Desenvolvimento de sistema para coleta de sinais de eletromiografia e detecção de movimento

Abstract

This work aims to develop an embedded device capable of collecting surface electromyographic signals, which are muscle contraction signals. The device was built using an ESP32 micro- controller and an AD8232 sensor. For data collection, three electrodes were positioned on the forearm to measure the hand-closing movement. Since visually analyzing the signals makes it difficult to confirm whether the collected data are truly electromyographic-and differentiating them is even more challenging due to the high noise level-a study was conducted using a database of electromyographic signals containing 10 types of hand movements. Based on this, a movement classification algorithm was developed to verify whether the signals collected by the embedded device effectively correspond to electromyographic signals. The device was able to collect signals at a sampling rate of 1000 Hz, where amplitude peaks were visible, characterizing muscle contractions. However, when the sampling rate was adjusted to 4000 Hz, the signals contained only noise. When validating the signals in the classification algorithm, all were correctly classified.