STM32 Self-Balancing Robot 🤖

Course: Embedded System Design (CE224.Q12)

Institution: University of Information Technology (UIT) - VNU-HCM

Instructor: Mr. Tran Ngoc Duc

📖 Overview

This project implements a Two-Wheeled Self-Balancing Robot based on the Inverted Pendulum mathematical model. The system is powered by the STM32F401CCU6 microcontroller and utilizes FreeRTOS for real-time multitasking.

The control system features a Cascade PID algorithm (Velocity & Angle loops) and a Complementary Filter for sensor fusion, ensuring stability and smooth operation. This project was developed for the Embedded System Design (CE224.Q12) course at the University of Information Technology - VNU-HCM.

🔗 Full Project Report: [https://drive.google.com/file/d/1FqUx3Trcz4oJ_H03-fAWJpXKTJU6ttZQ/view?usp=drive_link]

🛠 Tech Stack

Hardware

• Microcontroller: STM32F401CCU6 (Blackpill - Cortex M4).
• IMU Sensor: MPU6050 (Accelerometer + Gyroscope).
• Motor Driver: TB6612FNG.
• Actuators: 2x GA25-370 DC Gear Motors (with Quadrature Encoders).
• Power: 3S Li-ion Battery (11.1V - 12.6V) + LM2596 Buck Converter (5V/3.3V).

Software & Tools

• IDE: STM32CubeIDE.
• Config Tool: STM32CubeMX.
• OS: FreeRTOS (Real-time Operating System).
• Language: C, ARM Assembly (Startup code).
• Debugging: USB Virtual COM Port (CDC).

✨ Key Features

• Real-Time OS: Utilizes FreeRTOS to manage tasks (Sensor Reading, PID Calculation, Debugging) with hardware timer interrupts for precise sampling (1ms).
• Cascade PID Control:
  • Inner Loop (Angle): Maintains upright stability using "Derivative on Measurement" to eliminate "Derivative Kick."
  • Outer Loop (Speed): Controls position and velocity.
• Sensor Fusion: Implements a Complementary Filter ($\alpha=0.97$) to combine Accelerometer stability with Gyroscope responsiveness.
• Anti-Windup: PID integral term saturation to prevent overshoot.
• Hardware Abstraction: Optimized use of STM32 HAL and Low-Level Direct Register Access (Macros) for PWM generation.

🔌 Pin Configuration

Component	Signal	STM32 Pin	Timer/Function
MPU6050	SCL	PB6	I2C1
	SDA	PB7	I2C1
Left Motor	PWM	PA8	TIM1_CH1
	DIR	PB3	GPIO
	ENC A	PB14	TIM3_CH1
	ENC B	PB15	TIM3_CH2
Right Motor	PWM	PA9	TIM1_CH2
	DIR	PA3	GPIO
	ENC A	PB13	TIM4_CH1
	ENC B	PB12	TIM4_CH2
System	LED	PC13	Status Indicator

🚀 Installation & Setup

1. Clone the Repository:

   git clone [https://github.com/NguyenDinhNhatNguyen/Two-Wheeled-Self-Balancing-Robot]

2. Hardware Setup:
   • Connect the components according to the Pin Configuration table above.
   • Ensure the MPU6050 is mounted flat and secure.
3. Open in STM32CubeIDE:
   • File -> Open Projects from File System -> Select the cloned folder.
4. Build & Flash:
   • Connect the STM32F4 via ST-Link.
   • Click Run (Green Play Button).
5. Calibration:
   • Hold the robot upright (static) for the first 3 seconds after boot to allow Gyroscope Offset calibration.

📊 Control Logic (PID)

The system uses a custom PID implementation: $$Output = K_p \times e(t) + K_i \times \int e(t)dt - K_d \times Gyro_{filtered}$$

• Proportional (P): Corrects error immediately.
• Integral (I): Corrects steady-state error (Speed loop).
• Derivative (D): Uses raw Gyroscope data instead of Error Derivative to dampen oscillations and prevent spikes when the setpoint changes.

👥 Contributors

• Nguyen Viet Thien Nhan (23521086) - PID Algorithm, Motor Driver, Hardware Design.
• Nguyen Dinh Nhat Nguyen (23521043) - MPU6050 Driver, System Config (I2C, Timer), Hardware Assembly.

---

Ho Chi Minh City, December 2025