顾及分段相关性的恒星日滤波优化算法

Improved Sidereal Filtering Method Considering Segments' Correlation

多路径效应是高精度GPS动态变形监测中的主要误差来源之一。本文改进了常规恒星日滤波算法,提出了一种顾及分段相关性的恒星日滤波优化算法。本算法首先采用分段滑动窗口处理相邻两天的坐标残差时间序列,然后利用相关系数法求取各分段最佳相关延迟,对当天坐标残差序列进行恒星日滤波,得到削弱多路径效应后的坐标序列。研究结果表明,该算法较常规算法在N、E、U 3个方向上的滤波效果各提高7.7%、5.0%、11.2%,动态变形监测实验验证了本算法的有效性。

The effect of GPS signal multipath is considered as a major error sources for the high-precision GPS dynamic deformation monitoring.Based on the traditional sidereal filtering method,this paper proposes an improved sidereal filtering method considering segments＇ correlation.Firstly we perform a sliding average to the coordinates time series of adjacent two days,and then obtain the retardation who have a maximal correlation index,finally perform the sidereal filter to the coordinates time series of second day in each segments.Results show that the improved sidereal filtering method is more efficient than the traditional sidereal filtering method and the accuracy of N,E,U directions have increased7.7%,5.0%,and 11.2%respectively.The effectiveness of the algorithm is verified by the dynamic deformation monitoring experiment.