🎛️ EduController: Educational Framework for Control Theory and AI Control
🔗 Official Links
| Language |
GitHub Pages 🌐 |
GitHub 💻 |
| 🇺🇸 English |
 |
 |
| 🇯🇵 Japanese |
 |
 |
📘 Overview
EN:
EduController is a step-by-step, practical educational project that covers classical control, modern control, and AI-based next-generation control. It supports a wide range of topics from intuitive understanding of control theory in Python to HDL coding and LLM-integrated design.
🌐 Toward Next-Generation Control — FSM × PID × LLM
A three-layer hybrid control architecture with AITL Framework
Integrating FSM, PID, and LLM for intelligent control systems
🚀 Learn More
🧭 Structure Overview
| Track |
Overview (EN) |
| 🎛️ Control Theory Track (Part 01–05) |
Classical control, state-space, digital control, practical implementation |
| 🤖 AI Control Track (Part 06–08) |
Neural networks, reinforcement learning, data-driven control |
| 🧠 Integrated & Applied Control Track (Part 09–10) |
LLM-integrated control, inverted pendulum control |
📚 Chapter Structure
🎛️ Control Theory Track / Classical & Modern Control
| Chapter |
Title |
Summary |
| Part 01 |
Classical Control Theory
  |
Systematic study of PID control, time-domain and frequency-domain analysis & design. |
| Part 02 |
Modern Control Theory
|
Covers state-space representation, controllability, observability, pole placement, and observer design. |
| Part 03 |
Adaptive & Robust Control
|
MRAC, H∞ control, L1 control for robustness against parameter variations and disturbances. |
| Part 04 |
Digital Control & Signal Processing
|
Z-transform, discrete PID, digital filter design for implementation. |
| Part 05 |
Implementation & Applications
|
Python implementation, ROS practice, FPGA-based control for real systems.  |
🤖 AI Control Track / AI-based Control
| Chapter |
Title |
Summary |
| Part 06 |
Neural Network Control
|
NN-PID design, inverse model control using neural networks. |
| Part 07 |
Reinforcement Learning Control
|
Applying RL to inverted pendulum & vehicle control; implementing DDPG, PPO. |
| Part 08 |
Data-Driven Control
|
Model-free control using Koopman operator, system identification. |
🧠 Integrated & Applied Control Track / Integrated Control
| Chapter |
Title |
Summary |
| Part 09 |
Hybrid Control with LLM Integration
|
Three-layer architecture (FSM×PID×LLM) for next-gen control. |
| Part 10 |
Integrated Control of Inverted Pendulum
|
Integrated PID, LQR, RL, and HDL implementation for inverted pendulum control. |
| Directory |
Summary |
matlab_tools/
|
Visualization in Simulink, C code generation, HDL design. |
SoC_DesignKit_by_ChatGPT/
|
FSM, PID, LLM control templates, Verilog generation, testbench verification. |
🧭 Usage Flow Overview
These tools provide an end-to-end flow from model design to RTL verification.
Two types of C sources are supported (C generated from Simulink / handwritten C), enabling integration of PID/FSM/LLM control logic into HDL.
- Simulink or Handwritten C (matlab_tools/ etc.)
- Create a model in Simulink and generate fixed-point C code, or
prepare handwritten C step functions for FSM/LLM control.
- C code → HDL (SoC_DesignKit_by_ChatGPT/)
- Map C functions (PID / FSM / LLM kernels, etc.) to templates,
and automatically generate Verilog/SystemVerilog plus testbenches.
- Multiple C modules can be integrated within the same SoC (e.g., PID + FSM + LLM I/F).
- Simulation & Verification
- Verify functional equivalence between C implementation and RTL using the auto-generated testbench.
- Then proceed to synthesis and deployment in FPGA/ASIC flows as needed.
flowchart TB
A[Simulink Model]
A2[Handwritten C: FSM / LLM control]
B[C code - fixed-point]
C[SoC_DesignKit_by_ChatGPT Template Mapping]
D[RTL Generation : Verilog / SystemVerilog]
E[Testbench & Simulation]
F[FPGA / ASIC Synthesis]
A --> B
A2 --> B
B --> C
C --> D
D --> E
E --> F
| Project |
Links |
Description |
| 🎓 Edusemi-v4x |
|
Semiconductor design & process education (Python, sky130, OpenLane)
|
| 🤖 AITL-H |
|
Three-layer control framework (FSM×PID×LLM)
|
| 🧠 SamizoGPT |
|
Prompt design templates for ChatGPT (design assistance)
|
👤 Author
| Item |
Details |
| Name |
Shinichi Samizo |
| Education |
M.S. in Electrical and Electronic Engineering, Shinshu University |
| Career |
Former Engineer at Seiko Epson Corporation (since 1997) |
| Expertise |
Semiconductor devices (logic, memory, high-voltage mixed integration); Inkjet thin-film piezo actuators; Productization of PrecisionCore printheads, BOM management, and ISO training |
| ✉️ Email |
 |
| 🐦 X |
 |
| 💻 GitHub |
 |
📄 License
The default license is MIT, but the following specific directories/materials use a Hybrid License.
The default license is MIT, but specific directories/materials adopt a Hybrid License.
| 📌 Item |
License |
Description |
| Default |
MIT License |
Free to use, modify, and redistribute |
| Hybrid Scope |
Part05, Part09, Part10, matlab_tools, SoC_DesignKit_by_ChatGPT |
Depending on the nature of the materials, code, and diagrams, applies MIT License / CC BY 4.0 / CC BY-SA 4.0 / CC BY-NC 4.0 |
💬 Feedback
Propose improvements or start discussions via GitHub Discussions.
