基于SVM的飞行自组网分类切换MAC协议

SVM-based Classification Switching MAC Protocol for Flying Ad Hoc Network

飞行自组织网络(Flying Ad Hoc Network,FANET)具有节点高速移动、拓扑频繁变化的特点,单一媒体接入控制(Media Access Control,MAC)协议难以满足FANET的多样化业务需求,针对该问题,提出一种基于支持向量机(Support Vector Machine,SVM)的FANET分类切换MAC(SDS-MAC)协议。通过融合吞吐量、延迟和丢包率进而提出通信质量因子作为数据集标签的标定准则,解决了指标类型、量纲不一致的问题,以网络服务质量(Quality of Service,QoS)整体最优为目标构建基于SVM的MAC分类切换器。SDS-MAC考虑使用2种备选MAC协议联合控制,在动态变化的网络环境中,使全网无人机节点实时选择当下QoS整体更优的备选MAC协议。仿真实验表明,SDS-MAC与ML-MAC相比较,最佳协议运行时间比提高了2.3%,平均吞吐量提高了1.4%,平均端到端时延降低了24.9%,平均丢包率降低了15.2%。

Flying Ad Hoc Network(FANET)has the characteristics of high-speed node movement and frequent topology changes,and it’s difficult for a single Media Access Control(MAC)protocol to meet the diversified business needs of FANET.To deal with this problem,an SVM-based classification switching MAC(SDS-MAC)protocol for FANET is proposed.By fusing throughput,latency,and packet loss rate,the communication quality factor is proposed as the calibration basis for data set labels,which solves the problem of inconsistency between index types and dimensions,and an SVM-based MAC classification switcher is constructed with the goal of overall network Quality of Service(QoS)optimization.SDS-MAC considers the use of two alternative MAC protocols for joint control,so that the UAV nodes of the whole network can select the alternative MAC protocol with better QoS in real time in the dynamically changing network environment.Simulation experiments show that compared with ML-MAC,SDS-MAC improves the optimal protocol runtime ratio by 2.3%,the average throughput by 1.4%,the average end-to-end delay by 24.9%,and the average packet loss rate by 15.2%.