Augmented EKF based SLAM system with a side scan sonar

Abstract

In order to counteract and compensate the position drift in the motion of an autonomous underwater vehicle (AUV), the system named Simultaneous Localization and Mapping (SLAM) is necessary for an AUV equipped with a side scan sonar (SSS) and a inertial navigation system (INS). In this paper, we propose a special formulation of SLAM, which combines the position error model based on side scan sonar and the inertial navigation system of AUV as input data, instead of recomputing an estimate from fusing raw navigation sensor information with sonar data directly. This is accomplished by formulating a joint estimation problem, combining an error model of the native navigation system and object positions measured by means of the SSS. Up to now, Using the extended Kalman filter (EKF) to solve the SLAM problem of AUV is the most common means. However, with the increase of landmarks, the computational cost of EKF increases rapidly. The computation of augmented EKF method used in this paper increases linearly with the number of landmarks. Simulation result is presented to validate that our SLAM method reduces navigational errors successfully, at the same time computational complexity is limited to a reasonable level.