Training Strategies for Mitigating the Effect of Proportional Control on Classification in Pattern Recognition Based Myoelectric Control

ABSTRACT The performance of pattern recognition–based myoelectric control has seen significant interest in the research community for many years. Because of a recent surge in the development of dexterous prosthetic devices, determining the clinical viability of multifunction myoelectric control has become paramount. Several factors contribute to differences between offline classification accuracy and clinical usability, but the overriding theme is that the variability of the elicited patterns increases greatly during functional use. Proportional control has been shown to greatly improve the usability of conventional myoelectric control systems. Typically, a measure of the amplitude of the electromyogram (a rectified and smoothed version) is used to dictate the velocity of control of a device. The discriminatory power of myoelectric pattern classifiers, however, is also based largely on the amplitude features of the electromyogram. This work presents an introductory look at the effect of contraction strength and proportional control on pattern recognition–based control. These effects are investigated using typical pattern recognition data collection methods as well as a real-time position tracking test. Training with dynamically force varying contractions and appropriate gain selection is shown to significantly improve (p < 0.001) the classifier’s performance and tolerance to proportional control.