Research On In-motion Initial Alignment Technology Of Inertial Navigation System Based On Laser Doppler Velocimeter

Due to the advantages of high velocity measurement accuracy and fast dynamic response,the laser Doppler velocimeter(LDV)is expected to replace the odometer and form high-precision land integrated navigation system with strapdown inertial navigation system(SINS).At present,almost all land integrated navigation systems use the initial alignment method of static base alignment,which will greatly affect the mobility of the carrier and restrict the rapid response and precision attack ability in military operations.How to use external sensors to aid SINS to complete in-motion alignment has become a research hot spot in the field of integrated navigation.In this paper,the in-motion alignment method of LDV/SINS integrated navigation system is studied to improve the accuracy,speed and robustness of in-motion alignment.The main research contents done in the thesis are as follows:(1)On-line calibration method of LDV.A highly robust calibration method based on position observation is proposed to improve the velocity measurement accuracy of LDV,and then to improve the accuracy of in-motion alignment.Two groups of vehicle tests verify the accuracy and robustness of the proposed calibration method.When the sensor output is normal,the maximum horizontal position error of the two groups of tests is less than 0.04% of the total mileage.When the output of LDV and Global Positioning System(GPS)has outliers,the maximum horizontal position error of the two groups of tests is less than 0.06% of the total mileage.(2)Model improvement of in-motion alignment.The scale factor error and installation angle errors are added into the process model and measurement model of in-motion alignment,which is beneficial to improve the accuracy of in-motion alignment.(3)Filtering algorithm improvement of in-motion alignment.The square root of Unscented Quaternion Estimator(USQUE)is processed to ensure the stability of numerical values and the positive definiteness of covariance matrix.The measurement noise covariance matrix is adaptively processed to improve the accuracy,speed and robustness of in-motion alignment.A double gain structure is proposed to speed up the speed of the in-motion alignment by using the posterior measurement residual information in filter.An adaptive H-infinite filtering method is proposed to improve the robustness of in-motion alignment.Two groups of vehicle tests show that the alignment performance of the proposed in-motion coarse alignment scheme is better than that of the other comparison scheme,and the heading angle errors of the two groups of tests are less than 0.03° and 0.09° respectively after the 300 s coarse alignment process.