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

Neural networks for continuous online learning and control.

Min Chee Choy1, Dipti Srinivasan, Ruey Long Cheu

  • 1Department of Electrical and Computer Engineering, National University of Singapore, Singapore 117576, Singapore. engp1637@nus.edu.sg

IEEE Transactions on Neural Networks
|November 30, 2006
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

Objective evaluation of deep uncertainty predictions for COVID-19 detection.

Scientific reports·2022
Same author

Handling of uncertainty in medical data using machine learning and probability theory techniques: a review of 30 years (1991-2020).

Annals of operations research·2021
Same author

An Uncertainty-Aware Transfer Learning-Based Framework for COVID-19 Diagnosis.

IEEE transactions on neural networks and learning systems·2021
Same author

Health screening strategies for international air travelers during an epidemic or pandemic.

Journal of air transport management·2020
Same author

A Survey of Computational Intelligence Techniques for Wind Power Uncertainty Quantification in Smart Grids.

IEEE transactions on neural networks and learning systems·2019
Same author

Asynchronous State Estimation for Discrete-Time Switched Complex Networks With Communication Constraints.

IEEE transactions on neural networks and learning systems·2017

A new hybrid neural network model uses multistage online learning for traffic signal control. This advanced system significantly reduces vehicle delays and stoppage times in large-scale networks.

Area of Science:

  • Artificial Intelligence
  • Control Systems Engineering
  • Transportation Engineering

Background:

  • Distributed control problems with infinite horizons present significant challenges for real-time adaptation.
  • Traffic signal control requires dynamic adjustments to changing traffic volumes to prevent congestion.

Purpose of the Study:

  • To propose and evaluate a novel hybrid neural network (NN) model for solving infinite horizon distributed control problems.
  • To apply this model to real-time distributed traffic signal control in a large-scale network.

Main Methods:

  • Development of a hybrid NN model incorporating reinforcement learning and evolutionary algorithms for multistage online learning.
  • Implementation within a PARAMICS microscopic simulation of Singapore's Central Business District.

Related Experiment Videos

  • Comparison against existing traffic signal control algorithms and a simultaneous perturbation stochastic approximation-based neural network (SPSA-NN).

    • Main Results:

    • The hybrid NN model demonstrated significant improvements in traffic conditions.
    • Achieved a 78% reduction in total mean vehicle delay and an 84% reduction in total mean stoppage time compared to existing algorithms.
    • Outperformed the SPSA-NN in complex, large-scale simulations.

    Conclusions:

    • The hybrid NN model is effective for large-scale, distributed traffic signal control problems with infinite horizons.
    • The model's adaptive online learning capabilities ensure efficient traffic management.
    • The approach shows potential for broader applications in similar distributed control scenarios.