Integration of qualitative and quantitative spatial data within a semantic map for service robots

Abstract

A semantic map which comprises quantitative and qualitative space representations with semantic information seems to be an obligatory part of every robotic system designed to work hand in hand with humans. In this paper we present a semantic map architecture developed for the Courier service robot performing indoor delivery tasks. In our approach we propose to represent not only objects, but also spaces. With such a representation, the state of the world is a set of object-in-space relations, rather than object-object relations, which helps to reduce the total number of spatial relations.