Semantic Mapping

edit

Semantic maps describe the poses, types and properties of objects in the robot's environment. They are useful whenever a robot needs to interact with these objects.

As an introduction, you may want to have a look at the paper Autonomous Semantic Mapping for Robots Performing Everyday Manipulation Tasks in Kitchen Environments that describes both how these maps can be acquired and how they are internally represented. You can also create the internal representation manually, using the semantic map editor or a ROS service.