Wearable Activity Tracker

This project builds a complete IoT system that can detect and identify a person’s activity, such as walking or running, in real time. It uses a wearable device based on the RT-Thread RT-Spark board, which constantly collects movement data from its built-in accelerometer and gyroscope. These sensors track how the user moves.

Procedures:

For this project, the sensor configuration, initialization, and data acquisition process were carefully designed to ensure reliable and continuous motion sensing.The RT-Thread RT-Spark development board was first configured at the operating system level using menuconfig, where essential components such as the I2C device drivers, mpu6xxx, icm20608, i2c-tools, rw007 for Wi-Fi where added and enabled.

The main thread stack size was increased to accommodate mathematical operations and sensor processing, preventing stack overflow during runtime. These configurations ensured that the system could support sensor communication, data processing, and real-time execution.

The accelerometer and gyroscope were then configured to their default full-scale ranges (±2 g for acceleration and ±250 dps for angular velocity), providing sufficient sensitivity for human activity detection while maintaining low noise. The collected data is stored in a MongoDB Atlas database.

Screenshots of MongoDB:

Bar graph data indication:

• Gray: Standby
• Yellow: Walking
• Green: Running

Video Demonstration: https://drive.google.com/file/d/1KZiQ9KQjNT6ayoVs7HsH0jrIoZBdxVNH/view?usp=sharing

AI Tools Used During Development

• ChatGPT (OpenAI) – Used for code assistance, debugging, and explaining concepts.
• Google Gemini – Used for idea generation and documentation support.