Robust Neuro-Fuzzy Speaker Localization Using a Circular Microphone Array

A major application area of microphone array processing is the localization of sound sources, mainly of speaking persons. In contrast to most state-of-the-art approaches that are based on correlation measures, we propose a neurologically inspired system that generalizes findings about human spatial hearing to the multi-channel case. It mimics the processing in the human cochlea and the auditory mid-brain. To enhance the localization quality, a new spike generation approach is introduced, termed peak-over-average position (PoAP). A fuzzy combination is used to remove putative artifacts. In contrast to a human listener we employ multiple sensors to gain robustness in reverberant and noisy environments. Post-processing estimates the locations of concurrent speakers. The robustness of the proposed system is shown by comparison with the wellknown steered response power approach. Finally, we show the applicability of our realtime neuro-fuzzy model to the concurrent speaker localization task using real reverberant recordings.