According to studies confirming that robot assisted learning (RAL) can positively contribute to improving learners' motivation and achievement for language learning [1, 4], RAL is facing its diffusion through the demand of parents and government. However, the biggest obstacle to the long-term interaction between humans and robots is the robots' low-success rate of visual and voice recognition, as well as the limitation of artificial intelligence for the daily-life HRI [3]. This study demonstrated ROBOSEM's ability to sustain long term interaction between children and a robot in an elementary English class from the pilot studies with IROBIQ, called Langbot [1]. Five factors are of concern in sustaining the long-term interaction between children and ROBOSEM, as shown in Figure 2: (1) enhancing the recognition ability of ROBOSEM with class materials, such as marker hats, bracelet watches embodied RFID tags, Wiimocon, etc; (2) sharing the birth story of ROBOSEM, which works to increase children's tolerance toward weak recognition from the result of [2]; (3) making a favorable impression, such as by flashing children's faces on the screen; (4) telling the history of a child's personal learning activities; and (5) tele-operation by the intelligence of a human being.

Various studies propose that robots can be an effective tool for language teaching and learning. Especially they have been remarkably successful in elementary English classes [1][2][3][4]. The purpose of this study was to investigate some effects of a teaching assistant robot, Langbot, in elementary English classes in Korea. We adopted IROBIQ as Longbot for a pilot study.

Lately, there has been increasing field studies harnessing robots for educational purposes. Various studies into teaching assistant robots (or peer tutoring robots), in particular, are proving to be effective tools of language education because robots are remarkably successful in eliminating the affective filter. However, there is decreased in their interaction with children as time passes. The biggest obstacle to the interaction between human beings and robots in field studies using robots is its low-success rate of face and voice. Since the recognition of robots is weak in real environment such as classrooms, this causes some of the children to utterly give up interaction with robots or just ignore them. Hur and Han (2009) found that the robot storytelling was working to increase children' s tolerance toward the failure of recognition of a robot. Children want to know where the robot comes from, how old it is, who makes it, and when it was born, before they listen to it or its storytelling. The birth story of IROBIQ on an innate weak recognition was created as shown in Table 1. And then the movie on birth story of IROBIQ was developed as shown in Figure 2, and then the facial expression, action, display of IROBIQ were added to it. Figure 1 is the scene of telling the birth story.