A Fusion Structure of TRN and IBN with Adaptive Logic Based on Altitude Gradient

Abstract

We present a new navigation structure using dual databases for aerial vehicle, which include the simple switching technique. It is possible to use the various database for an alternative navigation system. We utilize the terrain referenced navigation (TRN) and image-based navigation (IBN). The TRN has been researched for fighter aircraft or cruise missiles. And the IBN is widely researched recently during the computer science technologies are rapidly grown up. Generally, the TRN is very effective on the environment with high altitude fluctuation with uniqueness. However, the IBN provides a very accurate pose of the vehicle for the environment with a lot of features in the input camera image. However, both algorithm has disadvantages that are greatly influenced by the operating environment. We propose a fusion structure using both algorithms with adaptive logic based on decision making logic. The decision variable includes the roughness and uniqueness of the environment. We use a stochastic linearization (SL) method in TRN also, which is widely known as an essential method on TRN using digital elevation map (DEM). In the case of the IBN, we use the speed up robust feature point (SURF) algorithm for the matching process, which is invariant to translation, rotation, scale, and illumination. Both measurement models are fused with the extended Kalman filter(EKF). We conducted simulations to validate the proposed structure using DEM and aerial projective map. Two kinds of databases are provided by the Ministry of Land, Infrastructure and Transport in the Republic of Korea.