Pattern Recognition-Based Prosthetic Arm

At Infinite Biomedical Technologies, a Johns Hopkins-affiliated research laboratory in the USA, I worked as a Project Manager and Embedded Software Engineer on the development of a pattern-recognition-based prosthetic arm.

I worked with a team to develop a standalone device capable of recording 8-channel EMG signals from a subject in real time and classifying the signals to determine the intended prosthetic hand position. The system was designed to provide non-invasive control of a prosthetic hand for transradial amputees.

We successfully implemented classification and control for multiple hand positions, including open, close, flex, extend, pronate, supinate, and hook.

The EMG patterns generated by the subject were analyzed by the embedded device, and the resulting classification was used to command the prosthetic hand to reproduce the subject's intended hand position.