Modified Adaptive Kalman Filtering with Multiple Fading Factors in 2-D Land Vehicle Navigation

Extended Kalman Filtering (EKF) and its modified techniques have been used worldwide for various vehicle navigation tasks. Several adaptive algorithms have been presented to overcome the issue of inaccurate a priori knowledge of the statistical noise of the system and measurement models in EKF. Among them, Single Fading factor Adaptive Kalman filter (SFAKF) and Multiple Fading factor Adaptive Kalman Filtering (MFAKF) techniques introduce the concept of fading factors to significantly reduce the influence of these abnormalities on the state estimation. This research work first demonstrates a successful implementation of SFAKF and MFAKF algorithms in 2-D land vehicle navigation. Then, a Modified Multiple Fading factor Adaptive Kalman Filtering (MMFAKF) algorithm has also been proposed which has modified the fading factor function to produce superior results over other competing algorithms, thus increasing reliability on estimation using real time data acquired from GNSS/ INS systems.