Abstract
This manuscript describes a method of adding sematic features on environmental maps based on extraction of human behavior with a mobile robot. Especially, this study assumes that semantic features are found around the areas where human groups are extracted. In this study, an extraction system of semantic features in the environmental maps using human positions is introduced. At first, the mobile robot performs a self-position estimation using the global occupancy grid map. At the same time, humans are extracted by a laser range sensor attached in the robot. Then, the number of humans is accumulated in each cell that the environmental maps are divided. Also, the system developed by ROS and RTM systems is introduced.
