A mechanical model of the human vocal system is being developed based on a mechatronics technology under a feedback control. While various ways of vocal sound production have been actively studied so far, a mechanical construction of the vocal system is considered to advantageously realize natural vocalization with its fluid dynamics. Motors are employed for the manipulation of the mechanical system. The system is able to adoptively learn the relations between motor control parameters and the produced vocal sounds using an auditory feedback with neural networks, by mimicking a human vocalization. This paper presents the construction of a talking robot and its singing performance, together with the adaptive control for the pitch learning. The talking robot generates vowel and consonant sounds of different pitches by dynamically controlling the vocal cords, vocal tract and nasal cavity.

We are developing a mechanical model of a human vocal system based on a mechatronics technology. Although various ways of vocal sound production have been actively studied so far, a mechanical construction is considered to advantageously realize natural vocalization with its fluid dynamics. In the voice generation, the analysis of the behaviors of the vocal cords and the vocal tract are required to be employed in the mechanical system. Furthermore, the fluid mechanics is less stable to make the control difficult. Several motors are employed to manipulate the mechanical vocal system. A neural network works to establish the relations between motor positions and the produced vocal sounds by an auditory feedback in the learning phase. In the speech performance, the mechanical system is able to utter while vocal pitches and phonemes are adaptively controlled by an auditory feedback control. This paper presents the construction of a vocal tract part and its adaptive control by the neural networks.