Development of an Inaudible Sound Communications System Based on a Machine Learning Approach

This study has developed a system that performs data communications using high frequency inaudible band of sound signals. Unlike radio communications using particular wireless devices, it only requires microphones and speakers employed in ordinary telephony communications. In this paper, we describe the possibility of a machine learning approach to identify the information embedded in sound signals. Our proposed technique identifies the symbols by reading sound spectrogram based on an image recognition approach. This paper describes some experimental results evaluating the performance of the proposed technique. It indicates that the proposed technique may have a certain appropriateness to design a classifier for the symbol identification in the inaudible sound communications.