Fuzzy and TSK Systems for Control, Regression, and Classification

This repository contains a collection of fuzzy logic and Takagi–Sugeno–Kang (TSK) modeling projects, developed in MATLAB / Simulink as part of academic coursework in control systems, fuzzy systems, regression, and classification.

The repository is organized as four independent projects, each fully self-contained, reproducible, and documented.

---

Repository Structure

/
├── data/
│ ├── airfoil_self_noise.dat
│ ├── epileptic_seizure_data.csv
│ ├── haberman.data
│ └── superconduct.csv
│
├── project1_dc_motor/
│ ├── src/
│ ├── figures/
│ └── README.md
│
├── project2_car_control/
│ ├── src/
│ ├── figures/
│ └── README.md
│
├── project3_regression/
│ ├── src/
│ ├── figures/
│ ├── final_figures/
│ ├── docs/
│ └── README.md
│
├── project4_classification/
│ ├── src/
│ ├── figures/
│ ├── docs/
│ └── README.md
│
└── README.md

---

Projects Overview

Project 1 – Fuzzy PI Control of a DC Motor

Folder: project1_dc_motor/

• Mamdani-type Fuzzy PI Controller
• DC motor speed control
• MATLAB & Simulink implementation
• Time-domain and frequency-domain analysis
• Bode diagrams and step responses

Key topics:

• Fuzzy control
• PI-equivalent fuzzy controllers
• Closed-loop stability and robustness

---

Project 2 – Fuzzy Logic Car Control with Obstacle Avoidance

Folder: project2_car_control/

• Mamdani Fuzzy Logic Controller (FLC)
• 2D vehicle navigation
• Obstacle avoidance with constant speed
• Steering control via heading correction

Key topics:

• Rule-based fuzzy control
• Multi-input decision making
• Autonomous navigation logic

---

Project 3 – TSK Fuzzy Regression Models

Folder: project3_regression/

• Takagi–Sugeno–Kang (TSK) fuzzy regression
• Multiple real-world datasets
• Comparison of:
  • Linear regression
  • ANFIS
  • TSK models with different rule counts
• Quantitative evaluation using:
  • MSE
  • RMSE
  • R²

Key topics:

• Data-driven fuzzy modeling
• Nonlinear regression
• Model comparison and validation

---

Project 4 – Fuzzy & Neuro-Fuzzy Classification

Folder: project4_classification/

• Fuzzy and ANFIS-based classifiers
• Binary and multi-class classification problems
• Evaluation on benchmark datasets
• Performance metrics:
  • Accuracy
  • Confusion matrices
  • Error rates

Key topics:

• Fuzzy classification
• Neuro-fuzzy systems
• Supervised learning with fuzzy models

---

Shared Datasets

All datasets used by Projects 3 and 4 are stored centrally in:

data/

This avoids duplication and ensures consistency across experiments.

---

How to Use

1. Clone the repository:

   git clone https://github.com/<your-username>/fuzzy-tsk-modeling.git
   

2. Open MATLAB.

3. Navigate to the desired project folder:

cd projectX_...

4. Follow the instructions in the project-specific README.md.

Each project can be executed independently.

Requirements

• MATLAB
• Fuzzy Logic Toolbox
• Control System Toolbox
• Simulink (for control projects)
• Statistics and Machine Learning Toolbox (for regression/classification)

Notes

• Each figures/ directory is documentation-only
• Only files under src/ are required to reproduce simulations
• All experiments are deterministic and fully reproducible
• No proprietary or hidden dependencies are used

Author

Ilias Korompilis

License

Academic and educational use only.