AUV navigation in the presence of a magnetic disturbance with an extended Kalman filter

Abstract

This paper presents a novel extended Kalman filter (EKF) used for navigation of an autonomous underwater vehicle (AUV). The AUV contains a magnetic compass, which is susceptible to magnetic disturbances, and an angular velocity sensor, which exhibits drift if solely integrated to estimate heading. To address these problems, the presented EKF fuses the information from these sensors in order to produce a more accurate estimate of heading, and learns a heading bias in real-time. The presented method has two distinct advantages. First, the heading bias can correct for errors from a poorly calibrated magnetic heading sensor. Second, the angular velocity information improves heading estimation, especially in the presence of a magnetic disturbance. Because the AUVs presented here are designed to acquire magnetic field measurements, this second advantage is of particular importance. In both simulation and experimental testing, the presented EKF learned a calibration bias for the magnetic heading sensor and improved heading estimation in the presence of a magnetic disturbance.