DC Motor Advanced Control (Pololu 37D + STM32F4)

State-feedback (LQR) and torque control for a brushed DC geared motor using an STM32F4 microcontroller. From modeling and identification to embedded implementation and validation.

---

🚀 Highlights

• Dual control layers: high‑performance position/velocity LQR + dedicated torque (current) loop.
• Observer + anti‑windup + discrete implementation for embedded real‑time execution (84 MHz Cortex‑M4).
• Full V‑Model engineering workflow (modeling, MIL / SIL / PIL / HIL traces) – see report.pdf.
• MATLAB / Simulink models mirrored by hand‑crafted C firmware (no auto‑code generation dependency for core logic).
• Data acquisition & rapid tuning scripts (UART → MATLAB live plotting).
• Modular encoder interface (quadrature Hall, 8400 counts / rev at output shaft with gearbox).

---

📋 Table of Contents

1. Motivation & Goals
2. System Overview (Simulink Models)
3. Hardware Kit
4. Control Design Overview
5. Repository Structure
6. Quick Start
7. Building & Flashing (STM32)
8. Results Snapshot

---

1. Motivation & Goals

Design and validate a robust, high‑fidelity control stack for a geared DC motor able to: (a) track position/velocity references with minimal overshoot and (b) deliver commanded torque (via current shaping) under load disturbances.

2. System Overview (Simulink Models)

The architectural block diagram is defined directly in the provided Simulink models. Refer to these files for the authoritative structure of the cascaded loops, observer integration, and PIL/SIL flow:

Position / Outer Loop Models:

• position/MIL_pos_pade_observer_discrete_antiwindup.slx – Full MIL validation (observer + anti‑windup + discrete implementation).
• position/SIL_PIL_pos.slx / position/SIL_PIL_extern_pos.slx – Software / Processor‑in‑the‑Loop integration with embedded code.

Torque / Inner Loop Models:

• torque/MIL_Torque_controller.slx – Torque (current) control design & tuning.
• torque/SIL_PIL_torque.slx / torque/SIL_PIL_torque_extern.slx – SIL / PIL validation of torque layer.
• torque/motor_model.slx – Electrical + mechanical plant (R, L, J, friction, gear reduction).

Support & Shared Components:

• quadratureEncoder_module/stm_qeplib.slx – Encoder (QEP) interface library.
• Inductance/Tuning_L.slx – Inductance identification workflow.
• position/observer.m, Inductance/Inductance.m, Sensor_calibration/sensor_calibration.m – Scripted identification / calibration steps feeding model parameters.

3. Hardware Kit

Component	Notes
STM32F401RE Nucleo	ARM Cortex‑M4 @ 84 MHz, UART, timers, ST‑Link
Pololu 37D Gearmotor	12 V brushed DC, 131.25:1 gearbox, 64 CPR internal encoder → 8400 counts/output rev
X-NUCLEO‑IHM04A1	Dual H‑bridge driver (direction + PWM)
ACS714 Current Sensor	Hall‑effect linear current sensing
Wheel + Mounting Kit	Mechanical load + fixture

4. Control Design Overview

Core steps (mirrored by scripts / models in repo):

• Parameter Identification: electrical (R, L via Inductance.m) & mechanical (gear ratio, friction, inertia) using captured current/voltage/speed datasets.
• State Model: continuous + discretized (zero‑order hold & Pade for delays where applicable).
• Outer Loop: LQR tuned for weighted compromise between settling time, steady error, and actuation energy.
• Inner Torque Loop: PI / LQR hybrid (depending on configuration) ensuring current → torque linearization.
• Observer: reconstructs unmeasured states (e.g. motor shaft speed) & compensates gearbox dynamics.
• Anti‑Windup: conditional integral clamping on saturation (implemented in discrete form for MCU tick).
• Validation Stages: MIL (pure Simulink) → SIL / PIL (SIL_PIL_* models) → on‑board tests logged over UART.

5. Repository Structure (Key Folders)

Path	Purpose
direct_coding/Position_Controller	Embedded firmware for position/velocity LQR controller
direct_coding/TorqueController	Embedded torque (current) control project
position/*.slx	Position control Simulink models (MIL / SIL / PIL)
torque/*.slx	Torque & motor electrical dynamics models
quadratureEncoder_module/	Reusable encoder (QEP) module (C++ / Simulink lib)
Inductance/	Inductance identification scripts & datasets
Sensor_calibration/	Current sensor calibration routine & baseline data
readFromCOM/	MATLAB scripts for serial streaming & logging
motor_data/	Captured reference vs response datasets
report.pdf	Full engineering report (derivations & analysis)

6. Quick Start

Prerequisites:

• STM32CubeIDE
• MATLAB / Simulink (Control System; versions 2020b+ tested)
• USB cable for Nucleo; lab power supply (≤ 12 V)

Clone:

git clone https://github.com/my-rice/DC-Motor-Control.git
cd DC-Motor-Control

Open the desired Simulink model (e.g. position/MIL_pos_pade_observer_discrete_antiwindup.slx) to explore the design.

7. Building & Flashing (STM32)

To test the hand-crafted embedded firmware:

1. Import direct_coding into STM32CubeIDE (File → Import → Existing Projects).
2. Configure clock & peripherals (already set in .ioc files; regenerate only if changing pins).
3. Build (Debug or Release).
4. Flash via ST‑Link (Run → Debug).
5. Open a serial terminal (115200 8N1 by default unless modified) or use the MATLAB script below.

To test the controllers with auto-generated code from Simulink:

1. Open the relevant SIL_PIL_*.slx model (e.g. position/SIL_PIL_extern_pos.slx).
2. Configure the Embedded Coder settings
3. Generate code and deploy to the Nucleo board via PIL workflow.
4. Open a serial terminal or use the MATLAB script for data logging.

8. Results Snapshot

The results are detailed in report.pdf with representative plots.

License

Licensed under the BSD 3-Clause License. See LICENSE for details.

---

🔧 Hardware Summary (Reference)

• Nucleo STM32F401RE: ARM Cortex‑M4 @ 84 MHz, UART, multiple timers → PWM generation for H‑bridge.
• Gearmotor: 131.25:1 reduction; effective resolution 8400 counts / rev (64 CPR × gear ratio).
• Driver (X‑NUCLEO‑IHM04A1): Dual full H‑bridge; directional control + enable/PWM pins.
• ACS714: Hall effect current sensing enabling torque estimation.

---

Feel free to open issues / PRs for improvements or questions.