Signal decomposition methods for non-contact vital-sign extraction using stepped-frequency continuous-wave radar

Non-contact methods that accurately estimate heart and respiration rates are an important goal in radar-based vital sign measurement that can be used for health surveillance without intrusion, sleep monitoring, and medical scenarios where contact sensors are not feasible. This paper conducts a comparative study of three signal decomposition techniques—empirical mode decomposition (EMD), ensemble empirical mode decomposition (EEMD), and variational mode decomposition (VMD)—for simultaneous estimation of cardiac and respiratory rates from stepped-frequency continuous-wave radar signals. The radar echo was processed by coherently averaging the range bins covering the subject’s chest to form respiratory and cardiac displacement signals that were decomposed into intrinsic mode functions. The dominant physiological mode in each target band (0.10–0.50 Hz for respiration, 0.80–2.00 Hz for the cardiac component) was selected by the in-band signal-to-noise ratio (SNR) with spectral entropy as the descriptor of spectral compactness. The signal processing pipeline was evaluated on two subjects at rest. The three methods agreed closely, with the respiratory rate and heart rate for Subject 1 being 18.8 breaths min –1 and 66.8 beats min −1 , respectively, and 14.4 breaths min –1 and 76.5 beats min −1 for Subject 2. VMD achieved the highest in-band SNR in both channels (23.89 dB for respiratory, 12.21 dB for cardiac), exceeding EMD and EEMD by 7 dB and 4.5 dB, with the lowest spectral entropy (0.215 and 0.457, a 29%–38% reduction). VMD also required only four modes versus five or six for the others, simplifying selection and improving reproducibility. Validation of the signal processing pipeline was done using a 1.5 Hz mechanical oscillating machine, which yielded a 1.33% relative error; a measurement uncertainty analysis (coverage factor k = 2) established the reported rate as 18.8 ± 0.7 breaths min –1 and 66.8 ± 1.2 beats min −1 for Subject 1, and 14.4 ± 0.7 breaths min –1 and 76.5 ± 1.2 beats min −1 for Subject 2.

Authors

Institutions

Publication Details

Journal
Measurement Science and Technology
Published
2026-11-19
DOI
https://doi.org/10.1088/1361-6501/aebd0e
Primary Topic
Non-Invasive Vital Sign Monitoring
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Signal decomposition methods for non-contact vital-sign extraction using stepped-frequency continuous-wave radar

Piyush K. Mathurkar, Abhay N. Gaikwad
Measurement Science and Technology
Non-Invasive Vital Sign Monitoring
article

Signal decomposition methods for non-contact vital-sign extraction using stepped-frequency continuous-wave radar

Piyush K. Mathurkar, Abhay N. Gaikwad
article en

Abstract

Non-contact methods that accurately estimate heart and respiration rates are an important goal in radar-based vital sign measurement that can be used for health surveillance without intrusion, sleep monitoring, and medical scenarios where contact sensors are not feasible. This paper conducts a comparative study of three signal decomposition techniques—empirical mode decomposition (EMD), ensemble empirical mode decomposition (EEMD), and variational mode decomposition (VMD)—for simultaneous estimation of cardiac and respiratory rates from stepped-frequency continuous-wave radar signals. The radar echo was processed by coherently averaging the range bins covering the subject’s chest to form respiratory and cardiac displacement signals that were decomposed into intrinsic mode functions. The dominant physiological mode in each target band (0.10–0.50 Hz for respiration, 0.80–2.00 Hz for the cardiac component) was selected by the in-band signal-to-noise ratio (SNR) with spectral entropy as the descriptor of spectral compactness. The signal processing pipeline was evaluated on two subjects at rest. The three methods agreed closely, with the respiratory rate and heart rate for Subject 1 being 18.8 breaths min –1 and 66.8 beats min −1 , respectively, and 14.4 breaths min –1 and 76.5 beats min −1 for Subject 2. VMD achieved the highest in-band SNR in both channels (23.89 dB for respiratory, 12.21 dB for cardiac), exceeding EMD and EEMD by 7 dB and 4.5 dB, with the lowest spectral entropy (0.215 and 0.457, a 29%–38% reduction). VMD also required only four modes versus five or six for the others, simplifying selection and improving reproducibility. Validation of the signal processing pipeline was done using a 1.5 Hz mechanical oscillating machine, which yielded a 1.33% relative error; a measurement uncertainty analysis (coverage factor k = 2) established the reported rate as 18.8 ± 0.7 breaths min –1 and 66.8 ± 1.2 beats min −1 for Subject 1, and 14.4 ± 0.7 breaths min –1 and 76.5 ± 1.2 beats min −1 for Subject 2.

Measurement Science and TechnologyVol. 37(46)
Vishwakarma Institute of Technology (IN), Babasaheb Naik College of Engineering, Pusad (IN)
Non-Invasive Vital Sign Monitoring
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon