Integration Design of Portable ECG Signal Acquisition with Deep-Learning Based Electrode Motion Artifact Removal on an Embedded System

For long-term electrocardiogram (ECG) signal monitoring, a portable and small size acquisition device with Bluetooth low energy (BLE) communication is designed integrated Nvidia Jetson Xavier NX for realizing the electrode motion artifact removal technique. The digitalized ECG codes are converted from front-end circuit, which contains several amplifiers filters in system. Thereafter, zero padding scheme applied each 10-bits data to separate them into two-bytes BLE transmission. Edge AI platform receives these transmitted removes (EM) noise using proposed memory shortcut connection-based denoised autoencoder (LMSC-DAE). simulation results demonstrate that algorithm significantly improves signal-to-noise ratio (SNR) by 5.41 dB under condition of SNRin = 12 dB, compared convolutional denoising long short-term (CNN-LSTM-DAE) method. practical test, an Arduino DUE employed generate interference controlling commercial digital-to-analog convertor. By combining non-inverting weighted summer, it can be verify reproducibility measurement clearly indicate LMSC-DAE has higher improvement SNR lower percentage root-mean-square difference than state-of-the-art Fully Convolutional Denoising Autoencoder (FCN-DAE).