Study on robust adaptive beamforming based on acoustic vector sensor array
Yaohui Lyu, Jiyuan Liu, Shanguo Gao, Jun Ling Song, Minghua Lü · OCEANS 2017 - Aberdeen · 2017
In order to detect quiet submarines, an important trend in underwater acoustics is the use of large arrays with many elements in an effort to detect weak signals. In recent years, vector sensors have been widely used in sonar equipment. Compared with sound pressure array, vector array can be used to expand array aperture to improve processing gain, and overcome the problem of left and right ambiguity. Adaptive beamforming (MVDR) has good resolution and interference suppression if the steering vector of the acoustic array is unbiased and a sufficient “snapshot” can be obtained for covariance matrix estimation. In practice, for a long array sonar observation data it is often inadequate (especially in the fast-moving target detection at close range should be chosen to make short-term window integral). In addition, the steering vector error caused by various causes will also degrade the performance of the adaptive beamforming algorithm. RCB is a promising robust adaptive beamforming algorithm. The algorithm directly links the calculation of adaptive diagonal element loading factor with the error of steering vector, which is easy to be operated and avoids the arbitrariness of other tolerance algorithms in the calculation of loading. In this paper, an improved RCB algorithm based on vector long linear array is proposed based on RCB algorithm. The algorithm performs the singular value decomposition of the data matrix, avoids the covariance matrix of the incomplete rank, and reduces the amount of computation and storage. The algorithm is insensitive to the steering vector mismatch and can estimate the target azimuth rapidly.