基于小样本无梯度学习的无线传感器网络分簇路由方法  

Clustering Routing Method for Wireless Sensor Networks Based on Small⁃Sample Gradient⁃Free Learning

作　　者：汪亮[1] 陈燕红 WANG Liang;CHEN Yanhong(School of Electrical Engineering,Hunan Mechanical&Electrical Polytechnic,Changsha Hu’nan 410151,China)

机构地区：[1]湖南机电职业技术学院电气工程学院,湖南长沙410151

出　　处：《传感技术学报》2023年第8期1303-1309,共7页Chinese Journal of Sensors and Actuators

基　　金：湖南省自然科学基金项目(2017JJ5028)。

摘　　要：针对无线传感器网络由于某些样本数量较少、网络节点间剩余能量差高,导致路由分簇困难、分类准确率低的问题,提出基于小样本无梯度学习的无线传感器网络分簇路由方法。采用条件生成对抗网络处理小样本数据,在有限样本中获取更丰富的信息,通过LEACH协议算法划分无线传感器网络中各传感器节点为簇,通过无梯度学习的GABP算法优化簇首节点和簇首数量,构建无线传感器网络分簇路由方法。实验结果表明,所提方法的簇首数量均分布在4个~7个处,最高网络节点间剩余能量差为0.0158,网络寿命达到241轮,因此,所提方法能够选择更理想的簇首数量、增强网络能耗均衡性、延长网络寿命。Aiming at the problems of difficult routing clustering and low classification accuracy in wireless sensor networks due to small number of samples and high residual energy difference between network nodes,a clustering routing method for wireless sensor networks based on small⁃sample gradient⁃free learning is proposed.Conditional generative adversarial network is used to process small⁃sample da⁃ta.The sensor nodes in the wireless sensor network are divided into clusters by using the LEACH protocol algorithm,the cluster head nodes and the number of cluster heads are optimized by using the GABP algorithm without gradient learning,and the cluster routing method of the wireless sensor network is constructed.The experimental results show that the number of cluster heads in the proposed method is distributed in 4-7 places,the maximum residual energy difference between network nodes is 0.0158,and the network life reaches 241 rounds.Therefore,the proposed method can select a more ideal number of cluster heads,enhance the balance of network energy consumption,and prolong the network life.

关 键 词：无线传感器 分簇路由 小样本无梯度学习 条件生成对抗网络 LEACH协议 簇首 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]