Active Ego-noise Control Based on Metamaterial in Small-size Robot

This paper proposes a hybrid ego-noise suppression method combining conventional active noise control (ANC) algorithm and acoustic metamaterials (AMMs). The ego-noise is harmful to the auditory system embedded in a small-size robot, but very hard to suppress. We adopted a multi-channel ANC system in which a microphone array is used both as reference mic and error mics with a beamforming algorithm. Regarding AMNI as a system with a special impulse response, it is used as an acoustic low-pass filter (0 similar to 1kHz). Modified versions of least mean square (LMS) algorithm are used in our work to ensure low computational requirement and real-time property filtered x LMS (FxLMS) for ANC system and constrained LMS for beamforming, respectively. This work shows the way of combining physical design and system design to overcome the shortcuts applied solely.