Sound source localization with microphone arrays using genetic algorithm

The most significant problem for speech detection and recognition systems in such environments with a lot of noise and reverberation is the degradation of signal arriving at the sensors, that is microphones. Most people place the microphone over the computer monitor, on the meeting table or on the wall. In these situations, the quality of the sound degrades in varying amounts depending on the environmental conditions. While using a close-talking microphone can be a solution to the problem, for most situations this is not practical or even not possible. Instead of a single microphone, using a microphone array to increase the power and quality of desired signal is quite popular in recent researches and applications in sound processing. A microphone array is a set of microphones placed geometrically at different locations and operating simultaneously. By delaying the signal received at each sensor (microphone) by certain amounts and summing them all together, it is possible to obtain a more powerful signal. This is called delay-and-sum beamforming using microphone arrays. In order to perform beamforming using microphone arrays, it is crucial to know the location of the sound source. But, in cases where there is more than one possible sound source, it is impossible to know the direction of the sound source in advance. In this study, it is aimed to determine the position of the sound source using real time methods. By performing delay-and-sum beamforming at possible source locations, it is assumed that the direction which results in with the highest signal to noise ratio (SNR) is considered to be the desired sound source location. Due to its well-known reputation in signal processing, Genetic Algorithm is used as optimization technique.