Framework for Wireless Acquisition of Surface EMG and Real-Time Control

Sammanfattning: Muscle driven devices are controlled or powered with muscle activation. These devices open up the possibility of offering patients with limited muscle function to automatically control assistive devices - for instance exoskeletons - with input from their own muscles. This solution would help a number of patient groups suffering from common conditions, such as spinal cord injuries, stroke and cerebral palsy. To use muscle activation as input it is necessary to have a way to communicate with the mus- cles. Electromyography (EMG) is a technology used to gain information about muscle function and activation. It is performed by measuring and analyzing electrical signals conducted by the muscles during activation. Activation and activation level can be seen from analyzed EMG signal. EMG signals are frequently measured and analyzed afterwards, however, to use it as a controlling an assistive devices, real time analysis is necessary. In this thesis real time acquisi- tion and analysis of EMG was performed. The measured signal was used as an input to control a simple MATLAB computer game. The EMG of a muscle of the forearm, Brachioradialis, was measured with Myon Aktos sys- tem. The measured signal was written to a server as soon as the measurements were acquired. MATLAB was used to connect to the server and performing the signal analysis. The analysis was kept simple in order to limit delay. The result showed that it was possible to acquire real time signal with this method. The delay was negligible, both for the testing and for the game play. Showing that it is possible to play a game with muscle activation supports the idea of a motor that can be controlled automatically with muscle input. Future work should focus on understanding movement intent with respect to EMG and on analyzing multiple signals from different muscles at the same time.