煤矿井下WSN容错性时间同步算法的实现

Implementation of fault-tolerance WSN time synchronization algorithm for coal mine

为了提高井下时间同步精度,针对无线传感器网络对节点同步精度要求高、鲁棒性强以及时间同步协议普遍容错性较低的特点,提出一种面向煤矿井下的无线传感器网络容错性时间同步算法。该算法结合煤矿井下结构的特殊性,采用动态选举时间基准节点和避免重复发送策略,增强了网络的可扩展性,对于新加入节点可迅速收敛。在洪泛时间同步协议算法的基础上,利用概率统计学中的残差分析理论改进线性回归算法并进行差错判断,剔除井下异常数据点对拟合曲线的影响,抑制由异常数据导致同步误差的大幅度跳变,提升无线传感器网络(WSN)同步算法的容错性和可靠性。实验证明,该算法可在网络拓扑不断变化的情况下达到一定的同步精度,可保持网络稳定性,满足井下同步精度需求。

According to the requirements of high precision and strong robustness among nodes,especially the characteristics of low fault-tolerance,an algorithm based on fault-tolerance was presented to enhance the synchronization accuracy for coal mine.Combined with the complex underground environment,the strategy using dynamic election of time reference node and avoiding repeated sending of packets was used to enhance network scalability.And the new nodes added into network can quickly converge with the new synchronization algorithm.With the analysis of flooding time synchronization protocol（FTSP）algorithm,the new algorithm introduces the residual analysis theory of probability and statistics to improve the linear regression algorithm and judge the errors so as to enhance the fault-tolerance of wireless sensor network（WSN）synchronization algorithm and the stability of the synchronous network.Experiments show that the synchronization accuracy is enough to meet the requirements of underground localization.