New anthropomorphic talking robot—investigation of the three-dimensional articulation mechanism and improvement of the pitch range

Abstract

We developed a new three-dimensional talking robot WT-6 (Waseda Talker-No. 6), which generates speech sounds by mechanically simulating articulatory motions as well as aeroacoustic phenomenon in the vocal tract. WT-6 consists of a 17-DOF mechanism (1-DOF lungs, 5-DOF vocal cords, 1-DOF jaws, 5-DOF tongue, and 4-DOF lips). It has 3-D lips, tongue, jaw, and velum, which form the 3-D vocal tract structure. It also has an independent jaw opening/closing mechanism, which controls the relative tongue position in the vocal tract as well as the oral opening. The previous robot, which had a 2-D tongue [J. Acoust. Soc. Am. 117, 2541 (2005)], was not enough to realize precise closure to produce humanlike consonants such as /s/ or /r/. The new tongue, which could be controlled to form the 3-D shape, makes it possible to produce more realistic vocal tract shape. The vocal cord model was also improved by adding a new pitch control mechanism pushing from the side of the vocal cords. The pitch range becomes broader than the previous robot, which is enough to reproduce normal human speech. Preliminary experimental results showed that synthesized speech quality improves for vowels /a/, /u/ and /o/. Some experimental results and video demonstration of the talking robot will be presented.