A study on the estimation method of noise covariance for underwater navigation systems

In this paper an adaptive receding horizon Kalman filter is newly derived for underwater navigation systems, which are subject to statistically unknown noises contained in both system and measurement equations. The adaptive receding horizon Kalman filter gets feedback on the method. The filter is applied to a navigation system based on a LBL (long baseline) system. Experiment and simulation show that the derived filter estimates covariance of the system and measurement properly.