Representing Workspace and Model Knowledge for a Robot with Mobile Sensors
Michael O. Shneier, Ernest Kent, P Mansbach
A representation is described for supplying a robot manipulator with information about its workspace. Information is obtained from sensors that move with the manipulator. Spatial information is stored in an octree, allowing fast computation of which parts of the workspace are occupied and which parts are navigable. Information about properties and features of objects is stored in a set of tables or attribute lists. This information is used to match objects in the world with stored models and to assign names to instances of objects and features. Recognized objects are stored in the same way as unrecognized objects, so that all operations on the workspace model are uniform. The two representations are linked to enabled objects to be located in space by name, by description, or by position, and to facilitate finding out what object occupies a particular volume in the workspace.
Proceedings of the 7th International Conference on Pattern Recognition
July 1, 1984
7th International Conference on Pattern Recognition
, Kent, E.
and Mansbach, P.
Representing Workspace and Model Knowledge for a Robot with Mobile Sensors, Proceedings of the 7th International Conference on Pattern Recognition, Montreal, CA, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=820263
(Accessed February 28, 2024)