Project Overview
The prototype I developed uses an STM32 Nucleo board with the SensorTile attachment. The video above first goes over training the device, which involves the learning of 6 different motions using the EmbeddedML library. It then shows a demo of the motions: moving in all directions as well as left and right-clicking.
The prototype works on all major desktop OSes (tested on Windows, macOS, and Ubuntu). The STM board outputs the recognized gesture through Serial communication, which is actively monitored by a Python script that can control the cursor movement.
Skills Used
C
Python
Embedded Systems
Internet of Things
Git
Machine Learning
Operating Systems