Cooperative works for agent robot and human using robot vision based on the model of knowledge, emotion and intention

In recent years, emphasis has been placed upon the development of a welfare support system that takes into account the aging society. Realization of cooperative work between welfare agent robots (following agent robots) that support a user and the user in a welfare system is indispensable. For that purpose, all agent robots have a common control mechanism, and cooperative work between the agents and the human is generated through the dynamic determination of each robot´s actions. Furthermore, even when an agent robot is performing a task, it is desirable for it to observe the user spontaneously. Robot vision is constructed using a model of knowledge, emotion and intention, from which welfare oriented active vision is realized. In this paper, we report a technique for realizing this robot-human cooperation, and demonstrate the validity of welfare oriented active vision