知能ロボットにおける問題解決と学習の統合

抄録

In the real world, there are many kinds of uncertainties in motions of robots. Moreover, robots cannot always get sufficient knowledge about tasks in advance. Therefore, intelligent robots must posses both problem solving ability to decide on proper motions with insufficient knowledge and learning ability to acquire knowledge from experiences. Effective integration of the two abilities is also important.
In this paper, we describe an experimental system named ARPEX-L (Advanced Robot Planning and Execution System with Learning Ability) . We applied ARPEX-L to two kinds of robot tasks, pick and place task and pushing block task. The former is an simple example of automatic robot programming. The latter is an attempt to make an actual robot learn by a symbolic method. Current defects of the system and future works are also described.