交互式多模型IEEE 1588时钟同步算法

时间同步是保证分布式测试数据有效的重要前提。时钟漂移对环境十分敏感,并没有准确的估计模型。提出了基于卡尔曼滤波器的交互式多模型IEEE1588时钟同步算法,该算法以一阶自回归模型和恒定加速模型对时钟漂移进行分别估计,运用交互式多模型系统对估计结果进行动态加权从而得到相对准确的时钟漂移。仿真实验结果表明,在时钟同步中引入该算法能够提高时钟同步精度。

Interactive Multi-Model IEEE 1588 Clock Synchronization Algorithm

Time synchronization is an important prerequisite for ensuring that distributed test data is valid. Clock drift is very sensitive to the environment and there is no accurate estimation model. A multi-model IEEE1588 clock synchronization algorithm based on Kalman filter is proposed. The algorithm uses a AR(1) model and a constant acceleration model to estimate the clock drift separately. The Interactive Multi-Model system is used to dynamically weight the estimation results to obtain a relatively accurate clock drift. Simulation results show that the introduction of this algorithm in clock synchronization can improve the clock synchronization accuracy.