GPS PHASE MEASURE CYCLE-SLIP DETECTING AND GPS BASELINE RESOLUTION BASED ON WAVELET TRANSFORM

Abstract

The errors in GPS measurement generally consist of systematic errors (such as clock bias, iono - sphere, troposphere effects etc.), random errors (such as measuring error) and outliers. Systematic errors can be canceled by differencing techniques or by adding parameters into the equation system. Outliers may be detected by adding parameters or by statistical methods such as expectation shifting or variance inflating.

Because wavelet analysis has many good natures both in the time domain and in the frequency domain, as can automatically zoom in or out with different scales (frequencies), the arbitrary details of a signal can be observed and analyzed by the aid of wavelet analysis. Based on the above features, wavelet analysis is reputed as “a mathematical microscope”.

The paper presents a wavelet based outlier detecting and filtering algorithm. GPS phase observables are decomposed into different components by wavelet transform. The content at scale 1 mainly consists of high frequency components including cycle-slips or outliers. Because the amount of cycle-slips is zoomed in by wavelet transform at the original location, so the locations of cycle slips are easily detected after wavelet transform.

Based on the properties of outliers, cycle slips and random noises under wavelet transform, a socalled á trouS reconstruction algorithm is selected. The noises in double difference phase observables are filtrated by removing some components from scale-land reconstructing original signal. Then an algorithm for GPS base-line resolution using phase double difference based on wavelet filtering is put forward in this paper. By using the proposed algorithm, the ambiguity searching space is reduced, and the validation of fixed ambiguities is improved as well. The efficiency for the proposed algorithm is approved by three examples.