Related Experiment Videos
Neural regulator design
1Department of Systems Engineering, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia
Summary
This study introduces a novel neural network controller for nonlinear systems, effectively managing both state and output feedback under various disturbances using advanced training methods.
Area of Science:
- Control Engineering
- Artificial Intelligence
- Nonlinear Systems
Background:
- Designing effective controllers for nonlinear plants presents significant challenges.
- Existing methods often struggle with complex dynamics and external disturbances.
- The integration of artificial intelligence offers potential solutions for advanced control strategies.
Purpose of the Study:
- To design and investigate a neural-net-based regulator for nonlinear plants.
- To address both state and output feedback control scenarios.
- To evaluate the controller's performance under deterministic and stochastic disturbances.
Main Methods:
- A Multilayered Feedforward Neural Network (MFNN) was employed as the nonlinear controller.
- The training methodology utilized the concept of Block Partial Derivatives (BPDs).
- The approach was applied to nonlinear plant control problems.
Main Results:
- The developed MFNN-based regulator demonstrated effective control for nonlinear plants.
- The controller showed robustness in handling both deterministic and stochastic disturbances.
- The BPDs training concept proved viable for MFNN controller development.
Conclusions:
- Neural network-based control offers a powerful approach for nonlinear systems.
- The MFNN with BPDs training is a promising technique for advanced regulator design.
- This method enhances control system performance in the presence of complex disturbances.