【摘要】 为了提高仿人机器人口形的真实度,呈现更为自然的人机对话,建立了仿人机器人头部系统。根据汉语语音特点,归纳了不同音节的视位。根据人面部肌肉解剖学,确定了控制机器人口形的五个控制点。通过对汉语发音研究,建立了发音口形拟合时间曲线,并且利用线性插值方法实现仿人机器人的动态口形控制。开发的语音系统实现语音识别和语音合成功能。实验结果表明,机器人可以识别出人说话内容并合成相应的对话。结合头部机械结构和控制系统,语音交互同时机器人能呈现出连续的、动态的口形,并且能够与语音相对应。更多还原

【Abstract】 In order to improve the authenticity of humanoid robot lip-shaped and perform a more natural humancomputer interaction,we established a humanoid robot head system. According to the characteristics of Chinese phonetics,the basic viseme of different syllable was summarized. According to the facial muscle anatomy,the control points of the robot lip-shape were determined. Through the study of Chinese pronunciation,the lip-shape fitting time curve was established,and the dynamic lip-shape control of the robot was realized with linear interpolation method.The developed voice system was equipped with voice recognition and speech synthesis. The experimental results show that the robot can recognize the speech content and synthesize the corresponding dialogue. By means of the mechanical structure of the head and the control system,a continuous dynamic lip-shape corresponding to human speech was presented.更多还原