MICROPHONE ARRAY OPTIMIZATION BY STOCHASTIC REGION CONTRACTION

The microphone array will be an important component in advanced speech-acquisition systems. A key issue is the attenuation of "noise" sources when pointing the array at a signal source. This paper is concerned with "optimal" microphone placement and gain for a linear, one-dimensional array often in a confined environment. A power spectral dispersion function (PSD) is used as a core element for a Min-Max objective function (PSDX). Derivation of the optimal spacings and gains of the microphones is a hard computational problem since the Min-Max objective function exhibits multiple local minima (hundreds or thousands). This paper addresses the computational problem of finding the global optimal solution of the PSDX function. A new method, stochastic region contraction (SRC), is proposed which achieves a computational speedup of 30-50 when compared to the commonly used simulated-annealing method. Moreover, SRC is ideally suited for coarse-grain parallel processing.