This paper presents a hybrid underwater navigation system for unmanned underwater vehicles. The navigation system consists of an inertial measurement unit (IMU) and a Doppler velocity log (DVL) along with the magnetic compass and depth sensor. A navigational system error model is derived to include the bias errors of accelerometers and gyros in the IMU, the scale errors of the DVL, and the bias errors of the compass and depth sensor. An extended Kalman filter is used for the error model in indirect feedback, where the measurement equation is composed of the navigational state errors and system parameters. This paper demonstrates the effectiveness of the inertial-Doppler navigation system through a lawn mowing mode in simulation and circular motion with rotating arm tests in the experiments. The hybrid navigation system equipped with a low-quality IMU sensor improves navigational performance, where the bias error of the accelerometers and the gyros of the IMU are 1.0 mg and 1.0 deg/h, respectively. The tracking error is less than 1.0 m for 10 min of circular motion from the experiments.

INTRODUCTION

Since a radio wave cannot pass through the operating environment of underwater vehicles, acoustic positioning systems such as LBL (long baseline), SBL (short baseline) and USBL (ultra short baseline) are widely used for underwater vehicles (Milne, 1983). These acoustic positioning systems have no accumulative error, while they have a low update rate and limited accuracy in position estimation. In general, the LBL system is widely used to localize underwater vehicles for its reliability and relatively higher accuracy, but we have to install acoustic sensors at a predetermined position on the sea bottom, and the working range is limited by the distance of the deployed sensors.

This content is only available via PDF.
You can access this article if you purchase or spend a download.