Data Fusion Algorithm for GPS/DR Integrated Vehicle Navigation System
Kou Yanhong Zhang Qishan Li Xianliang(Dept. of Electronic Engineering, Beijing University of Aeronautics and Astronautics)
An adaptive federated Kalman filter model for GPS/DR integrated vehicle navigation system was established. Attention was focused on the filter algorithm. To improve the precision and reliability, data fusion techniques such as subsystem state evaluation, adaptive information distribution, error compensation, iterative extended Kalman filter, resist outliers, and U D covariance decompose were used in the algorithm. To solve the problem of filtering divergence, a method to estimate the statistical feature of measurement noise was introduced. Theoretical analysis and semi physical simulation results demonstrated that the algorithm is efficient in precision, reliability, adaptivity and real time processing rate.