Abstracts Submitted: 510
Number of Users: 710
View Abstracts Submitted
Back to home Page
Acoustic Vector Sensor (AVS) is a device used to determine the Direction of Arrival (DoA) of an acoustic source by measuring the particle velocity and pressure. There are two types of AVS being used: Pressure gradient (P-P) and Inertial sensor (P-U ). In P-P based AVS, particle velocity is calculated by taking the pressure difference and pressure is calculated by taking the mean pressure value of two closely spaced hydrophone. In the P-U based method, particle velocity and pressure are calculated by using inertial sensor such as (accelerometer or geophone) and hydrophone respectively. In this paper, we have done simulation studies to compared the DoA estimation accuracy/errors of P-P and P-U based AVS by using different DoA estimation techniques such as beamforming techniques (Delay and Sum, MVDR, MUSIC and Cardioid) and Intensity-based method (Arc Tangent). Above comparisons has been done for different underwater scenarios by varying the sea-states and in the presence of Correlated and Un-correlated noise. Studies has been done for Correlated noise because of closeness of the hydrophones. Underwater ambient noise characteristics has been assumed to be colored Gaussian noise in nature. The transmitted signal from an acoustic source is assumed to be tonal of frequency range from 100 Hz to 5 kHz. Signal to Noise Ratio (SNR) has been varying from -10 dB to 20 dB and the distance between two closely spaced hydrophones has also been varying. The obtained results have been compared in terms of Root Mean Square Angular Error (RMSAE). We have compared different DoA estimation techniques and it is found that Arc Tangent technique outperforms other methods in terms of DoA estimation accuracy. The obtained accuracy of P-U based method is better than the accuracy of P-P based method. Performance of P-P method improves significantly on reducing the distance between two closely spaced hydrophones (under high SNR conditions). Also, it is observed that in presence of correlated noise, RMSAE of P-P method is much less as compared to the RMSAE of uncorrelated noise.
© Copyright 2017 All Rights Reserved