| 摘要: | 研究期間：10108~10207;Storytelling is a joyful and educational activity for most of children and has powerful effects on improving children’s skills in language expression (e.g., listening, reading, and comprehension). In addition, storytelling can foster children’s creativity and encourage social interaction. In our previous work, we have developed a robot-based learning companion named “singing and dancing bear”, which can accompany children for storytelling. More specifically, children can tell a story in front of a “singing and dancing bear” without the need of attaching some sensors on their bodies and then they act as listeners to watch how the “singing and dancing bear” vividly replays the story just told by them. We hope the “singing and dancing bear” can promote children’s interest at storytelling and improve their engagement. In many real applications, we found that there exist several drawbacks and limitations in the “singing and dancing bear”. The behavior of the bear were not as similar as the learner’s behaviors. For example, the bear was not able to applause, twist its waist, squat down, turn about, and so on. In addition, the storyteller has to directly face the camera; otherwise, the bear cannot effectively imitate the learner’s behaviors. Besides, we have not developed corresponding technologies to apply the bear in the cyber world. Therefore, we propose a three-year project to improve the “singing and dancing bear” both in its hardware and software in so the bear can be applied to accompany learners in a real-world environment. In the first year, we will try to improve the bear’s hardware by changing the 17-dgree freedom to 21-degree freedom in order to make the bear be able to imitate learner’s natural behaviors (e.g., twist its waist, squat down, turn about). In addition, we will develop an unsupervised learning algorithm to analyze the learner’s behaviors to find out what kinds of basic poses which the bear can not correctly imitate. Based on the analysis results, we can build a database for the use of the second-year project. In the second year, we will develop a neural-network-based algorithm to make the improved version of the bear be able to efficiently and effectively imitate the learner’s behaviors. In the third year, we hope to make the bear be a learning companion for children by incorporating an attention monitoring mechanism into the bear. In addition, to the storytelling activity, the bear can accompany children to study. The bear can remind the children of the attention status and reading distance in time when children are in appropriate situations. In addition, we will develop corresponding technologies to apply the bear in a cyber world. We really hope that we can accomplish a robot-based learning companion for children in storytelling and learning activities in this three-year project. |
