Localization Of Sound Sources By Beamforming

Beamforming is an array-based measurement technique for sound sourceslocation. Quickly, good resolution at high frequency and suited for long distanceand so on is its advantages, but bad resolution at low frequency, the limits oftraditional algorithm in moving sound source and the influence of disturbingsources on the other side of antenna limit its application, thus some new methodsare proposed in this paper.The Statistically Optimal Array Processing (SOAP) looking for weighting isproposed, and then the map of source distribution is reconstructed. The result showa better resolution at low frequency compared with Delay and Sum beamformer. Anumerical simulation and experiment are investigated to demonstrate the validity ofthis method.The Doppler effect result from the source motion with respect to the fixedreceiver array. In this paper, the non-stationary signal received by microphone arrayis changed into stationary signal considering the Doppler effect, then Delay andSum algorithm, SOAP algorithm and MUSIC algorithm are used to obtain the mapof source distribution. Simulation proves this method effective.A method using a double layer array is proposed. Conventional Beamformingusing a single layer array requires that all sound sources located on the same side ofthe array, while the proposed method is free of this restriction. According to thetheorem of plane wave and the superposition theorem of sound pressure, a formulais given to remove the influence of the disturbing source, and then the multiplesignal classification algorithm is used to obtain the map of source distribution. Anumerical simulation and an experiment are investigated to demonstrate thevalidity of this method.