The relation between the mechanic system and human has just been unilateral so far. This is the why people do not want to get familiar with multi-service robots. If the function of the emotion recognition is granted to the robot system, the concept of the design of the mechanic part will be changed a lot.
Pitch, Energy and Tempo extracted from the human speech are good and important factors to classify the each emotion, which are called prosodic features. There are some methods to arrange the trend of character followed by specific emotion. In this paper, prosodic feature vector and vector space model are proposed to cluster the boundary of emotion (neutral, happy, sad and angry). Prosodic feature vector is trained and classified using HMM (Hidden Markov Model) which is the powerful and effective theory to construct the statistical model. Vector space model can be showed in the shape of ellipsoid, which is applied by the difference of the standard deviation of prosodic factors during utterance according to each emotion. Even though some characteristics of the voice can not be concluded to the specific result because of the variety of expression, these two methods are efficient way to pass the limit of recognition.