Non Linear Optimum Filtering and Smoothing Based IMM in GPS Navigation
The information from Global Positioning System (GPS) has been increasingly used in many applications. Navigation is one of the important applications. To estimate the state of the vehicle, Kalman Filter (KF) is used traditionally. Since KF gives inaccurate results during non linear motion of the vehicle, an Unscented Kalman Filter has been introduced. An interacting multiple model algorithm is more efficient in dealing noise uncertainty than the conventional single model filter. In this paper, a novel estimator called Interacting Multiple Model Unscented Kalman Two Filter Smoother (IMM-UKTFS) is introduced to improve the navigation accuracy. The performance of the proposed method is verified by considering four types of models.