수중로봇의 정밀한 위치추정을 위한 USBL 사용 운항 중 헤딩 보정방법

Abstract

This paper proposes a method to perform continuous alignment during navigation by utilizing USBL position and the ICP algorithm to enhance the navigation performance of safety-support robots for underwater leisure activities. Conventional underwater navigation using USBL typically employed fusion methods based on INS (Inertial Navigation System) and Kalman filtering. However, the conventional fusion method using the Kalman filter causes unexpected position changes whenever USBL positioning data are received. This can adversely affect control for autonomous navigation. Moreover, inertial navigation systems without additional alignment processes after initial alignment significantly degrade the accuracy of fusion-based navigation when the initial heading estimation is inaccurate. To solve these issues, this paper proposes a method to continuously calculate and correct heading offsets during navigation by using USBL position and INS position after the initial alignment. The experimental results demonstrated that the navigation accuracy improved by up to 75.49% compared to the existing method. Additionally, the method was shown to enhance navigation performance even with reduced initial alignment time due to the continuous alignment process during navigation.