Understanding the Real World: Combining Objects, Appearance, Geometry and Topology for Semantic Mapping
2011 (English)Report (Other academic)
A cornerstone for mobile robots operating in man-made environments and interacting with humans is representing and understanding the human semantic concepts of space. In this report, we present a multi-layered semantic mapping algorithm able to combine information about the existence of objects in the environment with knowledge about the topology and semantic properties of space such as room size, shape and general appearance. We use it to infer semantic categories of rooms and predict existence of objects and values of other spatial properties. We perform experiments offline and online on a mobile robot showing the efficiency and usefulness of our system.
Place, publisher, year, edition, pages
Stockholm: KTH Royal Institute of Technology , 2011. , 20 p.
Computer and Information Science
IdentifiersURN: urn:nbn:se:kth:diva-34158OAI: oai:DiVA.org:kth-34158DiVA: diva2:419553
QC 201105272011-05-272011-05-272011-05-27Bibliographically approved