KR20210049467A - Fall detection method using smart terminal - Google Patents

Abstract

A fall detection method using a smart terminal is disclosed. The fall detection method using the smart terminal according to the present invention comprises the steps of: (a) sensing an acceleration pattern due to the movement of a user by extracting an acceleration value and a linear acceleration value using a three-axis acceleration sensor of the smart terminal and storing the extracted linear acceleration value in a buffer for a set period; (b) after analyzing the acceleration pattern in the step (a) and dividing a section suspected of the fall into a plurality of sections, determining whether all of a plurality of preset fall determination conditions are satisfied in the plurality of sections respectively; (c) sending a fall warning to the outside through the smart terminal when all of the fall determination conditions are satisfied in the section suspected of the fall through the step (b); and (d) transmitting fall occurrence information to an emergency contact prestored in the smart terminal in the case that a fall warning cancellation signal is not input for a set period by the user. According to the present invention, the pattern of acceleration according to the movement of the user sensed by the smart terminal is analyzed over time, and the section suspected of the fall is divided into three sections (segments) of 'prior-impact section', 'impact section' and 'post-impact section' such that it is determined whether each condition for the fall detection is satisfied, and if all are satisfied, it is finally determined as a fall, thereby reducing wrong fall detection probability.

Description

Fall detection method using smart terminal

In the present invention, even if the sampling period of the sensor increases, not only can a fall be efficiently detected, but also an acceleration pattern is analyzed, but after dividing the section suspected of falling into a plurality of sections, each condition for determining the fall is applied to all conditions. When this is satisfied, it relates to a fall detection method using a smart terminal that can reduce the probability of false fall detection by final determination as a fall.

Falls, which increase exponentially, are recognized as a major factor causing physical, mental, and economic problems in an increasing aging population worldwide.In the age of aging, the safety of the elderly and the safety of workers in construction sites and factories are highlighted. have.

In particular, when a person falls, that is, falls from a high place, or is found late, life problems occur. Therefore, it is required to detect a fall in the event of a fall and notify the surroundings with an alarm and vibration, and conventionally, a technology that determines a fall when a processed value such as measured acceleration by setting a threshold for fall detection exceeds a standard threshold was used. .

However, this technology has a problem that it consumes a lot of battery when it is applied to a smartphone because it senses a lot at the same time because the sampling cycle of the sensor is short, and when the acceleration is sensed using a smartphone, the acceleration value is temporarily referenced. If the threshold is exceeded, it is detected as a fall, so the probability of occurrence of a false fall is inevitably increased.

Korean Registered Patent Publication No. 10-1178936 (registered on August 27, 2012)

The present invention was devised to solve the above-described conventional problem, and even if the sampling period of the sensor increases, not only can a fall be efficiently detected, but also an acceleration pattern is analyzed, but the section in which a fall is suspected is divided into a plurality of sections. The object of the invention is to provide a fall detection method using a smart terminal that can reduce the probability of false fall detection by applying conditions for each fall determination, and finally determining the fall as if all the conditions are satisfied.

According to an aspect of the present invention, (a) an acceleration value and a linear acceleration value are extracted using a three-axis acceleration sensor of a smart terminal, and the extracted linear acceleration value is stored in a buffer for a set period of time, while an acceleration pattern due to user movement is stored. Sensing; (b) analyzing the acceleration pattern in step (a), dividing the suspected fall section into a plurality of sections, and determining whether all of the plurality of predetermined fall judgment conditions are satisfied in each of the plurality of sections; (c) sending a fall warning to the outside through the smart terminal when all fall determination conditions are satisfied in the suspected fall section through step (b); And (d) a fall detection method using a smart terminal comprising the step of transmitting the fall occurrence information to the emergency contact previously stored in the smart terminal when the fall alarm cancellation signal is not input by the user during the set time is provided.

The suspected fall section in step (b) includes a post-impact section, an impact section, and a pre-impact section sequentially divided from a current point in time when the acceleration pattern is sensed to a past point in time, and the post-impact section includes a user's movement. As a non-occurrence section, it is a section from the current point to the most recent point of motion judgment in which motion is determined within the suspected fall section, and the impact section is the maximum linear acceleration in a section within a preset maximum time for impact detection from the latest point of motion judgment. It is a section up to the point of occurrence of the value, and the pre-impact section is a section from the point of occurrence of the maximum linear acceleration value to the point at which the preset minimum acceleration detection maximum time is subtracted. A first fall determination condition to a third fall determination condition are provided as preset fall determination conditions for determining the fall of the child. In step (c), when all the first to third fall determination conditions are satisfied, a fall alarm is externally applied. Can be shipped to.

The step (b) includes the steps of: (b1) determining whether the linear acceleration value of the current viewpoint stored in the buffer is less than the motion reference acceleration value for determining the occurrence of the user's movement; (b2) When it is determined that the motion reference acceleration value is less than the motion reference acceleration value in step (b1), the linear acceleration value stored in the buffer from the current point in time to the past point is scanned to determine whether the linear acceleration value is equal to or greater than the motion reference acceleration value. step; (b3) Determining the movement in which the linear acceleration value is detected as the movement reference acceleration value or more through the step (b2) If the no-motion time between the latest point and the current point is more than the minimum time in the post-impact section, the first fall Determining that the determination condition is satisfied, and determining a period after the impact between the latest time point of the motion determination and the current time point; (b4) determining that the second fall determination condition is satisfied if the maximum linear acceleration value scanned within a preset maximum impact detection time range from the latest motion determination point is equal to or greater than a preset fall reference maximum acceleration value; (b5) determining, as an impact section, between the time point at which the maximum linear acceleration value in step (b4) occurs and the most recent time point for determining the motion; (b6) determining a pre-impact section between the time when the maximum linear acceleration value is generated and a preset minimum acceleration detection maximum time; And (b7) determining that the third fall determination condition is satisfied if the minimum linear acceleration value scanned within the pre-impact section is less than a preset fall reference minimum acceleration value.

According to the method for detecting a fall using the smart terminal of the present invention described above, the pattern of acceleration according to the user's movement sensed by the smart terminal is analyzed over time, and the section of the heart of the fall is'pre-shock section' and'impact section'. , It is possible to reduce the probability of false fall detection by dividing it into three segments of the'post-impact section' and determining whether the conditions for fall detection are satisfied, and finally determining the fall as if all are satisfied.

1 and 2 are flow charts showing a fall detection method using a smart terminal according to an embodiment of the present invention,
3 is a graph showing a state in which a fall heart section is divided according to an acceleration pattern analysis in a fall detection method using a smart terminal according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in more detail with reference to the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but will be implemented in a variety of different forms, only the present embodiments make the disclosure of the present invention complete, and the scope of the invention to those of ordinary skill in the art It is provided to inform you. In the drawings, the same reference numerals refer to the same elements.

A fall detection method using a smart terminal according to a preferred embodiment of the present invention analyzes the pattern of acceleration according to the user's movement sensed by the smart terminal over time, and analyzes the fall heart section as a'pre-shock section' and a'shock section'. , It is possible to reduce the probability of false fall detection by dividing it into three segments of the'post-impact section' and determining whether the conditions for fall detection are satisfied, and finally determining the fall as if all are satisfied.

Hereinafter, the present invention will be described in detail with reference to Examples.

1 and 2, a fall detection method using a smart terminal according to an embodiment of the present invention (hereinafter,'fall detection method') is largely an acceleration pattern sensing step (S100), a fall determination step (S200). , And a fall alarm sending step (S300) and a fall occurrence information transmitting step (S400).

First, an acceleration value and a linear acceleration value are extracted using a three-axis acceleration sensor of a smart terminal, and the extracted linear acceleration value is stored in a buffer for a set time while sensing an acceleration pattern due to a user's movement (S100).

Specifically, the three-axis acceleration values of the x, y, and z axes according to the user's movement are sensed in a sensing period (20 Hz or less) with low battery consumption through the x, y, z axis acceleration sensor of the smart terminal (S110).

Subsequently, a 3-axis acceleration value is extracted through a 3-axis acceleration sensor, and a linear acceleration value is extracted through this (S120). Thereafter, a linear acceleration value is stored in a buffer (not shown) for about several tens of seconds as a reference data value for determining and analyzing a fall, which will be described later (S130).

The relationship between the 3-axis acceleration value and the linear acceleration value can be summarized through the following equation.

: i번째 x,y,z 성분 가속도here,

: i-th x,y,z component acceleration

: i번째 가속도 값

: ith acceleration value

: i번째 선형 가속도 값

: i-th linear acceleration value

: i번째 구간 평균 가속도 크기

: Average acceleration size of the i-th section

: i번째 하이패스 필터 평균

: Average of the i-th high pass filter

: i번째 하이패스 필터 민감도 계수

: ith high pass filter sensitivity coefficient

) 을 하이패스 필터링하여 선형가속도 값(

)을 사용하며, 이는 가속도 센서의 편향 오차를 감소시키고 미동 감지를 위한 것이다.That is, in step S100, the acceleration magnitude value (

) By filtering the high pass to the linear acceleration value (

) Is used, which reduces the deflection error of the acceleration sensor and is for fine motion detection.

Next, as shown in Figs. 1 and 2, after analyzing the acceleration pattern in step S100 and dividing the suspected fall section into a plurality of sections, whether all of the plurality of predetermined fall judgment conditions are satisfied in each of the plurality of sections. It is determined (S200).

In the present invention, as shown in Fig. 3, the suspected fall section is sequentially divided into a plurality of sections from the current point in time when the acceleration pattern is sensed to the past point in time, and specifically, the post-impact section, the impact section, and the pre-impact section are Includes. In FIG. 3, the x-axis unit is sec (time), and the y-axis unit is m/sec ² (acceleration).

The post-impact section is the section in which the user does not move due to the impact after a fall.The acceleration pattern is analyzed in the reverse direction of time from the current point in time when the acceleration pattern is measured in the suspected fall section, and the motion is determined from the current point in time. It is a section. This section is a section in which there is no motion after a large impact, and whether or not the range of the linear acceleration value is within a predetermined range can be used as one of the reference values for determining a fall. Details will be described later.

The impact section is a section within a certain period of time (approximately 1 to 2 seconds) from the start point of the post-impact section (the most recent time point of motion determination), specifically, the linear acceleration value stored in the buffer is scanned within the section within the preset maximum impact detection time. After the maximum linear acceleration value is calculated, it is the section from the start point of the post-impact section to the point at which the maximum linear acceleration value occurs. This section is a section in which a large impact is substantially applied to the user by the fall, and the maximum linear acceleration value can be used as one of the reference values for determining the fall, and details will be described later.

The pre-impact section refers to a section within a certain time (approximately 1 to 2 seconds) from the start of the impact section, that is, the section from the time when the maximum linear acceleration value occurs to the time when the preset minimum acceleration detection maximum time is subtracted. The lowest linear acceleration value within this section is used as one of the reference values for determining a fall.

In the present invention, as a fall determination condition for determining a fall in the post-impact section, the impact section, and the pre-impact section, the first to the third fall judgment conditions are provided, and in step S300 described later, the first to the third fall judgment If all of the conditions are satisfied, it is determined that a fall has occurred and a fall warning is sent to the outside.

Hereinafter, step S200 (fall determination step) will be described in detail with reference to a flow chart.

As shown in FIG. 2, the step S200 includes detecting whether a motion has occurred (S201), searching for a non-motion section (S202), detecting whether a motion has occurred additionally (S203), determining a section after an impact (S204), and determining a section after the impact (S203). S205), determining the impact section characteristics (S206), determining the impact section (S207), determining the pre-impact section (S208), and determining the pre-impact section characteristics (S209).

First, it is determined whether or not there is a movement of the user by using the linear acceleration value of the current viewpoint stored in the buffer (S201).

)이 사용자의 움직임 발생 판단을 위한 움직임 기준가속도 값(

) 미만인지 여부를 판단하고, 미만인 것으로 확인되면 낙상후 충격에 의해 움직임이 없는 것으로 판단한다.Specifically, the linear acceleration value at the current point in time stored in the buffer (

) Is the motion reference acceleration value (

) Or not, and if it is found to be less than, it is judged that there is no movement due to the impact after the fall.

In step S201, if the linear acceleration value at the current point of time is determined to be greater than or equal to the motion reference acceleration value for determining the occurrence of the user's motion, it is determined that there is motion, the subsequent fall detection process is stopped, and step S100 is repeated.

)은 실험이나 다양한 운동 데이터를 이용한 기계 학습 등을 이용하여 기설정할 수 있다.Here, the motion reference acceleration value (

) Can be preset using an experiment or machine learning using various exercise data.

Next, the linear acceleration value stored in the buffer from the current point of time to the past point of time is scanned to search for a motion-free section (S202).

)이 사용자의 움직임 발생 판단을 위한 움직임 기준가속도 값(

) 이상인 것으로 확인되면 움직임이 발생한 것으로 판단한다.Next, while scanning the linear acceleration value stored in the buffer, it is additionally detected whether or not the user's movement occurs (S203). Specifically, the linear acceleration value detected while scanning (

) Is the motion reference acceleration value (

) If it is found to be abnormal, it is determined that movement has occurred.

)이 움직임 기준가속도 값(

) 이상으로 감지된 움직임 판단 최근시점(

)과 현재 시점(

) 사이의 움직임 미발생 시간(

)이 충격후 구간 최소시간(

) 이상인지 여부 확인을 통해 충격후 구간을 판단하고, 이상인 경우 제1 낙상판단조건을 만족하는 것으로 판단한다(S204).Next, as shown in FIG. 2, the linear acceleration value (

) Is the motion reference acceleration value (

) Or more detected motion judgment recent time point (

) And the current point in time (

) No movement time between (

) Is the minimum time after impact (

) The section after the impact is determined by checking whether it is abnormal, and if it is abnormal, it is determined that the first fall judgment condition is satisfied (S204).

)이 충격후 구간 최소시간(

) 미만인 경우, 이후의 낙상 검지 프로세스를 중지하고 S100 단계를 반복 실시한다.As a result of the determination in step S204, the time when no movement occurs (

) Is the minimum time after impact (

), the subsequent fall detection process is stopped and step S100 is repeated.

)은 실험이나 다양한 운동 데이터를 이용한 기계 학습 등을 이용하여 기설정할 수 있다.Here, the minimum time for the section after impact (

) Can be preset using an experiment or machine learning using various exercise data.

)이 충격후 구간 최소시간(

) 이상인 경우, 움직임 판단 최근시점(

)과 현재 시점(

) 사이의 구간을 충격후 구간으로 결정한다(S205).Next, as a result of the determination in step S204, the time when no movement occurs (

) Is the minimum time after impact (

) Or more, the latest point of motion judgment (

) And the current point in time (

) Is determined as the post-impact section (S205).

)으로부터 기설정된 충격감지 최대시간(

) 범위 이내에서 선형 가속도 값을 스캔하고 스캔된 값 중 최대 선형 가속도값(

)이 기설정된 낙상기준 최대 가속도값(

) 이상이면 제2 낙상판단조건을 만족하는 것으로 판단함으로써 충격 구간 특성이 있는지 여부를 판단한다(S206).Next, the latest point of motion judgment (

Maximum shock detection time preset from) (

) The linear acceleration value is scanned within the range, and the maximum linear acceleration value (

) Is the preset fall standard maximum acceleration value (

), it is determined whether or not there is an impact section characteristic by determining that the second fall judgment condition is satisfied (S206).

As a result of the determination in step S206, if it is not determined that there is an impact section characteristic, the subsequent fall detection process is stopped and step S100 is repeated.

)과 낙상기준 최대 가속도값(

)은 실험이나 다양한 운동 데이터를 이용한 기계 학습 등을 이용하여 기설정할 수 있다.Here, the maximum shock detection time (

) And the maximum acceleration value (

) Can be preset using an experiment or machine learning using various exercise data.

)이 기설정된 낙상기준 최대 가속도값(

) 이상인 것으로 판단되면, 최대 선형 가속도값(

)이 발생한 시점과 움직임 판단 최근시점(

) 사이를 충격 구간으로 결정한다(S207).Next, as a result of the determination in step S206, the maximum linear acceleration value (

) Is the preset fall standard maximum acceleration value (

) Or higher, the maximum linear acceleration value (

) Occurred and the latest time to determine the movement (

) Is determined as the impact section (S207).

)과 기설정된 최저 가속도 감지 최대 시간(

) 사이를 충격전 구간으로 결정한다(S208).Next, when the maximum linear acceleration value occurs (

) And the preset minimum acceleration detection maximum time (

) Is determined as the pre-impact section (S208).

)은 최대 선형 가속도값 발생 시점(

)에서 미리 설정한 최저 가속도 감지 최대 시간(

)을 뺀 시점(

)으로 결정된다.In other words, the start point of the mouth angle before impact (

) Is the point at which the maximum linear acceleration value occurs (

The maximum time to detect the minimum acceleration preset in) (

) Minus the time point (

).

)은 실험이나 다양한 운동 데이터를 이용한 기계 학습 등을 이용하여 기설정할 수 있다.Here, the minimum acceleration detection maximum time (

) Can be preset using an experiment or machine learning using various exercise data.

이 기설정된 낙상기준 최저 가속도값(

) 미만이면 충격전 구간 특성 조건으로서 제3 낙상판단조건을 만족하는 것으로 판단한다(S209).Next, the lowest linear acceleration value scanned within the pre-impact section (

This preset fall standard minimum acceleration value (

If it is less than ), it is determined that the third fall judgment condition is satisfied as a pre-impact section characteristic condition (S209).

As a result of the determination in step S209, if it is not determined that there is a characteristic of the pre-impact section, the subsequent fall detection process is stopped and step S100 is repeated.

)은 실험이나 다양한 운동 데이터를 이용한 기계 학습 등을 이용하여 기설정할 수 있다.Here, the fall criterion minimum acceleration value (

) Can be preset using an experiment or machine learning using various exercise data.

In the present invention, if all of the first fall judgment condition in steps S203 and S204, the second fall judgment condition in step S206, and the third fall judgment condition in step S209 are satisfied in sequence, the user is finally determined to have fallen, As such, it is possible to reduce the occurrence of fall detection errors by judging through a plurality of judgment procedures.

Next, as shown in FIG. 2, when all fall determination conditions (first to third fall determination conditions) are satisfied in a suspected fall section through step S200, a fall alarm is sent to the outside through a smart terminal (S300). ). For example, an alarm sound or an alarm vibration may be generated through a smartphone, and a third party in the vicinity may receive such alarm information and proceed with rapid processing for the safety of a person who has fallen.

Next, as shown in FIG. 2, when the fall warning cancellation signal by the user is not input to the smart terminal during the set time, the fall occurrence information is transmitted to the emergency contact previously stored in the smart terminal (S400).

In step S400, a fall alarm cancellation button may be displayed on the smart terminal in the form of a pop-up message, and if a button click signal for canceling the fall alarm is input to the smart terminal, it is determined that the user can move sufficiently even though a fall has occurred. And repeat step S100.

In step S400, a text message is automatically sent to an emergency contact network previously stored in a smart terminal such as a family member, an 119 emergency center, an acquaintance, etc., and a voice call may be made.

Although the present invention has been described with reference to the accompanying drawings and the above-described preferred embodiments, the present invention is not limited thereto, but is limited by the claims to be described later. Therefore, those of ordinary skill in the art can variously modify and modify the present invention within the scope of the technical spirit of the claims to be described later.

S100: Acceleration pattern sensing step
S200: Fall judgment stage
S300: Fall Alert Sending Stage
S400: Fall occurrence information transmission step

Claims (3)

(a) extracting an acceleration value and a linear acceleration value using a 3-axis acceleration sensor of a smart terminal, storing the extracted linear acceleration value in a buffer for a set time, and sensing an acceleration pattern due to user movement;
(b) analyzing the acceleration pattern of the step (a), dividing the suspected fall section into a plurality of sections, and determining whether all of the plurality of predetermined fall judgment conditions are satisfied in each of the plurality of sections;
(c) sending a fall warning to the outside through the smart terminal when all fall determination conditions are satisfied in the suspected fall section through step (b); And
(d) Fall detection method using a smart terminal comprising the step of transmitting the fall occurrence information to an emergency contact previously stored in the smart terminal when the fall alarm cancellation signal is not input by the user during the set time. The method of claim 1,
The suspected fall section of step (b) includes a post-impact section, an impact section, and a pre-impact section sequentially divided from a current point in time at which the acceleration pattern is sensed to a past point in time,
The post-impact section is a section in which the user's motion does not occur, and is a section from the current point in time to the most recent point in motion judgment in which motion is determined within the suspected fall section
The impact section is a section from the latest time point of the motion determination to the time point at which the maximum linear acceleration value occurs in a section within a preset maximum time for impact detection,
The pre-impact section is a section from the time when the maximum linear acceleration value occurs to the time when the preset minimum acceleration detection maximum time is subtracted,
A first fall determination condition to a third fall determination condition are provided as preset fall determination conditions for determining a fall in the post-impact section, the impact section, and the pre-impact section, and the step (c) includes the first to third fall judgment conditions. Fall detection method using a smart terminal, characterized in that when all fall judgment conditions are satisfied, a fall alarm is sent to the outside. The method of claim 2,
The step (b),
(b1) determining whether the linear acceleration value of the current viewpoint stored in the buffer is less than a motion reference acceleration value for determining the occurrence of the user's motion;
(b2) When it is determined that the motion reference acceleration value is less than the motion reference acceleration value in step (b1), the linear acceleration value stored in the buffer from the current point in time to the past point is scanned to determine whether the linear acceleration value is equal to or greater than the motion reference acceleration value. step;
(b3) Determining a motion in which the linear acceleration value is detected as the motion reference acceleration value or more through the step (b2) If the no-motion time between the latest point and the current point is greater than the minimum time in the post-impact section, the first fall Determining that the determination condition is satisfied, and determining a period after the impact between the latest time point of the motion determination and the current time point;
(b4) determining that the second fall determination condition is satisfied if the maximum linear acceleration value scanned within a preset maximum impact detection time range from the latest motion determination point is equal to or greater than a preset fall reference maximum acceleration value;
(b5) determining, as an impact section, between the time point at which the maximum linear acceleration value in step (b4) occurs and the most recent time point for determining the motion;
(b6) determining a pre-impact section between the time when the maximum linear acceleration value is generated and a preset minimum acceleration detection maximum time; And
(b7) Fall using a smart terminal comprising the step of determining that the third fall determination condition is satisfied if the minimum linear acceleration value scanned within the pre-impact section is less than a preset fall standard minimum acceleration value. Detection method.

Images