Development of a mechanical talking robot that produces natural vowels and consonant sounds

The purpose of this research is to clarify the human vocal mechanism from engineering viewpoints by simulating vocal movement with a robot, and to create a dynamic model. The authors developed an anthropomorphic talking robot WT-1 (Waseda Talker-No.1) in 1999. It simulates human vocal movement, and has articulators and vocal organs. WT-1 could speak Japanese vowels (|a|, |i|, |u|, |e|, |o|). However, its vowels weren't natural. Therefore, we developed a talking robot WT-1R (Waseda Talker-No.1 Refined) that improved WT-1 for the realization of natural vowels and consonant sounds in 2000. WT-1R has articulators (the 6-DOF tongue, 4-DOF lips, 1-DOF teeth, a nasal cavity and a 1-DOF soft palate) and vocal organs (the 1-DOF lungs, 2-DOF vocal cords). The total DOF of the robot is 15. This paper describes the improved mechanisms of WT-1R and the realization of the fluctuation for the natural voice and the clearness of vowels.