Research Of Airborne GPS Kinematic Position Velocity And Attitude Determination And Software Development

With the development of our national ecnomy and scientific techinology, the aerial survey has been widely used in the national production and defense construction. One of the most important problems in the aerosury is to precise determine the position, velocity and attitude of moving measure table. Therefore, some critical techinical issues are discussed in this paper. They include the error processing method, the kinematic positioning algorithm for long range flight, the methods of velocity and acceleration determination, and GPS attitude determination method based on adaptive robust kalman filter etc. The main works and contributions of this paper are summarized as follows:1. The current research situation of GPS kinematic positioning is presented based on the four basic elements of survey adjustment. They include the error processing strategy, the function and stochastic model, parameter estimation method and precision evalution criterions of GPS kinematic positioning. The studies of ambiguity solution are emphasized summarized.2. The tropospheric delay error in airborne GPS positioning is amply discussed. A new method of GPS tropospheric delay error correction is developed, and named as "Prejection Extension Method". The results show:the precision of tropospheric delay correction is always in the mm-level using this method, when the altitude of roving station from some hundred meters to some kilometers. Therefore, it is specially suitable for an airborne kinematic positioning.3. Based on the results of GPS data processing, some concludes of the relationship between the estimation of station elevation and that of tropospheric delay. When the estimated ZTDs are larger than the true ZTDs, the estimated elevation will be smaller than the true elevation; when the estimated ZTDs are smaller than the true ZTDs, the estimated elevation will be larger than the true elevation. Therefore, a mismodeling of tropospheric delay will result in a degradation of GPS positioning precision in height component.4. A numerical simulation is performed to analyze the influence of outliers on GPS kinematic relative positioning. The influence function of errors and the method of robust estimation are given. In order to deal with the outliers in GPS airborne kinematic relative positioning, a combining method of outlier detection and robust estimation is proposed. It includes two steps:firstly, the gross errors are detected and deleted based on the median method in the data preprocessing; then the residual outliers is controlled by the robust estimation in the parameter adjustment. To verify the validation of this method, several experiments are performed. The results show that the influence of the outliers can be effectively controlled and the computing speed is improved, when the combining method are adopted.5. GPS real time precise point positioning (RT-PPP) requires accurate and reliable prediction of satellite clock errors (SCE). This is one of the most challenging tasks in the development of a RT-PPP technique. By intensively analyzed recent years IGS SCE data, we find that a valid period of predicting GPS SCE is obviously dependent on the stability of the individual satellite clock. The most recent GPS satellites (BLOCK IIR and BLOCK IIR-M) are more stable than the former GPS satellites (BLOCK IIA). We can easily predict their next 6 hours SCE with LM and IGS ultra rapid products (URP). However, some of BLOCK IIA satellites clocks are instable. Therefore, we further analysis of their residuals. We find that its residuals show periodicity using spetrum analysis. Therefore, this part residuals can be further simulated using a periodicity correction model.6. Based on the analysis of a number of GPS satellite clock error data, a no-stationary time series model for real-time SCE prediction is developed. This prediction model includes:linear term, cyclic corrected terms and auto-regression term, which are used to represent SCE trend, cyclic parts and rest of the errors, respectively. Based on the numerical computations, the SCE predicted value standard deviation is about 0.5ns and mean error is about 1 ns, using the IGS final products as reference.7. A method of adaptively changing reference station is developed to make the single baseline model suitable for the long range airborne positioning. The basic idea is that the reference station will be automatically replaced by the nearest one, when the distance between kinematic station and original reference station is longer than the maximum distance decided by the software. The method of equivalently eliminating parameters is used to combine the observation data between new and original reference stations. To verify the validation of this method, GPS data of airborne campaign AlpinAero2008 is used. The results show that the proposed method can provide a more stable and consistent solution of kinematic positioning with about 2 cm precision level compared with those of the network solution.8. The central difference is a common method in the numerical differentiation. The derivative of a unknown function can be obtained using this method. Three points central differences method (CDM) has been widely used in GPS velocity and acceleration determination. However, this method will lead to a considerable error when the carrier has a strong maneuverability. The reasons have been analyzed intensively in this paper. And it is proved that the truncation errors will be decreased and observational errors will be increased when more points be used in CDM. For finding a optimal points of using CDM for GPS velocity determination, GPS data of airborne campaign MEXAGE2001 is used. The results show that if data interval is 1Hz, average velocity and acceleration of carrier is 20 m/s,0.2 m/s2 respectively, the nine points CDM is the best. Comparison of common using three points CDM, the precision of velocity determination is improved about 50%. And the improved level is more significant when the carrier has a strong maneuverability.9. In general, there are three methods of GPS velocity determination: the position derivation method, the carrier phase difference method and the Doppler method. The advantages and disadvantages of the three methods are analyzed intensively. And a new velocity determination method is devloped by combining carrier phase and Doppler observations, in order to overcome the disadvantages of existing velocity determination methods. The theory of this method is stricter than that of existing methods. And it makes full use of the information coming from the different types of GPS observation data. The experimental results show that the precision and reliability of velocity determination are obviously improved by using this method.10. The advantages and disadvantages of three existing methods of GPS attitude determination are analyzed intensively. A new method is presented, based on the adaptive robust kalman filter. The precision and reliability of the presented method are higher than that of the existing methods, because of intelligent using the information of prior kinematic model. The results show the precisions of attitude determination of this method are better than 0.01 degrees.11. A GNSS kinematic data processing software is developed, which has a proprietary intellectual property rights, named as XHGNSS V1.0. The overall design and characteristics of the software are outlined. The operation instruction and the major functional modules are introduced briefly. And the numerical examples of kinematic positioning, velocity determination and attitude determination using this software are given. The experimental results show the precisions of kinematic positioning, velocity determination and attitude determination of XHGNSS are better than 5 cm, 1cm/s,0.01 degrees, respectively.