Adaptive Beamforming for Sonar Audio
Chaoying Bao · 2005
Sonar audio is a major tool used by sonar operators to assist in classifying acoustic contacts. In this paper we discuss some issues that arise when adaptive beamforming is used for sonar audio. Frequency domain beamforming is used to reduce computational cost, and diagonal loading of the cross-spectral matrix is used to obtain the best quality output. The effectiveness of a robust Capon beamformer (RCB) with nonuniform loading is compared with that of a minimum variance distortionless response beamformer with uniform loading (MVDRUL). For the tests described in this paper, which involve a signal together with a strong interference and background noise, RCB produced the highest quality output and was more robust than MVDRUL. On the other hand, conventional beamforming (CBF) failed to provide a satisfactory output for any test.