Study On The Methodology And Applications For Multi-station And Multi-system GNSS/INS Precise Positioning

On July 31,2020,the global BeiDou navigation satellite system(BDS-3)has been officially established to provide global services,with a constellation of 30 satellites:three geostationary orbit(GEO)satellites,three inclined geosynchronous orbit(IGSO)satellites,and 24 medium earth orbit(MEO)satellites.The global users can use the multi-systems and multi-frequencies observations to improve the performance of the Positioning,Navigation,and Timing(PNT)services.With the continuous upgrading and modernization of GPS,GLONASS,Galileo,QZSS and NAVIC,the satellite navigation positioning system is developing towards the direction of multi-system and multi-frequency,which will present a new situation for "Letting a hundred schools of thought contend,a hundred flowers blossom".GNSS with multi-system and multi-frequency will bring more effective observations,which will help to enhance the satellite geometry,increase the fixed success rate of ambiguity,and improve the accuracy and reliability of GNSS positioning.However,GNSS signal is easy to be interfered.In some special scenarios,the signal will be loss of lock and disruption,which seriously affects the performance and availability of GNSS precise positioning.At the same time,the traditional GNSS data processing model still has a greater optimization and improvement.For example,the multi-baseline solution(MBS)model for multiple station is not perfect,the compatibility and interoperability of different satellite positioning systems need to be studied,and the combined navigation and positioning for GNSS and multi-sensor still need to be further studied.Therefore,in order to meet the demand of pricese PNT services,the research on high-precision GNS S positioning and GNSS/inertial navigation system(INS)integrated positioning has become an urgent requirement in the field of the navigation and location service,which has very important scientific significance and practical value.Aiming at the problems mentioned above,this theies is intended to:(1)In the aspect of GNSS precise positioning,this paper deeply studies the GNSS precise positioning model for multi-system and multi-station.On the basis of analyzing the traditional single baseline solution(SBS)model,this paper introduces the equivalent transformation theory,and establishes two rigorous data processing models for MBS,which expand the traditional GNSS data processing method.(2)In the aspect of multi-system GNSS interoperability,the tightly-coupled integration model for MBS is studied according to the observations of different satellite systems with overlapping frequency,and the differential inter-system biases(DISB)are estimated and analysed.Through the correction of the DISB parameters,the tightly-coupled integration positioning for multi-baseline and multi-system with overlapping frequency is realized to improve the positioning performance.(3)In the aspect of GNSS/INS integrated positioning,the INS is used to enhance the dynamic model of the RTK positioning.The RTK/INS tightly-coupled integration model for multi-baseline and multi-system is established and the performance is validated and analysed.The main work and contributions of this thesis are listed as follows:(1)The traditional GNSS relative positioning model is summarized systematically.the equivalent transformation model based on the un-differenced(UD)observation is developed in detail based on the equivalent transformation theory.By implementing two equivalent transformations,the satellite and receiver clock parameteres can be eliminated,and the equivalent transformation of the UD and double-differenced(DD)observation equations is obtained,which provides a rigorous model for the MBS in the same observation segment.The static experiment is used to evaluate the performance of the proposed method.The results show that the MBS can enhance the strength of the relative positioning model and improve the positioning accuracy and availability.The shorter the observation time is,the more significant the performance improvement is.For the 30-minutes scheme,there is a significant performance improvement by approximately 11%,10%and 14%in the north(N),east(E),and up(U)component,respectively.Meanwhile,the MBS has a higher internal coincidence accuracy.For the 30-minutes scheme,the closure errors of repeated baselines and triangle closed loops have a remarkable performance improvement by approximately(48%,59%,and 12%),and(54%,65%,and 10%)in the N,E,and U component,respectively.(2)Using the independent single-differenced(SD)observation between stations,the equivalent transformation is implemented to eliminate the difference parameters of the receiver clock errors.The equivalent transformation of the SD and DD observation equations is obtained,which provides a rigorous theoretical model for MBS.The SD equivalent observation equations further expand the traditional GNSS data processing method.In theory,the two models for MBS are absolutely equivalent to the traditional DD positioning model.If the optimal structure of least square spanning tree and the complexity of data processing are considered,the MBS model with SD observations is the optimal model for data processing of the multiple stations.Based on the SD equivalent observation model,the kinematic positioning model for multiple reference stations is further developed.The constraint equation and compression model of the multi-baseline real-time kinematic(MBRTK)positioning are given.The new algorithm can greatly enhance the model strength of the MBRTK,improve the success rate of ambiguity resolution,and further improve the accuracy and availability of the RTK positioning.The results of zero baselines and ultra short baselines show that the performance improvement of the MBRTK is about 8-21%,0-40%and 3-40%in N,E and U component,respectively.The results of the kinematic experiment show that MBRTK positioning has higher positioning accuracy and success rate than the tranditional single baeline RTK positioning.The additional reference stations will inevitably improve the model strength of the kinematic positioning,and improve the positioning accuracy and availability.(3)Based on the SD equivalent observation model,the calculation procedure of equivalent transformation matrix is deduced in detail.Considering the calculation properties of unit matrix and block diagonal matrix,a fast construction algorithm of equivalent transformation matrix is developed.Combined with the fast construction method of SD observation weight matrix of multiple stations,a fast construction method of the simplified equivalent observation equations(SOE)is proposed.Meanwhile,considering the block diagonal characteristics of the weight matrix of SD observations,the sequential Kalman filter(SKF)is used to estimate the unknown parameters.Thus,the simplified processing algorithm for multi baseline relative positioning is obtained.The results demonstrate that the sequential multiple baseline solution(SMBS)algorithm can significantly accelerate the processing procedure of RTK positioning and reduce the processing time for a single epoch.The results of the static experiment show that the SOE and SKF strategies have an average efficiency improvement by approximately 74.7%and 49.6%.Moreover,the results of kinematic experiment show that it takes about 0.298 ms to establish the equivalent observation equations by the MBS model but only about 0.117 ms for the SMBS algorithm.There is an efficiency improvement by approximately 60.6%for the SOE strategy.Meanwhile,it requires about 25.2 ms to complete the filtering calculation for the MBS model but only about 10.6 ms for the SMBS algorithm.An efficiency improvement by approximately 58.0%is achieved for the SKF strategy.Considered the real-time requirements for a single epoch,the new SMBS algorithm is highly efficient,especially for multi-baseline positioning with multiple satellites and stations.(4)According to the frequency characteristics of GPS/BDS/Galileo,the tightly-coupled integration model of MBS with overlapping frequencies is derived,and the equivalent observation equation of mixed constellation is obtained,which can implement the real-time estimation of the DISB parameters of different satellite systems.By correcting the DISB parameters,the model strength of the tightly-coupled integration for MBS positioning can be enhanced,and the postitionig performance can be further improved.The results of static experiment show that the performance of tightly-coupled integration with correction of DISB parameters is improved by 10-20%in the N,E and U component.The results of kinematic experiment show that the average performance improvement of the tightly-coupled integration for GPS/BDS-3/Galileo positioning is about 4.8%,0.0%and 9.7%in N,E and U component,and the success rate of ambiguity resolution(AR)is about 91.2%,which has an improvement by approximately 10%.(5)Combined with the SD equivalent observation model and INS positioning model,the RTK/INS tightly-coupled integration model for GPS/BDS/Galileo is derived,including the establishment of state equation and observation equation.The kinematic positioning is enhanced by the high-precision position and attitude information which is calculated by inertial measurement unit(IMU)in a short time.Meanwhile,the system errors of IMU can be corrected by GNSS positioning.The RTK/INS tightly-coupled integration model can be obtained and the processing flow is given in detail.The results of kinematic experiment show that the INS can procide a better predictive values of state parameters and the RTK/INS tightly-coupled integration positioning has a higher positioning accuracy and success rate of AR.Currently,China is actively promoting the construction of national PNT system.The precise GNSS positioning for multi-system and multi-station and GNSS/INS integrated positioning are the important issues of national PNT system.The further research,refinement and application of the above positioning model will promote the construction of China's PNT system to reach new heights.