基于条件生成对抗网络的无线传感网络多节点失效修复研究
Research on Multinode Failure Repair of Wireless Sensor Networks Based on Conditional Generative Adversarial Networks
王暾 1赵晓丽 1何苑 1郝梦岩1
作者信息
- 1. 长治学院 计算机系,山西 长治 046000
- 折叠
摘要
当前主流的传感节点失效修复主要通过纠删码完成,修复后节点具有更高的空间利用率,但无法有效提升网络寿命.为此,提出基于条件生成对抗网络的无线传感网络多节点数据重构方法,完成失效修复.感知无线传感网络节点,对失效节点展开裁决,确定失效节点位置,并重构节点内数据;将获取的失效节点用于条件生成对抗网络(CGAN)框架中生成器与节点替换网络的训练,通过训练好的生成器,以失效节点为条件,生成未失效节点;为提升修复性能,使用粒子群算法寻优节点替换网络参数,完成节点重构数据置换,实现失效节点的有效修复.结果表明:利用所提方法进行修复时,能耗最高仅为 17 J,剩余寿命最低可达到 300 h,连通度最高可达到 99.2%,具有较好的修复效果.
Abstract
The current mainstream sensor node failure repair is mainly achieved through erasure codes.After repair,nodes have higher space utilization,but network lifespan cannot be effectively improved.Therefore,a multi node data reconstruction method for wireless sensor networks based on conditional generative adversarial networks is proposed to complete failure repair.Wireless sensor network nodes are precieved,judgments on failed nodes are conducted,the location of failed nodes is determined,and data within the nodes are reconstructed.The obtained failed nodes are used for training generators and node replacement networks in the conditional generative adversarial network(CGAN)framework,and non failed nodes are generated based on the failed nodes through the trained generators.To improve repair performance,particle swarm optimization algorithm is used to optimize node replacement network parameters,complete node reconstruction data replacement,and achieve effective repair of failed nodes.The results show that when using the pro-posed method for repair,the highest energy consumption is only 17 J,the lowest remaining life can reach 300 h,and the highest connec-tivity can reach 99.2%,indicating a good repair effect.
关键词
无线传感网络/失效节点修复/条件生成对抗网络/节点失效判决/节点数据重构Key words
wireless sensor network/repair of failed nodes/conditional generation adversarial network/node failure judgment/node data reconstruction引用本文复制引用
基金项目
山西省高等学校科技创新项目(2020L0623)
山西省高等学校教学改革创新项目(J20231290)
出版年
2024
国家科技期刊平台
NSTL
万方数据