Abstract:
The desire to control and manipulate computers and human device interfaces (HDI) in more natural ways has been sought after for some time. This Engineering 090 project looks use hand movements and motions to control an HDI. This is done through processing, analysis and recognition of signals from surface electromyography (EMG) sensors and an accelerometer and gyroscopes to control an HDI. The HDI controlled in this project is a computer mouse and arrow keys on a keyboard via the Makey-Makey. This project implements electronic circuit design to filter the raw EMG signals to meaningful EMG signals, digital signal processing methods to characterize the EMG signals and machine learning techniques to classify the EMG signals into hand gestures using Artificial Neural Networks (ANN). Afterwards, based on the evaluated hand gesture or movement, signals are sent to a Makey-Makey to control the mouse and designated keys on the keyboard. The system, after implementation, is able to recognize with nearly no error four hand gestures within approximately an eight of a second. This result allows for real time accurate control of the desired HDI. This system in the future can be implemented to control and manipulate any HDI that can be driven by General Purpose In/Out (GPIO) signals. Finally, electromyography driven HDIs have the potential to control more complicated systems such as machines in hazardous areas or prosthetics, the possibilities are endless.
