Vision-based human augmented mapping for indoor environments

In this paper, we suggest a new method of vision-based human augmented mapping for indoor environments. It is a semi-autonomous approach using human-robot interation and can be an alternative to autonomous map building. The advantage of our approach is that the user can share the environments with the robot and insert semantic information to the environmental map. We apply PCA features for visual landmarks and employ hybrid map representation. Particularly, we define two types of nodes, U/R-nodes and divide the map building process into two, User's Guidance and Robot's Review. Experimental results in rectangular-shaped corridors show the feasibility of our method.