Building a Bridge Between AI and Robotics

About fifteen years ago, the hand eye system was one of the exciting research topics in artificial intelligence. Several AI groups attempted to develop intelligent robots by combining computer vision with computer controlled manipulator. However, after the success of early prototypes, research efforts have been splitted into general intelligence research and real world oriented robotics research. Currently, the author feels there exists a significant gap between AI and robotics in spite of the necessity of communication and integration. Thus, without building a bridge over the gap, AI will lose a fertile research field that needs real-time real-world intelligence, and robotics will never acquire intelligence even if it works skillfully. In this paper, I would like to encourage AI community to promote more efforts on real world robotics, with the discussions about key points for the study. This paper also introduces current steps of our robotics research that attempt to connect perception with action as intelligently as possible.