Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Speeding up the learning of robot kinematics through function decomposition.

Vicente Ruiz de Angulo1, Carme Torras

  • 1Institut de Robòtica i Informàtica Industrial, 08028 Barcelona, Spain. ruiz@iri.upc.edu

IEEE Transactions on Neural Networks
|December 14, 2005
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

IMU-based workspace area as a promising complementary tool to assess upper limb function in Neuromuscular diseases: A one-year follow-up.

Journal of neuromuscular diseases·2026
Same author

Imu-based kinematic analysis to enhance upper limb motor function assessment in neuromuscular diseases.

Journal of neuroengineering and rehabilitation·2025
Same author

Generating predicate suggestions based on the space of plans: an example of planning with preferences.

User modeling and user-adapted interaction·2023
Same author

Introducing CARESSER: A framework for in situ learning robot social assistance from expert knowledge and demonstrations.

User modeling and user-adapted interaction·2022
Same author

Evaluation of an exercise-enabling control interface for powered wheelchair users: a feasibility study with Duchenne muscular dystrophy.

Journal of neuroengineering and rehabilitation·2020
Same author

Growth Signatures of Rosette Plants from Time-Lapse Video.

IEEE/ACM transactions on computational biology and bioinformatics·2015

This study introduces a novel method to reduce the number of robot movements needed for learning inverse kinematics (IK). By decomposing IK into smaller functions, training time is significantly decreased, especially for high-precision tasks.

Area of Science:

  • Robotics
  • Machine Learning
  • Artificial Intelligence

Background:

  • Neural networks require extensive data for robot inverse kinematics (IK) approximation.
  • High precision in robot arm movements demands a large number of training samples.

Purpose of the Study:

  • To propose a method for reducing training data for robot IK.
  • To decrease the number of robot movements needed for learning IK.
  • To enable faster learning and relearning of IK for industrial robots.

Main Methods:

  • Expressing IK as a composition of learnable functions with reduced dimensionality.
  • Developing off-line and on-line training schemes for component functions.
  • Utilizing nearest neighbors and parameterized self-organizing maps for experimentation.

Related Experiment Videos

Main Results:

  • The proposed decomposition method significantly reduces the number of required training samples.
  • Time savings increase polynomially with the required precision.
  • Experimental validation confirms the efficiency of the decomposition approach.

Conclusions:

  • Decomposing IK into learnable functions is an effective strategy for reducing training data.
  • This method offers substantial time savings for learning robot IK, particularly for high-precision applications.
  • The approach is applicable to most industrial robots, enhancing learning efficiency.