🤖 AITL-H: Hybrid Structural Control Framework
⚠️ Note
The AITL-H project is currently in the conceptual and testing stage.
- Stability and response design rely on control theory (e.g., PID) as a prerequisite.
- FSM and LLM are designed as outer supervisory / support layers, and are not part of the stability-guaranteed control loop.
- The contents are conceptual ideas and partial PoC implementations, and are subject to change as further testing and development progress.
🆕 Update Log
| Date | Update | Reference |
|---|---|---|
| 2025-08-25 | 🚩 Added Humanoid Robot PoC (Flagship) to top | PoC Page |
| 2025-08-25 | 📑 Published 3 PoC reports (PWM Ripple / Thermal / Mission Energy) | Docs Index |
| 2025-08-25 | 🎤 Added template for presentation slides | Slides |
🔗 Official Links
| Language | GitHub Pages 🌐 | GitHub 💻 |
|---|---|---|
| 🇯🇵 Japanese |  |
 |
| 🇺🇸 English |  |
 |
🧭 Overview
| Item | Details |
|---|---|
| Name | AITL-H (Hybrid) |
| Objective | Establishing humanoid robot control methods using structural AI control |
| Core Principles | - FSM: instinctive behavior control via state transitions - PID: continuous control of physical quantities (angle, velocity) - LLM: intelligence through advanced decision-making, dialogue, and learning |
🧘 Three-Layer Architecture
| Layer | Function | Implementation |
|---|---|---|
| FSM Layer | Logic control based on state transitions | fsm_engine.py, fsm_state_def.yaml |
| PID Layer | Physical control of joints and motion quantities | pid_controller.py, pid_module.py |
| LLM Layer | Situation assessment, anomaly detection, language response | llm_interface.py, llm_logger.py |
Each layer is loosely coupled yet cooperative, allowing independent development and step-by-step integration.
flowchart TB
subgraph LLM["LLM Layer"]
L1[Decision-Making]
L2[Anomaly Detection]
L3[Language Response]
end
subgraph PID["PID Layer"]
P1[Continuous Control]
P2[Joint Angles / MIMO]
end
subgraph FSM["FSM Layer"]
F1[Logic Control]
F2[State Transitions]
end
LLM -->|Scenario / Commands| FSM
FSM -->|Mode Control / Gain Select| PID
PID -->|PWM / Control Signals| ACT["Actuators"]
ACT -->|Motion Response| SEN["Sensors (IMU, etc.)"]
SEN -->|Perception Feedback| LLM
🌏 Strategic Significance
AITL-H is not just a control architecture.
By integrating state feedback control and state transition control, and combining with LLMs and SystemDK, it achieves real-time optimal design under physical constraints.
- Industrial impact
- 94% reduction in fault response time (PoC evaluation)
- 8× faster reconfiguration of production lines
- 40% reduction in design change costs
- National significance
- Securing competitiveness in advanced-node semiconductors and industrial autonomous systems
- Gaining leadership in international standardization
This technology must be integrated now.
SystemDK is not unique to AITL-H but is an essential foundational technology for all advanced-node semiconductor designs.
🧪 PoC Related
| Title | Description | Path |
|---|---|---|
| 🚩 Humanoid Robot PoC (Flagship) | Integrated flagship with FSM × PID × LLM × State-Space × Energy Harvesting |   |
| 🧭 Gimbal Control (FSM + PID + LLM) | Educational PoC for hybrid closed-loop control | |
| ⚙️ Verilog Auto-Generation (FSM + PID) | YAML → C → Verilog conversion & verification | |
| 🛠 Auto Generator | Tools for YAML → C → Verilog conversion of FSM/PID configs | |
🗺️ Project Relationship Map
flowchart TB
EC["EduController
(Control Theory → AI Control)"]
AITLH["AITL-H
Hybrid Control & SystemDK"]
ESV["Edusemi-v4x
(SoC/RTL/Layout)"]
EC -- Teaching Feed --> AITLH
AITLH -- Methods & PoC Results --> ESV
EC -- Cross Reference --> ESV
A simple diagram showing the cross-reference among EduController ⇔ AITL-H ⇔ Edusemi-v4x.
🤖 ChatGPT Support Tools
Provided in accelerated_design/: Design support tools using ChatGPT
- State transition design support (prompt → FSM YAML automation)
- Test scenario / log visualization
- Automatic generation of design documents
A human-AI collaborative design framework.
🎛️ Connection with EduController
AITL-H is fully integrated with Chapter 9 (FSM × PID × LLM Hybrid Control) of the educational material EduController.
| Part | Content | Relation to AITL-H |
|---|---|---|
| Part 01–05 |
Classical to modern control theory (PID, state-space, etc.) | Theoretical basis of PID layer |
| Part 06–08 |
AI control (NN control, reinforcement learning, data-driven) | Supplementary knowledge for AI application design |
Part 09 |
FSM × PID × LLM integrated control | AITL-H architecture implemented as educational material |
🎓 Integrated Design Deployment with Edusemi-v4x
To extend towards SoC/RTL design, see the “Special Edition” of Edusemi-v4x, which covers:
| Chapter | Content | Link |
|---|---|---|
| Chapter 3 | SoC design with FSM × PID × LLM integrated control | |
| Chapter 4 | RTL → GDSII layout automation with OpenLane | |
| Chapter 5 | Physical verification and consistency check with DRC / LVS / DFM | |
📌 If you want to study physical constraints in more depth
Once you understand the flow from SoC design to physical verification, proceed to Special Edition Chapter 2a: Handling Thermal, Stress, and Noise Constraints in SystemDK.
📚 Related Projects
| Project | Description | Link |
|---|---|---|
| Edusemi-v4x | Semiconductor / SoC design educational material | |
| EduController | Control theory × AI control educational material | |
| SamizoGPT | Project Design Hub guide management | |
| AITL-Strategy-Proposal | AITL strategic proposals and policy recommendations | |
👤 Author
| 📌 Item | Details |
|---|---|
| Name | Shinichi Samizo Shinichi Samizo |
| Expertise | Semiconductor devices (logic, memory, high-voltage mixed integration) Inkjet thin-film piezo actuators Productization of printheads, BOM management, ISO training |
| 💻 GitHub |  |
📄 License
This project adopts a hybrid license
Depending on the nature of the educational materials, code, and figures, the following licenses apply.
| 📌 Item | License | Description |
|---|---|---|
| Code | MIT License | Free to use, modify, redistribute |
| Text materials | CC BY 4.0 | Attribution required |
| Figures & diagrams | CC BY-NC 4.0 | Non-commercial use only |
| External references | Follow original license | Clearly indicate source |
💬 Feedback
Please submit improvement suggestions or start discussions via GitHub Discussions.
