Research On Model Error Of Kalman Filter For Vehicle Integrated Navigation System

Real-time accurate positioning and navigation of vehicles is the key technology to realize intelligent driving.The vehicle integrated navigation system based on KF GNSS / INS has become an ideal navigation system because of its low cost and high-precision navigation output for a long time.Implementing the combined navigation of Kalman filter algorithm requires a navigation system model with sufficient accuracy as a premise.Therefore,from the perspective of inertial device modeling error,this paper proposes an improved KF algorithm,which realizes the simultaneous estimation of the navigation error state quantity and the scale factor model error,and performs real-time compensation for the navigation error state quantity.Corrected inertial device errors to reduce navigation errors during GNSS failures.In this paper,the positioning principle and error characteristics of GNSS are analyzed,and the necessity of establishing GNSS/INS integrated navigation is explained.Then,based on the basic principle of inertial navigation,it is determined that the measurement error of inertial device is the main reason affecting the calculation accuracy of inertial navigation,and the influence characteristics of inertial device error are studied in detail,which provides a basis for establishing the model error observer.Then introduced the relevant theoretical knowledge of Kalman filter,and studies the modeling of GNSS/INS integrated navigation system based on KF.In the modeling stage,INS differential equation is firstly determined.Then,after comprehensive consideration of the research objective,the accuracy and complexity of the model,the measurement error of IMU was reasonably modeled.Finally,the observation equation of the system was established by using the solution results of GNSS.Then,through formula derivation and theoretical analysis,the importance of estimating the scale factor model error to improve the model accuracy and reduce the estimation error of navigation state quantity is described in detail.Then,the decoupling estimation algorithm of state quantity and unknown deviation based on KF is introduced,and applied to integrated navigation system,the decoupling estimation algorithm of navigation error state quantity and scale factor model error is obtained,and the performance of the algorithm is analyzed.Finally,the simulation results show that under the influence of the scale factor model error source,the improved algorithm can effectively estimate and compensate the model error,and can make the sensor error estimation more accurate and stable.The model precision of integrated navigation system is improved after the sensor error estimation compensation.It can effectively reduce the divergence of integrated navigation results caused by sensor measurement errors when GNSS failure occurs,and maintain certain navigation accuracy.