Abstract of FKI-213-96

Document-Name:  fki-213-96.ps.gz
Title:          Why the Parti-game Algorithm Does not Work Satisfyingly
                for Manipulator Trajectory Generation
Authors:        Martin Eldracher, Ruediger Merklein 
                
Revision-Date:  1996/03/26
Category:       Technical Report (Forschungsberichte Künstliche Intelligenz)
Abstract:       Trajectory generation for manipulators can be performed in a
		very efficient manner, if a model of the environment is
		available. A classical approach to construct a world
		model is the cell decomposition method.  We describe in
		detail a recently published approach for planning in
		reinforcement-tasks, called Parti-game algorithm
		(Moore/Atkeson,1995), that also can be used for
		trajectory generation. Parti-game uses rectangular
		cells of varying size for modeling free-space in order
		to keep memory consumption and preprocessing time
		reasonable.  We present experimental results for
		kinematic trajectory generation for manipulators in two
		example environments and a real world environment.  We
		analyze the reasons for the disappointing behaviour of
		Parti-game even in examples still rather afar from real
		manipulators,  opposite to results with other recent
		approaches using graph-based algorithms.
Keywords:       manipulators, robotics, kinematic trajectory generation,
                parti-game, approximate cell decomposition
Size:           23 pages
Language:       English
ISSN:           0941-6358
Copyright:      The ``Forschungsberichte Künstliche Intelligenz''
                series includes primarily preliminary publications,
                specialized partial results, and supplementary
                material. In the interest of a subsequent final
                publication these reports should not be copied. All
                rights and the responsability for the contents of the
                report are with the authors, which would appreciate
                critical comments.

Gerhard Weiss