CaltechAUTHORS
  A Caltech Library Service

Motion planning in observations space with learned diffeomorphism models

Censi, Andrea and Nilsson, Adam and Murray, Richard M. (2012) Motion planning in observations space with learned diffeomorphism models. California Institute of Technology , Pasadena, CA. (Submitted) http://resolver.caltech.edu/CaltechCDSTR:2012.004

[img]
Preview
PDF - Submitted Version
See Usage Policy.

1421Kb

Use this Persistent URL to link to this item: http://resolver.caltech.edu/CaltechCDSTR:2012.004

Abstract

We consider the problem of planning motions in observations space, based on learned models of the dynamics that associate to each action a diffeomorphism of the observations domain. For an arbitrary set of diffeomorphisms, this problem must be formulated as a generic search problem. We adapt established algorithms of the graph search family. In this scenario, node expansion is very costly, as each node in the graph is associated to an uncertain diffeomorphism and corresponding predicted observations. We describe several improvements that ameliorate performance: the introduction of better image similarities to use as heuristics; a method to reduce the number of expanded nodes by preliminarily identifying redundant plans; and a method to pre-compute composite actions that make the search efficient in all directions.


Item Type:Report or Paper (Technical Report)
Group:Control and Dynamical Systems Technical Reports
Record Number:CaltechCDSTR:2012.004
Persistent URL:http://resolver.caltech.edu/CaltechCDSTR:2012.004
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:34529
Collection:CaltechCDSTR
Deposited By: Andrea Censi
Deposited On:27 Sep 2012 23:45
Last Modified:27 Dec 2012 02:45

Repository Staff Only: item control page