Research On GNSS Multi-frequency Loosely/Tightly Combination RTK Localization Algorithm

At present,the four major navigation systems in the world have completed the modernization construction,and each navigation system can already provide continuous navigation and positioning services to most users around the world.Multi-system integration greatly increases the number of satellites available for positioning,and the use of multiple frequency points also enables better compatibility and interoperability between systems.Therefore,this paper focuses on the basic principles of GNSS RTK positioning,and develops the research on GPS/BDS/Galileo/GLONASS four-system RTK positioning: for the case of short baseline RTK,an algorithm for estimating single-difference ambiguity using Kalman filter model is studied.Aiming at the situation of medium and long baseline RTK,an algorithm for estimating atmospheric delay error parameters using Kalman filter model is studied,and the feasibility of the two algorithms is verified by the measured data.For some complex environments,the number of available satellites in a single system is insufficient,making it difficult to complete RTK positioning.In order to solve this problem,this paper studies the GPS/BDS/Galileo three system co-frequency tight combination RTK positioning model,and proposes a method to reshape the parameters to be estimated.As well as L5/B2a/E5 a,the same and different types of receiver end system deviation and its stability issues are calculated,analyzed and discussed.Finally,the measured short baseline data and dynamic on-board data are used to perform tight combination RTK positioning performance and algorithm feasibility and reliability analysis.The research contents and results of this paper mainly include:1)The unified method of space-time reference for GPS,BDS,Galileo,and GLONASS systems is introduced,the GNSS RTK non-difference,single-difference and double-difference observation equations are deduced,and the available frequencies of the four main GNSS systems,the combined observation coefficients selection principle are introduced.2)Three main error sources in RTK positioning are introduced,and the mathematical models of the double-difference ionospheric delay error term and the double-difference tropospheric delay error term,which affect the positioning accuracy are deduced,and specific correction methods are given.Emphasis is placed on analyzing the quality of observational data.The specific method is to use the zero-baseline model to analyze the residuals of the dual-frequency/tri-frequency observations of the four major GNSS systems.The analysis shows that the noise levels of different systems are different.BDS-3 has the lowest observation noise,and the GLONASS system has the highest observation noise;the noise levels of BDS-2 satellites in different orbits are different;the noise level has an obvious trend of increasing with the decrease of the altitude angle.3)A Kalman filter model for estimating single-difference ambiguity for short baselines is studied.Using the measured data,seven system combinations of GPS,BDS,Galileo,GLONASS,GPS+BDS+Galileo,GPS+BDS+GLONASS,GPS+BDS+Galileo+GLONASS are used to conduct experiments to verify the feasibility of the algorithm and analyze positioning accuracy.The analysis shows that the multi-system combined RTK has higher positioning accuracy,and the GPS+BDS+Galileo+GLONASS combination has the highest positioning accuracy,which is much higher than any single-system mode.An RTK localization algorithm suitable for medium and long baselines is studied.The algorithm adopts the Kalman filter model for estimating atmospheric delay parameters,and uses the measured dual-frequency data of GPS/BDS-3/Galileo/GLONASS four systems to conduct experiments.The experimental results show that the positioning accuracy and ambiguity of the algorithm using the estimated atmospheric delay parameters are fixed.The rate is much higher than the strategy which is not estimating.4)A inter-system bias estimation method based on parameter reorganization is studied.Two types of identical frequencies of L1/E1/B1 C and L5/E5a/B2 a of GPS/BDS-3/Galileo system are analyzed through two scenarios of zero baseline and short baseline.,and analyze the multi-day stability of the system deviation..The results show that the mean values of ISCB and ISPB at the same frequency of receivers of the same type and version number are about 0,both of which show high stability;the mean values of ISCB and ISPB at some frequency points with different receiver types are not equal to 0,but still has high stability.Finally,the positioning performance of GPS/BDS-3/Galileo’s L1/E1/B1 C and L5/E5a/B2 a dual-frequency tight-combination and loosecombination RTK are compared through measured short baseline data and vehicle dynamic data.The experimental results show that when the number of available satellites is sufficient,the positioning accuracy of the two is equivalent,but when the number of satellites is insufficient,the positioning performance of the tight combination RTK is much higher than that of the loose combination RTK.Taking the GPS+BDS system combination as an example,when the cut-off altitude angle is 45°,the ambiguity fixation rate of the tight combination mode of the vehicle dynamic experiment is 33.3% higher than that of the loose combination mode.The fixed solution positioning accuracy of the tight combination mode is centimeter level,while the fixed solution positioning accuracy of the loose combination mode is meter level.Therefore,the tightly combination mode can effectively improve the RTK positioning accuracy in extreme scenarios with a small number of satellites,and has high applicability to scenes such as cities and canyons in real life.