Sound source localization in 3-D space by a triple-microphone algorithm

In this paper we present a novel technique which localizes sound source with acceptable accuracy and low time complexity. In this technique, contrary to former localization techniques in 3-D space, only three microphones are sufficient. Instead of searching the entire space domain for the sound source, we first estimate an approximate point of the sound source according to the time delays and the ratio of energy received by each pair of microphones. Then, we enhance our estimation by applying an error criterion to points in vicinity of the estimated point. Simulation results show performance of our work, especially when considering the processing time of algorithms as a metric.