Adaptive extended Kalman filter for aided inertial navigation system

Abstract

In this work inertial and GPS data fusion using EKF is presented. The EKF algorithm performance is very sensitive to it initial parameters choice. It's quite difficult to determine optimal parameters for Q and R matrices of EKF. This problem becomes more serious if low cost GPS receiver and MEMS IMU are used as integrated navigation system. The adaptive filter helps to solve this problem in optimal manner. The results of adaptive EKF performance during GPS signal outages are presented. The comparison of conventional and adaptive EKF for sensor data processing was made. It was shown that adaptive EKF outperforms conventional EKF in case of GPS signal outages.