Book chapter | Links: | BibTeX entry | |
---|---|---|---|
Authors: | Andrey Belkin | ||
Source: | C.-Y. Su, S. Rakheja, H. Liu (eds.), The 5th International Conference on Intelligent Robotics and Applications, Lecture Notes in Artificial Intelligence, vol. 7508 no. 7508, Springer, 2012. | ||
Pages: | 171-180 | ||
ISBN: | 978-953-51-0644-9 | ||
Modern autonomous robots are performing complex tasks in a real dynamic
environment. This requires real-time reactive and pro-active handling
of arising situations. A basis for such situation awareness and handling
can be a world modeling subsystem that acquires
information from sensors, fuses it into existing world description
and
delivers the required information to all other robot subsystems. Since
sensory information is affected by uncertainty and lacks for semantic
meaning, the employment of
a predefined information, that contains concepts and descriptions
of the
surrounding world, is crucial. This employment implies matching of
the
world model information to prior knowledge and subsequent complementing
of the dynamic descriptions with semantic meaning and missing attributes.
The following contribution describes a matching mechanism based on
the Kullback-Leibler
and Tanimoto distances and direct assignment of the prior
knowledge for the model complementation.