Indirect object search based on qualitative spatial relations

Abstract—Accomplishing object search tasks in human envi-ronments requires autonomous mobile robots to reason about potential object locations and to plan for the next best view accordingly. By using information about the 3D structure of the environment, knowledge about landmark objects and their spatial relationship to the sought object, the search can be improved by directing the robot towards the most likely object locations. In this paper we have designed, implemented and evaluated an approach for searching for objects on the basis of Qualitative Spatial Relations (QSRs) such as left-of and close-to. On the basis of QSRs between a landmark and the sought object we generate a Gaussian Mixture Model (GMM) for representing metric poses of potential object locations using a ternary point calculus. The GMM is employed within object search for planning the next best view. Preliminary results show that search methods based on QSRs are faster and more reliably than methods not considering them. I