How do four robots cooperate on creating an annotated map? The robots use the ERSP software for localization and mapping. The six teams work on these
Collecting information about moving objects.
Motion statistics is collected and the information is made available on the map. The subproject is handled by Jakob Sloth, Dan Jørgensen, and Søren Hauberg (Hauberg). See this page for details and status (in Danish).
Integration of maps.
The four robots each run vSlam. This sub project aims at integrating the four maps. See this page for details and status.
Recognition of objects on cluttered background. Is it possible to extract an object image as input to the recognizer subtraction of an image that includes the object from the image in which the object has been removed?
3-D landmark points.
Extraction of 3D points which characterize landmarks. Is it possible to extend the map to 3D?
Recognition of hand signs.
Hand gestures are used to communicate with the robots. How does one design a robust visual input language based on human gestures?
Localization of known objects.
Objects are to be localized on the map. The map is to be annotated when and where objects have been recognized.