Research On GNSS/INS Integrated Navigation Performance Improvement Technology

Based on the GNSS/INS integrated navigation positioning terminal architecture assisted with micro atomic clock,an overall scheme for improving the performance of the integrated navigation system is designed,the GNSS satellites selection algorithm in integrated navigation is studied,the GNSS/INS adaptive Kalman filter algorithm based on innovation(IAKF)is proposed and studied,and a real-time data acquisition and processing software based on the BD930 receiver chip is developed in this paper.On this basis,the vehicle-borne Strapdown GNSS / INS test platform is built and the onboard road test is carried out.The specific works are as fellows:(1)Firstly,the basic principle of GNSS/INS integrated navigation assisted with micro atomic clock is studied,and the two navigation modes of loose integration and tight integration of GNSS/INS integrated navigation are analyzed.Then,the requirement analysis of performance improvement of GNSS/INS integrated navigation assisted with micro atomic clock is given.Finally,the overall design scheme of performance improvement of GNSS/INS integrated navigation system assisted with micro atomic clock is developed.(2)According to the different requirements of GNSS/INS loose and tight integrated Navigation,the satellites selection algorithms suitable for loose integrated and tight integrated are proposed respectively.A two-dimensional convex hull satellites selection algorithm based on pseudo-range detection is proposed for loose integrated navigation,and a weighted H matrix satellites selection algorithm based on weighted GDOP is proposed for tight integrated navigation.Through experimental analysis,the availability and effectiveness of these two algorithms are verified.(3)The principle and characteristics of Kalman filter are analyzed,and the influence of Kalman filter parameters on filtering effect is analyzed.Then,the algorithm principle of fading Kalman filtering(FKF)is analyzed,and an adaptive Kalman filter algorithm based on innovation is proposed to solve the deficiency of FKF.Finally,through experimental analysis,it is verified that the IAKF algorithm proposed in this paper has better filtering accuracy and filtering stability than the KF algorithm for integrated navigation.(4)According to the requirements of integrated navigation,the real-time data acquisition and processing software of Trimble BD930 receiver chip is analyzed.The software is designed in detail by using modular method.Finally,the communication connection,data acquisition,receiver control,protocol analysis,data solution and data storage functions of the software are realized respectively,and MFC Graphical interface language is used to build the front-end interface of software users.On the basis of the above work,in order to verify the correctness and usability of the relevant software and algorithms in this paper,an experimental platform of vehicleborne Strapdown GNSS/INS integrated navigation assisted with micro atomic clock is built,and the data collection and data processing of on-board road test are completed.The experimental results show that the two-dimensional convex hull satellite selection algorithm based on pseudorange detection and the weighted H matrix satellite selection algorithm proposed in this paper can well identify and eliminate the pseudorange outliers.During the test period,in the presence of pseudorange interference,the usege of the above satellite selection algorithms can improve the positioning accuracy of the integrated navigation from hundreds of meters to meters.In addition,the IAKF algorithm proposed in this paper has better filter accuracy and filter stability than the KF algorithm for GNSS/INS loose integrated and GNSS/INS tight integrated navigation.Using the IAKF algorithm for GNSS/INS loose integrated navigation,the positioning error(longitude-latitude-height)and velocity error(ENU)are reduced by 52.00%,38.32%,22.80%,63.30%,57.16% and 3.03% respectively compared with using the KF algorithm.And using the IAKF algorithm for GNSS/INS tight integrated navigation,the positioning error(longitude-latitude-height)and velocity error(ENU)are reduced by 39.90%,47.78%,55.84%,14.58%,5.53% and 2.83% respectively compared with using the KF algorithm.