| At present,GNSS high-precision deformation monitoring mainly adopts the method of short baseline relative positioning.Based on the double-difference technology,the system errors related to the receiver and satellite end can be eliminated,but the multipath errors cannot be eliminated by difference,so the multipath errors become the main error source limiting the high-precision acquisition of deformation information.Therefore,it is of great significance to study how to reduce the influence of multipath error for improving the precision of GNSS deformation monitoring.In view of the above situation,this paper mainly studies from the following aspects:1.The observation model of GNSS is introduced in detail;The principle of multipath error generation is studied and the mathematical model of multipath error is derived.The influence factors related to the multipath error are studied,and the following conclusions are drawn by quantitative analysis: 1)When the distance is constant,the multipath phase delay is often affected by the incident Angle of the reflected signal,which is the most influential factor on the multipath error;2)The reflection coefficient of the reflector itself is mainly affected by the nature of the object,and will have a great impact on the amplitude of the multipath error.When other factors are the same,the size of the multipath error largely depends on the reflection coefficient;3)The distance between the reflector and the receiving antenna can attenuate the signal energy by increasing the distance,thus reducing the influence of multipath error.The experimental results show that when the reflection distance exceeds 50 m,the influence of multipath error can be basically ignored.2.In view of the mode aliasing and other problems existing in the empirical mode decomposition method,the stellar day filtering method of GNSS deformation monitoring was upgraded.An Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise-Principal is proposed Component Analysis,ICEEMDAN-PCA combined noise reduction method,GNSS multipath effect error is effectively corrected;Through simulation experiments and field data verification,EMD,EEMD,CEEMDAN and ICEEMDAN are compared and analyzed.It is found that ICEEMDAN has fewer components decomposed and better processing effect of white noise signal.A new scale extraction method is designed to select the modal components and reconstruct the multipath errors to weaken the multipath errors in GNSS coordinate sequences in real time.The experimental results show that the RMS of coordinate sequences in N,E and U directions is reduced from 2.5mm、2.0 mm and 6.2 mm to 2.1 mm、1.7 mm and 5.1 mm in EMD.EMD improved coordinate sequences in N,E and U directions by 39.8%,32.8%and 37.2%,while ICEEMDAN improved coordinate sequences in N,E and U directions by 48.9%,43.5% and 47.9%.The RMS results of ICEEMDAN method’s coordinate sequences in N,E and U directions are better and the improvement degree is also better,and GNSS coordinate sequences with higher precision can be obtained.3.Research and put forward the spatial domain modeling method of RBF neural network assisted by K-means clustering algorithm.Make use of the characteristics of selflearning of neural network without manual intervention,constantly learn and train a large number of original observation data,and then predict and correct multipath errors in real time through the trained model;The stellar daily filtering is compared with the RBF neural network algorithm assisted by K-means clustering algorithm,and the stability and reliability of this method to weaken GNSS multipath error in real time are analyzed from the statistical results of 7-day GNSS observation data of multi-system.Through the experimental data,it is found that the improvement degree of the coordinate sequence in E,N and U directions by the sideral day filtering method is 32.9%,33.5% and 30.3%,and the three data are 48.6%,41.7% and 48.7% respectively under the RBF neural network algorithm assisted by K-means clustering algorithm.This method can weaken the GNSS multipath error of multi-system in real time,with better observation accuracy and improvement effect.By this method,the influence of GNSS multipath error can be suppressed to a large extent.Figure [37],Table [19],Reference [88]... |
PDF Full Text Request