An improved time-domain multipath mitigation method based on the constraint of satellite elevation for low-cost single frequency receiver

Multipath effect seriously affects the accuracy of GNSS high precision positioning since it cannot be corrected by empirical model or eliminated by difference techniques. Sidereal filtering method can effectively mitigate multipath error for static positioning by using the characteristics of satellite time-domain repetition. The key of this method is to extract accurate multipath model on the reference day and mitigate multipath error by searching the multipath model accurately on the subsequent observation day. We analyze the relationship between the multipath error and satellite elevation first by using the data sets which collected from the Multi-GNSS Experiment (MGEX) stations. Based on the result, an improved time-domain sidereal filtering multipath mitigation method based on satellite elevation constraints is proposed. The information of satellite elevation is used to constrain the threshold strategy and decomposition layer of wavelet packets transform, which can effectively improve the denoising effect and extract a more precision multipath model. In addition, the performance of positioning of the proposed method at different mask angles is also evaluated. Real short baseline data sets are collected by low-cost single frequency receivers to evaluate the performance of the positioning of the proposed method, experimental results show that using the proposed method in multipath mitigation can improve approximately 25.00, 21.43 and 13.98% in the east, north and up components compared with the original results, respectively. Even compared with the traditional wavelet packets transform method, the improvement rate can reach approximately 10.78% for 3D positioning. Moreover, the proposed method can also improve the performance of positioning at different mask angles to a certain degree, especially for the low mask angle seriously affected by multipath effect. (C)2022 COSPAR. Published by Elsevier B.V. All rights reserved.