Navigation of autonomous underwater vehicle using extended Kalman filter

Abstract

To navigate the Autonomous Underwater Vehicle (AUV) accurately is one of the most important aspects in its application. A truly autonomous vehicle must determine its position which requires the optimal integration of all available attitude and velocity signals. This paper investigates the extended Kalman Filtering (EKF) method to merge asynchronous heading, attitude, velocity and Global Positioning System (GPS) information to produce a single state vector. Dead reckoning determines the vehicle's position by calculating the distance travelled using its measured speed and time interval. The vehicle takes GPS fixes whenever available to reduce the position error and fuses the measurements for position estimation. The implementation of this algorithm with EKF provides better tracking of the trajectory for underwater missions of longer durations. © 2010 Springer-Verlag.