PID Motor Control

Closed-loop PID motor controller on STM32 Nucleo-F411RE with real-time IMU feedback. Live tuning happens over UART, with step response capture to CSV for plotting.

Hardware

Wiring

STM32 Nucleo-F411RE          TB6612FNG
  PA6 (TIM3_CH1) ----------> PWMA
  PA5 ----------------------> AIN1
  PA8 ----------------------> AIN2
  PA9 ----------------------> STBY
  3.3V ---------------------> VCC
  GND ----------------------> GND
  external 5-12V -----------> VM

STM32                         MPU-6050
  PB8 (I2C1_SCL) --+-------> SCL
  PB9 (I2C1_SDA) --+-------> SDA
  3.3V -------[4.7K]--+      VCC
              [4.7K]--+      GND
  pull-ups to 3.3V           AD0 -> GND, address 0x68

STM32                         SSD1306 OLED
  PB8 (shared SCL) ---------> SCL
  PB9 (shared SDA) ---------> SDA
  3.3V ---------------------> VCC
  GND ----------------------> GND

USART2 on PA2/PA3 is routed through the ST-Link virtual COM port.

Architecture

TIM3_CH1 (20 kHz PWM) --> TB6612FNG --> DC Motor
                                          |
MPU-6050 (I2C, 14-byte burst read) <-----+
    |                                  (physical coupling)
    v
Complementary Filter (alpha=0.98)
    |
    v
PID Controller (1 kHz, TIM2 ISR)
    |  - Derivative on measurement
    |  - Low-pass filter on D term (N=10)
    |  - Integral clamping anti-windup
    v
Motor effort (-1.0 to +1.0) --> sign-magnitude drive

Main loop:
  - UART CLI for Kp, Ki, Kd, setpoint, and step response commands
  - OLED display for gains, setpoint, error, and output bar
  - CSV step response stream for offline plotting

Control Design

The PID loop runs at 1 kHz in TIM2’s update ISR. Three techniques keep the output clean:

Derivative on measurement. The D term differentiates the process variable, not the error signal. This eliminates the derivative kick that occurs when the setpoint changes abruptly.

Low-pass filter on D term. A first-order filter with N=10 smooths high-frequency noise from the IMU without adding significant phase lag at the control bandwidth.

Integral anti-windup. The integrator is clamped to the output range. When the motor saturates, the integrator stops accumulating, preventing overshoot on recovery.

UART CLI

The main loop runs a serial command interface for live tuning:

Step response data streams as CSV over UART for offline plotting with scripts/plot_step_response.py.

Build

Requires arm-none-eabi-gcc and CMake.

mkdir build && cd build
cmake -DCMAKE_BUILD_TYPE=Debug ..
make

Flash via ST-Link:

make flash
# or:
st-flash write pid-motor-control.bin 0x08000000

Connect a serial terminal at 115200 baud:

screen /dev/tty.usbmodem* 115200

Tuning Procedure

1. Start with kp 1.0, ki 0, kd 0.
2. Increase Kp until the system oscillates, then back off 20-30%.
3. Add Ki from a small value until steady-state error is eliminated.
4. Add Kd to dampen oscillations.
5. Use step <angle> to capture response data for plotting.

Milestones

• PWM output driving motor at variable duty cycle
• I2C reads from MPU-6050 raw accel and gyro data
• PID loop with proportional-only control
• Full PID with integral anti-windup and derivative filtering
• UART CLI for live tuning
• Step response capture and export
• OLED real-time display

Pin Summary

view on github