KR20240057478A - Smart Management System for Infants and Toddlers using Big Data-based Machine Learning Analysis Algorithm - Google Patents

Abstract

The present invention accurately detects abnormalities in infants and young children by taking into account changes in biometric data that may vary depending on the infant's physical characteristics or environmental factors, and provides big data that allows guardians to accurately recognize infants' risks early and prevent accidents. This is about a smart management system for infants and toddlers using a machine learning analysis algorithm.
The infant smart management system using a big data-based machine learning analysis algorithm according to the present invention to solve the above problems includes a biometric data acquisition unit; a data processing unit that processes biometric data acquired from the biometric data acquisition unit; and a monitoring unit that receives in real time the data transmitted from the data processing unit and monitors the status of the sensing object, wherein the monitoring unit stores the data transmitted from the data processing unit and turns it into big data, and artificially converts the accumulated data into big data. Through intelligent machine learning techniques, the biometric data of the detection target is repeatedly learned to learn sleep habits, sleeping posture, and body temperature changes. Based on the learned information, the monitoring threshold is automatically set for each detection target, and the threshold is reached. It is characterized by automatically transmitting biometric data to related organizations linked in advance and guiding them to take emergency measures appropriate for the emergency situation on site.

Description

Smart Management System for Infants and Toddlers using Big Data-based Machine Learning Analysis Algorithm}

The present invention relates to a smart management system for infants and young children using a big data-based machine learning analysis algorithm. More specifically, it relates to a smart management system for infants and young children using a big data-based machine learning analysis algorithm that monitors sleep status to prevent sudden death that may occur during sleep in infants. It is about a smart management system for infants and toddlers.

In general, in the case of infants and young children, it is difficult to accurately recognize their health and other conditions to their guardians due to their lack of language and cognitive abilities, and because of this, guardians are unable to recognize the risks of infants and young children at an early stage, resulting in various causes such as large and small diseases and sudden death. Fatal accidents continue to occur.

Therefore, in recent years, various devices have been developed to monitor the condition of infants and toddlers so that health abnormalities, etc. can be detected and taken action at an early stage. Conventional technologies for this purpose include the intelligent baby care device control method of Registered Patent Publication No. 1970919 (hereinafter referred to as (referred to as ‘patent document’) is disclosed.

The patent document includes a body consisting of a seat and a backrest, a control unit, a holder that automatically moves under the control of the control unit after placing the body, a short-distance communication unit for wireless communication between the control unit and an external smart device, A first step of communicating with a smart device through various sensors and a short-range communication unit in the control unit of the intelligent baby care device, which consists of a network communication unit for communication between the control unit and the guardian's smartphone; Step 2, where the control unit determines the result based on a signal received through communication; Step 3: transmitting the result determined by the control unit to the guardian's smartphone through the network communication unit; In the intelligent baby care device control method consisting of, in the first step, the control unit controls the baby's physical condition including body temperature, images taken of the baby, weight of the holder, voice, ambient temperature, oxygen concentration, harmful gases, or fine particles. Dust is received as a signal, and in the second step, the control unit can obtain results on body temperature, tilt, or pulse through signals about the baby's physical condition, and in the third step, the device is used not only through the guardian's smartphone, but also through , It can also be transmitted to the police, and a voice detection sensor is connected to the control unit. The control unit receives a signal from the control unit of the car through a short-range communication unit, and through the signal, it is determined that there are no parents around and the car is turned off. In this case, the baby's crying is determined through the signal from the voice detection sensor, and if the crying continues for more than a certain period of time, the guardian's smartphone or the police is notified, and a temperature detection sensor is connected to the control unit. The ambient temperature is determined through a signal from the temperature sensor, and if the ambient temperature is below a certain temperature, which is the set minimum temperature, or above a certain temperature, which is the set maximum temperature, the thermoelectric element operates, and the control unit has an oxygen concentration sensor. It is connected, and an emergency oxygen box is installed at the bottom of the intelligent baby care device. The control unit determines the oxygen concentration through a signal from the oxygen concentration sensor. If it is determined to be below a certain oxygen concentration due to airtightness and insufficient oxygen, the emergency oxygen box By operating, the control unit controls the actuator coupled to the opening and closing valve of the emergency oxygen box when operating the emergency oxygen box, and controls the actuator to move backwards, thereby opening the oxygen supply door pipe of the emergency oxygen box, A harmful gas detection sensor and a fine dust sensor are installed in the control unit. The control unit receives a signal from the control unit of the car through a short-distance communication unit, and when the window of the car is opened through a signal, the harmful gas detection sensor detects harmful gas. It detects fine dust through a fine dust detection sensor, and when harmful gas or fine dust exceeds a certain level, it notifies the guardian's smartphone or the police.

However, the above patent document detects abnormal signs of infants and children riding in a vehicle and guides them to the guardian's smartphone, etc., by detecting the infant's inclination, body temperature, sound, pulse, oxygen concentration, harmful gas or fine dust, etc. It is a problem that detects the condition and determines whether there is an abnormality, or simply due to external environmental factors that may occur while driving the vehicle, misdetection of body temperature, sound, and tilt may occur, and trust in infant monitoring decreases due to repeated misdetection. There is.

Therefore, when acquiring biometric data of infants and toddlers and detecting abnormalities, abnormalities in infants can be accurately detected by taking into account changes in biometric data that may vary depending on the infant's physical characteristics or environmental factors, and through this, guardians can reduce the risks to their infants. The development of an improved infant management system is required to prevent accidents by accurately recognizing them at an early stage.

KR 10-1970919 B1 (2019. 04. 15.) KR 10-1021155 B1 (2011. 03. 03.) KR 10-2020-0001331 A (2020. 01. 06.) KR 10-1999453 B1 (2019. 07. 05.) KR 10-1755594 B1 (2017. 07. 03.) KR 10-2016-0069725 A (2016. 06. 17.)

The present invention was created to solve the problems of the conventional infant and young child smart management system as described above. The problem that the present invention aims to solve is to consider changes in biometric data that may vary depending on the infant's physical characteristics or environmental factors. The goal is to provide a smart management system for infants and young children using a big data-based machine learning analysis algorithm that accurately detects abnormalities in infants and through which guardians can accurately recognize infants' risks early and prevent accidents.

The infant smart management system using a big data-based machine learning analysis algorithm according to the present invention to solve the above problems includes a biometric data acquisition unit built in a smart cradle to acquire biometric data of the detection target; a data processing unit that processes biometric data acquired from the biometric data acquisition unit; and a monitoring unit that receives in real time the data transmitted from the data processing unit and monitors the status of the detection object, wherein the monitoring unit stores the data transmitted from the data processing unit and turns it into big data, and artificially converts the accumulated data into big data. Through intelligent machine learning techniques, the biometric data of the detection target is repeatedly learned to learn sleep habits, sleeping posture, and body temperature changes. Based on the learned information, the monitoring threshold is automatically set for each detection target, and the threshold is reached. It is characterized by automatically transmitting biometric data to related organizations linked in advance and guiding them to take emergency measures appropriate for the emergency situation on site.

And the smart cradle of the present invention includes a housing portion of a predetermined size with an open upper surface; a first cushion part made of elastic material that is inserted into the housing part and installed to expose a predetermined width along the upper circumference; a head seating portion located on one side of the upper surface of the first cushion portion and supporting the head of the sensing target; a body seating portion of a predetermined size located on the other side of the upper surface of the first cushion portion and supporting a body portion of the sensing object; and a second cushion part made of an elastic material positioned between the first cushion part and the body seating part and protruding to have a "U" shape, wherein the biometric data acquisition unit includes the head seating part and the body seating part. Another feature is that it is accommodated inside the housing portion.

In addition, in the present invention, the body seating portion is composed of an air cell having a plurality of grid-shaped isolated partition spaces, and a compressor is installed in the housing portion to supply air to the air cells of the body seating portion, so that the air cells are charged. Another feature is that the inclination of the body seating portion is changed by air to selectively induce a change in the posture of the sensing object.

In addition, the present invention includes a data storage unit where the monitoring unit stores data transmitted from the data processing unit; A machine learning unit that repeatedly learns the data stored in the data storage unit through machine learning; An analysis unit that detects abnormalities in the detection target by analyzing real-time data transmitted through the data processing unit and data learned through the machine learning unit; And another feature includes a control unit that selectively controls the operation of the smart cradle or transmits biometric data to a related organization according to the analysis results of the analysis unit.

In the present invention, the data processing unit amplifies a signal indicating sleep breathing of the detection target among the acquired biometric data, and the monitoring unit analyzes the sleep breathing data amplified and transmitted from the data processing unit to determine the sleep apnea section and Another feature is to shorten the sleep apnea time of the detection target or prevent the occurrence of sleep apnea by predicting the sleep apnea section according to the analyzed sleep apnea section and controlling the operation of the smart cradle. .

According to the present invention, biometric data according to the physical characteristics or environmental factors of the detection target is continuously collected and converted into big data, and this big data is repeatedly learned based on machine learning, which has the advantage of enabling customized monitoring for each detection target. .

In addition, it has the advantage of detecting sleep apnea of the monitored subject and inducing a change in sleeping position when necessary, thereby artificially eliminating the sleep apnea section and preventing sudden death of the monitored subject.

Figure 1 is a configuration diagram showing an example of a smart management system for infants and toddlers using a big data-based machine learning analysis algorithm according to the present invention.
Figures 2 and 3 are configuration diagrams showing an example of a smart cradle according to the present invention.
Figure 4 is a configuration diagram showing an example of a data processing unit according to the present invention.

Hereinafter, the present invention will be described in detail according to the accompanying drawings showing preferred embodiments.

The present invention accurately detects abnormalities in infants and young children by taking into account changes in biometric data that may vary depending on the infant's physical characteristics or environmental factors, and through this, guardians can accurately recognize infants' risks early and prevent accidents. The present invention is intended to provide a smart management system for infants and young children using a big data-based machine learning analysis algorithm. As shown in FIG. 1, the present invention consists of a biometric data acquisition unit 10, a data processing unit 20, and a monitoring unit 30. It comes true.

The biometric data acquisition unit 10 is configured to acquire biometric signals of the detection target. The biometric data acquisition unit 10 is built into clothes, bracelets, anklets, foot wraps, hand wraps, etc. worn by the detection target, or is installed in a smart device. It can be built into the cradle (1).

As shown in Figures 2 and 3, the smart cradle (1) has a housing part (1A) of a predetermined size with an open top, and is inserted into the housing part (1A) to expose a predetermined width along the upper circumference. A first cushion part (1B) made of an elastic material, a head seating part (1C) located on one side of the upper surface of the first cushion part (1B) to support the head of the sensing object, and the first cushion part A body seating portion (1D) of a predetermined size located on the other side of the upper surface of (1B) to support the body part of the sensing object, and a “U” shape located between the first cushion portion (1B) and the body seating portion (1D). It may be configured to include a second cushion portion 1E made of an elastic material that protrudes to have a shape.

In addition, the biometric data acquisition unit 10 is accommodated inside the housing unit 1A across the head seating unit 1C and the body seating unit 1D, and is configured to collect necessary biometric signals, respectively. At this time, the detection target The sound detection unit 12 and the apnea detection unit 14, which measure sounds such as breathing and snoring, are installed close to the head seating unit 1C, and the body temperature detection unit 11 and pulse detection unit 13 ), it can be installed close to the body seating portion (1D) and configured to reliably acquire biological signals.

In addition, the body seating unit 1D may be composed of a posture change inducing unit 16 made of air cells having a plurality of grid-shaped isolated partition spaces, and the housing unit 1A is provided with air cells of the body seating unit 1D. A compressor that supplies air can be installed, and through this configuration, when necessary, the air cell of the posture change inducing part 16 is charged with air to change the inclination of the body seating part (1D), thereby changing the posture of the sensing object. It can be guided as much as possible.

At this time, the compressor is accommodated and installed inside a sound-insulating case having a predetermined size, and through this, the noise generated during operation of the compressor can be reduced.

In addition, the housing unit 1A may include a motion detection unit 15 that detects the movement of the sensing object and an emergency alarm unit 17 that generates an alarm when an emergency situation occurs, and the body temperature detection unit 11 , the signals detected by the sound detection unit 12, pulse detection unit 13, apnea detection unit 14, motion detection unit 15, posture change induction unit 16, and emergency alarm unit 17 are data described later. The signal is processed through the processing unit 20 and transmitted to the monitoring unit 30.

The data processing unit 20 removes noise included in the various biometric signals acquired by the biometric data acquisition unit 10, amplifies the signals, and then converts the analog signals into digital signals, and biometric data converted into digital signals in this way is transmitted to the monitoring unit 30, which will be described later.

This data processing unit 20 includes a signal processing unit 21 and a wireless communication unit 22, as shown in FIG. 4, and the signal processing unit 21 includes a buffer (voltage follower) and a high frequency filter (High Pass Filter, HPF). , it can be configured to include an amplifier (AMP), a low-pass filter (LPF), and an offset. Through this, the loss of input bio-signals is minimized and noise in a specific band is removed to enable accurate analysis. do.

Among the biometric signals (biological data) acquired at this time, in the case of the sleep breathing signal detected by the apnea detection unit 14, the strength of the signal is amplified by 2 to 3 times, and through this, the monitoring unit 30 detects fine breathing and apnea. It is processed so that sections can be clearly distinguished.

The monitoring unit 30 may be configured as a monitoring app installed on a smartphone, and the monitoring unit 30 receives data transmitted from the data processing unit 20 in real time and monitors the status of the detection object.

As shown in FIG. 1, the monitoring unit 30 includes a data storage unit 31 that stores data transmitted from the data processing unit 20, and repeats the data stored in the data storage unit 31 through machine learning. A machine learning unit 32 that learns, and an analysis unit 33 that detects abnormalities in the detection target by analyzing real-time data transmitted through the data processing unit 20 and data learned through the machine learning unit 32. And a control unit 34 that selectively controls the operation of the smart cradle (1) or transmits biometric data to a related organization according to the analysis results of the analysis unit (33).

And the analysis unit 33 analyzes the sleep breathing data amplified and transmitted from the data processing unit 20 to determine and analyze the sleep apnea section, predicts the sleep apnea section according to the analyzed sleep apnea section, and operates the control unit 34. By controlling the operation of the smart cradle (1), it can be controlled to shorten the sleep apnea time of the detected object or to prevent the occurrence of sleep apnea.

At this time, the control unit 34 controls the operation of the compressor provided in the smart cradle 1 to change the shape of the posture change inducing part 16 made of air cells, and the change in shape of the posture change inducing part 16 causes the body seating part. As the tilt of the body, such as the body of the sensing object placed in (1D), changes, sleep apnea symptoms are naturally alleviated and prevented.

In addition, if sleep apnea symptoms are not alleviated or sleep apnea symptoms persist due to a change in the shape of the posture change inducing unit 16, the control unit 34 generates an alarm through the emergency alarm unit 17 or releases the alarm on a smartphone, etc. Information can be provided so that the user can recognize it, and through this, the user can be encouraged to directly check the status of the detection target.

In addition, the control unit 34 detects peak values from sounds collected during sleep through the sound detection unit 12, compares them with a reference value, and patterns the detected peak values to predict whether sleep is abnormal.

At this time, if the detected peak value is less than the reference value, the reference value is reset, the case where the peak value is not detected within the set period is determined, the peak value that occurs outside the set period is treated as noise, and the peak value is within the range of the set pattern. It is determined that the detection target is in a deep sleep state.

In addition, when the peak value is not detected, it is judged as an apnea state or a deep sleep state by comparing it with the threshold time, and the peak value outside the period processed as noise is judged to be noise caused by movement such as tossing and turning of the sensing object, and the sleep state is determined as in Table 1. (W, LS, DS, REM) are judged.

division Waking up (W) Light sleep state (LS) Deep sleep state (DS) REM sleep state HR average height W minus 5 to 20 b/min LS minus 5 to 10 b/min Between W and LS Variability greatness littleness very small greatness rhythm regular regular regular very irregular HR changes or events plenty part doesn't exist Almost none and short BM great movement has exist has exist doesn't exist doesn't exist small movements has exist has exist doesn't exist doesn't exist twitch doesn't exist doesn't exist doesn't exist has exist

Based on the biometric data acquired as above, the sleep state is patterned and converted into data, this data is repeatedly and cumulatively collected and turned into big data, and the sleep state pattern of the big data detected target is repeatedly learned through artificial intelligence machine learning techniques. You will establish customized sleeping habits.

If the sleeping habits of the sensing target are customized as described above, abnormalities in the sleeping state of the sensing target can be accurately detected at an early stage, and if necessary, a change in sleeping posture can be induced or the user can be guided. By automatically contacting related organizations (119, hospitals, etc.) linked in advance, it is possible to reduce emergency situations such as sudden death of the detection target.

Above, it was explained that it is only about early detection of abnormalities such as sleep apnea in sleeping habits, but in addition, the sleeping posture and body temperature changes of the detection target are additionally learned, and through this, the optimized monitoring threshold value is automatically set for each detection target. And, when the threshold is reached, biometric data is automatically transmitted to related organizations linked in advance, so that emergency measures appropriate for the emergency situation can be carried out on site.

Meanwhile, the monitoring unit 30 is composed of a smartphone app and provides real-time guidance by displaying the analyzed sleep state patterns, posture, and body temperature changes in graphs, etc., and provides guidance on sleep abnormalities such as sleep apnea through vibration and sound. , It can be configured to guide the health index by converting the health status analyzed based on the biometric data of the detection target into a comprehensive score.

Additionally, the monitoring unit 30 may be configured to register, change, and delete detection targets through settings.

In the above, it was explained that the patterned accumulated data of the detection target is converted into big data and repeated learning is performed through artificial intelligence machine learning techniques. However, if sufficient biometric data of the detection target is not secured during initial operation, Sleep pattern data is extracted from the average data stored in the cloud server, and the extracted data is compared and analyzed with biometric data acquired in real time from the detection object to determine sleep abnormalities, etc. After the biometric data of the detection object has accumulated sufficiently, It can be configured for customized operation by repeatedly learning the accumulated biometric data of the detection target through artificial intelligence machine learning techniques.

As described above, the present invention continuously collects biometric data according to the physical characteristics or environmental factors of the detection target and turns it into big data, and this big data is repeatedly learned based on machine learning to enable customized monitoring for each detection target. do.

In addition, it detects sleep apnea of the monitored subject and induces a change in sleeping position when necessary, thereby artificially eliminating the sleep apnea section and preventing sudden death of the monitored subject.

Above, for convenience of explanation, reference numerals and names have been given to the drawings showing preferred embodiments and the configurations shown in the drawings. However, this is an embodiment according to the present invention, and the shapes shown in the drawings and the names given are The scope of the rights should not be construed as limited, and changes to various shapes that can be predicted from the description of the invention and simple substitution to a configuration that performs the same function are within the scope of changes that can be easily carried out by a person skilled in the art. You will see this as extremely self-evident.

1: Smart cradle 1A: Housing part
1B: First cushion part 1C: Head seating part
1D: Body seating part 1E: Second cushion part
10: Biometric data acquisition unit 11: Body temperature detection unit
12: Sound detection unit 13: Pulse detection unit
14: Apnea detection unit 15: Motion detection unit
16: Posture change induction unit 17: Emergency alarm unit
20: data processing unit 21: signal processing unit
22: wireless communication department 30: monitoring department
31: data storage unit 32: machine learning unit
33: analysis unit 34: control unit

Claims (5)

A biometric data acquisition unit (10) built into the smart cradle (1) to acquire biometric data of the detection target;
a data processing unit (20) that processes the biometric data obtained from the biometric data acquisition unit (10); and
a monitoring unit 30 that receives data transmitted from the data processing unit 20 in real time and monitors the status of the detection object;
Including,
The monitoring unit 30,
The data transmitted from the data processing unit 20 is stored and converted into big data, and the accumulated data is used to repeatedly learn the biometric data of the detection target through artificial intelligence machine learning techniques to learn sleep habits, sleeping posture, and body temperature changes, Based on the learned information, the monitoring threshold is automatically set for each detection target, and when the threshold is reached, the biometric data is automatically delivered to the relevant organizations linked in advance, so that emergency measures appropriate for the emergency situation can be carried out on site. A smart management system for infants and toddlers using a big data-based machine learning analysis algorithm.
In claim 1,
The smart cradle (1) is,
A housing portion (1A) of a predetermined size with an open upper surface;
a first cushion part (1B) made of elastic material that is inserted into the housing part (1A) and installed to expose a predetermined width along the upper circumference;
a head seating portion (1C) located on one side of the upper surface of the first cushion portion (1B) to support the head of the sensing object;
a body seating portion (1D) of a predetermined size located on the other side of the upper surface of the first cushion portion (1B) and supporting the body part of the sensing object; and
a second cushion part (1E) made of elastic material that protrudes to have a “U” shape and is located between the first cushion part (1B) and the body seating part (1D);
Including,
The biometric data acquisition unit 10,
A smart management system for infants and toddlers using a big data-based machine learning analysis algorithm, characterized in that it is accommodated inside the housing portion (1A) across the head seating portion (1C) and the body seating portion (1D).
In claim 2,
The body seating portion (1D),
It is composed of air cells with a plurality of grid-shaped isolated compartment spaces,
In the housing portion 1A,
A compressor is installed to supply air to the air cell of the body seating portion (1D), and the tilt of the body seating portion (1D) is changed by the air charged in the air cell to selectively induce a change in the posture of the sensing object. A smart management system for infants and toddlers using a big data-based machine learning analysis algorithm.
In claim 1,
The monitoring unit 30,
a data storage unit 31 that stores data transmitted from the data processing unit 20;
A machine learning unit 32 that repeatedly learns the data stored in the data storage unit 31 through machine learning;
An analysis unit 33 that detects abnormalities in the detection target by analyzing real-time data transmitted through the data processing unit 20 and data learned through the machine learning unit 32; and
A control unit 34 that selectively controls the operation of the smart cradle (1) or transmits biometric data to a related organization according to the analysis results of the analysis unit 33;
A smart management system for infants and toddlers using a big data-based machine learning analysis algorithm, characterized in that it includes.
In claim 1,
The data processing unit 20,
Among the acquired biometric data, a signal indicating sleep breathing of the detection target is amplified,
The monitoring unit 30,
By analyzing the sleep breathing data amplified and transmitted from the data processing unit 20, the sleep apnea section is determined and analyzed,
Big data that predicts the sleep apnea section according to the analyzed sleep apnea section and controls the operation of the smart cradle (1) to shorten the sleep apnea time of the detection target or prevent the occurrence of sleep apnea. Infant smart management system using based machine learning analysis algorithm.

Images