skip to main content

Attention:

The NSF Public Access Repository (NSF-PAR) system and access will be unavailable from 11:00 PM ET on Friday, July 12 until 2:00 AM ET on Saturday, July 13 due to maintenance. We apologize for the inconvenience.


Title: DeepVS: A Deep Learning Approach For RF-based Vital Signs Sensing
Vital signs (e.g., heart and respiratory rate) are indicative for health status assessment. Efforts have been made to extract vital signs using radio frequency (RF) techniques (e.g., Wi-Fi, FMCW, UWB), which offer a non-touch solution for continuous and ubiquitous monitoring without users’ cooperative efforts. While RF-based vital signs monitoring is user-friendly, its robustness faces two challenges. On the one hand, the RF signal is modulated by the periodic chest wall displacement due to heartbeat and breathing in a nonlinear manner. It is inherently hard to identify the fundamental heart and respiratory rates (HR and RR) in the presence of higher order harmonics of them and intermodulation between HR and RR, especially when they have overlapping frequency bands. On the other hand, the inadvertent body movements may disturb and distort the RF signal, overwhelming the vital signals, thus inhibiting the parameter estimation of the physiological movement (i.e., heartbeat and breathing). In this paper, we propose DeepVS, a deep learning approach that addresses the aforementioned challenges from the non-linearity and inadvertent movements for robust RF-based vital signs sensing in a unified manner. DeepVS combines 1D CNN and attention models to exploit local features and temporal correlations. Moreover, it leverages a two-stream scheme to integrate features from both time and frequency domains. Additionally, DeepVS unifies the estimation of HR and RR with a multi-head structure, which only adds limited extra overhead (<1%) to the existing model, compared to doubling the overhead using two separate models for HR and RR respectively. Our experiments demonstrate that DeepVS achieves 80-percentile HR/RR errors of 7.4/4.9 beat/breaths per minute (bpm) on a challenging dataset, as compared to 11.8/7.3 bpm of a non-learning solution. Besides, an ablation study has been conducted to quantify the effectiveness of DeepVS.  more » « less
Award ID(s):
2119299
NSF-PAR ID:
10356921
Author(s) / Creator(s):
; ; ; ;
Date Published:
Journal Name:
In 13th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics (BCB ’22)
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Basic vital signs such as heart and respiratory rates (HR and RR) are essential bio-indicators. Their longitudinal in-home collection enables prediction and detection of disease onset and change, providing for earlier health intervention. In this paper, we propose a robust, non-touch vital signs monitoring system using a pair of co-located Ultra-Wide Band (UWB) and depth sensors. By extensive manual examination, we identify four typical temporal and spectral signal patterns and their suitable vital signs estimators. We devise a probabilistic weighted framework (PWF) that quantifies evidence of these patterns to update the weighted combination of estimator output to track the vital signs robustly. We also design a “heatmap” based signal quality detector to exclude the disturbed signal from inadvertent motions. To monitor multiple co-habiting subjects in-home, we build a two-branch long short-term memory (LSTM) neural network to distinguish between individuals and their activities, providing activity context crucial to disambiguating critical from normal vital sign variability. To achieve reliable context annotation, we carefully devise the feature set of the consecutive skeletal poses from the depth data, and develop a probabilistic tracking model to tackle non-line-of-sight (NLOS) cases. Our experimental results demonstrate the robustness and superior performance of the individual modules as well as the end-to-end system for passive and context-aware vital signs monitoring. 
    more » « less
  2. Non-contact vital signs monitoring (NCVSM) with radio frequency (RF) is attracting increasing attention due to its non-invasive nature. Recent advances in COTS radar technologies accelerate the development of RF-based solutions. While researchers have implemented and demonstrated the feasibility of NCVSM with diverse radar hardware, most efforts have been focused on devising algorithms to extract vital signs, with limited understanding about the effects of radar configurations. The deficiency in such understanding hinders the design of software defined radar (SDR) optimally customized for NCVSM. In this work, we first hypothesize the effects of FMCW radar configurations using signal-to-interference-plus-noise ratio (SINR) based signal modeling, then we conduct extensive experiments with a COTS FMCW radar, TinyRad, to understand how various parameters impact NCVSM performance compared to a medical device. We find that a larger bandwidth or higher transmitting power in general improves vital sign estimation accuracy; however, coherent processing of consecutive chirps (time diversity) or multiple receiving antennas (space diversity) does not improve the performance. Observations on the baseband (BB) signal show that coherent processing contributes to a higher amplitude but similar phase patterns, whose periodic changes are the key in extracting vital signs. 
    more » « less
  3. Continuous monitoring of respiration provides invaluable insights about health status management (e.g., the progression or recovery of diseases). Recent advancements in radio frequency (RF) technologies show promise for continuous respiration monitoring by virtue of their non-invasive nature, and preferred over wearable solutions that require frequent charging and continuous wearing. However, RF signals are susceptible to large body movements, which are inevitable in real life, challenging the robustness of respiration monitoring. While many existing methods have been proposed to achieve robust RF-based respiration monitoring, their reliance on supervised data limits their potential for broad applicability. In this context, we propose, RF-Q, an unsupervised/self-supervised model to achieve signal quality assessment and quality-aware estimation for robust RF-based respiration monitoring. RF-Q uses the recon- struction error of an autoencoder (AE) neural network to quantify the quality of respiratory information in RF signals without the need for data labeling. With the combination of the quantified sig- nal quality and reconstructed signal in a weighted fusion, we are able to achieve improved robustness of RF respiration monitor- ing. We demonstrate that, instead of applying sophisticated models devised with respective expertise using a considerable amount of labeled data, by just quantifying the signal quality in an unsupervised manner we can significantly boost the average end-to-end (e2e) respiratory rate estimation accuracy of a baseline by an improvement ratio of 2.75, higher than the gain of 1.94 achieved by a supervised baseline method that excludes distorted data. 
    more » « less
  4. While radio frequency (RF) based respiration monitoring for at- home health screening is receiving increasing attention, robustness remains an open challenge. In recent work, deep learning (DL) methods have been demonstrated effective in dealing with non- linear issues from multi-path interference to motion disturbance, thus improving the accuracy of RF-based respiration monitoring. However, such DL methods usually require large amounts of train- ing data with intensive manual labeling efforts, and frequently not openly available. We propose RF-Q for robust RF-based respiration monitoring, using self-supervised learning with an autoencoder (AE) neural network to quantify the quality of respiratory signal based on the residual between the original and reconstructed sig- nals. We demonstrate that, by simply quantifying the signal quality with AE for weighted estimation we can boost the end-to-end (e2e) respiration monitoring accuracy by an improvement ratio of 2.75 compared to a baseline. 
    more » « less
  5. Using wireless signals to monitor human vital signs, especially heartbeat information, has been intensively studied in the past decade. This non-contact sensing modality can drive various applications from cardiac health, sleep, and emotion management. Under the circumstance of the COVID-19 pandemic, non-contact heart monitoring receives increasingly market demands. However, existing wireless heart monitoring schemes can only detect limited heart activities, such as heart rate, fiducial points, and Seismocardiography (SCG)-like information. In this paper, we present CardiacWave to enable a non-contact high-definition heart monitoring. CardiacWave can provide a full spectrum of Electrocardiogram (ECG)-like heart activities, including the details of P-wave, T-wave, and QRS complex. Specifically, CardiacWave is built upon the Cardiac-mmWave scattering effect (CaSE), which is a variable frequency response of the cardiac electromagnetic field under the mmWave interrogation. The CardiacWave design consists of a noise-resistant sensing scheme to interrogate CaSE and a cardiac activity profiling module for extracting cardiac electrical activities from the interrogation response. Our experiments show that the CardiacWave-induced ECG measures have a high positive correlation with the heart activity ground truth (i.e., measurements from a medical-grade instrument). The timing difference of P-waves, T-waves, and QRS complex is 0.67%, 0.71%, and 0.49%, respectively, and a mean cardiac event difference is within a delay of 5.3 milliseconds. These results indicate that CaridacWave offers high-fidelity and integral heart clinical characteristics. Furthermore, we evaluate the CardiacWave system with participants under various conditions, including heart and breath rates, ages, and heart habits (e.g., tobacco use). 
    more » « less