Development of mechanical model of human tongue by artificial muscle with human-like sensory system

作为机械语音产生模型(会说话的机器人)的研究，我对舌头机制进行了改进，使其更接近人类。舌头是由软材料制成的，我体验了电磁铁驱动舌头的机构，开发了适合于软材料的舌形测量传感器系统，由于空间有限，很难附着在软材料上，传感器系统采用了光反射器。光反射模块由LED和两种类型的光晶体管组成，以扩大测量范围。在对传感器进行评价时，首先建立已知形状的舌形模型，利用传感器系统对舌形进行测量，并将计算数据与已知数据进行比较。结果表明，该传感器能够以足够的精度测量物体的形状。其次，根据传感器数据再现了机器人舌头的形状，并对产生的声音进行了生产率评估。结果表明，该系统对舌形的再现具有一定的实用价值。

As research of mechanical speech production models (talking robots), I improved the tongue mechanism to be more human-like. The tongue is made of soft material, and I experienced mechanism with electromagnets to actuate the tongue, and developed sensor systems to measure the tongue shape which is adaptable for soft material.The sensor systems were adopted the photo reflectors, because of the space limitation and the difficulty to attach to the soft material. A photo reflector module is consisted of LED and two types of photo transistors to broaden the measurement range. In evaluation of the sensor, first I set a tongue shape model which shape was already known, measure the shape with the sensor systems and compared the calculated data and known data. The result showed the sensors could measure the shape with enough accuracy. And as second, I reproduced the robot tongue shape by the sensor data and evaluated the productivity from the produced sound. From the result, the system was useful to reproduce the tongue shape.