Abstract
In this paper, a random microphone array with optimized layout is developed for locating and separate noise sources. The sensor locations are determined, with the aid of simulated annealing method, to minimize the maximum sidelobe level of the farfield beampattern. A two-stage algorithm is formulated on the basis of the spherical-wave model. In the localization stage, the active source regions are located by using the frequency-domain delay-and-sum (DAS) algorithm. In the separation stage, source signals are extracted by solving an inverse problem, based on the source positions identified in the stage one. Equivalent sources are assumed to be fewer (overdetermined) than or more (underdetermined) than microphones. Tikhonov regularization is employed to solve the inverse problem and calculate the acoustic variables including sound pressure, particle velocity, sound intensity, and sound power based on the equivalent source model (ESM). Numerical and experimental results are presented.