On detecting and adaptive timing for electromyogram based control signals

The purpose of this study was to develop and evaluate a signal-processing algorithm for controlling electromechanical devices by use of electromyogram (EMG) signals acquired with surface electrodes. The goal was to provide a robust trigger pulse with an appropriate duty cycle that, to some extent, reflects the duration of the muscle contraction. The EMG signal was amplified, digitized, and processed by an 8-bit microprocessor. The Multiplication of Backward Differences (MOBD) algorithm, originally designed to detect the QRS complexes in the electrocardiogram, was used to detect the on-set of the EMG signal. An adaptive refractory period scheme was devised to adjust the trigger pulse width based on EMG activities after the on-set point. This paper describes the development of the microprocessor based system and the preliminary test results.