Electromyography (EMG) measures the electrical activity of skeletal muscles and is widely used in the medical field. Much research has been done to characterize the muscle signals. Such research has been to control robotic arms or develop new ways to interact with the machines in our everyday life. However, the equipment used in this research is costly or not commercially available. During the WINLAB 2013 summer internship, hobbyist EMGs were used to see if it could be possible to interact with machines using muscle signals at a very minimal cost. EMG hardware available from Advancer Technologies was used with the Arduino microcontroller to measure predetermined hand movements. These movements consisted of lifting, curling, and applying pressure to the index and middle fingers. The movements were recorded and then processed offline. Different machine learning algorithms were applied to the processed data and the results seemed to be comparable to those of studies using much more sophisticated hardware.

Last modified 4 years ago Last modified on Oct 30, 2013, 11:37:47 PM