基于无线携能传输和多级边缘卸载的空地协作巡检算法  

Air-ground Cooperative Inspection Algorithm Based on Wireless Power Transfer and Multi-level Edge Offloading

作　　者：陈智雄[1,2] 杨家伟 肖楠 田新成 CHEN Zhixiong;YANG Jiawei;XIAO Nan;TIAN Xincheng(Department of Electronics and Communication Engineering,North China Electric Power University,Baoding 071003,Hebei Province,China;Hebei Provincial Key Laboratory of Electric Power Internet of Things Technology(North China Electric Power University),Baoding 071003,Hebei Province,China;Tangshan Power Supply Company of State Grid Jibei Electric Power Co.,Ltd.,Tangshan 063000,Hebei Province,China)

机构地区：[1]华北电力大学电子与通信工程系,河北省保定市071003 [2]河北省电力物联网技术重点实验室(华北电力大学),河北省保定市071003 [3]国网冀北电力有限公司唐山供电公司,河北省唐山市063000

出　　处：《电网技术》2022年第10期3961-3969,共9页Power System Technology

基　　金：国家自然科学基金青年基金资助项目(61601182);中央高校基本科研业务费(2021MS070)。

摘　　要：变电站采用智能机器人和无人机可实现高效、自动设备巡检。地面机器人在地上和室内近距离巡检方面具有优势;无人机更加灵活,巡检范围和效率更大,但是易受供能等限制。为了充分发挥空地联合巡检的优势,文章提出一种基于无线携能传输和多级边缘卸载的地面机器人和无人机协作巡检算法。首先针对典型变电站场景,给出各级设备在本地计算和卸载时的能耗、速率和时延计算方法,并建立无线携能传输和无人机中继条件下的多级任务卸载模型。接着兼顾时延和能耗要求,将最优化巡检问题描述为马尔科夫决策过程,提出一种基于Q-Learning的最佳任务卸载算法。仿真对比验证了论文算法的有效性与可靠性,通过灵活的卸载算法可实现系统综合性能最大化。The usage of intelligent robots and drones could make equipment inspections more efficient and automatic.Ground robots have several advantages in close inspections on the ground and indoors.UAVs are flexible,with greater inspection range and efficiency,but there are restrictions such as energy supply.In order to give full play to superiority of air-ground joint inspection,this paper proposed a ground-based robot and UAV cooperative inspection algorithm,which based on wireless power transfer and multi-level edge unloading.Firstly,in connection with the typical substation scenario,calculation methods of energy consumption,rate and time delay of all levels of equipment in local calculation and unloading were proposed,and a multi-level task unloading model under the conditions of wireless power transfer and UAV relay was established.Then,this study taking the requirements of time delay and energy consumption into account,the optimization inspection problem was described as a Markov decision process,and an optimal task offloading algorithm based on Q-Learning was proposed.The simulation comparison verified the effectiveness and reliability of this algorithm,and the flexible unloading algorithm could maximize the overall performance of the system.

关 键 词：智能巡检 移动边缘计算 无线携能传输 Q-LEARNING 资源优化 

分 类 号：TM721[电气工程—电力系统及自动化]