Performance comparison of nonlinear estimation techniques in terrain referenced navigation

This paper studies a terrain referenced navigation which estimates a status of aircraft by using terrain elevation database and radar altimeter measurements. Terrain referenced navigation algorithm usually divides into two modes which are acquisition mode and tracking mode. In the tracking mode, states are estimated in real-time and extended Kalman filter(EKF) is a well-known navigation method for estimation. However, when conventional EKF algorithm is applied in terrain referenced navigation, estimation error can diverge because of nonlinearity of terrain elevation. To remedy this divergence problem, various nonlinear estimation techniques have been studied like stochastic linearization, bank of Kalman filters and unscented Kalman filter(UKF). In this paper, above three nonlinear estimation techniques are introduced and applied in terrain referenced navigation. Simulation results show improved navigation performance of three methods by comparing it with conventional EKF. Finally, pros and cons of each nonlinear estimation methods are analyzed by comparing their computation time and navigation performance.