A collection of embedded systems projects built around the STM32 microcontroller family, spanning custom bootloaders, real-time multi-threaded systems, motor control, ADC interfacing, and on-device AI inference.

All firmware is written in C and built with STM32CubeIDE / STM32CubeMX unless noted.

A custom bootloader for the STM32G4 series that receives a compiled .bin binary wirelessly over USART2 via an HC-05 Bluetooth module.

How it works:

1. On reset, the bootloader runs and waits for an incoming .bin over USART2.
2. The received binary is buffered in RAM.
3. The bootloader erases and re-flashes the application region starting at 0x08008000.
4. Execution jumps to the application's reset handler.

Key peripherals: USART2, HC-05 (UART-Bluetooth bridge), Flash HAL

A FreeRTOS-based multi-threaded firmware for a Battery Management System and Vehicle Control Unit. Designed for a high-reliability braking system with multiple concurrent sensor and communication tasks.

Peripherals used:

• UART4, USART2 (serial comms)
• ADC1, ADC2 (pressure sensing)
• FDCAN3 (CAN bus communication)
• USB (data logging or host interface)
• GPIO (actuator control, status signaling)

Architecture: Each subsystem (pressure monitoring, CAN TX/RX, UART logging) runs as an independent FreeRTOS task with defined priorities.

STM32-based firmware for driving and monitoring a 3-phase inverter in a Hyperloop pod drivetrain.

Features:

• Generates 3 PWM signals + 3 complementary (inverted) PWM signals via TIM1 for 3-phase inverter gate control
• PWM frequency controls motor speed; duty cycle controls amplitude (torque/force)
• Inverter health monitoring with fault detection
• Status reporting and fault signaling via GPIO and UART

STM32 firmware developed for the Aeron Polux Hyperloop pod (version 2.1). Exact subsystem details are folder-level; see source files inside Aeron Polux 2.1/.

Demonstrates simultaneous multi-channel ADC acquisition on STM32, likely using DMA for continuous non-blocking reads.

Use case: Sensor fusion, analog data aggregation.

An early attempt at running an AI/ML inference model on STM32 for soil moisture-based irrigation prediction. Uses sensor inputs to decide irrigation triggers without a host processor.

• IDE: STM32CubeIDE
• HAL: STM32 HAL / LL drivers
• RTOS: FreeRTOS (BMS_VCU project)
• Language: C (C99)
• Programmer: ST-LINK / custom BT bootloader

1. Clone the repo:

   git clone https://github.com/ChaitanyaParate/stm32-Projects.git

2. Open the desired project folder in STM32CubeIDE (each folder is a self-contained CubeIDE project).
3. Build via Project > Build All or Ctrl+B.
4. Flash using ST-LINK or, for BT_program_upload, use the Bluetooth bootloader flow.

Chaitanya Parate
B.Tech Computer Science, MIT World Peace University, Pune
GitHub