Measuring multi-site pulse transit time with an AI-enabled mmWave radar
Jiangyifei Zhu, Kuang Yuan, Akarsh Prabhakara, Yunzhi Li, Gongwei Wang, Kelly Michaelsen, Justin Chan, Swarun Kumar · 2025
Read Measuring multi-site pulse transit time with an AI-enabled mmWave radar by Jiangyifei Zhu, Kuang Yuan, Akarsh Prabhakara, Yunzhi Li, Gongwei Wang, Kelly Michaelsen, Justin Chan, Swarun Kumar (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-18141.txt.
Introduction
Jiangyifei Zhu, with seven co-authors, measures pulse transit time at several body sites at once with a single radar and no body contact. Transit time tracks arterial stiffness and relates inversely to diastolic blood pressure. Contact sensors are the norm, so a contactless multi-site tool is new.
Beamforming and deep learning read the tiny motions from one millimeter-wave radar. Across three body pathways, the readings matched contact references well, and blood pressure estimates met standard device guidelines in the study. Health sensing readers get a path toward unobtrusive heart monitoring that could run during sleep or rest.
Catalog record
| CREATOR | Jiangyifei Zhu, Kuang Yuan, Akarsh Prabhakara, Yunzhi Li, Gongwei Wang, Kelly Michaelsen, Justin Chan, Swarun Kumar |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 79 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-18141
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Measuring multi-site pulse transit time with an AI-enabled mmWave radar Jiangyifei Zhu, Kuang Yuan, Akarsh Prabhakara, Yunzhi Li, Gongwei Wang, Kelly Michaelsen, Justin Chan, Swarun Kumar Abstract Pulse Transit Time (PTT) is a measure of arterial stiffness and a physiological marker associated with cardiovascular function, with an inverse relationship to diastolic blood pressure (DBP). We present the first AI-enabled mmWave system for contactless multi-site PTT measurement using a single radar. By leveraging radar beamforming and deep learning algorithms our system simultaneously measures PTT and estimates diastolic blood pressure at multiple sites. The system was evaluated across three physiological pathways – heart-to-radial artery, heart-to-carotid artery, and mastoid area-to-radial artery – achieving correlation coefficients of 0.73–0.89 compared to contact-based reference sensors for measuring PTT. Furthermore, the system demonstrated correlation coefficients of 0.90–0.92 for estimating DBP, and achieved a mean error of -1.00–0.62 mmHg and standard deviation of 4.97–5.70 mmHg, meeting the FDA’s AAMI guidelines for non-invasive blood pressure monitors. These results suggest that our proposed system has the potential to provide a non-invasive measure of cardiovascular health across multiple regions of the body. Introduction Frequent monitoring of cardiovascular health can facilitate timely interventions including lifestyle changes such as exercise and diet, as well as pharmacologic changes like antihypertensive drugs [ 1 ] . Increased arterial stiffness is associated with a higher risk of cardiovascular conditions including hypertension [ 2 ] , stroke [ 3 ] , coronary artery disease [ 3 ] , sleep apnea [ 4 , 5 ] , and Alzheimer’s disease [ 6 ] , and can be quantified using pulse transit time (PTT) , the time it takes for a pulse wave to travel between a proximal and distal site within an artery [ 7 ] . In addition, pulse transit time is indirectly related to diastolic blood pressure (DBP) and has been explored as the basis for cuff-less blood pressure monitoring [ 8 , 9 , 10 , 11 , 12 ] . The ability to measure PTT contactlessly at multiple sites along the body has the potential to support pre-screening of cardiovascular diseases and long-term tracking of disease progression, particularly for high-risk populations such as older adults and individuals with chronic conditions [ 13 ] . This is because PTT varies across different body sites and site-specific difference are influenced by underlying physiologic stress mechanisms [ 14 ] . This suggests that multiple PTT measurements may provide a more comprehensive assessment of cardiovascular hemodynamics than a single measurement [ 15 ] . Arterial stiffness can be measured using contact-based methods, including pulse tonometry for arterial pressure assessment [ 16 ] , photoplethysmography (PPG) for detecting light absorption changes [ 17 ] , inertial measurement units (IMU) for capturing pulse motion [ 18 ] , and electrodes or wearable sensors [ 19 , 20 , 9 , 21 , 22 , 23 , 24 , 25 ] , which may cause skin irritation and discomfort. Contactless measurement systems have leveraged cameras [ 26 , 27 , 28 , 29 , 30 ] to measure PPG, and laser-doppler vibrometry [ 31 ] to measure the surface displacement at multiple arterial sites. However, such optical-based systems may not work well in low-lighting conditions or when objects like clothing block the device’s view. More recently, radar-based approaches have been proposed, leveraging millimeter-wave (mmWave) [ 32 , 33 , 34 , 35 , 36 ] , or microwave [ 37 ] , to measure the seismocardiogram (SCG) at the chest and ballistocardiogram (BCG) at the head. However, existing radar-based systems either require multiple sensor units to measure multiple arterial points [ 32 , 33 ] , which can be challenging to deploy and precisely position, or are limited to measuring a single PTT pair [ 35 , 38 , 39 ] . Here, we present PolyPulse , a proof-of-concept AI-enabled mmWave radar system that contactlessly measures pulse transit time and estimates DBP along three key physiological pathways along the upper body. Our system transmits chirp signals toward a seated individual and performs fine-grained beamforming on the reflected signal to capture the subtle bodily movements caused by cardiac activity. The design leverages a neural network model to process the reflection waveforms from the apex of the heart, the mastoid area at the head, the radial artery at the wrist, and the carotid artery at the neck. Using these signals, the model jointly estimates pulse transit times and diastolic blood pressure for physiological pathways that originate from the heart and head and extend to distal points at the wrist or neck. Our wireless radar system is compact and measures 16.5 \times 14 cm, which can fit within the form factor of smart home devices. Given the increasing prevalence of mmWave radars on consumer devices, our proposed multi-site PTT system can make remote monitoring of cardiovascular health at home more comprehensive and accessible. Results Concept and prototype Figure 1 : Overview of the mmWave radar system for multi-site pulse transit time (PTT) measurement. a, Rhythmic contractions of the heart generate pulse waves that propagate across the arterial system. (Image source: https://smart.servier.com/category/anatomy-and-the-human-body/cardiovascular-system) b, The mmWave radar detects minute surface displacements by measuring phase variations in reflected signals over time. c, The mmWave radar’s beamforming algorithm targets four key physiological sites to measure PTT: the apex of the heart (green) and the mastoid area (yellow) as proximal reference points, and the radial artery (blue) and carotid artery (red) as distal arterial sites. The waveforms represent the phase signals from these sites, captured by the mmWave radar positioned at a fixed distance beneath the subject. d, Our system estimates pulse transit time across three pairs: PTT_{SCG\rightarrow wrist},PTT_{SCG\rightarrow neck},PTT_{BCG\rightarrow wrist} . This is achieved by first identifying key cardiac waveforms features: the aortic opening in the SCG signal, the J peak in the BCG signal, and the foot of the carotid and radial arterial pulse waveform. PTT is then computed as the time difference between these signal features at the proximal reference points and those at distal arterial sites. As the heart beats, it generates pulse waves that propagate through arterial pathways in the body. This pulsatile flow causes rhythmic expansion and contraction of the arterial walls, which in turn results in minute superficial displacements of the skin (Fig. 1 a) [ 40 ] . PolyPulse transmits FMCW signals from 77–81 GHz towards a seated individual and analyzes both the phase and magnitude of the reflected signals at key physiological sites where a cardiac signal can be measured. The phase variations of the reflections correspond to superficial displacement of the skin at distal sites ( Fig. 1 b, c), while the magnitude variations correspond to changes in signal absorption due to blood flow that occur throughout each cardiac cycle. Our system measures cardiac signals by capturing reflections from two proximal reference points and two distal arterial points. For the proximal reference point, our radar captures reflections from the apex of the heart at the chest to obtain the SCG signal, and the mastoid area at the head to obtain the BCG signal respectively. For the distal reference points, we measure reflections from the right radial artery at the wrist and the right carotid artery at the neck. To