EMG Signal Interference Minimization Proposal Using a High Precision Multichannel Acquisition System and an Auto-Calibrated Adaptive Filtering Technique

Abstract

Within the biomedicine scope, electromyography or EMG signals are known as an electrical impulses record produced by the human brain with the intention of activating the body muscles movement. Usually, these signals are acquired using surface electrodes placed on the skin in the place where muscle activity is intended to record. However, the acquisition of these signals is usually affected by unwanted interference or “Noise” that come from different sources such as the static produced by skin contact with the electrodes and electromagnetic interference from the system power supplies and the environment where measurements are made, etc. Therefore, and also considering that this work proposes to use a high-resolution system for the acquisition of EMG signals, and this high resolution involves a high sampling frequency, which at the same time causes the system to become more vulnerable to the previously aforementioned interferences, therefore, this work proposes as a solution to use an auto-calibration system that allows the acquisition system to learn the interferences produced in the acquisition of the EMG signals before making the measurement, in order to try to eliminate them when they are subsequently acquired, the filtering technique was proposed in previous work. The proposed efficiency evaluation metric is known as the signal-to-noise ratio (SNR) compared before and after using the proposed auto-calibrated filtration system. This system allows acquiring an electromyography signal with a minimum noise level, which subsequently allows to faithfully use this type of acquired signals in systems of extraction of characteristics and classification of EMG signals.