JoaoESmoreira

Home Projects Blog Emacsphere

Fuzzy and Neuro-Fuzzy Control of Dynamic Systems

The project focuses on the design, implementation, and testing of both fuzzy controllers (Mamdani and Sugeno) and neuro-fuzzy systems for modeling dynamic processes.

The full repository and documentation can be found here: Fuzzy and Neuro-Fuzzy Control of Dynamic Systems Repository.

Tools and Technologies

  • MATLAB
  • Simulink
  • Fuzzy Logic Toolbox
  • ANFIS (Adaptive Neuro-Fuzzy Inference System)

Learning Objectives

  1. To design and test fuzzy controllers (Mamdani and TSK types).
  2. To model a dynamic process using clustering and optimization techniques.

Part A – Fuzzy Control

The first part consists of implementing two fuzzy controllers, Mamdani and Sugeno, using MATLAB’s Simulink environment and the Fuzzy Logic Toolbox.

Requirements

  • MATLAB with the Fuzzy Logic Toolbox.
  • Implementation of:
    • Mamdani controllers with 9 and 25 rules.
    • Sugeno controllers with 9 and 25 rules.
  • The system’s dynamics are defined by a given continuous transfer function.
  • Analysis of controller performance regarding:
    • Reference tracking.
    • Load and actuator disturbance compensation.
    • Control effort minimization.

Tools and Methods

  • fuzzyLogicDesigner or fuzzy command for building the FIS.
  • Simulink for simulation and visualization of system behavior.
  • Performance evaluation based on error integration and control effort.

Deliverables

  • A PDF report describing the controllers and their performance.
  • The Simulink diagram (.slx).
  • Controller files (.fis).

Part B – Neuro-Fuzzy Systems

The second part focuses on building neuro-fuzzy models of dynamic systems using clustering and adaptive learning techniques.

Methodology

  • The dynamic process is modeled based on input-output time series data.
  • Training is performed using the ANFIS architecture with clustering and optimization.
  • The model aims to approximate the nonlinear function: \[ y(k) = f(y(k-1), y(k-2), y(k-3), u(k-1), u(k-2), u(k-3)) \]

Tools and Techniques

  • anfisedit or neuroFuzzyDesigner for the ANFIS GUI.
  • Clustering methods:
    • Subtractive Clustering
    • Fuzzy C-Means (FCM)
  • Data partitioning:
    • 70% for training
    • 30% for testing
  • Optimization performed by hybrid or backpropagation methods.

Deliverables

  • A brief report describing the neuro-fuzzy models and performance results.
  • Plots of membership functions for each input variable.
  • Fuzzy model files (.fis), MATLAB training/testing scripts (.m), and Simulink diagrams (if used).

Notes

  • Use appropriate sampling times to ensure numerical stability.
  • Avoid reproducing assignment text; focus on analysis and interpretation.
  • Reports and files must be submitted in a single compressed archive named:
    • ML2023FuzzyPLxGy.zip
    • The report file: ML2023FuzzyPLxGy.pdf

Footer

Copyright © 2025 Joao ES Moreira

The contents of this website are licensed under the Creative Commons Attribution-NoDerivatives 4.0 International License (CC-BY-ND 4.0).

The source code of this website is licensed under the MIT license, and available in GitHub repositor. User-submitted contributions to the site are welcome, as long as the contributor agrees to license their submission with the CC-BY-ND 4.0 license.