Data Preprocessing Of GPS Plane Control Network And Its Stability

Since 1990s, GPS positioning technique has been widely used in establishing many kinds of control networks. Because of complex observation conditions, GPS signal always encounters several kinds of disturbances coming from error sources, such as ionosphere refraction, troposphere refraction, multi-path effect, GPS receiver's random noise, etc. Therefore, it has been the main issue for GPS date preprocessing to find these errors, reduce them and improve the precision of baseline vector.The first grade plane control network is the datum of construction surveying. Based on the long used period of plane control network for large engineering, the stability of control points will directly influence the construction surveying and engineering construction quality. Therefore, developing a stable processing methods and their software developments will have great significance in reality.In this thesis, GPS and wavelet analysis have been employed to develop GPS data preprocessing methods and investigate the stability of plane control network stability analysis. Some real engineering cases were used to validate the proposed methods.The Sutong bridge GPS control network's surveying data was utilized to analyze the error of start and end points and surveying time influence to the baseline vector. Prime 4 hours observation data was selected and calculated, results indicate that selecting prime observation period of time can secure high quality Baseline.An analysis software was programmed in Visual basic and Matlab. The original arrying signal of the GPS plane control network from Sutong bridge and Gezhou dam were extracted. The advantage of multiscale of wavelet, multiresolution, flexility and movelity were combined with db3 wavelet function to complete a three-level wavelet decomposition for the original carrying signal, single-difference observed vale, and double-difference observed results. The effectiveness of alleviating noise with respect to the characteristic of high frequency and different difference method was investigated. The investigation demonstrated that wavelet analysis has great theoretical significance and practical value to decompose and assemble GPS data. It can be easily seen that wavelet analysis is a very important mathematical tool to GPSdata processing.A denoising program was developed in this study. The original observed data of Gezhou dam's baseline was analyzed. Three kinds of methods were used to denoise,which are different decomposition level, choice of different threshold value and combination, and different wavelet functions. An optimal reducing noise method was found based on the comparative study. The results show that an appropriate denoising method can effectively lift up the precision of the baseline, the standard error is 17.71mm without reducing noise and 4.60mm after reducing noise.The stability of plane control network and its analyzed methods were investigated. The stability analysis software for average gap method, t-check method, and deformation error ellipse method was developed. It was used to obtain the baseline of two times' data of Sutong bridge's plane control network. The Cosa GPS processing system developed by Wuhan University was used to conduct two-dimensional accumulated adjustment. Also applied my personally developed stability package to test the stability of the plane control network. A consistent conclusion was reached based on the compared results and analyses of the three methods.