U-GPF INFORMATION FUSION ALGORITHM FOR GPS/DR INTEGRATED POSITIONING SYSTEM

摘要

U-GPF is proposed for GPS/DR integrated positioning system to improve its performance.It is based on the Gaussian Particle Filter (GPF) and Unscented Kalman Filter (UKF).UKF is used to calculate the estimate parameters value and covariance matrix in the observation update, and the distribution function is sampled as the importance density function for GPF.Simulation results show that U-GPF and UKF has similar accuracy on the Gaussian noise, but they are better than Extended Kalman Filter (EKF).However, for the non-Gaussian noise, U-GPF has higher accuracy than UKF and EKF.The collected real data is applied to validate the U-GPF and the results are consistent with the theory analysis and simulation result.