RU206144U1 - Data registration and transmission unit for a human sleep monitoring system - Google Patents

Abstract

The utility model relates to electronic technology and specialized computer devices and can be used to create devices that control the parameters that characterize a person's sleep. The technical result achieved by the implementation of this utility model consists in expanding the functionality of processing the output signals of sensitive sensors by monitoring a wide range of parameters during a person's sleep, such as respiratory rate, heart rate, sleep phases and position of a person in a dream, etc. increasing the accuracy of their measurements and achieving compactness of the proposed unit. The technical result is achieved due to the fact that the unit for recording and transmitting data for a human sleep monitoring system contains a housing made in the form of a plastic molded form, inside which there is a central processing unit (CPU), an analog-to-digital converter, a unit for interacting with a piezoelectric sensor, providing additional amplification and filtering of the signal, a module with a battery, a radio modem and a non-volatile memory module, while additionally containing a unit for interacting with a bend sensor. 4 c.p. f-ly, 3 dwg

Description

The utility model relates to electronic technology and specialized computer devices and can be used to create devices that control the parameters that characterize a person's sleep. The proposed technical solution is a specialized device of electronic and computer technology, containing elements (blocks) and connections between them, which are in functional and constructive unity and placed in a limited space with the possibility of execution in a single housing.

The following technical solutions are known from this prior art.

There is a known sleep monitoring system (patent CN208837915), which includes a sensitive device made in the form of a piezoelectric film sensor, an analog processing unit that amplifies and filters the output signal of the piezoelectric sensor, an analog-to-digital converter (ADC) for converting the amplified and filtered analog signal into digital signal. The digital signal processing module analyzes the received data from the ADC, including information about breathing, heartbeat and movements of the user during sleep, and transfers the data using the communication module to the information storage module. The analog and digital nodes of the device are powered by the power module.

Known device for monitoring biophysical parameters of a person during sleep (patent CN203914900), consisting of a housing made of high-strength metal. Inside the housing there is a sensitive sensor for determining the heart rate, preferably a MEMS accelerometer, and a central control unit, which consists of a series-connected signal amplifier, an analog-to-digital converter, a microprocessor and a wireless communication unit.

The wireless communication unit can be configured to transmit data via ZigBee, Z-Wave or Bluetooth.

The housing is also equipped with a power supply module for supplying power to the aforementioned components.

A monitoring system (patent CN106725339) is known, which includes a piezoelectric film sensor, a signal amplification and filtering unit, a central signal processing unit and a communication unit, characterized in that the signal amplification and filtering unit includes an integral amplifying subunit, a high-frequency filtering subblocks, a low-frequency filtering unit amplifying subunit connected in series, wherein the integral amplifying subunit is connected to the output contact of the piezoelectric sensor, and the low-frequency filtering amplifying subunit is connected to the input to the central signal processing unit.

Known sleep monitoring system (patent CN107595251), which allows you to determine the basic vital parameters of a person during sleep.

Piezoelectric film sensors with a signal amplification module are built into the pillow to detect breathing and heart rate signals of the user, as well as body movement during sleep. The declared sensors with an amplification module are connected to the DSP controller via a USB cable. The controller filters the amplified analog signal received from the piezoelectric sensor using a filtering module, after which the received signal is digitally processed using an ADC when implementing the fast Fourier transform method for analyzing the frequency spectrum in order to isolate the frequency and amplitude of respiration, heartbeat and body movement ...

Then the processed information about the user's parameters is transmitted wirelessly via Bluetooth to the mobile device and displayed in the WeChat application. An SD card is additionally used to store information about the user's parameters during sleep.

These devices allow you to register the necessary vital parameters for monitoring a person's sleep, but they do not have the ability to determine a person's posture during sleep, to register the number of night rises, which does not allow comprehensive monitoring of a person's sleep indicators, and also have limitations associated with the ability to obtain data only from one person, while when two people are on the registered surface, the device that controls the parameters will receive inaccurate information.

Nevertheless, the technical solution described in patent CN107595251 is the closest analogue to the claimed utility model and can act as a prototype.

The task of this utility model is to create a compact unit for recording and transmitting data for a sleep monitoring system with the possibility of analog and digital processing of the output signals of the piezoelectric sensor and bending sensor and the implementation of mathematical algorithms in order to determine additional parameters, thereby providing a comprehensive monitoring of the biophysical parameters of a person's sleep, as well as with the possibility of implementing a configured system for monitoring human parameters during sleep for accurate and reliable recording of data received by the system simultaneously from two users in the case of embedding the system in a double zone, and the data recording and transmission unit is designed to be conveniently attached directly to the tape containing the declared sensitive sensors.

The technical result achieved by the implementation of this utility model consists in expanding the functionality of processing the output signals of sensitive sensors by monitoring a wide range of parameters during a person's sleep, such as respiratory rate, heart rate, sleep phases and position of a person in sleep, etc. while increasing the accuracy of their measurements and achieving the compactness of the proposed unit.

The technical result is achieved due to the fact that the unit for recording and transmitting data for a human sleep monitoring system contains a housing made in the form of a plastic molded form with an indicator made in the form of an LED, a control button and a connector for connecting a charger, while inside the housing there is a central a processor (CPU), an analog-to-digital converter, a unit for interacting with a piezoelectric sensor, providing additional amplification and filtering of the signal, a module with a battery, a radio modem and a nonvolatile memory module, characterized in that the housing in its inner part additionally contains a unit for interacting with the sensor bending, a module for the configuration of the type of sleeping area, which defines a single or double sleeping area, and a module for determining the side of a double area.

In a preferred embodiment of the unit, the unit for interacting with the piezoelectric sensor includes a signal preprocessing unit, a signal filtering unit, a signal amplification unit, and these units are connected in series with each other, while the unit for interaction with the piezoelectric sensor additionally contains a reference voltage generator, connected to the inputs of the preprocessing unit and the amplification unit, as well as a digital potentiometer connected to the central processor and the amplification unit.

In another preferred embodiment of the unit, the unit for interacting with the bend sensor includes a preprocessing unit, a sensor channel switching unit made in the form of an analog multiplexer, and a matching unit, these units being connected in series with each other.

In another preferred embodiment of the unit, a 2.4 GHz radio modem that supports BluetoothLowEnergy technology is built into the CPU.

In another preferred embodiment of the block, the analog-to-digital converter is built into the CPU.

In another preferred embodiment, an accelerometer is built into the unit.

The possibility of simultaneous analog and digital processing of the output signals of the piezoelectric sensor and the bend sensor with the implementation of various configurations of the declared unit associated with varying the number of users and binding to the sleeping area, as well as integration into the software add-on unit that calculates additional human parameters based on the information received from the declared sensitive sensors, provide a comprehensive monitoring and analysis of human parameters during sleep, which allows more accurate diagnostics of the user.

Further, the utility model will be described with reference to the drawings:

Fig. 1 is a hardware diagram of a unit for recording and transmitting data for a system for monitoring human sleep parameters, where 1 is a central processing unit (CPU), 2 is a unit with an amplifier for interacting with a piezo sensor, 3 is a unit for interacting with a bend sensor, 4 is an indicator, made in the form of an LED, 5 - a control button, 6 - a configurator module of the sleeping area type, 7 - a module for determining the side of a sleeping area, 8 - a non-volatile memory module, 9 - a low-frequency quartz resonator;

fig. 2 is a block diagram of analog processing of a piezoelectric sensor signal, where 10 is a piezoelectric sensor, 11 is a preprocessing unit, 12 is a filtering unit, 13 is an amplification unit, 14 is a reference voltage generator (GON), 15 is a digital potentiometer;

fig. 3 is a block diagram of analog processing of a bending sensor signal, where 16 is a bending sensor, 17 is a preprocessing unit, 18 is a channel switching unit, 19 is a matching unit.

The considered block for recording and transmitting data for a sleep monitoring system allows you to connect such sensors for recording parameters as:

piezo-sensitive tape (piezo sensor) to determine the user's pulse rate and breathing rate;

multichannel ribbon bend sensor that detects the user's posture or position during sleep.

Taken together, when processing the parameters obtained from these sensors, it also becomes possible to determine the user's sleep phase, which is important when processing the obtained parameters to identify potential health problems and issue recommendations for adjusting the diet, lifestyle or timely warning a person about the need to visit doctor. Such measures allow timely signaling of serious illnesses, such as nocturnal sleepiness, snoring, cardiovascular diseases, diseases of the nervous system (panic attacks, restless legs syndrome), sleep disturbances and promptly transcending their treatment. And also, information about the position of a person in a dream can identify or prevent diseases of the spine (for example, cervicothoracic osteochondrosis) in the early stages.

For analog processing of the output signals of the claimed sensitive sensors, blocks for interaction with the piezoelectric sensor and the bend sensor are located in the data recording and transmission unit.

A block with an amplifier for interaction with a piezoelectric sensor (2) provides the conversion of the mechanical effect exerted on the piezoelectric sensor by the user into an electrical signal, amplification and filtering of the signal to expand the dynamic range of the signal, as well as transmission of the signal to the input of the analog-to-digital converter (ADC), built into the central processor (1).

The unit for interaction with a multichannel bending sensor (3) converts the resistance coming from several outputs of the bending sensor into an electric voltage, commutes the converted signals and transfers them to the ADC input.

An analog-to-digital converter (ADC) is used to convert the output signal from the piezoelectric and bend sensor units into a digital signal and transmit the received signal to the central processor (1).

The central processor is designed to generate a signal to activate the polled sensors and process the received signals, control the periphery of the registration unit (indicators, vibration motor), transmit data and interact with a smartphone through a radio modem built into the central processor (1).

The radio modem operates at a frequency of 2.4 GHz and supports Bluetooth LowEnergy technology to transmit information about the user's biophysical parameters and / or other information measured or calculated by the processor unit to an electronic device (for example, a smartphone).

For the functioning of the central processor, the data recording and transmission unit also includes:

nonvolatile memory, which is a memory device in the form of a digital microcircuit intended for storing system and / or user information;

low-frequency quartz resonator (9) - an electronic component - an additional source of clock for the central processor, necessary for the operation of the real-time clock.

To control the data transmission recording unit, the housing contains:

an indicator element made in the form of a set of LEDs (4), controlled by the central processor and designed to signal various statuses of the device;

mechanical momentary button (5) designed for various user actions (calibration, hard reset, etc.).

For more accurate processing of the received signals and expanding the processing capabilities of signals associated with receiving data from two users, the following have been introduced into the data recording and transmission unit:

a configuration module for the type of sleeping area, which determines whether the sleeping area is single or double. This module is necessary for the correct display of the appearance of the bed and the indicator of the occupancy of the bed in the mobile application;

a configuration module for the side of the sleeping area, which allows you to determine which side of the sleeping area is a specific user of the sleep monitoring system (in the case of a double area).

The device has the ability to operate both from a storage battery (AKB), which ensures the autonomous operation of the device, and from an external source, and the switching of the device power from the battery to an external power source and vice versa (depending on whether the external source is connected) is carried out at the expense of the node, containing a power switch. The charging is carried out at the expense of the charger.

The supply of low voltage (5 V) for charging the built-in battery is provided through the USB port of the data logging and transmission unit.

The circuit unit of the voltage divider, the output of which is connected to the central processor, is designed to detect the presence of an external power supply.

Also, as part of the block under consideration, it is proposed to use an additional sensor - an accelerometer in order to implement energy saving algorithms. The following approach to the implementation of energy saving is assumed: the accelerometer provides the ability not only to measure acceleration along three axes of Cartesian coordinates, but also to implement its own simple logic, for example, tracking the excess of the acceleration threshold along one or several axes. The result of such tracking is the formation of a logic level at a special output of the accelerometer microcircuit. Thus, by setting a certain acceleration threshold, the central processor (1) can turn off the most power-consuming components of the system, such as blocks associated with the piezoelectric sensor and the bend sensor (2, 3), and wait for a signal from the accelerometer, which will be triggered at the moment when the user lies on the mattress.

The implementation of the utility model is implemented as follows (figure 1).

The data recording and transmission unit is fixed on the output connectors of the piezoelectric sensor and the bend sensor. Using the button (5) on the body of the unit, it is activated, the indicator (4) lights up, the unit goes into standby mode for connecting the user's smartphone, while interaction with the smartphone is carried out via Bluetooth. Also, button (5) can be used to calibrate the bend sensor and reset the device to factory settings.

After connecting the smartphone, the user adjusts the sensors relative to the sleeping area, for this, the type of sleeping area is selected in the application on the smartphone, if the sleeping area is intended for one person, in the configuration module of the sleeping area type (6), the single user mode is activated, if it is double, then the mode two users.

In the case of a double zone, a separate set of declared sensitive sensors is provided for each user, located on the left and right sleeping zones and connected to independent registration and transmission units, characterized by a configurator of the side of the sleeping area.

To bind the smartphone of each user to the side of the sleeping area, it is necessary to activate the configuration module of the side of the sleeping area (7), for which each of two users must, in accordance with the recommendation of the smartphone, take a lying position on the selected side of the sleeping area, while the data from the bend sensor enters the module configuration of the side of the sleeping area (7) and the module fixes this data. During sleep, the analog signal from the piezoelectric sensor goes through the block with an amplifier (2) to the central processor (1) through an analog-to-digital converter. The signal from the bend sensor is similarly fed to the central processor (1) through the unit (3). The processing of the parameters characterizing a person's sleep, obtained from the piezoelectric sensor and the bending sensor, is carried out taking into account the setting and calibration using the configurator module of the sleeping zone type (6) and the module for determining the side of the sleeping zone (7).

The process of analog signal processing implemented in a microcontroller associated with a piezoelectric sensor is shown in Fig. 2.

At the hardware level, an analog signal goes through three stages of processing:

matching - converting the charge generated by the sensor into an electrical voltage;

anti-aliasing filtering to eliminate the effect of aliasing;

amplification to increase the dynamic range of the signal.

The dotted line indicates the logical placement of elements outside or inside the processor unit.

The signal from the piezoelectric sensor (10) is fed to the preprocessing unit (11) with a zero gain (repeater on an operational amplifier). At the output of the block, an undistorted signal with a bias is observed, created using a reference voltage generator (14), implemented on a linear voltage regulator microcircuit. Bias is required because the system uses unipolar power and the piezo output is bipolar. The presence of the offset allows you to shift the average (zero) signal level approximately in the region of half of the supply voltage and thus fix the full swing of the signal.

From the output of the preprocessing unit (11), the signal enters the filtering unit (12).

Block (12) is built on the basis of a second-order active filter according to the Sallen-Key scheme and is designed to suppress the signal in the output spectrum, which can significantly reduce the effect of network pickups on the signal and eliminate the aliasing effect. Since the filter has zero gain, it does not require a voltage reference.

After filtering, the signal enters the amplification unit (13), designed to expand the dynamic range of the useful signal. The block is built on the basis of a non-inverting amplifier circuit based on an operational amplifier. Since the signal arrives at the input of the block with a bias, the feedback circuit is connected to the reference voltage. The gain can be varied by software using a digital potentiometer (15), which is an integrated circuit with a digital SPI interface and connected to a central processing unit (CPU). After amplification, the signal is fed to the input of an analog-to-digital converter (ADC) built into the central processing unit (CPU). Further signal processing is performed using digital methods.

The analog signal processing in the microcontroller associated with the bend sensor is shown in FIG. 3.

The bend sensor (16) is connected to a preprocessing unit (17), which converts the resistance of the sensor segments into an electrical voltage. The circuitry unit is implemented on the basis of voltage dividers, the number of which is determined by the number of sensor channels.

From the preprocessing unit, the signals go to the channel switching unit (18), which is an analog multiplexer that switches the signal from several inputs to one output in accordance with the state of the address inputs, which are set by the central processing unit (CPU).

From the output of the switching unit, the signal enters the matching unit (19), implemented as a voltage follower on an operational amplifier. The signal then goes to an analog-to-digital converter (ADC) built into the central processing unit (CPU).

The composition of the described architecture is dictated by the classical methods of analog circuitry.

After analog processing, the signals of the declared sensitive sensors undergo digital processing using an ADC built into the central processing unit (CPU).

The digital processing of the signals from the piezoelectric sensor and the bend sensor is carried out through the implementation of mathematical computational methods in the software modules of the ADC of the registration and transmission unit.

To extract the pulse and respiration values from the output signal of the unit associated with the piezoelectric sensor, the following steps are performed:

a digital IIR filter (filter with infinite impulse response) of low pass for suppressing the pulse signal and isolating the respiration signal;

calculation of the respiration spectrum by the fast Fourier transform (FFT) method in a window of 4096 points and calculation of its frequency from the maximum of the spectral component;

IIR digital high-pass filter for respiratory signal suppression;

decomposition into experimental modes (using the EEMD method), as a result of which the signal is decomposed into several components (modes). Thus, it is possible to get rid of noise and quite effectively implement the selection of the pulse signal against the background of the respiration signal remaining after the first stage;

calculation of the spectrum of the pulse signal by the FFT method in a window of 4096 points and the calculation of the heart rate.

These steps are given as an example, since signal processing methods may vary.

Practice has shown that the process of analog processing of the output signal of the piezoelectric sensor is quite sensitive to external disturbances (network noise, sudden movements of the user, characteristics of the sleeping area), which leads to a high degree of variability of the signal shape both in amplitude and in time. Therefore, to solve this problem, several additional steps are used to digitally process the arrays of heart rate and respiration rates before the user sees the numerical values in the mobile application:

discarding invalid values in retrospect from 10 dimensions. The rejection criterion is going beyond the threshold equal to the average value in the specified window of 10 points. If the change in the current value in absolute value exceeds the threshold by 3 units, then the value is replaced by the average;

multistage averaging of the obtained curve using a sliding window consisting of three points: the current point, one previous and one subsequent value is taken. If the change in modulus of the current value exceeds the average by 0.1 units in the window of the third point, then it is replaced by the average.

The first step is implemented in the registration and data transfer unit, and the second step is implemented on the side of the mobile application.

These steps are given as an example, since signal processing methods may vary.

To expand the spectrum of analyzed human sleep parameters, the data recording and transmission unit includes a hardware add-on that implements mathematical methods and algorithms for more accurate registration of parameters obtained from sensitive sensors and determination of additional parameters based on data received from sensors, which increases the functionality of the proposed unit.

To determine the user's sleep phase in the registration and transmission unit, a mathematical algorithm is implemented based on a software add-on, which is based on the data of the piezoelectric sensor and assumes the counting of movements per unit of time. If the signal of the piezoelectric sensor exceeds a certain threshold, which occurs only during movements, then the movement is recorded. The intensity of movement is the higher, the longer the signal of the piezoelectric sensor is below the threshold value.

To determine the user's position on the sleeping area during sleep, digital processing of the output signal of the unit associated with the bend sensor is carried out using a software component that implements the machine learning method, the logic of which is based on the concept of a person's posture, defined as a stable horizontal position of the human body in time on beds.

For this, at the first stage, the model is trained, in the process of which a known position on the mattress is put in accordance with a specific set of measured values of the bend sensor. After that, a model is synthesized, which subsequently independently classifies the poses when an array of bending resistances is fed to its input. One example of this process includes the following steps:

collecting training data and comparing them with known poses, whereby known poses can be classified as follows:

on the back

on the left side

on the right side

on belly

training the model using the CreatML framework and the RandomForest algorithm with 3000 iterations;

Applying the trained model to classify poses based only on the flexure sensor readings.

Thus, the use of this useful model will make it possible to carry out a comprehensive assessment of users' sleep, improve the accuracy of control of biophysical parameters characterizing a person's sleep, due to the possibility of simultaneous analog and digital processing of the signals of the piezoelectric sensor and bending sensor, integration into a software add-on unit that provides the determination of additional indicators based on information from the declared sensitive sensors, and providing the possibility of a more precise adjustment of the processing of the data received from the piezoelectric sensor and the bend sensor, allowing to obtain data when the sensors are installed on various types of sleeping zones.

Claims (5)

1. The block for recording and transmitting data for a human sleep monitoring system contains a housing, while inside the housing there are a central processing unit (CPU), an analog-to-digital converter, a block for interacting with a piezoelectric sensor, which provides additional amplification and filtering of the signal, a module with a battery and a module non-volatile memory, characterized in that it further comprises a unit for interacting with the bend sensor. 2. The unit for recording and transmitting data for the human sleep monitoring system according to claim 1, characterized in that the unit for interaction with the piezoelectric sensor includes a signal preprocessing unit, a signal filtering unit, a signal amplification unit, and these units are connected in series with each other , while the unit for interaction with the piezoelectric sensor additionally contains a reference voltage generator connected to the inputs of the preprocessing unit and the amplification unit, as well as a digital potentiometer connected to the central processor and the amplification unit. 3. The unit for recording and transmitting data for the human sleep monitoring system according to claim 1, characterized in that the unit for interacting with the bend sensor includes a preprocessing unit, a sensor channel switching unit made in the form of an analog multiplexer, and a matching unit, when This said blocks are connected in series with each other. 4. The unit for recording and transmitting data for a human sleep monitoring system according to claim 1, characterized in that it includes a radio modem. 5. The unit for recording and transmitting data for a human sleep monitoring system according to claim 1, characterized in that an accelerometer is built into the unit.

Images