In this thesis, it is shown that the myoelectric signal exhibits a deterministic structure during the initial phase of a muscle contraction. Furthermore, it is demonstrated that this structure is reproducible for a given contraction type and distinct from one type of contraction to another. This thesis describes an approach to multifunction myoelectric control based on the classification of the initial patterns. Features are extracted from several time segments of the myoelectric signal to preserve pattern structure. These features are then classified using an artificial neural network. Results are presented to support this approach.

This work has resulted in a novel approach to the control of a multifunction myoelectric prosthesis. The new control scheme increases the number of functions which can be controlled by a single channel of myoelectric signal but does so in a way which does not increase the effort required by the amputee. The control signals are derived from natural contraction patterns which can be produced reliably with little subject training. The system can be adjusted to the individual and retrained, if necessary, to meet the changing needs of the amputee. As well, the control system can adapt to slow varying changes in the control patterns. The state selection delay is below 250 ms and the new control scheme integrates function selection with proportional speed control.