Principal Components Analysis Tuning for Improved Myoelectric Control

Abstract

Information extracted from signals recorded from multi-channel surface myoelectric signal (MES) recording sites can be used as inputs to control systems for powered prostheses. For small, closely spaced muscles, such as the muscles in the forearm, the detected MES often contains contributions from more than one muscle; the contribution from each specific muscle being modified by a tissue filter between the muscle and the detection points. In some cases the contributions from very small/deep muscles are masked by those from larger/superficial muscles. In such circumstances, subtle changes in muscle activations associated with different movements may not be easily detectable. In this work, the measured raw MES signals are rotated by class specific rotation matrices to spatially decorrelate the measured data prior to feature extraction. This tunes the pattern recognition classifier to better discriminate the test motions. Preliminary work indicates that this additional preprocessing step significantly reduces classification errors.