On effectiveness of subarray adaptive beamforming for suppressing near-field moving noise

The performance of a hull-mounted sonar array is adversely affected by the presence of own-ship self noises such as propeller noise, machinery noise and flow noise. Among those self noises, the effect of flow noise is hardest to combat due to its near-field and non-stationary nature. In this paper, we investigate the effectiveness of using two stage subarray adaptive beamforming for reducing the effect of flow noise. Three approaches are considered. These are adaptive-conventional, adaptive-adaptive and conventional-adaptive beamforming. The performances of these three subarray beamforming approaches are evaluated against the performance of a full array adaptive beamformer which provides a benchmark in terms of the array gain. A series of moving monopoles is used to simulate the near-field and non-stationary nature of flow noise. The investigation shows that all three subarray adaptive beamforming approaches achieve higher array gains than a full array adaptive beamformer. The reasons for this are analysed. It is found that the time required for full array adaptive beamforming is too long to achieve a good performance in a fast time-varying environment. Among the three subarray adaptive beamforming approaches, the performance gain obtained by the conventional-adaptive approach over a full array adaptive beamformer is moderate, less than 2 dB in average. In contrast, the gains achieved by the adaptive-conventional and adaptive-adaptive approaches are substantially higher, about 12 dB in average. This difference in performance indicates the importance of having adaptive processing at the first stage of subarray beamforming.