一种面向分簇实时监测应用的WSN时间同步算法

A WSN Time Synchronization Algorithm Oriented Clustering Real-time Monitoring Application

为了建立一种适用于分簇实时监测无线传感器网络的时间同步算法,同时建立具有较高精度的时钟偏移补偿模型,首先在分析分簇无线传感器网络特性的基础上,提出了一种基于分簇网络路由协议的跨层式同步拓扑构建方法;通过利用簇头路由信息和由邻居握手协议建立的簇内节点间的邻接关系,分别建立基于簇间双向组播和簇内双向广播的同步机制;最后结合参数估计理论构建基于线性模型的非簇头节点时钟偏移估计方法以及基于最大似然估计的簇头节点时钟相位补偿方法。理论及实验结果表明:该算法不仅保证了同步拓扑的有效性,而且使得同步开销较HRTS算法降低33%,较TPSN算法降低88%;单跳平均误差与TPSN算法相差仅6.36μs,较RBS算法提高12.87μs。

In order to establish a WSN time synchronization algorithm oriented clustering real-time monitoring application,and build a more precise model of clock offset compensation,based on the analysis of the characteristic of clustering WSN,a cross-layer synchronization topology construction method with the foundation of routing protocol of clustering network was proposed.Then the synchronization mechanism based on Inter-Cluster Bidirectional Multicast Protocol and Cluster Bidirectional Broadcast Protocol was constructed by using of cluster-head information and adjacent relation of non cluster-head,which was set up by Neighborhood Handshake Protocol.Finally,the parameter estimating theory was introduced to structure clock offset compensation based on linear model and maximum likelihood estimation.Theoretical and experimental results showed that the algorithm can guarantee the effectiveness of synchronization topology.The synchronization overhead is reduced by 33% compared to the HRTS algorithm,and 88% compared to the TPSN algorithm.The single-hop average error of the proposed method is different from TPSN algorithm by only 6.36 μs,and improved by 12.87 μs compared with the RBS algorithm.