Jove
Visualize
Contact Us

Related Experiment Videos

The balance between initial training and lifelong adaptation in evolving robot controllers.

Joanne H Walker1, Simon M Garrett, Myra S Wilson

  • 1Department of Computer Science, University of Wales, Aberystwyth SY23 3DB, UK. jnw@aber.ac.uk

IEEE Transactions on Systems, Man, and Cybernetics. Part B, Cybernetics : a Publication of the IEEE Systems, Man, and Cybernetics Society
|April 11, 2006
PubMed
Summary

Robots can achieve lifelong adaptation in changing environments using evolutionary methods embodied on the robot. This approach combines initial training with continuous evolution for improved real-world performance.

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

[Genotypes of Staphylococcus aureus strains with methicillin resistant phenotype].

Revista medica de Chile·2007
Same author

How do we evaluate artificial immune systems?

Evolutionary computation·2005
Same author

On the use of qualitative reasoning to simulate and identify metabolic pathways.

Bioinformatics (Oxford, England)·2005
See all related articles
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

Area of Science:

  • Robotics
  • Artificial Intelligence
  • Evolutionary Computation

Background:

  • Designing robots for real-world performance requires adaptability to dynamic environments.
  • Robots must improve in stable conditions and adapt to changes.
  • Lifelong learning and adaptation are key challenges in robotics.

Purpose of the Study:

  • To investigate the use of evolutionary methods for lifelong robot adaptation.
  • To demonstrate how evolutionary processes can be embodied directly on a robot.
  • To highlight the synergistic role of initial training and continuous adaptation.

Main Methods:

  • Experimental validation of evolutionary algorithms for robot adaptation.
  • Implementation of an embodied evolutionary process on a robotic platform.

Related Experiment Videos

  • Integration of a distinct training phase with lifelong evolutionary adaptation.

    • Main Results:

    • Evolutionary methods enable robots to achieve lifelong adaptation.
    • Embodied evolutionary processes on the robot facilitate continuous performance improvement.
    • Initial training significantly enhances the effectiveness of lifelong adaptation.

    Conclusions:

    • Evolutionary computation offers a viable pathway for lifelong robot adaptation.
    • Embodied evolution on the robot is a practical approach for real-world challenges.
    • A combination of structured training and evolutionary adaptation is crucial for robust robotic systems.