Fuzzy rule based voice emotion control for user demand speech generation of emotion robot

Abstract

The emotional function of the human mind has an important role for decision-making, memory, action, and good communication or so. Especially, emotion characteristics of voice are very import for warm communication, successful business, human-to-human good relationship, and good care for children and silver ages. On the other hand, recently, service robot market such as educator, helper, secretary, deliver, and guider has been growing up because of old population and complicated social situation. In that case the emotion function is needed in those areas. The emotion characteristic of voice depends on pitch contour, acoustic energy, vocal tract features, speech energy or so. Therefore we need to consider on how we have to apply and implement emotion function of voice for service robot. However, its implement for robot is very difficult and recognition is also not easy because of various emotion patterns in voice. This paper suggests method of voice emotion generation for user demand emotion talk in service robot. Fuzzy rule based approach is introduced to generate emotion for user demand emotional function by controlling pitch contour, acoustic energy, vocal tract features, and speech energy.