基于最大熵谱估计的GPS坐标时间序列噪声分析

NOISE ANALYSIS ON GPS COORDINATE TIME SERIES BASED ON MAXIMUM ENTROPY SPECTRUM ESTIMATION

通过提取苏通大桥GPS坐标序列,对不同时间观测的同一监测点时间序列采用db6小波函数进行五层分解并去噪处理,得到噪声序列。然后采用最大熵谱估计分析GPS坐标分量的信号噪声特性,并与传统的谱分析方法相比较。结果表明,GPS坐标时间序列主要含有白噪声与闪烁噪声,随机漫步噪声不明显。因此,在分析GPS观测得到的时间序列时,应当顾及这种噪声特性,适当地加权以得到比较切合实际的结果。

According to the extraction of GPS coordinate series in Sutong Bridge,db6 wavelet function decomposes the same monitoring point time series for different time observations into five layers and denoising processes,and gets noise sequences. Then we analyze the signal noise characteristics of GPS coordinate component by using the method of maximum entropy spectrum estimation,and compared with the method in traditional spectrum analysis. The results indicate that GPS coordinate time series mainly contain white noise and flicker noise,the random walk noise is not obvious. So that,analyzing the time series obtained by the GPS observation,the noise property should be taken into account and the observations should be weighted correctly in order to fit the actual situation.