A leveling algorithm for an underwater vehicle using extended Kalman filter

Abstract

Recently, the strapdown Attitude Reference System (ARS) became popular as an economic system for a small, light, low-cost system like an underwater vehicle. The ARS provides attitude information updated from the initial attitude. So, the initial attitude errors have a great effect on the ARS. In this paper, the leveling algorithm for compensating initial attitude errors using inertial sensors and a speed log is presented. The meaning of leveling in this paper is to acquire the two attitude angles of roll and pitch of the vehicle during its motion. The linear system model for the leveling is derived in order to apply extended Kalman filter (EKF) which is known to have many desirable properties. The simulation shows that the leveling algorithm using EKF is adequate by virtue of its property of decreasing attitude errors.