Bio-inspired directional microphone for robotic hearing system

The primary focus to be implementing a social robot is to communicate with human beings. A part of this communication is designed using direct voice commands; as a result, one of the challenges is to formulate localization of incoming voice/sound. The literature of this scope is abundant with various approaches using pair/array of the microphones, however, a couple of constraints are still being updated, such as the accuracy and computation time. Because the higher computation time likely to render the lower accuracy. To this particular purpose, a cosine dependent sound source localization (SSL) with varying azimuth angles (0 degrees-360 degrees) is presented using a bio-inspired MEMS directional microphone. The developed microphone was measured in an anechoic chamber. The directivity was found in a full compliance with the cosine dependent ideal pressure-gradient sensor. With the satisfactory directionality, the SSL was formulated by considering one diaphragm of the developed directional microphone, and the maximum SSL deviation relative to the incoming sound was found to be 3.8 degrees which is better than the TDOA based SSL algorithms. With the higher accuracy-in addition to its low noise-the developed work can be extended as a robotic hearing system.