RU167288U1 - DEVICE FOR AUTOMATED DETERMINATION OF SLEEP STRUCTURE - Google Patents

Abstract

A device for the automated determination of sleep patterns, containing a bio-radar, characterized in that the bio-radar is a short-range radar with a continuous signal and a quadrature detector, the output of the bio-radar is connected to the input of the primary filtering unit for bio-radar signals, its output is connected to the input of the artifact detection unit, the output of which connected to the input of the block between the artifact periods, connected to the input of the phase reversal elimination unit, which is connected to the input m block normalization of artefact periods, the output of which is connected to the filter block of the respiratory signal and the filter block of the heartbeat signal, the output of the filter block of the respiratory signal is connected to the block for determining respiratory cycles and their parameters, the output of which is connected to the input of the block for extracting signs of the respiratory signal, the output of the signal filter block the heartbeat is connected by the output of the heart rate determination unit, and its output is connected to the input of the heartbeat signal extraction unit, the signal from which is the signal from the block of extraction of signs of a breathing signal, they are fed to the input of the block of normalization of signs, the output of which is connected to the input of the primary classification block connected to the input of the secondary classification block, the signal from which is fed to the interface with the output device.

Description

Technical field

The utility model relates to medical equipment and can be used to automatically determine the structure of sleep.

State of the art

From the prior art it is known to use bioradars for detecting movements and cardiorespiratory activity (US 2012/0022348 A1, publ. 01/26/2012; RU 2463949, publ. 20.10.2012; WO 2008057883 A3, publ. 31.07.2007), as well as a device for recognizing patterns of a bio-radar signal (RU 142167 U1, publ. 06/20/2014), but these technical solutions are not directly intended to determine the structure of sleep.

Closest to the claimed technical solution is the invention according to US patent application US 20140088373 SYSTEM AND METHOD FOR DETERMINING SLEEP STAGE (IPC A61B 5/00; A61B 5/11, published 03/27/2014), which describes the determination of sleep patterns using a sensor that records breathing and movement, as a sensor can be used, including bioradar.

The specified invention has the following disadvantages: data on the heart rate of the sleeping person are not used; signs are limited by the frequency of respiratory movements, the amplitude of the respiratory movements, the duration of the periods of movement, the spectral density of the power of the frequencies of respiration and amplitudes of respiration, the entropy of similarity; in the case of using a bioradar, it is necessary to determine the approximate distance between the bioradar and the user to set the range gate.

Utility Model Disclosure

The problem to which the claimed technical solution is directed is to improve the quality of sleep pattern recognition based on bio-radar monitoring.

The technical result is an improved quality of determining the structure of sleep based on bio-radar monitoring.

The result is ensured by the fact that the device for the automated determination of the sleep structure contains a bio-radar, which is a short-range radar with a continuous signal and a quadrature detector, the output of the bio-radar is connected to the input of the primary filtering unit for bio-radar signals, its output is connected to the input of the artifact detection unit, the output of which connected to the input of the block of inter-artifact periods, connected to the input of the phase reversal elimination unit, which is connected to the house of the unit for normalizing the inter-artifact periods, the output of which is connected to the filtering unit for the respiratory signal and the filtering unit for the heartbeat signal, the output of the filtering unit for the respiratory signal is connected to the unit for determining respiratory cycles and their parameters, the output of which is connected to the input of the unit for extracting signs of the respiratory signal, the output of the signal filtering unit the heartbeat is connected by the output of the heartbeat detection unit, and its output is connected to the input of the heartbeat signal extraction unit, the signal from which and the signal l with a block feature extraction respiration signal received at the input unit normalization characteristics, whose output is connected to the input of the primary classification unit connected to the input of the secondary classification, the signal from which is fed to an output device for interfacing with the block.

List of figures

FIG. 1. Block diagram of a device for non-contact determination of sleep structure: 1 - bioradar; 2 - block primary filtering bioradar signals; 3 - block definition of artifacts; 4 - block selection of artifact periods; 5 - block reversal phase reversal; 6 - block normalization of artifact periods; 7 - block filtering the breathing signal; 8 - unit for determining respiratory cycles and their parameters; 9 - block filtering respiratory cycles; 10 - block extraction of signs from a breathing signal; 11 - block filtering the heartbeat signal; 12 - block determining heart rate; 13 - block filtering heart rate; 14 - block extracting signs from a heartbeat signal; 15 - block normalization of signs; 16 - block primary classification; 17 - block secondary classification; 18 - block interface with the output device, 19 - output device.

FIG. 2. Sleep monitoring using the device simultaneously with polysomnographic research.

Utility Model Implementation

Sleep disorders are widespread in the population and often remain undiagnosed. Insomnia (insomnia) is one of the most common diseases, ~ 30% of the adult population reports problems with insomnia, and ~ 10% about chronic insomnia. Obstructive apnea, one of the disorders of breathing in a dream, is observed in 9-21% of women and 24-31% of men, while about 80% of cases of apnea remain undiagnosed. In addition, sleep disturbances significantly increase the risk of accidents and traffic accidents.

The use of devices for long-term monitoring of sleep with the goal of an objective daily assessment of the duration, structure and parameters of sleep, as well as the presence of respiratory disorders in a dream, is of interest both for the early diagnosis of sleep disorders (which is especially important in high-risk groups) and for monitoring the effectiveness of therapy . In addition, such monitoring can help in organizing a sleep-wakeful regime, increase the patient's attention to his health and motivate him to observe sleep hygiene. An objective daily assessment of the quality and duration of sleep can help in organizing a sleep-wake regimen, increase the patient’s attention to his health, and motivate him to observe sleep hygiene.

Thus, there is a need for tools for long-term home monitoring of sleep for both healthy people, for the purpose of prevention, and for people with sleep disorders, for monitoring the dynamics of the disease and correcting therapy.

Of particular interest is the ability to monitor sleep using non-contact methods that can provide the patient with maximum comfort. One of these methods is bioradiolocation (BRL). DRL is a method of remote detection and diagnosis of people, based on the modulation of the radar signal by vibrational movements and organ movements (contractions of the heart muscle, translational-recurrent movements of the chest during breathing, etc.).

The functioning of the device for the automated determination of sleep patterns is as follows (Fig. 1): a bioradar (1) is a short-range radar with a continuous signal and a quadrature detector. Bioradar radar (1) provides registration of vibrational movements and movements of human organs, namely the movements of the body and limbs, chest vibrations caused by respiratory movements and pulsations of the heart.

The output of the bioradar is from 2 to 16 digitized signals. Each signal represents an I or Q quadrature at a specific frequency. Thus, a total of 1 to 8 frequencies can be used in a given frequency range.

The primary filtering unit for bio-radar signals (2) provides band-pass filtering of digital signals coming from the radar in the range from 0.01 Hz to 5 Hz. The input of the bio-radar signal filtering unit (2) is connected to the output of the bio-radar (1), and the output of the unit is connected to the artifact detection unit (3). The artifact detection unit (3) provides the definition of artifacts by comparing the following parameters or a subset of them, calculated using a sliding window, with threshold values: signal energy; signal power spectral density; signal entropy; distance calculated using the dynamic timeline transformation method.

The output of the artifact definition block (3) is connected to the block for selecting artefact periods (4). The block for the selection of inter-artifact periods (4) provides a transition from several bio-radar signals (2-16) to one by selecting one bio-radar signal for each of the inter-artifact periods. The choice is made on the basis of comparing the following parameters of bio-radar signals during inter-artifact periods or a subset of them: signal entropy, maximum signal energy, distance calculated using the dynamic time transformation method.

The output of the block for the selection of inter-artifact periods (4) is connected by the phase reversal elimination block (5). The phase reversal elimination block (5) ensures the same orientation of the respiratory peaks throughout the recording due to the reversal of the artifact periods with an incorrect orientation. The correct orientation of the inter-artifact periods is determined by analyzing the height and width of the respiratory peaks.

The output of the phase reversal elimination block (5) is connected to the normalization block of the artifact periods (6). The block of normalization of artifact periods (6) provides the same average amplitude of artifact periods during recording. Normalization is performed by subtracting the average signal value during the artifact period and dividing by the standard deviation of the signal during the artifact period. The output of the unit for normalizing the inter-artifact periods (6) is connected to the filtering unit of the respiratory signal (7) and to the filtering unit of the heartbeat signal (11).

The respiratory signal filtering unit (7) provides band-pass filtering of the bio-radar signal in the range from 0.1 to 0.6 Hz. The output of the respiratory signal filtering unit (7) is connected to the unit for determining respiratory cycles (DC) and their parameters (8). The unit for determining the respiratory cycles and their parameters (8) provides identification of each respiratory cycle and its following parameters: the coordinates of the apex, the coordinates of the beginning of the respiratory cycle, the coordinates of the end of the respiratory cycle, amplitude and width. The output of the respiratory cycle determination unit and their parameters (8) is connected to the respiratory cycle filtration unit (9). The respiratory cycle filtration unit (9) removes falsely defined respiratory cycles. Respiratory cycles are deleted if their amplitude or width is greater than or less than the threshold values calculated based on neighboring respiratory cycles.

The output of the respiratory cycle filtration unit (9) is connected to the feature extraction unit from the respiratory signal (10). The feature extraction unit from the breath signal (10) provides feature extraction for each epoch (30 second signal interval). Signs are derived based on the power spectral density; respiratory rate; entropy of patterns, approximating entropy; a statistical description of the parameters of the respiratory cycles (mean, standard deviation, interquartile range, median of amplitudes and widths, etc.); the similarity of the breathing pattern of the era with the breathing pattern of neighboring eras, estimated using a dynamic time and frequency scales; standard deviation of DC correlation; respiratory volume and flow rate during inspiration, exhalation, and full DC, etc. The output of the respiratory cycle filtration unit (9) is connected to the normalization unit for signs (15).

The heartbeat signal filtering unit (11) provides band-pass filtering of the bioradar signal in the range from 0.7 to 2 Hz. The output of the heartbeat signal filtering unit (11) is connected to the heart rate detecting unit (12). The heart rate determination unit (12) provides identification of the heart rate and the distance between them (RR-intervals). The input of the heart rate determination unit (12) is connected to the heartbeat signal filtering unit (11), and the output of the unit is connected to the heartbeat filtering unit (13). The heart rate filtering unit (13) provides the removal of RR-intervals with a length greater than or less than the threshold values calculated on the basis of neighboring RR-intervals. The input of the heart rate filtration unit (13) is connected to the heart rate determination unit (12), and the output of the heart rate filtration unit (13) is connected to the feature extraction unit from the heartbeat signal (14). A feature extraction unit from a heartbeat signal (14) provides feature extraction for each era (30 second signal interval). Signs are extracted based on temporal, spectral and non-linear methods for analyzing heart rate variability. The input of the unit is connected to the heart rate filtering unit (13), and the output is connected to the feature normalization unit (15).

The feature normalization block (15) ensures the normalization of each of the extracted features during the recording. Normalization is carried out by subtracting the average value of the characteristic during the recording and dividing by the standard deviation of the characteristic during the recording. The inputs of the block (15) are connected to the output of the block for extracting signs from the heartbeat signal (14) and the output of the block for extracting signs from the breathing signal (10), and the output of block (15) is connected to the primary classification block (16).

The primary classification block (16) provides an estimate of the probability of eras belonging to a particular stage of sleep or wakefulness based on an analysis of a characteristic description. The input of the primary classification block (16) is connected to the normalization block of signs (15), the output of the block is connected to the secondary classification block (17). The secondary classification block (17) provides a classification of each epoch as belonging to a particular stage of sleep or wakefulness based on a probability analysis for the target epoch and its neighbors obtained in the previous block (16). The output of the unit (17) is connected to the interface unit with the output device (18).

The interface unit (18) with the output device transmits a sleep structure (a set of classified eras) to the output devices using a USB cable or a radio interface. As a radio interface, Bluetooth or Wi-Fi technology can be used. The output device may be a personal computer or smartphone. The input of the block (18) is connected to the secondary classification block (17), the output of the block (18) is connected to the output device (19).

The proposed device operates as follows. The user positions the device so that the bio-radar antennas are aimed at the place of sleep. The distance from the device to the user during sleep should not exceed 3 m. There should be no other moving objects except the user within the glow of the antenna. The user turns on the device and goes to sleep. The bio-radar signal, reflected from the surface of the user's body, is modulated by displacements of the body surface caused by cardiac contractions, respiratory movements and motor activity. The reflected signal hits the receiving antenna of the bioradar and is digitized by it. After which the signal undergoes preliminary processing and respiratory cycles and moments of cardiac contraction are localized on it (blocks 2–9, 11–13). Then, each 30-second interval (epoch) is described by a vector of signs extracted from the signal of heart rate variability (10) and breathing pattern (14). These signs are normalized (15) and are used for the primary classification (16) during which the probabilities that this era belongs to one of the stages of sleep (REM phase, superficial sleep, deep sleep) or wakefulness are assessed. Then, each era is described by the probabilities obtained at the previous stage for her and her immediate neighbors. This description is used for secondary classification, during which a decision is made to what stage of sleep or wakefulness each epoch belongs to. Information about the decisions taken from the interface unit (18) is transmitted to the output device. As a result, after waking up, the user can see the structure of his sleep during the night on the output device.

The operability of the developed device for the automated determination of sleep patterns was evaluated by comparing the sleep patterns obtained with it with the “gold standard” for determining sleep patterns — polysomnography. The experiments on parallel synchronous data recording by this device and polysomnograph were carried out at the Federal State Budgetary Institution “SZFMITs im. V.A. Almazova »Ministry of Health of Russia. The experiments were conducted for 32 subjects (12 of whom are men, average age 44.22 ± 15.44). Experiments have shown that this device allows you to determine the structure of sleep with an average measure of inter-expert consent kappa kappa of 0.5. In FIG. Figure 2 shows a photo of an experiment (monitoring sleep using a device simultaneously with polysomnography). A device for the automated determination of sleep patterns allows you to determine wakefulness, REM sleep phase, superficial sleep (1-2 stages of the slow sleep phase) and deep sleep (3rd stage of the slow sleep phase).

The technical result achieved is to determine the structure of human sleep based on bio-radar monitoring.

Claims (1)

A device for the automated determination of sleep patterns, containing a bio-radar, characterized in that the bio-radar is a short-range radar with a continuous signal and a quadrature detector, the output of the bio-radar is connected to the input of the primary filtering unit for bio-radar signals, its output is connected to the input of the artifact detection unit, the output of which connected to the input of the block between the artifact periods, connected to the input of the phase reversal elimination unit, which is connected to the input m block normalization of artefact periods, the output of which is connected to the filter block of the respiratory signal and the filter block of the heartbeat signal, the output of the filter block of the respiratory signal is connected to the block for determining respiratory cycles and their parameters, the output of which is connected to the input of the block for extracting signs of the respiratory signal, the output of the signal filter block the heartbeat is connected by the output of the heart rate determination unit, and its output is connected to the input of the heartbeat signal extraction unit, the signal from which is the signal from the block of extraction of signs of a breathing signal, they are fed to the input of the block of normalization of signs, the output of which is connected to the input of the primary classification block connected to the input of the secondary classification block, the signal from which is fed to the interface with the output device.

Images