Collision Avoidance with Bipartite Arrangements (bibtex)
by Gottfried, Björn
Abstract:
We demonstrate that qualitative spatial reasoning techniques provide useful means for autonomous systems in the context of collision avoidance. A qualitative spatial representation is presented which relies on quite abstract distinctions which describe how two objects can be aligned relative to each other. Possible alignments involve left and right, front and back, during, contains, and overlaps. It is shown how such qualitative relations already allow to make important decisions, for instance, in order to predict collisions. While the proposed calculus guarantees that predictions about collision-free situations never lead to collisions, predictions about collisions do not necessarily lead to collisions (though they occur more likely when they are predicted). We shall learn that it is by no means necessary to utilise precise measurements in order to predict collisions.
Reference:
Gottfried, Björn, "Collision Avoidance with Bipartite Arrangements", In CIKM-WS on Knowledge Representation for Autonomous Systems, 2005.
Bibtex Entry:
@INPROCEEDINGS{Gottfriedh,
  author = {Gottfried, Bj{\"o}rn},
  title = {{Collision Avoidance with Bipartite Arrangements}},
  booktitle = {CIKM-WS on Knowledge Representation for Autonomous Systems},
  year = {2005},
  editor = {Balakirsky, Stephen and Schlenoff, Craig},
  abstract = {We demonstrate that qualitative spatial reasoning techniques provide
	useful means for autonomous systems in the context of collision avoidance.
	A qualitative spatial representation is presented which relies on
	quite abstract distinctions which describe how two objects can be
	aligned relative to each other. Possible alignments involve left
	and right, front and back, during, contains, and overlaps. It is
	shown how such qualitative relations already allow to make important
	decisions, for instance, in order to predict collisions. While the
	proposed calculus guarantees that predictions about collision-free
	situations never lead to collisions, predictions about collisions
	do not necessarily lead to collisions (though they occur more likely
	when they are predicted). We shall learn that it is by no means necessary
	to utilise precise measurements in order to predict collisions.},
  owner = {pmania},
  timestamp = {2012.11.06},
  url = {http://portal.acm.org/results.cfm?coll=guide{\&}dl=GUIDE{\&}query=Bj{\%}F6rn+Gottfried{\&}short=1}
}
Powered by bibtexbrowser