卫星物联网中基于能量感知的自适应节能路由策略  

Adaptive energy-efficient routing strategy based on energy aware in satellite Internet of Things

作　　者：杨桂松[1] 陶挺 何杏宇 杜平 YANG Guisong;TAO Ting;HE Xingyu;DU Ping(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China;College of Communication and Art Design,University of Shanghai for Science and Technology,Shanghai 200093,China;Shanghai Institute of Satellite Internet Engineering Co.,Ltd./Shanghai Satellite Network Research Institute,Shanghai 201210,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093 [2]上海理工大学出版印刷与艺术设计学院,上海200093 [3]上海卫星互联网研究院有限公司/上海市卫星互联网重点实验室,上海201210

出　　处：《智能计算机与应用》2024年第6期127-133,共7页Intelligent Computer and Applications

基　　金：南通市科技局社会民生计划项目(MS12021060);敏捷智能计算四川省重点实验室开放式基金资助项目;浦东新区科技发展基金产学研专项(PKX2021-D10)。

摘　　要：随着5G技术的发展和6G技术的研究,低轨卫星网络在卫星物联网中的地位越发重要,而作为网络核心技术的路由策略仍面临一些挑战。本文针对低轨卫星网络拓扑结构动态变化、链路的不连续性、卫星能量供应受限等问题,为了及时感知星间链路状态和卫星的能量并选择正确的路由,将卫星物联网中的路由选择问题转化为马尔可夫决策过程下的最优策略问题,提出一种基于Dueling DQN(Dueling Deep Q Network)的自适应节能路由算法。该算法通过改进DQN中神经网络的架构,大幅度地提升了学习的效果。仿真结果表明,与传统的DQN算法相比,该算法能有效降低系统能耗,均衡网络负载,提高网络吞吐量。With the development of 5G technology and the research of 6G technology,LEO satellite networks have become increasingly important in the satellite Internet of Things.However,as the core technology of the network,routing strategies still face some challenges.In this paper,an intelligent sensing routing algorithm based on Dueling DQN(Dueling Deep Q Network)is proposed to address issues such as dynamic changes in the topology of low earth orbit satellite networks,link discontinuity,and limited satellite energy supply,which could sense the status of the inter-satellite link and the energy of the satellite in time and select the correct route,and the routing problem in the satellite Internet of Things is transformed into the optimal strategy problem in the Markov decision-making process.This algorithm greatly improves the learning effect by improving the architecture of neural networks in DQN.Simulation results show that compared with traditional DQN algorithm,this algorithm can effectively reduce system energy consumption,balance network load,and improve network throughput.

关 键 词：卫星物联网 路由算法 能量感知 深度强化学习 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]