KR20230108853A - vulnerable social group danger recognition detection system based on multiple sensing - Google Patents

Abstract

The present invention relates to a multi-detection based vulnerable social group danger recognition detection system, which is installed in a residence of a vulnerable social group and configured to detect a state of the residence in real time and transmit, if a result of analysis conducted based on the detected information indicates a dangerous situation, information about the dangerous situation to guardians or related organizations, thereby making it possible to quickly respond to the situation. In addition, the present invention includes: a sensing means for detecting biometric information, behavior, and residence states of a person to be monitored; and a control means for analyzing a dangerous situation of the person to be monitored through sensed information input from the sensing means and generating an alarm signal according to an analysis result. In the present invention, when an abnormal situation is recognized in any one among pieces of sensed information sensed through multi-detection, analysis is conducted by combining the abnormal situation with a sensed value of related sensed information so that a dangerous situation can be analyzed more accurately. In other words, by means of the present invention, warning alarms can be generated more accurately. As a result, errors in crisis responses, such as emergency dispatch, can be reduced, thereby reducing the occurrence of unnecessary responses or enabling rapid responses based on accurate situation determination.

Description

Vulnerable social group danger recognition detection system based on multiple sensing}

The present invention is installed in the residence of the vulnerable class to detect the monitored person and the residence status in real time, and if the result of analysis based on the detected information is a dangerous situation, the information is transmitted to the guardian or related institution to quickly respond It relates to a multi-sensing based risk perception detection system for vulnerable groups.

Due to recent nuclear familyization, low birth rate and aging population, the number of elderly people living alone is increasing. Since there is a limit to how to know even if an elderly person living alone is in danger or if an abnormality occurs, the need to establish a social safety net for the welfare of the elderly and the vulnerable is emerging in the government and welfare-related institutions.

In the prior art, in order to detect the status of the vulnerable class in real time, an IoT-based system including various sensors is installed and operated in a remote location.

The system determines a dangerous situation when one or more of the values sensed through each sensor is determined to be abnormal information.

In this case, the abnormal information may include an act of sitting in a seat, an act of lying on a bed, and an outsider's fall, but the system cannot distinguish between these and analyzes invalid garbage information to determine the situation.

In this case, a lot of information is input from various sensors, and since the system analyzes all the information, not only the processing speed is slow, but also the reliability of the analysis result is low.

In addition, when the analysis is based on one sensor value, unnecessary dispatch may occur because the act of lying on the bed and the fall cannot be distinguished.

In addition, since it is not possible to distinguish dangerous situations such as fainting after a fall and trauma in detail, it is difficult to cope with accurate situation judgment.

KR 10-2167296 B1

In order to solve the above-mentioned problems, the present invention receives multiple detection information from detection sensors provided in a residence, and when abnormal information is detected by any one of the sensors, an event is generated to determine whether the event is dangerous, and the event occurrence time An object of the present invention is to provide a multi-sensing based risk recognition detection system for vulnerable groups that analyzes multi-sensing information of a time corresponding to the multi-sensing system to more accurately determine a dangerous situation.

Another object of the present invention is to detect garbage information that is not valid for analysis among sensing information input through multi-sensing, and apply only valid sensing information to an artificial neural network to improve the reliability and processing speed of analysis results. Multi-sensing based vulnerable class It is to provide a risk recognition detection system.

A multi-sensing based risk perception detection system for the vulnerable class according to an embodiment of the present invention includes a sensing means for sensing biometric information, behavior and residence status of a person to be monitored; and control means for analyzing a dangerous situation of the person to be monitored through the sensing information input from the sensing means and generating an alarm signal according to the analysis result.

The detecting unit may include a biometric information collection module for sensing biometric information of a person to be monitored; a behavior detection module for detecting a behavior of a subject to be monitored; a door detection module for detecting whether a door is opened or closed in a residence; and a power detection module for detecting power consumption in the residence.

The control means includes an information input module receiving sensing information from the sensing means; a validity judgment module that receives sensing information from the information input module and determines validity of the sensing information; an information storage module for receiving and storing sensing information transmitted from the information input module, and deleting invalid sensing information from among the sensing information after a predetermined period of time; It receives valid sensing information from the information storage module, classifies the sensing information into biometric information, behavioral information, and residence status information of the monitored person using an artificial neural network, and combines abnormal information among the classified information to combine the monitored person's dangerous situation. an information analysis module for analyzing; and a result output module for outputting the risk situation analysis result derived through the information analysis module and a warning signal accordingly.

The information input module groups the detected information received from the detecting unit according to a predetermined time range and transmits the groups to the validity judgment module and the information storage module.

The validity judgment module individually analyzes a plurality of sensed values constituting the sensed information and transmits sensed time information of the corresponding sensed information to the information storage module when any one of them is abnormal.

In addition, the validity of the sensed information is first determined based on the biometric information among the biometric information, behavioral information, and residence state information of the subject to be monitored.

The information storage module transmits detection information corresponding to the detection time information received from the validity determination module to the information analysis module.

In the case of a general vertical detection sensor, it is not easy to distinguish between a situation in which the subject is simply lying on a bed or the floor and a situation in which the subject falls on the floor. As a result, an alarm may be generated even when an actual danger situation does not occur.

However, when an abnormal situation is recognized in any one of the sensing information detected through multiple sensing, the multi-sensing based vulnerable class risk perception detection system according to an embodiment of the present invention is analyzed by combining with the sensing value of the related sensing information. By doing so, a more accurate risk situation analysis can be made.

That is, according to the present invention, an alarm can be generated more accurately, and thus, an error in a crisis response situation such as an emergency dispatch can be reduced, thereby reducing unnecessary response situations or rapid response according to accurate situation judgment.

In addition, among the sensing information detected through multi-sensing, garbage information that is not valid for analysis is detected using the validity judgment module, and only the sensing information that is valid for analysis is applied to the artificial neural network, thereby increasing the reliability and processing speed of the analysis result. can improve

In addition, various effects identified directly or indirectly through the present specification may be provided.

1 is a block diagram showing the configuration of a vulnerable group risk perception detection system using multiple sensors according to an embodiment of the present invention.
2 is a diagram illustrating a state in which a monitored person is being sensed by a sensing means installed in a residence of a person to be monitored according to an embodiment of the present invention.
3 is a flowchart illustrating a detection process using a multi-sensing based risk-awareness detection system for vulnerable groups using multiple sensors according to an embodiment of the present invention.
4 is a flowchart illustrating a process of determining validity of sensing information according to an embodiment of the present invention.
5 is a flowchart illustrating a process of classifying and analyzing surveillance information according to an embodiment of the present invention and outputting a result.
6 is a graph illustrating heart rate per minute and body temperature information of a monitored person according to an embodiment of the present invention.
7 is a graph illustrating heart rate per minute and motion speed information of a monitored person according to an embodiment of the present invention.
8 is a graph illustrating heart rate per minute, movement speed, and vertical movement direction information of a monitored person according to an embodiment of the present invention.
9 is a graph illustrating heart beats per minute, left and right movement directions, and door opening/closing information of a monitored person according to an embodiment of the present invention.
10 is a graph illustrating heart rate per minute and power consumption information of a monitored person according to an embodiment of the present invention.
11 is a graph illustrating door opening and closing, power consumption, and left and right movement direction information of a monitored person according to an embodiment of the present invention.
12 is a diagram illustrating a case in which each sensed information according to an embodiment of the present invention can be combined.

Hereinafter, some embodiments of the present invention will be described in detail through exemplary drawings.

In adding reference numerals to the components of each drawing, the same numerals are used as much as possible even though they are displayed on different drawings for the same components.

FIG. 1 is a block diagram showing the configuration of a system for detecting risk perception of a vulnerable group using multiple sensors according to an embodiment of the present invention, and FIG. 2 shows a detection means installed in a residence of a subject according to an embodiment of the present invention A diagram showing a state in which the monitored person is being sensed is shown.

Referring to these drawings, the system for detecting risk awareness for vulnerable groups using multiple sensors according to an embodiment of the present invention includes a sensing means 100, a control means 200 and a central server 300.

The detecting unit 100 senses the biometric information, behavior, and residence status of the person to be monitored.

The sensing unit 100 includes a biometric information collection module 110, an action detection module 120, a door detection module 130, and a power detection module 140.

In addition, the detection means 100 is a fire detection module installed on one side of the residence to detect the state of the residence; temperature sensing module; humidity sensing module; illuminance detection module; smoke detection module; Or a gas leak detection module; may include one or more of.

In addition, the sensing means 100 may calculate sensing information linked with metadata including time information and location information.

The time information may include detection time information, day of the week information, and the like, and the location information may include residence ID information and module ID information.

The biometric information collection module 110 is attached to the body of the monitored person and collects the monitored person's biometric information, and may detect at least one of heart rate, electrocardiogram, blood pressure, body temperature, and brain wave of the monitored person.

The biometric information collection module 110 may be configured as a wearable device, and the device may have a notification cancellation button.

When the monitored person is in a dangerous situation, information on the dangerous situation and an alarm signal are transmitted to the central server 300 by the control means 200. Information about this may be transmitted to the central server 300 through the notification cancellation button.

The behavior detection module 120 includes a vertical sensor installed on one side of the person's residence to sense the person's motion in the vertical direction and a horizontal sensor to detect the motion in the left and right directions.

The vertical sensing unit and the horizontal sensing unit may each include two pyroelectric infrared sensors.

In this case, the two vertical sensing units are positioned apart from each other by a predetermined distance in the vertical direction, and the horizontal sensing units are positioned apart from each other in the left and right direction by the predetermined distance.

In addition, through the waveform generated by the behavior detection module 120, the up, down, left and right movement direction, speed, body temperature, and how far the monitored person is located can be detected through the behavior detection module 120. .

The door detection module 130 detects the opening and closing of a door installed in a residence.

The power detection module 140 detects the power consumption of the residence.

The control unit 200 includes an information input module 210, an validity judgment module 220, an information storage module 230, an information analysis module 240, and a result output module 250.

The information input module 210 receives sensing information in which biometric information, behavioral information, and residence state information of the monitored person are combined from the sensing unit 100 .

Then, the detection information is grouped and stored for each preset time range, and transmitted to the validity judgment module 220 and the information storage module 230.

The time range is 1 minute, and may be set based on heart rate information per minute of the monitored person detected by the biometric information collection module 110 .

The validity judgment module 220 judges the validity of the received sensed information, but if one or more of the biometric information, behavioral information, and residence state information of the monitored person is determined to be non-corrected information, the corresponding sensed information is analyzed. It is judged as valid sensing information.

In detail, the sensing information that is a combination of heart rate, body temperature, up and down movement direction, left and right movement direction, movement speed, location, whether the door is opened or closed, and power consumption of the monitored person transmitted from the information input module 210 is individually analyzed to detect the corresponding detection. judge the validity of the information;

In this case, the validity determination module 220 may preferentially determine the validity of the detected information based on the heart rate per minute information of the monitored person detected by the biometric information collection module 110 .

When the validity judgment module 220 determines that the received sensed information is valid for analysis, an event is generated to determine whether it is a dangerous situation. Then, event occurrence time information, which is detection time information of the corresponding detection information, is transmitted to the information storage module 230 .

The information storage module 230 stores all detection information received from the information input module 210, but the detection information that does not correspond to the event occurrence time information received from the validity determination module 220 is a separate garbage ( Garbage) information.

In addition, the garbage information is deleted after a predetermined time period, but the time period may be set to 8 hours, which is an appropriate sleep time for the monitored person.

In addition, the information storage module 230 transmits the event occurrence time information received from the validity determination module 220 and corresponding detection information to the information analysis module 240 .

The information analysis module 240 applies an artificial neural network to the sensed information transmitted from the information storage module 230, and classifies the sensed information into biometric information, behavioral information, or residence state information of the subject.

In detail, the sensing information may be classified into heart rate, vertical motion direction, left and right motion direction, motion speed, body temperature, location, whether the door is opened or closed, and power consumption of the person being monitored by the sensing means 100 .

Then, the information analysis module 240 analyzes the dangerous situation of the monitored person by combining the detected information determined as abnormal information among the classified detected information.

In this case, the information analysis module 240 may combine sensing information having a predetermined time range into one sensing information.

For example, sensing information grouped by 1 minute from the information input module 210 may be continuously combined into one sensing information having a range of 1 hour.

In this case, the information analysis module 240 may receive garbage information from the information storage module 230 in order to continuously combine detected information.

The result output module 250 transmits information as a result of the analysis of the dangerous situation of the monitored person derived through the information analysis module 240 and a corresponding alarm signal to the central server 300 .

Then, valid detection information according to the analysis result information is transmitted to the validity determination module 240 to be used as learning information for validity determination.

The central server 300 may refer to a server or a client that receives result information and an alarm signal from the control means 200 and is installed in a guardian or related institution.

3 is a flow chart illustrating a detection process using a multi-sensing based risk-awareness detection system for vulnerable groups using multiple sensors according to an embodiment of the present invention.

Looking at this, the sensing means 100 composed of one or more sensors detects the biometric information, behavior, and residence status of the person to be monitored (S100).

The information input module 210 receives sensing information in which biometric information, behavioral information, and residence state information of the monitored person are combined from the sensing unit 100 (S200).

And the information input module 210 transmits the combined sensing information to the validity judgment module 220 and the information storage module 230 (S300).

The validity determination module 220 determines the validity of the sensed information (S400).

The information storage module 230 classifies sensing information corresponding to invalid information among stored sensing information as separate garbage information, and deletes the garbage information after a predetermined time has elapsed (S500).

The information storage module 230 transmits sensing information corresponding to valid information to the information analysis module 240 (S600).

The information analysis module 240 classifies the sensed information into biological information, behavioral information, and residence state information of the monitored person using an artificial neural network, and analyzes the risk situation of the monitored person by combining abnormal information among the classified information. (S700).

The result output module 250 transmits the risk situation analysis result derived through the information analysis module 240 and a warning signal according thereto to the central server 300 (S800).

4 is a flowchart illustrating a process of determining validity of sensing information according to an embodiment of the present invention.

Looking at this, in the process (S400), the effectiveness judgment module 220 analyzes whether the heart rate per minute of the monitored person is out of the normal range (S410).

In the above process (S410), the validity judgment module 220 may analyze whether the heart rate per minute is out of the normal range according to the sex and age of the person to be monitored.

At this time, if the heart rate per minute of the person to be monitored is out of the normal range, the corresponding sensed information is determined to be valid information.

And, if the monitored person's heart rate per minute is normal in the process (S410), it is analyzed whether the monitored person's heart rate per minute is an abnormal pattern (S420).

The abnormal pattern may mean whether it has been detected 30 times or more in 10 seconds.

That is, even if the heart rate per minute of the monitored person is normal, if the detected information is detected more than 30 times in 10 seconds, the corresponding detected information is determined as valid information.

When the monitor's heart rate per minute is determined to be normal in step S410 and it is determined that no abnormal pattern is detected in the monitor's heart rate per minute in step S420, the action detection module 120 detects the It is analyzed whether the movement speed value of the monitored person is '0' for a preset time (S430), or the value representing the door opening/closing detected by the door sensing module 130 after the monitored person enters and exits a specific place is determined for a preset time. It may be analyzed whether it is '0' (S440) or whether the power consumption sensed from the power detection module 140 has not changed for a preset time or is detected as less than or equal to a preset power consumption (S450).

When any one of the values sensed by the action detection module 120, the door detection module 130, and the power detection module 140 is determined to be abnormal information, the corresponding detection information is determined to be valid information.

5 is a flowchart illustrating a process of classifying and analyzing surveillance information according to an embodiment of the present invention and outputting a result.

Looking at this, in the process (S700), the information analysis module 240 divides the sensed information transmitted from the information storage module 230 into biometric information, behavioral information, and residence state information of the subject (S710).

Then, it is analyzed whether an abnormal pattern of the heart rate per minute of the monitored person is detected more than a preset number of times during a preset time (S720).

In the process (S720), when the abnormal pattern is not detected more than a predetermined number of times, two or more of the values detected by the behavior detection module 120, the door detection module 130, and the power detection module 140 are abnormal information. Cognitive analysis is performed (S730).

If the abnormal pattern is detected more than a preset number of times in the process (S720), whether one or more of the values detected by the action detection module 120, the door detection module 130, and the power detection module 140 is abnormal information. Analyze (S740).

If two or more of the detected values are determined to be abnormal information in the process (S730), the abnormal information values are combined and analyzed (S750).

The values detected in the steps S730 and S740 mean the movement direction, movement speed, position and body temperature of the monitored person, whether the door is opened or closed, and power consumption.

In the process (S800), the result output module 250 transmits the result information analyzed by combining two or more abnormal information values and a corresponding alarm signal to the central server 300 (S810).

Alternatively, if the heart rate of the person to be monitored is determined to be abnormal information in step S740, the result output module 250 transmits information on the heart rate of the person to be monitored and an alarm signal accordingly to the central server 300. (S820).

The abnormal information on the heart rate in step S820 may mean a case where the heart rate per minute is out of the normal range or an abnormal pattern is detected more than a preset number of times during a preset time.

For example, if an abnormal pattern is detected 10 times for 10 minutes, it may be analyzed that the heart rate of the monitor is abnormal.

6 is a graph illustrating heart rate per minute (dotted line) and body temperature information (dashed line) of a person being monitored according to an embodiment of the present invention.

Looking at this, the information analysis module 240 combines the heart rate per minute and body temperature information of the monitored person, which appears to be abnormal for a predetermined time, among the classified sensed information.

The monitor's heart rate is detected as 80 times per minute or more, so it can be regarded as tachycardia, and the body temperature is detected as 40°C or more, so it can be regarded as high temperature. Therefore, it can be determined as a dangerous situation by the information analysis module 240 .

7 is a graph illustrating information on heart beats per minute (dotted line) and movement speed (dashed line) of a person being monitored according to an embodiment of the present invention.

Looking at this, the information analysis module 240 combines heart rate per minute information that appears abnormal for a predetermined time among the classified sensing information and movement speed information among information that can be combined with heart rate information per minute.

In this case, if only the movement speed information is examined, it may be determined that the monitored person is simply slow.

However, as the monitored person's movement is detected as being remarkably slow for a predetermined time period and the monitored person's heart rate is detected as less than 60 beats per minute, the monitored person's movement speed may also be determined as abnormal information.

In this case, the monitored person may be seen in a state of bradycardia, and at the same time, since the movement appears to be remarkably slow, the information analysis module 240 may determine a dangerous situation.

In addition, even when no movement appears, it may be determined as a dangerous situation by the information analysis module 240 .

FIG. 8 shows a graph illustrating information on heart beats per minute (dotted line), movement speed (dashed line), and up and down movement direction (solid line) of a person to be monitored according to an embodiment of the present invention.

Looking at this, the information analysis module 240 combines the heart rate per minute of the monitored person, which appears to be abnormal for a predetermined time, among the classified sensing information, and the movement speed and direction information among information that can be combined.

In this case, since it is seen that the monitored person moves downward rapidly and the heart rate per minute is instantaneously increased, the fall can be determined by the information analysis module 240 .

After the fall is determined, when the movement speed of the monitored person is detected to be significantly slow and the heart rate per minute is continuously detected to be out of the normal range for a predetermined time, the information analysis module 240 may determine a dangerous situation. .

FIG. 9 shows a graph illustrating information on heart beats per minute (dotted line), left and right movement directions (solid line), and door opening and closing (dashed line) of a person being monitored according to an embodiment of the present invention.

Looking at this, the information analysis module 240 combines the heart rate per minute of the monitored person and door opening/closing information and left and right movement direction information among information that can be combined with the heart rate per minute and door opening/closing information that appear to be abnormal for a predetermined time among the classified sensing information.

In this case, it is detected that the door is not reopened for a predetermined time after the person to be monitored enters a specific place such as a bathroom or a veranda, and the heart rate per minute of the person to be monitored is outside of the normal range. Therefore, the information analysis module 240 can be judged as a dangerous situation by

At this time, whether or not the monitored person enters the place can be determined through left and right movement direction information.

When it is detected that the monitored subject has not entered or entered a specific place by the behavior detection module 120, the heart rate information of the monitored subject may be analyzed in combination with information other than door opening/closing and left/right movement direction information.

10 is a graph illustrating information on heart beats per minute (dotted line) and power consumption (dashed line) of a person being monitored according to an embodiment of the present invention.

Looking at this, the information analysis module 240 combines the heart rate per minute and power consumption information of the monitored person, which appear to be abnormal for a predetermined time, among the classified sensing information.

In this case, it is detected that the power consumption is less than 60 Wh or there is no change during the preset time and the heart rate per minute of the monitored person is detected to be out of the normal range, so the information analysis module 240 may determine a dangerous situation.

In this case, when it is detected that the movement of the monitored person is significantly slow or does not move during the predetermined time period, the motion speed information of the monitored person may be combined with the heart rate per minute and power consumption information.

FIG. 11 shows a graph illustrating information about door opening and closing (dotted line), power consumption (dashed line), and left and right movement direction (solid line) of a subject according to an embodiment of the present invention.

Looking at this, the information analysis module 240 combines door opening/closing and power consumption information that appear to be abnormal for a predetermined time among the classified sensing information and left and right movement direction information of the monitored person among combinable information.

In this case, it is detected that there is no change in power consumption for a preset time and it is detected that the door is not reopened for a preset time after the monitored person enters and exits a specific place. can be judged

As shown in FIGS. 6 to 11 , the information analysis module 240 may perform analysis by combining two or more pieces of separated sensed information.

12 is a diagram illustrating a case in which each sense information according to an embodiment of the present invention can be combined.

Looking at this, the heart rate information of the monitored person may be combined with one or more of the motion direction, speed, body temperature, door opening/closing, and power consumption information of the monitored person.

Further, the door open/close information may be combined with one or more of power consumption and movement direction information.

The scope of the present invention should be construed as including all changes or modified forms derived based on the technical idea of the present invention in addition to the embodiments disclosed herein.

100: detection means
110: biometric information collection module
120: behavior detection module
130: door detection module
140: power detection module
200: control means
210: information input module
220: validity judgment module
230: information storage module
240: information analysis module
250: result output module
300: central server

Claims (5)

a sensing means for sensing the biometric information, behavior, and residence status of the person to be monitored; and
Control means for analyzing the danger situation of the person to be monitored through the detection information received from the detection means and generating an alarm signal according to the analysis result;
The control means,
an information input module that receives sensing information from the sensing means;
a validity judgment module that receives sensing information from the information input module and determines validity of the sensing information;
An information storage module for receiving and storing sensing information from the information input module and deleting invalid sensing information from among the sensing information after a predetermined period of time;
Receives valid sensing information from the information storage module, classifies the sensing information into biological information, behavioral information, and residence state information of the monitored person using an artificial neural network, and combines abnormal information among the classified information to risk the monitored person. An information analysis module for analyzing the situation;
A multi-sensing based vulnerable group risk perception detection system, characterized in that it comprises a result output module for outputting a dangerous situation analysis result derived through the information analysis module and an alarm signal accordingly. According to claim 1,
The validity judgment module individually analyzes a plurality of detection values constituting the detection information and transmits detection time information of the corresponding detection information to the information storage module when any one is abnormal;
The information storage module transmits detection information corresponding to the detection time information received from the validity determination module to the information analysis module. According to claim 1,
The information input module,
The multi-sensing based vulnerable group risk perception detection system, characterized in that the detection information received from the detection means is grouped by a preset time range and transmitted to the validity determination module and the information storage module. According to claim 1,
The validity judgment module,
A multi-sensing based vulnerable class risk perception detection system, characterized in that the validity of the sensing information is first determined based on the biometric information of the monitored person's biometric information, behavioral information, and residence state information. According to claim 1,
The detection means is
A biometric information collection module for detecting the biometric information of the subject;
A behavior detection module for detecting the behavior of the subject;
A door detection module for detecting whether the door is opened or closed in the residence;
A risk-awareness detection system for the vulnerable class based on multi-sensing including a power detection module that detects power consumption in residence.

Images