Bearing estimation in very shallow waters with an AUV mounted Acoustic Vector Sensor

This paper reports first in-field results of a passive acoustic system installed on an Autonomous Underwater Vehicle (AUV) for tracking low frequency underwater sources in very shallow waters. The problem of passive acoustic detection and Direction-of-Arrival estimation has been approached by using a DIFAR (Directional Frequency Analysis and Recording) vector sensor, which consists of an omnidirectional hydrophone and two dipole sensors oriented orthogonally on the horizontal plane. A mechanical and electrical design study has been conducted, identifying the most appropriate technical solution for the implementation of the DIFAR vector sensor on board of an AUV by minimizing the effects of interference between the vehicle and the payload. Bearing estimation algorithms have been implemented both in time and frequency domain. Tests in real scenarios have been conducted. Results of these measurements show that the bearing estimates are consistent with the reference ground truth. The computationally light methods implemented and the low-cost instrumentation used for data processing are suitable for being used in real-time processing on heterogeneous autonomous vehicles.