利用EMD阈值消噪改进MW组合周跳探测性能

Improving the Performance of MW Combined Observation on Cycle Slip Detection Using EMD Threshold De-Noising

分析MW组合观测值的噪声特性,评价其实际周跳探测性能,提出利用EMD阈值消噪方法进行降噪。对于EMD的端点效应问题,在MW递推均值上增加虚拟噪声构成虚拟观测值对数据进行延拓。估计MW组合经EMD分解后各层分量的阈值,削弱噪声影响。分析消噪后含周跳MW组合所发生的变化,指出周跳具有"扩散"现象,提出基于消噪后MW组合的探测方案。

The noise level of MW combination is sometimes high, causing the possibility of miss detec- tion when using the TurboEdit algorithm. The noise property of MW is analyzed and the actual ability of cycle slip detection is evaluated. The EMD threshold de-noising method is proposed to reduce noise level of MW. To deal with the end effect of EMD, the data to be decomposed is extended by a set of data, which is formed by adding virtual noise on the time average value of MW. The thresholds of the IMFs generated by EMD of MW are estimated and used to threshold the noise. The new characteristic of de-noised MW with cycle slip is analyzed. It is found that the cycle slip will spread to neighboring data and a detection scheme based on the de-noised MW is proposed.