Embodiment of an Agent using Anthropomorphization of an Object

Abstract

Recently, many home appliances have acquired multiple functions due to the development of ubiquitous technology. For example, refrigerators are able to provide users the expiration dates of foods to a user from the RFID tags attached to the food package (Chen et al., 2007). The development of such intelligent environments increases the quantity of information passed from appliances to users. For users, it has become more difficult to understand all the appliances’ functions and information. Thus, it is important for users to know what functions they should be using and what information they should focus on. An anthropomorphic agent including virtual agent and real world agent like communication robot is one way of relating this enormous amount of information to users. It can enhance this information by speaking, pointing, and using gestures on screen or in the real world. There have been many studies on the communication methods and gestures of anthropomorphic agents. For example, Scheutz studied the influence of spoken words of robots on the cooperation task between human and robot (Scheutz et al., 2008). Breazeal studied how users accept emotions expressed via a robot's facial movements using the facial robot Kismet (Breazeal, 1997). Imai et al. succeeded in joint attention between a user and the robot Robovie and attracting the user attention toward a poster using the robot's face and arm direction (Imai et al., 2003). Shinozawa et al. noted in their experiment (Shinozawa et al., 2005) that using a communication robot is a better way for providing real world information than a virtual agent. This study also suggested that a communication robot is useful for describing the information and functional capabilities of home and office equipment. However, the agent’s ability to direct the users' attention towards an artifact sometimes fails because the agent gains more of the user's attention than the target. Presentations given by robots at exhibitions sometimes fail because the people are attracted to the robots more than what the robot is explaining. Murakawa et al. noted that a sales-robot in a shop does not ensure that the amount of goods sold will increased even though it does attract the attention of customers and that they did look at the goods (Murakawa and Totoki et al., 2006). These failures occurred because of an underestimation of peoples' curiosity in robots. Fukayama et al. noted in their experiment using a virtual agent that the agent sometimes draws attention and disturbs the user's attention on the task before them (Fukayama et al., 2003). They 8