基于EMD和小波的GPS/BDS变形监测中的多路径误差削弱方法

摘要

This paper is concerned with the sidereal day repeat characteristics of the multipath error and high frequency random noise of calculated coordinates for real-time deformation monitoring.First,we apply the EMD and EMD combined with wavelet threshold de-noising methods,to the existing coordinates multipath time series model.Then,based on this model and through sidereal filtering,we weaken the multipath error of subsequent coordinates,which has a strong correlation with the model.The measured data processing results show that the EMD method can remove the high frequency random noise sequence and weaken the effect of multipath error of GPS/BDS real-time monitoring effectively,improving by 50％ the accuracy of real-time monitoring.The EMD and wavelet combined method is superior to single EMD filtering method.%针对GPS/BDS实时监测坐标序列中多路径误差的周日重复特性和高频随机噪声,分别采用EMD以及EMD与小波阈值去噪相结合的方法对现有坐标序列构建多路径时序模型,并通过恒星日滤波削弱后续坐标序列中具有强相关性的多路径误差.实测数据的处理结果表明,EMD可以很好地去除GPS/BDS实时监测序列中的高频随机噪声并削弱多路径误差的影响,提高实时监测精度50％左右,EMD和小波组合方法较单EMD效果稍好.