| Because of many advantages, such as all-weather, high precision, automation, high benefit and so on, GNSS have been widely used in geodesy, geodynamics, earthquake and geological hazard, deformation monitoring and other academic research and engineering applications, and it greatly promotes the development and application of surveying technology. With the increasingly extensive application of GNSS, the need of high precision positioning is increasing, and GNSS error processing has been paid more and more attention. The GNSS signal is influenced by three kinds of error sources:error related to the satellite, error related to signal propagation and error related to the receiver. Error related to the satellite and receiver include clock error, antenna phase deviation, orbit error, hardware delay, observation noise, etc; the error in the process of signal propagation include the space environment (ionosphere, troposphere, etc.) propagation error and the ground station environment (multipath reflection, vegetation and antenna snow cover, water vapor and volcanic plume, etc.) propagation error. For the ionosphere, troposphere and satellite clock error, they can be reduced by system error model, parameter estimation and differential technique, But for the error caused by multipath, antenna snow cover and other ground station environment, due to the complex local characteristics and weak spatial correlation, there are currently no an effective correction model or universally applicable data processing method.Changes of the ground station environment will cause variation of signal strength, polarization characteristics, propagation direction and path, resulting in positioning error. For example, the influence of geodetic GNSS receiver carrier phase multipath is about centimeter level,but under the environment of water surface,pseudo-range multipath error can be up to 7 meters,and the influence of receiver antenna snow on the the PPP can be up to several centimeters or more. The errors caused by ground station environment directly influence the accuracy and reliability of precise navigation and positioning, precise deformation monitoring, structural vibration monitoring and earthquake coseismic analysis, etc, it has become the main source of error and the key factors restricting the high precise positioning and its application. When analyzing GNSS earthquake coseismic in high sampling rate, it is difficult to model due to multipath error and earthquake frequency overlap. GNSS continuous operation sites applied to the study of global geophysical have significant position error, because it is inevitably affected by snow and other harsh environment in many parts of the world. How to effectively extract or eliminate the influence of station environment error has become an international hot research topic in recent years.When GNSS receiver provide pseudo-range and carrier phase, it also provide SNR (Signal-to-noise Ratio) data to measure the quality of received signal. SNR is an important indicator of signal quality of GNSS receiver, which includes observation quality information,and it is also sensitive to station observation environment, SNR changes are closely related to station environment, such as season, temperature, soil moisture and snow cover. So GNSS SNR data has unique and effective value in observation quality evaluation, adjustment random model construction, station environment error processing, GNSS SNR data is getting more and more attention.Existing research show that, the multipath error usually regarded as GNSS signal noise contains useful station environment information. GNSS reflected signal is closely related to multipath environment, station environment such as soil moisture, vegetation and snow cover will change the characteristics of ground reflector, and then change SNR parameter such as SNR frequency, amplitude, phase,et al. Relationship between multipath reflection signal and environment parameters is builded to inverse the parameter. Snow, soil and vegetation water is an indispensable important component in the terrestrial water cycle, and has an important influence on the climate and ecosystem, and its dynamic changes are inseparable from environment and climate. GNSS remote sensing technology based on multipath effect provides a new and high efficiency way for monitoring soil humidity, snow thickness and vegetation growth. The technology has broad prospects in research and application of GNSS because it make full use of the existing GNSS CORS network.In this dissertation, in-depth discussion of station environment error characteristics and the relationship between SNR data and station environment is conducted. Based on these discussions, station environment error extraction and correction method using GNSS SNR data are studied, and the inversing method of soil moisture by the SNR data is comprehensively discussed. The main contents, relative conclusions and results in this dissertation are as follow:1. Systematic discussion of the physical mechanism, geometric model and spatial and temporal characteristics of GNSS multipath environment error is conducted. With experiments on the simulated and measured data, the time-frequency characteristics of multipath environment error in the observation and Coordinate domain are studied. The results showed that the repetition period of multipath error environment is about 236s; From the aspect of magnitude, these errors would not exceed some theoretical values. pseudo-range multipath error is generally not more than one symbol width, and phase multipath error is not over 1/4 wavelength of the carrier. Meanwhile, multipath environment error also has certain energy concentration distribution. These time-frequency characteristics provide theoretical basis for the use of digital signal processing methods to extract station environment error. Besides, the characteristics and model of SNR under multipath reflection environment are discussed in detail,and the influence of station environment on SNR value is analysed. The correlations between SNR value and station environment provide theoretical evidence for processing method of environment error and environmental parameters estimation using SNR data.2. The relationship between GNSS SNR and carrier phase multipath error are studied and discussed, how to correct multipath error in carrier phase by the GNSS SNR is realized. and the analysis of LC combination observation residuals using measured data is conducted. The results show that, multipath error of carrier phase observation can be corrected to some extent by SNR, and LC observables residuals have energy between 0.0001 and 0.0005Hz,which is consistent with the multipath error frequency.3. The algorithm for detecting outlier of GNSS site coordinate sequence caused by snow and ice by SNR observations are presented, with the algorithm, the gross errors in position time series by SNR can be identified and handled. The GPS site data observed by Plate Buandary observation(PBO) in American is used to test and analyze the algorithm. As is shown in this case, under the assumption of site linear tectonic motion, RMS value of linear fitting for GPS site coordinate sequence is reduced from 0.29 cm,0.16 and 1.67cm (E, N, V) to 0.12,0.11 and 0.44cm, the algorithm can effectively improve accuracy of positioning.4. The functional models of SNR multipath are studied,and SNR data selection strategy and effective inversion area are discussed in detail.On this basis, the GNSS soil moisture inversion package GNSS_SMI based on Matlab is designed and realized. With the GPS observation, measured and simulated soil moisture data, The inversion results of soil moisture based on SNR are compared and analyzed, and the correlation between the multipath phase and the soil moisture are quantified. The results of experiments show that, the effective inversion region of soil moisture is a group elliptic related to the height of antenna, satellite elevation angle and azimuth angle, L2C SNR data which is consistent to multipath reflection model is more beneficial to humidity inversion, And relative phase delay φ of SNR is an important indicator of soil moisture, the exponential function can better reflect the mapping relationship between delay phase and soil water content.5. Considering the short-time changes of season, weather, vegetation, slope,etc. have little influence on SNR phase parameters, the method of soil moisture estimation based on sliding time window is proposed. The soil moisture is retrieved by three methods including the full time data, the sliding time window prediction and the sliding time window interpolation, and the inversion results are compared and analyze. The mean correlation coefficient are respectively 0.717,0.832 and 0.952. Compared with the full time inversion method, The mean correlation coefficient is increased by 16.2%and 32.9%, the error of L1 norm of the window prediction and interpolation method is decreased by 39.8%and 62.0%, and the error of L2 norm is decreased by 17.4%and 54.6%, As the results of experiments shows, the time window modeling soil moisture inversion can effectively simulate the short change of station reflection environment,and thus improve the inversion precision. Although the error of Window interpolation method can be reduced to an ideal effect, but it is difficult to realize near real-time estimation of soil moisture; the method of the sliding time window prediction has slightly lower precision than the sliding time window interpolation, but it can be used for near real-time estimation of soil moisture... |
PDF Full Text Request