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

Self-Organized Behavior Generation for Musculoskeletal Robots.

Ralf Der1, Georg Martius2

  • 1Institute for Computer Science, University of Leipzig Leipzig, Germany.

Frontiers in Neurorobotics
|April 1, 2017
PubMed
Summary

This study introduces a self-learning robot controller that discovers object properties through physical interaction, enabling unique movement patterns for each object. This neurocontroller advances robot adaptability and understanding of sensorimotor control.

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

Sensing multi-directional forces at superresolution using taxel value isoline theory.

Nature communications·2025
Same author

Emergence of natural and robust bipedal walking by learning from biologically plausible objectives.

iScience·2025
Same author

Machine Learning of a Density Functional for Anisotropic Patchy Particles.

Journal of chemical theory and computation·2024
Same author

When to Be Critical? Performance and Evolvability in Different Regimes of Neural Ising Agents.

Artificial life·2022
Same author

Guiding the design of superresolution tactile skins with taxel value isolines theory.

Science robotics·2022
Same author

Editorial: Complexity and Self-Organization.

Frontiers in robotics and AI·2021

Area of Science:

  • Robotics and Control Systems
  • Artificial Intelligence
  • Dynamical Systems Theory

Background:

  • Traditional robot control struggles with complex, unknown dynamics.
  • Self-organized control offers scalability, robustness, and resilience for such systems.
  • Recent advancements in extrinsic differential plasticity provide a novel learning mechanism.

Purpose of the Study:

  • To present a self-learning neurocontroller utilizing extrinsic differential plasticity.
  • To apply this controller to an anthropomorphic robot arm interacting with objects of unknown dynamics.
  • To investigate the emergence of object-specific sensorimotor patterns and hand-eye coordination.

Main Methods:

  • Implementation of a self-learning neurocontroller based on extrinsic differential plasticity.
Keywords:
anthropomimeticlearningmusculoskeletalrobot controlself-explorationself-organizationtendon-driven

Related Experiment Videos

  • Experimental setup with an anthropomorphic robot arm and various attached objects (pendulum, bottle, wheel, brush).
  • Integration of camera-based hand coordinate feedback for hand-eye coordination.

    • Main Results:

    • The robot autonomously learns distinct sensorimotor patterns for different objects (e.g., circular motion for pendulum, shaking for bottle).
    • The robot demonstrates object recognition through learned dynamical patterns, discovering object affordances.
    • Spontaneous self-organization of hand-eye coordination is observed when visual feedback is included.
    • Behaviors emerge from operating at the border of instability, utilizing limit cycle attractors and attractor morphing.

    Conclusions:

    • The self-learning neurocontroller effectively enables robots to adapt to and learn from unknown object dynamics.
    • This approach facilitates the discovery of object affordances and robust sensorimotor pattern generation.
    • The findings offer insights into human sensorimotor control and subjective feelings, with potential feedback for robotics research.