| Since the establishment of Global Positioning System(GPS),the navigation and positioning theory has made considerable progress.Thereinto,the precise kinematic relative positioning technique has been widely applied in deformation monitoring,precise agriculture,attitude determination and formation flying,etc with the advantages of real-time,all-weather,high-precision and reliability positioning.The existing kinematic relative positioning theory and method can guarantee the positioning accuracy in most occasion,while in some cases where there are lots of shelters like buildings or trees,or the vehicle is in high dynamic,or the reflection of GNSS signal is severe near water,or the non-geodetic GNSS receivers or antennas are employed,it is hard to or even cannot get reliable and precise positioning results by single positioning system or classic dynamic relative positioning theory.With the improvement and development of GNSS,the four global systems including GPS,GLONASS,Beidou and Galileo can provide positioning service in the whole world by 2020,which makes it possible for precise kinematic relative positioning in complex observation environment.Therefore,the paper will set up the mathematical model suitable for single/combined GNSS kinematic relative positioning and put forward the processing methods for complex observation environment.First of all,in the light of FDMA adopted in GLONASS,the combined positioning model suitable for single GLONASS and multi-GNSS systems is put forward.Then,based on the existing classic kinematic relative positioning theory,the paper focuses mainly on the refinement of stochastic model and robust estimation according to the data characteristic in the complex observation condition and non-geodetic low-cost GNSS receiver.In addition,the software of multi-GNSS kinematic relative positioning is also developed and with measured data,its efficiency and precision are analyzed in detail.The main work and contributions of this thesis are as follows:(1)The mathematical model of relative positioning as well as the processing methods in different baseline length are first illustrated.For the particularity of GLONASS,the treatment of inter-frequency bias and single-differenced ambiguity of reference satellite is improved on the basis of original single-differenced filtering method.For the sake of the demand in the combined multi-GNSS kinematic relative positioning,the corresponding parameter estimation approach is established which can deal with single/multiple frequency and single/multiple GNSS system observation.The results show that the model is effective in single/combined GNSS positioning,the positioning accuracy of GLONASS is comparable to that of other systems in short baseline observing and the fixing rate of GLONASS ambiguity is likewise improved.(2)Considering that there may inevitably exist numerous cycle slips and gross error as the complex observation environment changes rapidly,especially when GNSS signal is serverely obstructed,it is necessary to make full use of the existing data to acquire the optimal positioning results as the number of observation is limited and its accuracy is unknown beforehand.The paper starts from the stochastic model of positioning,concludes the advantages and disadvantages of satellite elevation stochastic model and Signal Noise Ratio(SNR)stochastic model,comes up with the real time pseudorange noise calculation method of relative positioning and compares these three models by measured data.The results show that,the caculation method of real time pseudorange noise can reflect the real quality of observation.Thus,the float solution of Kalman Filtering is improved as well as the final positioning results as the reduction of robust iteration.(3)In complex observing condition,it is essential to iteratively exclude the impacts of gross error and cycle slips which can not be detected before parameter estimation.Hence,based upon classic IGGIII robust method,the improved scheme first reducing or zeroizing the weight of observation with gross errors.Then for the undetected cycle slip,it reinitializes the ambiguity of this problematic observation which is treated as gross error in two consecutive epochs.The results show that the improved scheme can effectively inhibit the influence of gross error and promptly handles the undected cycle slips in parameter estimation.Its robust effect is better that that of IGGIII.(4)On the basis of combined multi-GNSS relative positioning mathmatical model illustrated in this paper,the software for GNSS precise kinematic relative positioning is developed and tested with the measured observation from multi-GNSS,low-cost GNSS receiver and complex kinematic observation environment.The results show that the software has the capability of processing multi-GNSS observation and yields the solution whose 3D accuracy is better than 0.1m in complex kinematic observation environment. |
PDF Full Text Request