The project that we were assigned to was about the movement of an arm prosthesis by capturing EMG signals. This project had already been in development for several years. For our thesis, we took the same title as one of the students who worked on the project before us, Bren Van Caesbroeck, because we had to base our project on his work. We did this with the permission of our promoter.
Our project consisted of two parts:
First, we were asked to create a better control algorithm for capturing EMG signals. We had to do this by creating a new diagram in Simulink. This proved to be too difficult for us, so we were given another task.
The second task was to find alternatives for the neural network model created by Bren Van Caesbroeck. We had to implement different kinds of backpropagation methods in MATLAB, to create neural networks. For each neural network we took a different amount of epochs (i.e. iterations) and different learning rates. We then had to compare the error percentage of all these different methods, to see which one of them gave the best result.
If you want to cite this thesis in your own thesis, paper, or report, use this format (APA):
DARDENNE, A., NELISSEN, J. (2010). Arm prosthesis control based on surface EMG signal extraction, classification and decomposition techniques.
Unpublished thesis, Xios, IWT.