论文部分内容阅读
In the field of noise identification with microphone arrays, the conventional delay-and-sum (DAS)beamforming is the most popular signal processing technique. However, acoustic image results which aregenerated by DAS beamforming are easily influenced by background noise, particularly for in situ wind tunneltests. Even when the arithmetic averaging is used to statistically remove the interference from the backgroundnoise, the results are far from perfect, for the interference from the coherent noise sources are still present. Inaddition, DAS beamforming based on arithmetic averaging fails to deliver real-time computation capability. A newapproach is introduced in this paper. This is so-called observer-based beamforming method which has a recursiveform similar to the state observer in classical control theory, with a real-time computation capability. In addition,coherent background noise can be gradually rejected in iterations. Theoretical derivations of the observer-basedbeamforming algorithm are carefully developed in this paper. Two numerical simulations demonstrate the perfectcoherent noise rejection and real-time computation capability of the observer-based beamforming, which thereforecan be regarded as an attractive alternative to DAS beamforming.