A machine learning approach for feature-sensitive motion planning

Marco Morales, Lydia Tapia, Roger Pearce, Samuel Rodriguez, Nancy M. Amato

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Although there are many motion planning techniques, there is no method that outperforms all others for all problem instances. Rather, each technique has different strengths and weaknesses which makes it best-suited for certain types of problems. Moreover, since an environment can contain vastly different regions, there may not be a single planner that will perform well in all its regions. Ideally, one would use a suite of planners in concert and would solve the problem by applying the best-suited planner in each region. In this paper, we propose an automated framework for feature-sensitive motion planning. We use a machine learning approach to characterize and partition C-space into regions that are well suited to one of the methods in our library of roadmap-based motion planners. After the best-suited method is applied in each region, the resulting region roadmaps are combined to form a roadmap of the entire planning space. Over a range of problems, we demonstrate that our simple prototype system reliably outperforms any of the planners on their own.

Original languageEnglish (US)
Title of host publicationAlgorithmic Foundations of Robotics VI
EditorsMichael Erdmann, Mark Overmars, A. Frank van der Stappen, Hsu Hsu
Pages361-376
Number of pages16
StatePublished - Dec 1 2005
Externally publishedYes

Publication series

NameSpringer Tracts in Advanced Robotics
Volume17
ISSN (Print)1610-7438
ISSN (Electronic)1610-742X

Fingerprint

Motion planning
Learning systems
Planning

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Artificial Intelligence

Cite this

Morales, M., Tapia, L., Pearce, R., Rodriguez, S., & Amato, N. M. (2005). A machine learning approach for feature-sensitive motion planning. In M. Erdmann, M. Overmars, A. F. van der Stappen, & H. Hsu (Eds.), Algorithmic Foundations of Robotics VI (pp. 361-376). (Springer Tracts in Advanced Robotics; Vol. 17).

A machine learning approach for feature-sensitive motion planning. / Morales, Marco; Tapia, Lydia; Pearce, Roger; Rodriguez, Samuel; Amato, Nancy M.

Algorithmic Foundations of Robotics VI. ed. / Michael Erdmann; Mark Overmars; A. Frank van der Stappen; Hsu Hsu. 2005. p. 361-376 (Springer Tracts in Advanced Robotics; Vol. 17).

Research output: Chapter in Book/Report/Conference proceedingChapter

Morales, M, Tapia, L, Pearce, R, Rodriguez, S & Amato, NM 2005, A machine learning approach for feature-sensitive motion planning. in M Erdmann, M Overmars, AF van der Stappen & H Hsu (eds), Algorithmic Foundations of Robotics VI. Springer Tracts in Advanced Robotics, vol. 17, pp. 361-376.
Morales M, Tapia L, Pearce R, Rodriguez S, Amato NM. A machine learning approach for feature-sensitive motion planning. In Erdmann M, Overmars M, van der Stappen AF, Hsu H, editors, Algorithmic Foundations of Robotics VI. 2005. p. 361-376. (Springer Tracts in Advanced Robotics).
Morales, Marco ; Tapia, Lydia ; Pearce, Roger ; Rodriguez, Samuel ; Amato, Nancy M. / A machine learning approach for feature-sensitive motion planning. Algorithmic Foundations of Robotics VI. editor / Michael Erdmann ; Mark Overmars ; A. Frank van der Stappen ; Hsu Hsu. 2005. pp. 361-376 (Springer Tracts in Advanced Robotics).
@inbook{6232925d38554ce994e95f3333823fd7,
title = "A machine learning approach for feature-sensitive motion planning",
abstract = "Although there are many motion planning techniques, there is no method that outperforms all others for all problem instances. Rather, each technique has different strengths and weaknesses which makes it best-suited for certain types of problems. Moreover, since an environment can contain vastly different regions, there may not be a single planner that will perform well in all its regions. Ideally, one would use a suite of planners in concert and would solve the problem by applying the best-suited planner in each region. In this paper, we propose an automated framework for feature-sensitive motion planning. We use a machine learning approach to characterize and partition C-space into regions that are well suited to one of the methods in our library of roadmap-based motion planners. After the best-suited method is applied in each region, the resulting region roadmaps are combined to form a roadmap of the entire planning space. Over a range of problems, we demonstrate that our simple prototype system reliably outperforms any of the planners on their own.",
author = "Marco Morales and Lydia Tapia and Roger Pearce and Samuel Rodriguez and Amato, {Nancy M.}",
year = "2005",
month = "12",
day = "1",
language = "English (US)",
isbn = "9783540257288",
series = "Springer Tracts in Advanced Robotics",
pages = "361--376",
editor = "Michael Erdmann and Mark Overmars and {van der Stappen}, {A. Frank} and Hsu Hsu",
booktitle = "Algorithmic Foundations of Robotics VI",

}

TY - CHAP

T1 - A machine learning approach for feature-sensitive motion planning

AU - Morales, Marco

AU - Tapia, Lydia

AU - Pearce, Roger

AU - Rodriguez, Samuel

AU - Amato, Nancy M.

PY - 2005/12/1

Y1 - 2005/12/1

N2 - Although there are many motion planning techniques, there is no method that outperforms all others for all problem instances. Rather, each technique has different strengths and weaknesses which makes it best-suited for certain types of problems. Moreover, since an environment can contain vastly different regions, there may not be a single planner that will perform well in all its regions. Ideally, one would use a suite of planners in concert and would solve the problem by applying the best-suited planner in each region. In this paper, we propose an automated framework for feature-sensitive motion planning. We use a machine learning approach to characterize and partition C-space into regions that are well suited to one of the methods in our library of roadmap-based motion planners. After the best-suited method is applied in each region, the resulting region roadmaps are combined to form a roadmap of the entire planning space. Over a range of problems, we demonstrate that our simple prototype system reliably outperforms any of the planners on their own.

AB - Although there are many motion planning techniques, there is no method that outperforms all others for all problem instances. Rather, each technique has different strengths and weaknesses which makes it best-suited for certain types of problems. Moreover, since an environment can contain vastly different regions, there may not be a single planner that will perform well in all its regions. Ideally, one would use a suite of planners in concert and would solve the problem by applying the best-suited planner in each region. In this paper, we propose an automated framework for feature-sensitive motion planning. We use a machine learning approach to characterize and partition C-space into regions that are well suited to one of the methods in our library of roadmap-based motion planners. After the best-suited method is applied in each region, the resulting region roadmaps are combined to form a roadmap of the entire planning space. Over a range of problems, we demonstrate that our simple prototype system reliably outperforms any of the planners on their own.

UR - http://www.scopus.com/inward/record.url?scp=84880152114&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84880152114&partnerID=8YFLogxK

M3 - Chapter

AN - SCOPUS:84880152114

SN - 9783540257288

T3 - Springer Tracts in Advanced Robotics

SP - 361

EP - 376

BT - Algorithmic Foundations of Robotics VI

A2 - Erdmann, Michael

A2 - Overmars, Mark

A2 - van der Stappen, A. Frank

A2 - Hsu, Hsu

ER -