Embedded EOG-based BCI System for Robotic Control Using TinyML

Learn about an innovative Brain-Computer Interface (BCI) system presentation that explores using Electrooculography (EOG) for robotic control applications. Discover how this embedded system leverages eye movements to assist people with severe motor disabilities, particularly those with locked-in syndrome (LIS), in controlling Electrically Powered Wheelchairs and interacting with smart environments. Explore the technical implementation of a wearable BCI system that operates on a small SoC powered by MCU, utilizing TinyML for efficient processing on resource-constrained devices. Examine the three-block processing system, including pre-processing, event detection, and event classification, which enables accurate recognition of voluntary and involuntary eye movements. Learn how the system achieves 99.3% classification accuracy using a 1-dimensional CNN model, successfully controlling three degrees of freedom in wheeled robots through eye movements like winks and blinks.