In recent years, speech recognition systems have been employed in vehicle instrumentation in a wide variety of environments. Speech recognition plays an important role in the chief engineering systems on ships, in this type of system, speech recognition supports the engine room controls, and lower than 0-dB signal-to-noise ratio (SNR) operability is required. In such a low SNR environment, a noise signal can be misjudged as speech, dramatically decreasing the recognition rate. Therefore, this study focuses on a recognition system that uses body-conducted signals. Since noise is not introduced within body-conducted signals that are conducted in solids, even within sites which are low SNR environments such as engine rooms, Ms necessary to construct a system with a high recognition rate. However, when constructing such systems, in order to create models specialized for body-conducted speech, it is necessary to use learning data consisting of sentences that must be read numerous times. Therefore, in the present study we applied a method in which the specific nature of body-conducted speech is reflected within an existing speech recognition system with only small numbers of vocalizations. Because two of the prerequisites for operation within a site such as an engine room where noise exists are both "hands-free" and "eyes-free" operations, we also investigated the effects of making this type of system wireless.
展开▼
