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

Biological arm motion through reinforcement learning.

Jun Izawa1, Toshiyuki Kondo, Koji Ito

  • 1Sensory and Motor Research Group, Human and Information Science Laboratory, NTT Communication Science Laboratories 3-1, Morinosato-Wakamiya, 243-01, Atsugi-shi, Japan. izawa@idea.brl.ntt.co.jp

Biological Cybernetics
|August 17, 2004
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

Preoperative posterior pelvic tilt is associated with residual low back pain and inferior patient-reported outcomes after total hip arthroplasty.

International orthopaedics·2026
Same author

Association between intraoperative irrigation volume and postoperative abdominal abscess in perforated peptic ulcer.

BMC surgery·2026
Same author

OsHAK1 is a major transporter mediating the uptake of monovalent thallium ions in rice under potassium-deficient conditions.

Plant & cell physiology·2026
Same author

Development and internal-external validation of a machine learning model to predict the risk for postoperative adverse events in patients with peptic ulcer perforation: a secondary cohort study of the perforated peptic ulcer analyzing project study.

International journal of surgery (London, England)·2026
Same author

Temporal Changes in Finger and Upper Limb Usage During Hospitalization After Distal Radius Fracture Surgery: A Case Report.

Cureus·2026
Same author

Effect of low-temperature thermal dechlorination on polychlorinated naphthalenes in municipal solid waste incineration fly ash: Concentrations, congener profiles, and dioxin-like toxicity.

Chemosphere·2026

This study introduces a novel reinforcement learning method to control biological systems with redundant actuators. The approach optimizes muscle activation for smooth, efficient movement without prior dynamic knowledge.

Area of Science:

  • Robotics and Control Systems
  • Computational Neuroscience
  • Biomechanical Engineering

Background:

  • Biological systems often feature redundant actuators, such as muscles, complicating direct control.
  • Applying reinforcement learning (RL) to these systems is challenging due to the high dimensionality and redundancy of the muscle activation space.

Purpose of the Study:

  • To develop an optimal learning control method for biological systems with redundant actuators using reinforcement learning.
  • To address the difficulties in applying RL to biological control systems stemming from muscle activation redundancy.

Main Methods:

  • A hierarchical control strategy was implemented, dividing the control input space into prioritized subspaces.
  • Reinforcement learning search noise was initially confined to a high-priority subspace, focusing on impedance control.

Related Experiment Videos

  • Learning constraints were progressively relaxed, expanding the search space to a lower-priority subspace as learning advanced.

    • Main Results:

    • The proposed method successfully enabled smooth reaching motions in a simulated biological system.
    • The reinforcement learning approach achieved effective control without requiring prior knowledge of the arm's dynamics.
    • The prioritized subspace learning effectively managed actuator redundancy.

    Conclusions:

    • This hierarchical reinforcement learning approach offers a viable solution for controlling redundant biological actuators.
    • The method facilitates the development of adaptive and efficient motor control in robotics and neuroscience.
    • Optimal learning control for complex biological systems can be achieved by managing control input space complexity.