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

Model reference direct adaptive control of nonlinear plants using neural networks

M Bahrami1, K E Tait

  • 1School of Electrical Engineering, University of New South Wales, Kensington, Australia.

International Journal of Neural Systems
|March 1, 1994
PubMed
Summary

A novel supervised steepest descent learning scheme for neural network controllers eliminates the need for error backpropagation and plant identification. This direct adaptive control method demonstrates satisfactory performance in nonlinear plant simulations.

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

Evaluating Dose-response of Cataract Induction in Radiotherapy of Head and Neck Cancers Patients.

Journal of biomedical physics & engineering·2021
Same author

Evaluation of the effect of BMSCs condition media and methylprednisolone in TGF-β expression and functional recovery after an acute spinal cord injury.

Bratislavske lekarske listy·2019
Same author

Empowerment of women and mental health improvement with a Preventive approach.

Journal of education and health promotion·2018
Same author

An efficient size-dependent shear deformable shell model and molecular dynamics simulation for axial instability analysis of silicon nanoshells.

Journal of molecular graphics & modelling·2017
Same author

Exchange, correlation, and scattering effects on surface plasmons in arm-chair graphene nanoribbons.

Optics express·2017
Same author

Adipose derived mesenchymal stem cells express keratinocyte lineage markers in a co-culture model.

Cellular and molecular biology (Noisy-le-Grand, France)·2016

Area of Science:

  • Control Engineering
  • Artificial Intelligence
  • Machine Learning

Background:

  • Traditional methods for neural network-based plant control often require a separate plant identification stage.
  • Existing adaptive control schemes may involve complex error backpropagation algorithms.
  • Direct adaptive control offers a potentially more efficient approach to managing nonlinear systems.

Purpose of the Study:

  • To propose a new learning scheme for multilayer feedforward neural networks as direct adaptive controllers.
  • To develop a method that avoids the error backpropagation requirement in neural network training.
  • To demonstrate the advantages of this scheme over existing methodologies for nonlinear plant control.

Main Methods:

  • A supervised steepest descent learning algorithm is employed for training the neural network controller.

Related Experiment Videos

  • The proposed scheme facilitates direct adaptive control without a preliminary plant identification phase.
  • The neural network architecture and training methodology are detailed.

    • Main Results:

    • The learning scheme successfully trains multilayer feedforward neural networks for direct adaptive control.
    • Simulations using model reference control of nonlinear plants indicate satisfactory performance.
    • The absence of error backpropagation and plant identification simplifies the control process.

    Conclusions:

    • The suggested learning scheme provides an effective and simplified approach to direct adaptive control using neural networks.
    • This method surpasses existing techniques by eliminating the need for plant identification and backpropagation.
    • The findings support the practical application of this direct adaptive control strategy for nonlinear systems.