Research On Error Compensation And Anti-interference Of GNSS/MIMU Integrated System

GNSS/MIMU integrated system has characteristics of low-cost,high-autonomy,high-precision and good-dynamic.So it becomes one of the mainstream solutions for the low-end navigation.In this thesis,the problem of error-compensation and anti-interference in GNSS/MIMU system is studied.It mainly covers the micro inertial device error calibration and modeling method,MIMU attitude fusion technology and the solution of vehicle motion interference problem,the fast solution method of nonlinear carrier phase attitude equations,the processing method of reducing noise correlation based on ambiguity function,design of double-difference-carrier-phase smoothing filter model for MIMU aiding GNSS attitude determination,the novel scheme of integrated attitude measurement based on dual extended kalman filter,etc.This thesis is composed of six chapters to describe the above research in detail.The main research results are summarized as follows:1.In the low-frequency modeling of random error,the wavelet multi-resolution method is used to solve the problem of high-frequency interference when the sample correlation is extracted,and the optimal sample length is determined by the sample variation coefficient method.Finally,smoothing,filtering and prediction of measurement data are carried out by setup model through kalman filter.The whole modeling process of random error is improved.2.For the low-cost MIMU attitude measurement system,the dual extended kalman filter attitude fusion algorithm is designed.The measured values of accelerometers and magnetic sensors are used as kalman filter states,and the linearization of the observation equation is avoided.Therefore,the method is more intuitive than the Euler angle or the four element method,and the structure is clear and easy to control.In addition,the state convergence model of vehicle motion is designed with methods of gyro compensation and statistical analysis so that the problem of motion acceleration interference in attitude measurement is solved.3.Regarding to the complication of the solving process of the GNSS carrier phase attitude equation caused by nonlinear factors,the rotation matrix method is proposed based on the space geometry model.This method is simpler,faster and more efficient comparing with the analytical method and more advantageous in the correlation analysis of noise error.4.The ambiguity function method(AFM)has more freedom in the calculation of the integer ambiguity,but its solution of the success rate is relatively low since it lacks the analysis of the relationship between the noise error and the satellite space geometry.In this thesis,the method of ambiguity function reduction correlation is proposed to solve these defects.Compared with the LAMBDA method,the computational complexity is greatly reduced,more freedom was given to the algorithm and programming in the same reasonable decorrelation process is much easily to be implemented.5.In MIMU aiding GNSS attitude determination,the smoothing filter model based on MIMU is designed to suppress the noise interference of GNSS double difference carrier phase,which improves the accuracy of the attitude equation solution.Low-cost MIMU under its own conditions provides reasonable depth assistance to the data source part of the GNSS attitude measurement system instead of the constraint condition for integer ambiguity searching,the attitude angle provided by the attitude and heading reference system.Finally,the novel scheme of low-cost GNSS/MIMU integrated attitude determination system is designed,and the performance of the system is verified by experiments.