Abstract: With the large-scale grid integration of renewable energy sources and the accelerating intelligent transformation of new power systems, the Power Internet of Things (PIoT) has become pivotal for achieving power system intelligence. However, PIoT Devices (PIoTD) in remote areas face numerous challenges, including insufficient network coverage, limited energy harvesting, and poor communication conditions. To address these issues, this study proposes an AI-based ″Cloud—Edge—Device″ cooperative framework that employs multi-Unmanned Aerial Vehicle Simultaneous Wireless Information and Power Transfer (UAV-SWIPT) technology to continuously supply energy to energy-constrained PIoTD. By deploying UAV equipped with SWIPT services in low-altitude networks within Space—Air—Ground (SAG) integrated networks, the framework assists SAG-PIoT devices with energy supplementation and communication relay. Furthermore, to optimize multi-UAV collaboration and enhance data relay, transmit power allocation, Global Energy Efficiency (GEE), and PIoTD association scheduling, a multi-agent Deep Reinforcement Learning (DRL) algorithm is introduced to address the challenges of incomplete global information and high-dimensional variable coupling in dynamic environments. The simulation results demonstrate that the proposed algorithm achieves faster convergence and superior energy efficiency compared to several benchmark algorithms. In terms of max-min transmission rate maximization, the proposed algorithm attains the highest minimum transmission rate, reaching 1.75×10⁵ bit/s. Additionally, the optimal power-splitting ratio for SWIPT is approximately 0.7, at which the GEE is maximized.

Key words: integrated Space—Air—Ground （SAG） network, Power Internet of Things (PIoT), Simultaneous Wireless Information and Power Transfer (SWIPT), Global Energy Efficiency (GEE), reinforcement learning

摘要： 随着可再生能源的规模化并网和新型电力系统智能化转型的加速，电力物联网（PIoT）已成为实现电力系统智能化的关键。然而，偏远地区电力物联网设备（PIoTD）面临着网络覆盖不足、能量收集受限、通信条件差等诸多问题。为解决这些问题，提出一种基于人工智能的“云-边缘-设备”合作框架，采用多无人机无线携能通信（UAV-SWIPT）技术来为能量受限的PIoTD持续提供能量。通过空天地网络中低空组网的UAV搭载SWIPT服务，辅助空天地一体化电力物联网（SAG-PIoT）设备进行能量补充和通信中继。此外，为优化多无人机的协同工作，提升数据中继、发射功率分配、全局能耗效率（GEE）及PIoTD关联调度，引入一种多智能体深度强化学习（DRL）算法，以解决动态环境下全局信息不完整和高维度变量耦合问题。仿真结果表明，相比于其他几种基准算法，所提算法收敛速度更快，能耗效率表现更优。在最大化最小传输速率方面，该算法的最小传输速率最高，达到了1.75×10⁵ bit/s。此外，SWIPT功率分割的最佳比例在0.7左右，此时GEE最高。

关键词: 空天地一体化网络, 电力物联网, 无线携能通信, 全局能耗效率, 强化学习

CLC Number: 

• TP391