Related Experiment Video
Updated: Aug 5, 2026
09:42
Acquisition and Semi-Automated Analysis of Respiratory Muscle Surface Electromyography
Published on: January 24, 2025
Millimeter-Wave Radar-Based ECG Reconstruction Using Respiratory Harmonic Suppression and CA-WTBNet
Bowen Xiao1, Chuyi Zhou2, Lu Wang1
1School of Mechanical and Electrical Engineering, Chengdu University of Technology, Chengdu 610059, China.
Bioengineering (Basel, Switzerland)
|July 28, 2026
Summary
This study introduces a novel millimeter-wave radar system for non-contact electrocardiogram (ECG) reconstruction. The advanced method significantly improves accuracy by suppressing respiratory interference and enhancing heartbeat signal analysis.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Radar Technology
Background:
- Millimeter-wave radar offers non-contact cardiac monitoring potential for electrocardiogram (ECG) signal reconstruction.
- Existing radar-based ECG methods struggle with incomplete heartbeat information extraction and feature modeling, limiting accuracy.
Purpose of the Study:
- To develop an advanced millimeter-wave radar-based ECG reconstruction method.
- To enhance reconstruction accuracy by addressing limitations in current radar-based approaches.
Main Methods:
- Implemented a respiratory-harmonic-suppressed multi-channel signal-processing frontend using maximal overlap discrete wavelet transform.
- Developed the CA-WTBNet deep reconstruction network integrating channel-attention, convolutional residual modules, Transformer blocks, and bidirectional LSTM.
- Utilized a respiratory harmonic detection and heart-rate frequency protection strategy.
Main Results:
- Achieved high performance on a public dataset: Pearson correlation coefficient of 0.9641, normalized root mean square error of 0.0458, R-peak F1 score of 0.9956, and R-peak timing error of 3.13 ms.
- Demonstrated superior performance compared to existing methods, with a 0.53% improvement in Pearson correlation coefficient and a 10.20% reduction in normalized root mean square error.
Conclusions:
- The proposed millimeter-wave radar ECG reconstruction method significantly improves accuracy and performance.
- The integrated signal processing and deep learning network effectively extracts cardiac information while suppressing interference.
