测绘学报

• 学术论文 •    

参考经验模态分解-独立分量分析及其在GPS多路径误差处理中的应用

罗飞雪1,2,戴吾蛟3,唐成盼2,黄大伟2,伍锡锈2   

  1. 1. 湖南农业大学
    2. 中南大学
    3. 中南大学测绘与国土信息工程系
  • 收稿日期:2011-05-13 修回日期:2011-10-26 出版日期:2012-06-25 发布日期:2012-06-25
  • 通讯作者: 戴吾蛟

EMD-ICA with Reference Signal Method and Its Application in GPS Multipath

  • Received:2011-05-13 Revised:2011-10-26 Online:2012-06-25 Published:2012-06-25

摘要: 多路径误差在绝大部分应用中,其测站间相关性很弱,不能通过差分观测得到有效削弱,成为高精度GPS动态变形监测中的主要误差源。当天线与周围环境变化很小或不变时,多路径误差具有很强的重复性,利用周日重复性建模是改正多路径误差的一种有效方法。但随着时间的推移,多路径效应的重复性会逐渐降低。此外,多路径效应误差占据的频率范围比较宽,容易与结构的低频振动位移混合在同一频率带内,而传统信号处理方法多从频域进行处理,对于频率混叠这类情形的效果不甚理想。为此,针对GPS多路径误差的特点,结合具有良好多尺度分解能力的经验模式分解方法(EMD)和具有良好盲分离能力的独立分量分析(ICA)方法,提出了带参考信号的EMD-ICA方法来削弱多路径效应的影响,应用实例的结果证明了该方法的有效性。

Abstract: Multipath error is a main influence of high-precision GPS dynamic deformation monitoring as it is difficult to be mitigated by difference observation in most of applications. If the local environment of the antenna is unchanged from one day to another, the multipath error will nearly be the same. So establishing the error correction model with repeatability characteristics is a valid method to mitigate multipath error. However, the repetition of multipath error reduces along with time passing. What is more, as the frequency range of multipath error is wide, it may mix with the low- Frequency of vibration signals. The traditional methods usually process in frequency domain, so that it can’t deal with such error. Therefore, a method called EMD-ICA with reference signal is proposed. It combines Empirical Mode Decomposition (EMD) with independent component analysis (ICA). The test shows the effectiveness of this method.