Extended Kalman filter applied in the navigation of an AUV

Abstract

This work deals with the navigation problem of an autonomous underwater vehicle. Two state estimators are proposed like solution, using sensor fusion based in Extended Kalman Filter. The state estimators use measures of the following sensors: an inertial measurement unit, a Doppler effect velocity sensor, a depth sensor and a compass. The first state estimator, estimate the attitude independently of the velocity and depth estimation. In the second estimator, a coupling in velocity and attitude equations is considerate in the Extended Kalman Filter. To design and test the proposed state estimators, was employed the database of the Pirajuba autonomous underwater vehicle, This database contains the record of the vehicle sensors during sea tests. The results of a numeric simulation with this database validate the proposed state estimators in this work. Finally was made a comparative analysis of these state estimators.