基于公交车辅助的城市车载自组织网络数据分发策略  

Data Distribution Strategy of Urban Vehicle Ad hoc Network Based on Bus Assistance

作　　者：李杏姣 赵毅峰 符琦[2] 蒋云霞[2] LI Xingjiao;ZHAO Yifeng;FU Qi;JIANG Yunxia(Center of Teacher Teaching Development,Guangdong University of Education,Guangzhou 510303,China;School of Computer Science and Engineering,Hunan University of Science and Technology,Xiangtan 411201,China)

机构地区：[1]广东第二师范学院教师教学发展中心,广东广州510303 [2]湖南科技大学计算机科学与工程学院,湖南湘潭411201

出　　处：《湖南科技大学学报（自然科学版）》2021年第3期82-87,共6页Journal of Hunan University of Science And Technology：Natural Science Edition

基　　金：湖南省教育厅科技重点项目资助(19A174)。

摘　　要：针对城市车载自组织网络传统路由协议网络开销增长过快的问题,提出了一种基于深度学习的改进数据分发策略.该策略以DSDV路由协议为研究基础,引入具有固定行驶路线的公交车为辅助节点进行数据快速分发策略分析,同时通过梯度下降法对基于分簇的公交车辅助数据分发评价指标进行训练,以获取局部最优的数据分发路径选择策略,从而达到提升数据分发速度,减小网络开销的目的.实验数据表明:基于公交车辅助和节点分簇的DSDV数据分发策略数据路由开销和端到端平均时延等方面均比传统的DSDV协议具有更好的性能,以及道路环境的适应性.Aiming at the problem that the network overhead of traditional routing protocols in urban Vehicle Ad hoc Network(VANET) is growing too fast, an improved data distribution strategy based on deep learning was proposed. Based on the research of DSDV routing protocol, some bus nodes were introduced as the secondary nodes in the strategy, for fast data distribution policy analysis. At the same time, the gradient descent method was used for bus auxiliary data distribution based on clustering evaluation index for training, so as to obtain the local optimal data distribution path selection strategy, to promote the speed of data distribution and to reduce the network overhead. The experimental data showed that the DSDV data distribution strategy based on bus-assisted and node-clustered has better performance than the traditional DSDV protocol in terms of data routing overhead and end-to-end average delay, as well as adaptability to road environment.

关 键 词：车载自组织网络 路由协议 梯度下降法 DSDV 

分 类 号：TN913[电子电信—通信与信息系统]