Abstract:
Myoelectric control of powered prostheses is a field of rehabilitation engineering that has received wide attention in the recent decades. In this thesis a historical perspective of the studies in the field is given. The physiological properties of muscles are reviewed. The linear models, algorithms for identifying model parameters, and basic pattern recognition considerations are outlined. A scheme to extract motion information from a single surface EMG channel is discussed. The results obtained in performance tests are given. Suggestions for future research topics are made. Major computer programs used are given in the appendix.
