Multiple sources localization with microphone array: A subarray approach

The problems of location and distance estimations with planar microphone array for multiple point sources are addressed in the present study. Though traditional array processing methods show the reliability in estimation of direction-of arrival (DOA) of the sources, such techniques are incapable to provide source distances which requires for three dimensional (3-D) target coordinates. With planar array processing, azimuth and elevation angles are always obtainable but no longer to extend into more details in 3-D analysis. A new approach to deal with this problem is proposed by estimating multiple sources' DOA with multiple planar sub-arrays. A least square solution based on the estimated DOA of the sources is derived to give the source distances with minimum two sub-arrays in arbitrary microphone configurations. The benefits of the proposed method include a extension of the capability of planar array processing to a higher dimensional analysis. Moreover, its adaptability with any DOA estimation methods makes it convenient to use. The new approach is examined with conventional delay and sum beamforming and MUSIC methods under noisy conditions by both simulations and experiments. [K.C.T. is supported by a studentship from the Hong Kong Polytechnic University.]