| Beam-forming algorithm is very important in sonar, radar and radio communication system. The emergence of microphone arrays (or sensors) makes up for the deficiencies in a single microphone. It can achieve the formation of directional beam through beam-forming algorithms according to the collected spatial information. In this way, we can enhance the signal in the specified direction (or the desired direction), and eliminate the signal in the undesired directions.In this thesis, microphone arrays and beamform algorithm are introduced and the basic theory, the design methods and the results of the simulation experiments of the broadband beam former are all analyzed and discussed in the case of far-field and near-field. The designed beamformer could enhance the signals in expected direction and weaken the noises in unexpected diretions, therefore, the signal could be sampled in the specific direction.At present, there are many beam-forming algorithms, but most are designed for far-field narrow-band signal. This paper considers the theory and design of a broadband far-field and constant beam-width beam-former. The minimum variance distortion-less response (MVDR) method, space re-sampling method and the overall weighted method are used to achieve a far better far-field broadband constant beam-width beam-forming.Generally, we sample the signals directly by the apparatus, in the case of high SNR, that is no problem. But if in the case of low SNR, we will design some kind filters for denoising. At present, there are many such algorithms, but they have many limitations. If we sample signals with the method proposed in the paper, the SNR will be improved at the sample stage.Through the results of the simulation experiments, we could infer that this method proposed in the paper could be used in the specific direction singal sample. The signal whose frequency is from 800 Hz to 3200 Hz could become constant beam-width though the designed beam-former and the relative frequency bandwidth is 1.789. |
PDF Full Text Request