抄録
Speech recognition systems have recently been introduced in a wide variety of environments such as vehicle instrumentation. Speech recognition plays an important role in ships' chief engineer systems. In such a system, voice recognition supports engine room controls, where operability at lower than 0 dB SNR is required. In such a low SNR environment, noise signals may be misjudged as speech, drastically decreasing the recognition rate. Speech recognition systems operating in low SNR environments have therefore not received much attention. This study focused on a recognition system that uses body-conducted signals. Such signals are seldom affected by background noise, so a high recognition rate can be expected in low SNR environments such as engine rooms. To examine the possibility of building a body-conducted speech recognition system operating under an SNR environment lower than 0 dB, we investigated locations for effective signal detection, examined feature vectors derived from those locations, and performed a simplified experiment with a commercial recognition system in our college's training ship, theOshima Maru, so that we could verify its effectiveness.
