基于注意力机制的毫米波雷达ECG信号反演算法研究

Reconstruction of ECG Signals from Millimeter Wave Radar Based on Attention Mechanism

关键词:  毫米波雷达  心电图  卷积神经网络(CNN)  注意力机制  小波变换

Key words:  Millimeter Wave Radar  ECG  Convolutional Neural Network (CNN)  attention mechanism  wavelet transform

摘要:

针对毫米波雷达反演高精度的人体心电图(ECG)信号，本文提出一种融合小波变换和注意力(Attention)机制的AM-CNN-GRU神经网络的ECG信号重建方法。首先对雷达信号进行预处理，从中提取与生理特征相关的相位形变信号，并使用滑动窗口去除直流干扰；然后利用小波变换对形变信号进行分解，提取出与ECG信号显著相关的小波基分量，并在CNN-GRU网络的基础上，对小波分解的信号引入通道注意力机制，输入到网络中训练得到ECG信号。实验数据验证表明，本文提出的模型相比传统的CNN-BiLSTM模型，在模型参数更轻量化的同时，取得了更高精度的ECG信号反演结果。

Abstract:

Aiming at high-precision reconstruction of Electrocardiogram (ECG) signals from Millimeter Wave Radar,a reconstruction method based on AM-CNN-GRU network combining wavelet transform and attention mechanism is proposed.Firstly,the radar signal is preprocessed to extract the phase deformation signal related to physiological characteristics,and the sliding window is used to remove the DC interference.Then,the deformation signal is decomposed by wavelet transform to extract the wavelet base component exhibiting a significant relationship with the ECG signal.While the CNN-GRU network model is built as the main architecture,wavelet decomposed signals are imported as input signals after application of the channel attention mechanism,and then the ECG signal is obtained.The experimental results showed that the proposed model was lighter in model parameters and achieved results with higher precision than the CNN-BiLSTM model.