A relationship between ability of perception and learning efficiency

“Robot learning” is researched well in recent years. In each research, various types of robot are used. Here, robot’s sensor is selected and is set by each researcher. If each robot has different sensor, it perceives different environment. This fact causes that whether a given task can be achieved or not depends on kinds of robot’s sensor, even if in same environment. In this paper, we investigate a relationship between number of kinds of robot’s sensors and robot’s learning efficiency. We experiment to investigate it on computer simulation. And we discuss a relationship between ability of perception and learning efficiency.