multi-layered structure of SAGIN make it vulnerable to wiretapping that can compromise data integrity and pose significant challenges in managing energy consumption across all communication layers. To address these security challenges, this study explores an uplink physical layer security (PLS) within the rate-splitting multiple access (RSMA) framework for SAGIN. A joint optimization problem of secrecy rate and energy consumption subject to security and power constraints is formulated. To address this problem, we propose multi-agent deep reinforcement learning based algorithms, multi-agent twin delayed deep deterministic policy gradient (MATD3) and multi-agent deep Q-learning (MADQL). These algorithms aim to maximize a reward function which is designed based on the optimization model's objective, by controlling the transmit power of RSMA sub-signals and user-receiver associations. In addition, we employ fixed and adaptive penalty mechanisms in the proposed MATD3 and MADQL algorithms to avoid excessive restriction on exploration and instability during training. Extensive simulation results demonstrate that the proposed RSMA-MATD3 outperform existing multiple access schemes, including non-orthogonal multiple access (NOMA) and frequency division multiple access (FDMA), by achieving improved secrecy rates while maintaining similar energy consumption across various comparable metrics such as maximum transmit power constraint and number of user equipment (UE). Additionally, it is verified for both MATD3 and MADQL that the proposed adaptive penalty within the reward function enables higher secrecy rates at lower transmit power, while the fixed penalty achieves similar performance at higher transmit power.