Maximum likelihood estimation to denoising channels in beamforming circular array

Abstract

The delay and sum beamforming is the most simple technique in direction of arrival (DOA) Estimation. Although its performance on spatial discrimination is poor, compared to other beamforming, delay and sum still is used in large operating sound navigation and ranging (SONAR) because of its low computational cost. A circular hydrophone array (CHA), commonly used in SONAR system, is an attractive alternative to provide a more uniform directive response over all azimuth angle. This array is analyzed here, working with experimental data, acquired in a acoustic tank and in the sea. Maximum likelihood estimation (MLE) is applied to denoising noisy channels, summing up them after in delay-and-sum. First of all, a noise in an acoustic tank is considered to represent the hydrophones, cables, and acquisition system noises. Then, an environmental noise is collected in the sea. The MLE use both of them to calculate the weights in the beamforming. A boat is used to running around the array, and the DOA of the uniformly weight and MLE in delay-and-sum shows the performance improvement.