Sleep study is of major importance for the assessment of sleep apneas, sleep stages or the diagnosis of sleep disorders. Although there exists a variety of tools for the evaluation of sleep, including the gold standard polysomnography (PSG) or high-tech wearable devices, the last decade there is also an increasing interest in the use of ultra wide-band (UWB) radar sensors for non-contact medical studies. The objective of this study is to develop a pre-clinical environment for the measurement of two important vital signs of the human functions, i.e., the respiratory and heartbeat rates, through UWB radar recordings of chest motion. For this purpose, at first we composed a simulating chest and heart motion architecture which represents these two basic body functions in parallel. After that, with the employment of a UWB-radar, we performed extensive recordings of simulator's displacements, and we concluded to the mathematical estimation of the respiratory and heartbeat rates, compared to the initial frequencies given to the simulated procedure. Our experimental results prove that we can estimate both rates in an accurate manner and even if these pre-clinical tests are made under ideal conditions, the UWB-radar can further be used for clinical assessment, with promising perspectives.
Vital signs such as the respiration and heartbeat rates are
among the most important signals that provide
signifiImpulse response ultra wideband (IR-UWB) radar sensors
have gained the researchers interest as they constitute a
non-contact manner of detecting the vital signs [
The traditional manner of monitoring the sleep is through the well-established polysomnography (PSG) tool. However, PSG requires numerous leads to capture the whole body activity during sleep, which raise the patients’ discomfort, due to the restricted movement [2]. Moreover, most of the times in order to evaluate the patients’ sleep quality, the examinations should be of long nEvelop-O
0000-0003-2823-5584 (A. Pentari); 0000-0002-8057-5546 (D. Manousos); 0000-0002-0316-5252 (T. Kassiotis); © 2023 Copyright for this paper by its authors. Use permitted under Creative ws.org) duration, resulting to long recordings and thus, making their analysis dificult. On the other hand, IR-UWB radar sensors have been proven able tools for the study not only of the sleep apneas but also for the detection of sleep
The purpose of this study is to evaluate the IR-UWB radar’s ability to capture the motion of a simulator, constructed by the members of our laboratory, which simuexperimental procedure is to locate the radar in front of the simulator, under various experimental conditions, and evaluate how accurate are the radar’s respiratory and heartbeat rates estimation, compared to the fixed given values of simulator’s motion, i.e., to the experimental conditions. As the environmental conditions of our experiments were ideal, implying that no other target existing in the room and the noise was limited, the mathematical approaches followed for the estimation of the vital signs were simple, fast, and well-established. 2.
The vital signs’ information includes the respiration and heartbeat of the human target. This information is contained in the periodic expanding and contracting of the chest cavity. However, the first step is to describe how we that, we can estimate the quantities of interest, i.e., the 0000-0002-5102-6185 (G. Rigas); 0000-0001-8454-1450 (M. Tsiknakis) take the vital signals via the radar’s acquisitions. After respiratory rate and heartbeat.
The IR-UWB tool, via the radar pulses, aims to detect and quantify the periodic expanding of the chest. Thus, an important parameter in the construction of the vital signal is the distance between the radar antenna and the human chest, which changes over time. Previous work has proved that the distance could be represented as follows, () = 0 + sin(2 ) + ℎ sin(2 ℎ)
(1) where, 0 is the nominal distance, , ℎ are the mean values over the range of possible displacements of the activities, respectively. Moreover, as , ℎ we denote the respiratory and heartbeat frequencies, respectively, which are the quantities that we want to estimate. denoised as our experiments are held under ideal circumstances.
The x vital signal, it is a combination of two quantities, as shows the following equation, i.e., it contains the efects of both respiratory and heartbeat:
−1 1 √ =0 x , ℎ() =
∑ u( − − ( , ℎ)) where, as we denote the following delay: ( , ℎ) = 0 + sin(2 )
+ ℎ sin(2 ℎ)
with, denoting the electromagnetic wave speed and as u we denote the decluttered signal (i.e., after loop-back ifltering on the raw signal r. The and ℎ are the rates
Respiration Rate (RR): Regarding the RR, the most well-established method of its computation is through the power spectrum. Thus, we first apply the Fourier transform to the extracted signal r and then we take its power spectrum. as this is presented in Fig. 2. As it is (3) (4) chest cavity that caused by respiration and heartbeat of respiratory and heartbeat, respectively.
−1 1 √ =0 r() = ∑ s( − , )
(2) in our experimental case the ”clean” signal x is already x , ℎ() accomplished by additive noise n() . However, the the power spectrum’s amplitude, but in to the range 0.8 − 1.4
. , the highest spectrum’s amplitude is the respiratory rate.
It is worth to notice that, the respiratory rate was set to the range 0.2 − 0.4 processes.
Heartbeat Rate (HR): The HR is a more dificult quantity to be estimated, as the heartbeat is captured more dificult than the respiratory from the chest motion through the radar. However, in terms of our analysis, as the environment was ideal and the evaluation of the , in the simulated experimental
Both of the mentioned processes are usually accom- forward motion, backward, increase/decrease step time, plished by a lowpass filter, for the noise reduction and increase/decrease distance, normal/abnormal breathing the more accurate estimation of the RR and, again, before and, breathing with apnea and abnormal frequency. the application of the Fourier transform, the radar signal passes through a highpass filter for the HR estimation.
In order to implement the algorithms for estimating heart and respiration rates using the UWB radar, a ground truth was necessary. To this approach we created an artificially predefined dataset based on the physiological principles of the heart/breathing system. Because UWB radar technology is based on detecting the reflectivity and displacement of a collection of body points from the radar, such an approach is a mechanical system for simulating chest wall motion that incorporates both respiratory and heart rate. As for the respiratory rhythm, this is evident from the apparent movement of the chest during breathing which according to this movement ranges from 4-12mm, with a frequency range of 0.2-0.34Hz (1220 breaths per minute). However, the movement of the chest surface, in addition to the assessment of the respiratory rate, also contributes to the assessment of the heart rate since, depending on the phase of the heart’s operation, a displacement of 0.2-0.5mm is induced on the chest surface with a frequency range of 1-1.34Hz (60-80 beats per minute) [4].
In this direction, a mechanical approach to the displacement of the thoracic surface is the system illustrated in Fig. 3. This system consists of two stepper motors with a discrete resolution of 200 steps/rev (steps/rotation) which, in combination to an “8-step micro-stepping mode”, through the stepper driver, can reach a resolution of 1600 steps/rev. The rotary movements of the motors are converted into linear movements via a trapezoidal screw with an 8mm “Lead” which corresponds to a maximum resolution of 0.005mm per step. The operation of the system has been configured in such a way that it is possible to adjust the amplitude and oscillation of the moving part or even implement a non-periodic motion using a “look-up table” based on the desired amplitude and speed of motion. The range of motion is calculated by the number of rotations, while the frequency of oscillation is calculated by the time it takes the moving part to make a cycle, i.e., the period. The whole system was designed and printed in pieces by a 3D printer.
A pre-defined set of control functions was implemented in order to create scenarios for breathing and heart rate. For instance, we can create scenarios for steady or dynamic frequency regarding breathing rates and heart rates, as well as for emulating special events like abnormalities on heart/respiratory rates (e.g. apnea event). The emulator communicates through a simple serial interface and its functions include: stop motion,
Ultra-Wideband (UWB) radar technology is capable of accurately detecting vital signs such as respiration and heart rate, making it a valuable tool for monitoring the health and well-being of individuals in medical and fitness applications. The LT102 radar module is a readyto-use UWB radar system for indoor environments that combines high-quality antennas, advanced signal processing capabilities, and communication interfaces into a single unit. It is designed to comply with regulatory standards and is customizable for various applications such as presence detection and breath analysis. The LT102 module is powered through a USB connection and uses a USB full speed (virtual com port) for communication and also has an auxiliary connector that can be used as general-purpose input/outputs (GPIOs) or as an additional communication interface [5].
The LT102 system uses the direct readout of backscattered pulses as its operating principle. The system emits pulses (Fig. 3) which travel through space and hit any targets within the active area of the radar. These targets then reflect a portion of the incoming energy (echoes) back to the radar module (Fig.4). The receiver then converts the incoming signal into digital data, which is provided to the micro-controller unit (MCU) for processing according to the specific application.
The table below lists the general specifications of the device. These include the typical detection range, maximum power consumption, operating frequency, integrated antenna, and communication interfaces.
The Electrical Specifications table provides the range of operating conditions and requirements for the device.
Specifically, all of our experimental measurements were of 120 seconds duration, with a sampling frequency equal to 8.9. According to the simulator’s frequency of “chest” and “heart” motions, they were set equal to 0.25 Hz and 1.34 Hz, respectively.
In Table 4, we present the estimated respiratory and heartbeat rates derived from our algorithmic process, based on the signals acquired through the interaction of the simulator and the radar sensor. As mentioned in the methodology, the estimation of the RR and HR was based on the application of the Fourier transform and then, by taking the power spectrum. The respiratory rate was derived by limiting the frequencies to the range 0.2-0.4Hz whilst, the heartbeat rate was computed by, ifrst, passing the radar signal through a highpass filter 3.2. Experimental Results of a passband frequency equal to 0.8Hz and then, taking extracting the power spectrum. The frequency range to Regarding our experimental procedure, multiple simu- which we searched for the highest signal’s energy was lated radar-based recordings were extracted. Analyti- from 0.8 to 1.4 Hz. cally, in Table 3, we present the conditions under which Based on the estimations we observe the following: our experiments were carried out.
In this study we aimed to evaluate the IR-UWB radar sensor’s capabilities on capturing the RR and HR frequencies in diferent experimental conditions. Specifically, we examined a variety of diferent experimental scenarios, which concerned radar’s recordings from simulated motions of chest and heart. In terms of our evaluations, we constructed a simulator of moving functions, which had the ability to change angle and distance. This tool constituted the radar’s target. To conclude, through our experimental evaluations we observed that the target’s angle and distance from the radar play significant role to the robust and accurate estimation of RR and HR. Finally, the acquired signals are also afected by the conditions under which the target exists.
Our future goal is to examine the IR-UWB radar sensor on clinical evaluations. Our purpose is to examine if we can monitor sleep through the radar and pass to clinical assessment, concerning the sleep apneas and the sleep stages. To this end, the PSG tool will be the comparative method.
This research was funded by the European Regional Development Fund of the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship, and Innovation, under the call RESEARCH-CREATE–INNOVATE (project name: HealthSonar, project code:T1EDK-03990).
We would like to express our gratitude to all those who have supported and contributed to the development of this paper. We would like to extend our appreciation to our colleagues and collaborators for their invaluable input and suggestions. Special thanks to Georgios Christodoulakis, for his significant contribution to the construction of the simulator tool.