Research On Filtering Algorithm Of GPS/SINS Tightly-coupled Integrated Navigation

Because the navigation function of GPS and SINS has strong complementary,the combination of them can enhance advantages and avoid disadvantages.The GPS/SINS integrated navigation system has superior performance,is widely used in various countries around the world,become the hotspot and keystone of the high accuracy and reliability of navigation systems research.In the GPS/SINS integrated navigation system,The study of filtering algorithm and its engineering application is the key link in the process of improving system performance.Firstly the research background of this article is presented: gives the overview of the current research of integrated navigation system and the filtering algorithms.Describes the common coordinate system and coordinate system conversion relationship;Respectively expounds the working principle and error analysis of GPS and SINS.On the platform of matlab to simulate the vehicle trajectory,builds the model of the SINS inertial measuring element group and simulates.Only uses the SINS for navigation positioning,analyzes the advantages and disadvantages,to verify the correctness of the model established in this paper.Set up the model of the GPS/SINS tightly-integrated navigation system,then the state equation and measurement equation are given,designs the navigation and positioning system simulation process,laid the foundation for the analysis of filtering algorithm below.Then introduces the work principle and algorithm process of KF,analyzes three nonlinear filtering algorithms:EKF,UKF and PF.Then summarizes the advantages and disadvantages of the filtering algorithms and their application models.Focuses on the research of accuracy and computational-complexity.Based on the flops theory,analyses the complexity of KF,EKF,UKF and PF algorithm,deduces the calculation formula of accurate complexity.Then simulates to compare the state of the mean square error value,computational-complexity,and status tracking capability,verifys the correctness of the analysis for filtering algorithm.Provides a theoretical basis for the integrated navigation filtering algorithm selection.Analyzes the nonlinearity of measurement equation in tightly-integrated navigation in this paper: In tightly-integrated navigation system,if the position error is less than km order of magnitude,the error introduced by the linearization of measurement equations can be ignored.Based on this conclusion,linearizes the measurement equation and builds a hybrid model consisting of a nonlinear equation of state and linear measurement equation.Based on the ideas of decomposition of the structure,optimizes the UKF algorithm,combined with KF and sequential processing,a sequential SUKF algorithm is proposed,using UKF to do state-step-prediction,using KF to update the measurement,and the sequential processing method to improve measurement update process.Finally,the EKF,UKF algorithm and the sequential SUKF were simulated in GPS/SINS tightly-integrated navigation system,then analyzes and compares the navigation parameter estimation error of three filtering algorithm.The results show that: the accuracy of the sequential SUKF algorithm proposed in this paper is the same as the UKF algorithm,but real-time performance of the algorithm was increased by 30% compared to UKF.