Pattern recognition based prosthetic arm

The following section entails my work at Infinite biomedical technologies (a John Hopkins-affiliated research lab in the USA). Here I worked as a project manager, and with my team developed a stand-alone device that non-invasively records 8-channels of the subject’s EMG signals (in real-time) to differentiate, determine and control prosthetic hand positions. We successfully implemented hand-open, close, flex, extend, pronate, supinate and hook hand positions.
  The images below (as well as the YouTube video) display the working of our interface for hand open and close positions. The EMG patterns generated when the subject opens or closes their hand is analyzed by our device and based on the classification result the robotic hand mimics the subject’s hand position.