Influence of Sensor Design on Bio-Inspired, Adaptive Acoustic Sensing

Bio-inspired acoustic sensing is one way to address the problems of current speech processing systems. These include successful and efficient speech recognition in situations with a low signal-to-noise ratio and multiple sound sources, and dynamic adaptation to changing listening conditions. To achieve this, following the biological models, signal processing (frequency decomposition and dynamic compression) as well as dynamic adaptation of the detection and processing characteristics are integrated into the sensor stage. Such a bio-inspired acoustic sensor system will be presented in the talk. At its heart are active silicon beams with integrated deflection sensing and actuation in combination with real-time feedback. Results from electroacoustic and vibrometric measurements as well as FEM simulations of different sensor designs will be presented and it will be discussed how a specially tailored design of the active silicon beams can avoid artifacts caused by additional resonances in the audible frequency range. Furthermore, two operating modes of the sensor are compared, that are analogous to pressure gradient and pressure microphones.