The authors are developing a vocalization training system for the auditory-impaired. The training system employs a talking robot which has mechanical organs like a human. With an adaptive learning strategy using an auditory feedback control, the robot autonomously learns the vocalization to speak like a human, and then reproduces the speech articulation from inputted vocal sounds. In the previous study, the training system for 5 Japanese vowels was constructed. The effectiveness was assessed by a training experiment conducted in Kagawa Prefectural School for the Deaf, and significant results were obtained. In the next step, the training system for consonant vocalization is studied. The plosive sounds such as /p/, /t/ and /k/ are produced by sudden opening and closing motions of a mouth, and it is not an easy task to reproduce the vocalization based on the auditory feedback learning. To solve this problem, visual information is employed to reproduce the plosive sound vocalization by the talking robot. In this study the learning method of the plosive sounds is introduced. The reproduced robotic vocalizations are evaluated by an experiment, and we validated that the robot successfully reproduced the vocalizations of the able-bodied.
展开▼
机译：作者正在开发用于听觉障碍的发声培训系统。培训系统采用谈话机器人，具有像人类的机械器官。使用使用听觉反馈控制的自适应学习策略，机器人自主地学习声音化以像人类那样说话，然后再现来自输入的声音的语音阐述。在以前的研究中，建立了5个日本元音的培训系统。通过在Kagawa县学校进行聋人进行的培训实验评估了有效性，并获得了显着的结果。在下一步中，研究了辅音发声的培训系统。通过嘴巴突然打开和闭合动作产生诸如/ p /，/ t /和/ k / / k /的爆震声，并且基于听觉反馈学习再现发声并不是一项容易的任务。为了解决这个问题，使用可视信息来通过谈话机器人重现娱乐声音发声。在这项研究中，介绍了娱乐声音的学习方法。再现的机器人发声通过实验评估，我们验证了机器人成功再现了能够拥有的声音。
展开▼
