Integrated geophysical navigation for autonomous underwater vehicles

Abstract

While GPS is not available underwater, underwater vehicles need an alternative algorithm to compensate for the integration error of the inertial navigation system and to find an accurate position. This study deals with geophysical navigation, which references terrain elevation, gravity, and magnetic fields. Representing the source of each anomaly as a bundle of line sources, the magnetic and gravity anomaly field was generated. Downward continuation with a planar assumption was applied to the surface distribution of anomalies to generate a layer of estimated underwater anomaly. To prevent the divergence of the filter because of the continuation error, a bias filter was applied. An augmented state filter and a two-stage filter were used as bias filters. The covariance of bias change was set proportion to the root mean square of innovation. The simulation analysis showed that the stability of the filter was improved through estimation of bias. By combining the Rao–Blackwellized terrain referenced navigation, integrated geophysical navigation was tested from the environment generated from real terrain information.