To improve the performance of GPS single point positioning under flood environment based on multi-CNR weight strategy and multipath mitigation

Multipath effect largely affects the performance of GPS positioning since it is related to the surrounding environment of the GPS station and cannot be effectively corrected by the empirical model. In particular, the multipath effect increases during flood due to the reflection and refractive of water performing higher than that of dry soil. Thus, in order to improve the performance of SPP (signal point positioning) during flood, the multipath effect should be considered and mitigated. The quantities related to multipath effect are analyzed by using data sets collected during flood, which includes the CNR (carrier-to-noise ratio) and MP (pseudorange multipath). In addition, the performance of SPP during a flood is evaluated, results show that the decreased percentages of SPP in the three components are approximately 25.00 % for E, 17.91 % for N, and 21.48 % for U, respectively. To mitigate the influence of multipath induced by flood, the correlation between the CNR, MP, and elevation angle (EA) is investigated in theory. Results indicate that the correlation between CNR and MP is much greater than the correlation between MP and EA. Thus, a new method based on an improved multi-CNR weight strategy and the sidereal filtering method is proposed to mitigate the multipath effect. The performance of the proposed method is assessed by using real data sets. Experimental results show that the average correlation between the multipath correction model and multipath error can reach approximately 0.65. After applying the proposed method, the accuracy of MP and SPP can be effectively improved, and the average improvement rate of the MP and three components are approximately 16.84 % and 14.76 % compared with the original results, respectively. It should be noted that this proposed method is also suitable for static SPP applications in other multipath environments, such as the environment obscured by trees.