基于Bootstrap采样粒子滤波的WirelessHART时间同步算法﹡

针对无线HART传感器网络时间同步精度较低、能耗过大等问题，提出了一种基于Bootstrap采样的粒子滤波时间同步算法。在未知网络延迟分布的情况下，为了减少观测次数，对发送端和接收端的时间戳观测值进行Bootstrap采样，采用混合粒子滤波算法，获得精确的时钟偏移，从而不仅降低了无线HART传感网络时间同步误差，而且使能耗减小。最后，实验表明，对于无线HART网状分层网络，当观测量达到10以上时，粒子滤波算法获得的时间偏差的均方误差约是最大似然估计算法的50%，而基于Bootstrap采样的粒子滤波算法获得的时间偏差的均方误差约是最大似然估计算法的35%，仿真的结果验证了该方法的可行性和有效性。

Time synchronization algorithm based on Bootstrap sampling and particle filter for wireless HART networks

For the issues of excessive energy consumption and the low accuracy of time synchronization in the entire wireless HART sensor network, and a novel time synchronization algorithm based on particle filter is proposed by Bootstrap sampling. Under the scenario of delay distribution in unknown network, in order to reduce the times of observations, the hybrid particle filter algorithm is used to obtain precise clock offset by the timestamp observations from the sender and the receiver are sampled by the Bootstrap sampling, It not only reduces the time synchronization error of wireless HART sensor network, but also decreases the energy consumption. Finally, the experiments show that the mean square error values of time deviation obtained from the particle filter algorithm are just 50% of the algorithm of maximum likelihood estimation when observed quantity is higher than 10 for wireless HART mesh hierarchical network. However, the same mean square error values of time deviation obtained from Bootstrap sampling based particle filter algorithm are only 35% of the algorithm of maximum likelihood estimation, and the simulation results show the feasibility and validity of the algorithm.