Final Year Project

Project Title

Transformers at the Edge: EMG-Based Hand Prosthetics

Description

This project develops a hardware-accelerated system that tracks continuous hand motion from bio-signals generated during muscle activity. It leverages a lightweight and quantized Transformer model, which is pre-trained on non-invasive surface EMG data to learn human kinematics. The model is deployed onto a low-cost Xilinx Zedboard FPGA to tap its inherent parallelism and achieve the ultra-low latency required for real-time interaction via high-speed processing, and the output kinematics are streamed to Unity for 3D visualization. The purpose is to deliver a scalable edge-AI framework for advanced prosthetics using hardware-software co-design, bridging the gap between simulated biomechanics and real-world hardware deployment.