Research On Expression Model Of Robot's Humanoid Face

Future robot, especially those mix in with human being daily just as nurse robot, entertainment robot, family robot and so on, should have the configuration and function adapt to human being and environment, and could work with person in harmony in many case, therefore, the ability to communicate effectively with a human become a absolutely necessary factor. The communication so called refers to not only transfer of general information but also more humanistic intercourse of emotion.It is known that face expression take a very important role of in the way of emotion expression. If robot have a appearance similar to human's, and could express its emotion through lifelike expression just as a human adapt to different environment, it could be changed more humanistic, the communication between human and robot could became more nature, easier, more fluency and more friendly, then robot could become our friend really.We have designed a machnical humanoid head-robot in order to study and develop a humanoid robot having the ability to express human-like emotion through expression and communicate naturally with a human. The head-robot was drived by step motors, the movement of which all feature points constructed the robot's expression. The design of whole face was based on Ekman' FACS theory and had the distributing and movement of human facial muscle for reference. We respectively completed the movement modules' design of eyebrows, eyes, mouth and jaw with a modularization idea. Furthermore, we added three DOFs of robot's neck in order to make whole head more lively.The process for human being from feeling the stimulation from environment, generating emotion, then to generating expression is a very complex system that can be affected by many factors. None of those factors could be easily measured and described so that it is a very difficult work to map this system clearly in the way of traditional mathematics and engineering. This study used fuzzy rules-based control system as the control modal from emotion to expression, which connected effectively, the emotion generated from emotion system and mechanical structure of head-robot. It not only mapped emotion intensity onto the displacement of each face feature point qualitatively and quantificationally, but also consulted the generation of expression when in blend emotion state.Furthermore, this paper studied simulation method of robot expression for the need to next work. We combined Pro/engineer and 3DS MAX to model and simulate the robot expression.