ball-2
Studies in Semantic Modeling
of Real-World Objects using
Perceptual Anchoring
ANDREAS PERSSON
Information Technology
Örebro Studies in Technology 83 I ÖREBRO 2019
2019
AN
DR
EA
S P
ER
SS
ON
Stud
ie
s i
n S
em
an
tic M
od
eli
ng o
f R
ea
l-W
orl
d O
bje
cts .
..
andreas persson is a former student of the Robotics and Intelligent Systems Master’s Programme at Örebro University, Örebro, Sweden. In 2012 he continued his academic studies as a graduate student at the Centre for Applied Autonomous Sensor Systems (AASS) of Örebro University. His research interests include, besides per-ceptual anchoring: knowledge representation, human-robot interaction, machine learning, and computer vision. For an autonomous agent to handle various types of real-world situated scenarios similarly as a human would, the agent must maintain a consonance between the perceived world (through sensory capabilities) and its internal representa-tion of the world in the form of symbolic knowledge. An approach for modeling such internal representations of objects is through the concept of perceptual anchoring, which, by definition, handles the problem of creating and maintain-ing the correspondence between symbols and sensor data that refer to the same physical object in the external world. This thesis aims to present studies in the use of perceptual anchoring for addressing the problem of real-world semantic world modeling. The proposed method introduces an anchoring architecture which emphasizes, in particular, sensor-driven acquisition of perceptual sensor data and the use of object attribute values in order to correctly anchor objects, and thereby maintain a consistent representation of the external environment.
issn 1650-8580 isbn 978-91-7529-283-0
