jasa logo

Acoustic Vector Sensors Underwater Communications in the Makai Experiment

F.A. Bozzi fabozzi(at)ualg.pt
ans
S.M. Jesus sjesus(at)ualg.pt
LarSys, Universidade do Algarve, Campus de Gambelas, PT-8005-139 Faro, Portugal.

Comments: download pdf.
Ref.: in Proceedings of Meetings on Acoustics (POMA), Vol.44(1), 070011, October 2021.

Abstract
The performance of a vector sensor array (VSA) receiving communication signals in the Makai experiment is quantified. Makai experiment was a four-week field experiment, which took place off the coast of Kauai Island, Hawai, in 2005. The data analyzed here refers to the communication test carried out on September 23rd. The VSA is composed of four accelerometer-based vector sensors, which was tied to a drifting research vessel in a shallow water area. A moored bottom source was used to transmit signals from 8k to 14kHz frequency range.

The present study explores the VS beam steering method. In this method, pressure and particle velocity channels are weighted-combined leading to directional gain. DoA estimation uses the Minimum Variance Distortionless Response (MVDR) beamforming as a pre-processing for the VS beam steering. Noise normalization uses the Maximum Likelihood Estimation (MLE) for denoising. Then, the communications chain for coherent Phase Shift keying (PSK) modulation is composed of synchronization and Doppler compensation, and a single Decision Feedback Equalizer. Pressure and particle velocity components of VS have been used in beamforming or diversity techniques. These techniques act somehow in opposite ways, depending on the channels cross-correlation assumption. In this work, a beamforming technique for communications is addressed.

Communication results are analyzed by the bit error rate (BER) for the pressure- only array, a single VS, and the VSA. The ranges analyzed (230m and 910m) present delay-Doppler spread that can achieve 20 symbol intervals and 10Hz. These characteristics make synchronization and Doppler tracking a challenge in coherent modulation. Thus, velocity channels are used for improvement. Bearing fluctuation issue is noticed in the operating frequency. Fluctuation of 90 degrees is verified, and its impact on demodulation is high-lighted. Denoising has shown benefits since the VSA is under the ship noise and the channel-gain of the acquisition system is unknown. The study shows that even with the bearing fluctuation, a single VS can estimate the source direction with an accuracy of 50 degrees. It is shown that a single VS provides similar communication performance comparing to four pressure sensors. The BER for the shorter range varies from 1% to 5%, depending on the number of sensors used. For the long-range, BER does not achieve values lower than 9%. It is noticed that in a multipath environment, steering to the DoA may not lead to the lowest error for communications.