A novel navigation method for autonomous underwater vehicle in the middle water column

The navigation problem of autonomous underwater vehicles (AUVs) is one of the main reasons that restrict the development of all underwater platforms. In the middle water column where the distance from the seafloor exceeds the Doppler Velocity Logs (DVL) bottom lock range, the methods of navigations are mainly relying on acoustics and inertial measurement units (IMUs). This paper proposes a novel navigation method in the middle water column. In this method the sea water is stratified into several reference layers in the vertical depth direction and each reference layer is assumed to have the same current information in a short time. Based on this assumption, the current model can be greatly simplified. With reobservation of these reference layers with the ADCP (Acoustic Doppler Current Profile), the current velocity and vehicle velocity can be estimated by fusion of other low cost sensors such as attitude and depth sensors during the descent. The integration of vehicle velocity can be used for position estimation directly in case of emergency circumstances, and it can provide velocity aiding for the IMU or other methods to increase navigation accuracy. The method has been functional tested in a field experiment in the Thousand Island Lake. The results show that the navigation method is effective.