KR101797854B1 - System and method for fall detection using smart band - Google Patents

Abstract

A method and system for determining a fall using a smart band is disclosed. The fall determining method includes the steps of: receiving inertia information according to a movement of a pedestrian from a wearable device worn by a pedestrian; calculating a SVM (Signal Vector Magnitude) value based on the inertia information; Determining whether or not the pedestrian has fallen based on the pattern information based on the Fourier transformed information and the pattern information indicating the movement of the predetermined pedestrian can do.

Description

TECHNICAL FIELD [0001] The present invention relates to a method and system for determining a fall using a smart band,

Embodiments of the present invention relate to a method and system for determining whether a wearer falls or not by using a wearable device.

More particularly, the present invention relates to a technique for determining whether or not a pedestrian has fallen by using information sensed by a wearable device worn on the body of the pedestrian.

Due to the recent increase in average life expectancy and declining fertility rate, aging has become a social problem all over the world. According to the "Future Population Estimation" data of the Korea National Statistical Portal, the proportion of elderly people aged 65 or older in Korea is 12.7% as of 2014. In Korea, it has already entered the "aging society" in 2000, and it is expected that the proportion of elderly people will reach "aged society" with 14.3% in 2018 and 20.8% in 2026 . Elderly people are exposed to various environmental risks. Especially, there is a higher risk of accidents than other age groups, and in case of an accident, sudden physical disability and health deterioration are accompanied by disabilities in daily life. The proportion of these falls in elderly persons is increasing with age. According to a survey conducted by the research institute, about 28-30% of elderly people aged 65 or older, 35% elderly people aged 70 or older, 32 ~ 42% elderly people aged 75 or older, and an annual fall rate of 50% .

For the elderly, falls cause psychological damage due to social psychological dysfunction as well as physical damage such as immaturity. And in the worst case, fall complications can lead to death. There are two major ways to detect fatal fall in elderly people. One method is to detect the presence or absence of fall using image information, and the other method is to attach a sensor to a fallen pedestrian to detect the presence or absence of a fall. In the case of the method of detecting fall using the image information, there is a problem of invasion of privacy because it uses the image information of the pedestrian.

In the method of detecting the presence or absence of a fall using the sensor, the presence or absence of a fall is detected by monitoring various posture changes and activity of the pedestrian while the sensor is attached to a specific body part such as a waist, a chest, and a wrist. It is preferred that the sensor is attached to the wrist in the form of a wristwatch for the convenience of wearing the pedestrian. However, in the case where the sensor is attached to the wrist, there is a problem that the fall detection rate is very low due to the free movement of the wrist. In other words, it is often detected as a fall even if it is not a fall.

Accordingly, there is a need for a technique for increasing the accuracy of determining the presence or absence of a fall while providing a device for sensing falls with a small wrist band type such as a wristwatch to reduce or minimize the user's discomfort.

 Korean Patent Laid-Open No. 10-2012-0114629 discloses a technique of setting a safe area in which a patient should be positioned within a background image, comparing the background image with an image input in real time, / RTI >

[1] M.J. Mathine, A. C. F. Coster, N. H. Lovell, and B. G. Celler., "A pilot study of human movement in the home using accelerometry". J. Telemed, Telecare, vol. 10 pp. 144-151, 2004. [2] D. M. Karantonis, M. R. Narayanan, M. Mathie, N. H. Lovell, Branko G. Celler, "Implementation of a Real-Time Human Movement Classifier Using a Triaxial Accelerometer for Ambulatory Monitoring" IEEE Trans. on information technology in Biomedical, vol. 10, no.1, January 2006. [3] A.M. Turing (1950), Computing Machinary and Intelligence, Mind 49: 433-460.

The present invention is intended to improve the accuracy of fall detection while manufacturing a wearable device equipped with a sensor for sensing movement, movement, or attitude change of a pedestrian in a manner minimizing inconvenience to the wearer.

In the present invention, even if the wearable device is manufactured as a wrist band type, the presence or absence of a fall is determined once again using the pattern matching for the fall result detected using the information sensed by the sensor, .

In the present invention, when a fall message is determined by using a push message function provided by a cloud server, information (a message or a message) indicating a fall condition of a pedestrian to a pedestrian's guardian, emergency agency, Emergency call, etc.).

The fall determining method includes the steps of: receiving inertia information according to a movement of a pedestrian from a wearable device worn by a pedestrian; calculating a SVM (Signal Vector Magnitude) value based on the inertia information; Determining whether or not the pedestrian has fallen based on the pattern information based on the Fourier transformed information and the pattern information indicating the movement of the predetermined pedestrian can do.

According to an aspect, the pattern information indicating the movement of the pedestrian may include a plurality of patterns classified into different specific frequency bands according to the movement of the pedestrian.

According to another aspect of the present invention, the step of performing the Fourier transform includes a step of primarily determining whether a fall of the pedestrian is based on the calculated SVM value and a predetermined threshold value, And performing a Fourier transform on the SVM values for the predetermined time period.

According to another aspect, the step of determining whether or not the pedestrian falls may re-determine whether or not the pedestrian primarily determined based on the pattern information falls.

According to another aspect, the step of calculating the SVM value includes the steps of storing the SVM value calculated based on the inertia information received from the wearable device at a predetermined reference time in a buffer, and performing the Fourier transform May perform Fourier transform on the SVM values corresponding to the predetermined sampling size among the SVM values stored in the buffer.

According to another aspect, when it is determined that the pedestrian has fallen, information informing the fall of the pedestrian may be provided to at least one of the guardian terminal, emergency terminal, and emergency agent terminal previously registered through the cloud push server.

According to another aspect, in the case where it is determined that the pedestrian has fallen, a push message informing the fall of the pedestrian may be transmitted to at least one of the registered guardian terminal, the emergency terminal, and the emergency agent terminal.

According to another aspect of the present invention, the wearable device may include an inertia measurement sensor that is manufactured as a wrist band type, is worn by the pedestrian, and measures a motion of the pedestrian.

The fall determination system includes an information receiving unit that receives inertia information based on the movement of a pedestrian from a wearable device worn by a pedestrian, a calculation unit that calculates a SVM (Signal Vector Magnitude) value based on the inertia information, And a fall determining unit for determining whether or not the pedestrian has fallen based on the pattern information based on the Fourier transformed information and the pattern information indicating the movement of the predetermined pedestrian, .

According to an aspect of the present invention, the pattern information indicating the movement of the pedestrian may include a plurality of patterns that are divided into different specific frequency bands according to the movement of the pedestrian.

According to another aspect of the present invention, the Fourier transform unit primarily determines whether a fall of the pedestrian is based on the calculated SVM value and a predetermined threshold value, and when the pedestrian is primarily determined that the pedestrian has fallen, The Fourier transform can be performed on the SVM values for a predetermined time.

According to another aspect, when it is determined that the pedestrian has fallen, information informing the fall of the pedestrian may be provided to at least one of the guardian terminal, emergency terminal, and emergency agent terminal previously registered through the cloud server.

According to another aspect of the present invention, the wearable device is manufactured as a wrist band type, is worn by the pedestrian, and can measure a movement of the pedestrian according to the movement of the pedestrian.

According to another aspect of the present invention, the wearable device includes an inertia measurement sensor for generating inertia information by detecting movement of a pedestrian, a wireless communication module for transmitting the inertia information to the fall determination system to determine whether the pedestrian is falling, And a power supply unit for supplying power to the inertial measurement sensor and the short-range communication module.

According to another aspect, in the case where it is determined that the pedestrian has fallen, a push message informing the fall of the pedestrian may be transmitted to at least one of the registered guardian terminal, the emergency terminal, and the emergency agent terminal.

According to the present invention, a wearable device equipped with a sensor for sensing movement, movement, or attitude change of a pedestrian is manufactured in a form minimizing the inconvenience of a pedestrian, It is possible to increase the accuracy of the presence or absence of fall finally determined by re-determining the presence or absence of fall once again using the pattern matching with respect to the fall result judged on the basis.

In addition, when a fall is judged by using a push message function provided by the cloud server, information (a message or an emergency call, etc.) informing the pedestrian's fallen state to the guardian of the pedestrian, the emergency institution, So that emergency measures due to falls can be performed quickly.

FIG. 1 is a diagram showing a configuration of an entire network for determining whether or not a pedestrian has fallen, according to an embodiment of the present invention.
2 is a diagram showing a configuration of a wearable device in an embodiment of the present invention.
3 is a block diagram illustrating an internal configuration of a fall determination system in an embodiment of the present invention.
4 is a flowchart showing a fall determination method in an embodiment of the present invention.
5 is a view showing a screen displayed on the fall determination system 300 according to an embodiment of the present invention.
FIG. 6 is a diagram for explaining an operation in which an FFT transform is performed to determine whether or not to fall secondarily in an embodiment of the present invention. FIG.
FIG. 7 is a diagram for explaining an operation of determining whether a fall occurs using pattern matching based on an FFT in an embodiment of the present invention. Referring to FIG.
8 is a view for explaining an operation of providing information informing occurrence of a fall of a pedestrian in cooperation with a cloud server in an embodiment of the present invention.

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

The present invention relates to a technique for determining the presence or absence of a fall of a pedestrian by using inertia information sensed and received from a wearable device worn by a pedestrian such as an elderly person with a frequent fall or a leg-incompetent person. Particularly, the present invention relates to a technology for re-determining whether or not a fall occurs by using pattern matching when judging a fall by using inertia information instead of simply determining inertia information using only inertia information. For example, the present invention relates to a technique of once again determining whether or not a fall occurs by using a characteristic having a specific frequency band according to a posture of a pedestrian, such as running, stepping, stepping, falling,

In the present embodiments, a pattern may be defined as a collection of features or features of individual objects. A feature can be defined as an aspect, a property, or a characteristic capable of discriminating an object that is a target of pattern recognition. That is, pattern matching can mean one of the problems in cognitive science and artificial intelligence that finds a specific object with a characteristic. For example, it may be determined whether a fall has occurred to a pedestrian using features (frequency bands) that represent the fall.

In the present embodiments, the operation of determining whether or not a pedestrian has fallen through information received from the wearable device and pattern matching is performed by a user terminal such as a smart phone, a personal computer (PC), a tablet PC And may be implemented in the form of a service application and may be executed in a user terminal.

FIG. 1 is a diagram showing a configuration of an entire network for determining whether or not a pedestrian has fallen, according to an embodiment of the present invention.

1 may include a wearable device 110, a fall determination system 120, a server 130, a cloud server 140, and a terminal 150 of an emergency agent, guardian, and paramedics previously registered.

The wearable device 110 includes an inertial measurement unit (IMU) for sensing movement, movement, posture change, etc. of the pedestrian 101, and transmits inertia information sensed by the inertial measurement sensor to the wireless communication module To the fall determination system 120. For example, inertia information sensed through a short-range wireless communication module such as a Bluetooth module may be transmitted to the fall determination system 120.

The fall determination system 120 receives inertia information from the wearable device 110 for each predetermined reference time, temporarily stores the inertia information in a storage device (not shown) such as a buffer, It is possible to judge whether or not to fall. If it is determined that the fall is occurred, the fall determination system 120 can re-determine whether or not the fall occurs by using the pattern matching.

At this time, if the second determination result is also a fall, the fall determination system 120 may transmit information indicating that the pedestrian 101 has fallen to the server 130. For example, identification information (e.g., pedestrian ID, telephone number, terminal unique number, etc.) of the pedestrian registered in the server 130 and a flag flag and the like may be transmitted to the server 130 together.

The server 130 may manage and monitor at least one fall determination system in which a service application for determining a fall is installed and driven. For example, the server 130 may store and manage the identification information of fall determination systems operated by the service application in association with the phone number of the guardian, the telephone number of the emergency agency, and the telephone number information of the emergency unit.

When a fall occurs to the pedestrian 101, the server 130 interlocks with the cloud push server 140 provided by Google, Naver, etc., and transmits a push message informing fall through the push server 140 in advance To the terminal 150 of the guardian, the emergency institution, and the emergency unit. Then, the terminal 150 of the guardian, the emergency institution, and the rescue party receiving the push message informing the fall responds to the fall determination system 120 carried by the pedestrian 101 and calls the pedestrian 101 or people located around the pedestrian It is possible to check the fall situation of the pedestrian 101 and to cause the pedestrian 101 to take emergency measures.

2 is a diagram showing a configuration of a wearable device in an embodiment of the present invention.

2, the wearable device 110 is an example of a wristwatch-type band type, which corresponds to the embodiment. The wearable device 110 may include an ankle band type, a waistband And the like.

Referring to FIG. 2, the wearable device 110 may be manufactured as a band type worn on the wrist. That is, the wearable device 110 may be manufactured in the form of a wristwatch, and may include a short range wireless communication module 111, an inertial measurement sensor 112, and a power supply unit 113 for supplying power.

For example, the short-range wireless communication module 111 may include a wireless communication module such as Zigbee, Bluetooth, and NFC, and the inertial measurement sensor 112 may include EBIMU-9DOFV2 of E2BOX. The inertial measurement sensor 112 such as the EBIMU-9DOFV2 may correspond to a small AHRS (Attitude Heading Reference System) module having a three-axis acceleration sensor, a three-axis gyroscope sensor, and a three-axis geomagnetic sensor. The acceleration sensor supports data update and output speeds up to 1000 Hz and can output the pure acceleration value from which the gravity component is removed. In the data output mode, ASCII output mode, Hex (binary) output mode, and Polling output mode are supported. Hereinafter, an ASCII output mode is used as an example as shown in Table 1 below . The sensitivity of the acceleration sensor by gravity acceleration is 2000 Dps, and the output rate can be set to be 50 times per second.

SOL DATA1 SP DATA2 SP ... SP DATEN EOL * ascii data 1 , ascii data 2 , ... , ascii data n CR LF

The power supply unit 113 may include a lithium ion polymer battery. The operating power of the three-axis acceleration sensor is 4.5 V supplied from the battery. In order to remove the noise included in the noise output from the acceleration sensor, a digital low pass filter (hereinafter referred to as " May be included. For example, the digital LPF can only pass signals in the 50 to 256 Hz band. Since the size of the EBIMU-9DOFV2 module used as the inertial measurement sensor 112 is very small, it can be easily manufactured as a wearable device for sensing the movement of the pedestrian occupying a small space.

FIG. 3 is a block diagram illustrating an internal configuration of a fall determination system in an embodiment of the present invention, and FIG. 4 is a flowchart illustrating a fall determination method in an embodiment of the present invention.

3, the fall determination system 300 may include an information receiving unit 310, a calculation unit 320, a Fourier transform unit 330, and a fall determination unit 340. 4 may be performed by the information receiving unit 310, the calculating unit 320, the Fourier transforming unit 330, and the fall determining unit 340, which are components of FIG. have. The wearable device 110 may include a wireless communication module 111, an inertial measurement sensor 112, and a power supply unit 113. The wearable device 110 may be mounted on the wrist of a pedestrian to detect movement of the pedestrian, May be transmitted to the fall determination system 330.

In operation 410, the information receiving unit 310 may receive the inertia information sensed by the wearable device 110 through the wireless communication module.

For example, the information receiving unit 310 may include a Bluetooth communication module such as BLE 4.0, and may receive inertia information from the wireless communication module 111 of the wearable device through the Bluetooth communication module. The inertia information is information sensed by the inertia measurement sensor 112, and may include acceleration information, direction information, and the like according to the motion of the pedestrian. The information receiving unit 310 may receive the inertia information from the wearable device 110 every predetermined reference time. For example, the information receiving unit 310 receives the inertia information in units of 0.02 seconds, and temporarily stores the inertia information in the buffer to convert the SVM value.

In step 420, the calculation unit 320 may calculate a Signal Vector Magnitude (SVM) value based on the inertia information received from the wearable device 110. [

For example, the calculation unit 320 may convert the inertia information received in the information receiving unit 310 in units of 0.02 seconds, for example, the acceleration information into the SVM value and temporarily store the SVM value in the buffer. To calculate the SVM value, the above-mentioned [1] MJ . Mathine , AC F Coster , NH Lovell, and BG Celler ., "A pilot study of long term monitoring of human movement in the home using accelerometry ". J. Telemed , Telecare , vol. 10 pp. [4] DM Karantonis , MR Narayanan , M. Mathie , N. Lovell, and Branko G. Celler , "A Real-Time Human Movement Classifier Using a Triaxial Accelerometer for Ambulatory Monitoring", IEEE Trans . on information technology in Biomedical, vol. 10 , no.1, January 2006. SVM can be used. For example, the SVM can be expressed as Equation 1 below.

In Equation (1), x _i is an acceleration signal vector value in forward and backward directions, y _i is an acceleration signal vector value in left and right directions, and z _i is an acceleration signal vector value in up and down directions. In Equation (1), a threshold may be set through repeated experiments to recognize a fall according to the SVM value of the signal vector magnitude.

In operation 430, the Fourier transform unit 330 may perform a Fourier transform on the SVM value. At this time, Fast Fourier Transform (FFT) can be used.

In step 431, the Fourier transform unit 330 may compare the SVM values for which the inertia information has been transformed with the predetermined threshold value to determine whether or not a fall has occurred primarily. In this case, the threshold value can be defined as the SVM value at the fall through repetitive experiments such as 1.7 g and 1.8 g, and the above-mentioned [3] AM . Turing (1950), Computing < / RTI > Machinary and Intelligence, Mind 49: 433-460. The threshold value determination method in FIG.

As a result of the comparison, if the SVM value is less than the threshold value, it can be determined that the fall is not issued. On the other hand, if the SVM value is greater than the threshold value, it can be determined that a fall has occurred first, and a second fall determination can be performed to verify once again whether the fall occurs.

In step 432, if the SVM value is greater than the threshold value, the Fourier transform unit 330 may perform FFT on the SVM values corresponding to the predetermined sampling size among the SVM values temporarily stored in the buffer. The Fourier transform unit 330 may perform FFT transform on SVM values to determine whether or not the SVM values indicating the motion change of the pedestrian are composed of certain frequency components.

For example, the Fourier transform unit 330 can FFT-transform 75 SVM values for 1.5 seconds to generate 75 FFT-transformed representative components. Since the fall occurs in about 1.5 seconds, the data sampling size can be set to 75 and 75 SVM values can be converted to FFT.

In step 440, the fall determining unit 340 can determine whether or not the pedestrian has fallen based on the pattern information based on the FFT information and the pattern information indicating the movement of the pedestrian determined. That is, the fall determining unit 340 can determine whether or not the pedestrian falls by comparing the pattern information represented by the 75 FFT values corresponding to the converted sampling size with the pattern information indicating the movement of the predetermined pedestrian.

For example, the fall determining unit 340 extracts frequency and amplitude from the converted 75 FFT values, and compares the pattern information indicating the movement of the pedestrian, thereby determining whether the pedestrian is running, It is possible to judge whether or not there is a fall, such as whether it is running in place or a real fall occurred. Hereinafter, a detailed operation for determining whether or not a fall occurs secondarily will be described later with reference to FIGS. 6 and 7. FIG.

 If it is determined in step 450 that a fall has occurred secondarily, the information receiving unit 310 can transmit information indicating that a fall has occurred to the pedestrian.

For example, the information receiving unit 310 may transmit the information indicating the fall through the Google cloud message to the guardian of the pedestrian, the emergency institution, or the terminal of the emergency unit. Hereinafter, the detailed operation of transmitting the information indicating the fall will be described later with reference to FIG.

5 is a view showing a screen displayed on the fall determination system 300 according to an embodiment of the present invention.

FIG. 5 shows an example of a screen provided when a service application for detecting whether or not a fall occurs is performed in the fall determination system 300. FIG.

In FIG. 5, the service application can be manufactured to run on a fall determination system such as a smart phone based on an Android OS (Android Operating System) to determine whether a fall exists. The fall determination system 300 in which the service application is operated can operate as a gateway that performs an emergency call function to determine whether there is a fall in real time rather than experimental data and notify the fact of a fall. For example, when a fall occurs to a pedestrian, the fall determination system 300, the server 130, and the cloud push server 140 are automatically interlocked with each other, so that information or a message informing the fall is transmitted to a terminal of an emergency institution, a pedestrian, .

Referring to FIG. 5, the screen 500 includes display information 501 for sensing whether a fall occurs, display information 502 for entering a guardian contact to notify occurrence of fall at the occurrence of a fall, display information 503), and display information 504 associated with the help.

The display information 501 is for controlling ON / OFF of the fall detection function. When the pedestrian selects the display information 501 on the screen 500, the display information 501 is ON and the fall detection function Can be set. For example, when the display information 501 is selected, the information receiving unit 310 can transmit an ON command to the inertia measurement sensor 112 through the wireless communication module 111 of the wearable device 110. [ Then, the inertia measurement sensor 112 may transmit the sensed inertia information to the fall determination system 300 through the wireless communication module 111 according to the movement of the pedestrian.

When the display information 501 is again selected while the display information 501 is set to ON, the fall detection function is set to OFF and the fall detection function can be changed from ON to OFF on the screen 500 . Then, the information receiving unit 320 can transmit an off command to the inertial measurement sensor 112 through the wireless communication module 111, and the inertial measurement sensor 112 can stop the motion change sensing of the pedestrian.

When the display information 2 (502) is selected, a screen for inputting the contact information of the guardian to notify the fact of fall at the time of fall can be displayed on the fall determination system 300. For example, a telephone number of a family member, a friend, or the like can be input as a contact of a guardian.

When the display information 3 (504) is selected, a screen for guiding a method for preventing fall can be provided. When the display information 4 (504) is selected, a screen for guiding help explaining the method of using the service application May be displayed on the fall determination system 300.

FIG. 6 is a diagram for explaining an operation in which an FFT transform is performed to determine whether or not to fall secondarily in an embodiment of the present invention. FIG.

6 shows a conceptual diagram of an Async Task. Inertia information received from the wearable device 110 is separated into a main thread 601 and a work thread 602 by using an Async Task, .

Referring to FIG. 6, the inertia information received from the wearable device 110 may be converted into SVM values and temporarily stored in the buffer. At this time, the Fourier transform unit 330 may add the 50 SVM values to the array list per second. Then, the Fourier transform unit 330 may perform FFT transform on 75 data units corresponding to the sampling size, that is, 75 SVM values, in units of 75 in the array list. As described above, in the main thread 601, the inertia information is converted into the SVM values, and the work thread 602 can be separated to perform the FFT transform on a sampling size basis. That is, it is possible to independently perform the FFT transform for the secondary determination of the fall.

FIG. 7 is a diagram for explaining an operation of determining whether a fall occurs using pattern matching based on an FFT in an embodiment of the present invention. Referring to FIG.

FIG. 7 is a graph showing the frequency bands during running, stair climbing, jumping, and falling, and may have a specific frequency band that is distinguished from each other according to walking, such as running, stepping, and running. Accordingly, the predetermined pattern information for fall determination may include patterns 710 to 740 of a specific frequency band corresponding to each of running, stepping, jumping, and falling as shown in FIG.

FIG. 7 shows a specific frequency band appearing for each specific action based on an experiment for finding a threshold value of the SVM and a feature required for pattern matching. The subjects were 20 people who were between 20 and 30 years old regardless of their gender. They selected those who had no inconvenience in walking, no discomfort in daily life, Experiments were conducted to determine specific frequency bands.

Table 2 below shows information related to the volunteers participating in the experiment, and Table 3 shows the list of activities of daily living behavior.

gender life span) age Key (cm) Weight (kg) man 13 23.8 ± 1.9 172.8 ± 2.9 69.5 ± 9.5 Woman 7 21.7 ± 4.9 162.2 ± 4.9 55.6 ± 8.2

As a result of the experiment to set the threshold value of the SVM for the applicants having the conditions of Table 2 and Table 3, the SVM value as shown in the following Table 4 can be obtained.

According to Table 4, it is difficult to compare SVM values in running, stair climbing, and jumping behavior with actual SVM values when the threshold is set at 1.7. Accordingly, it is possible to obtain a specific frequency band corresponding to running, stair climbing, and jumping by FFT-transforming experimental data (SVM values) for running, stepping, and running. For example, it can be confirmed that a frequency band of 20 Hz or more occurs mainly in the rear position and a frequency band of 3 Hz or less in the case of falling.

Referring again to FIG. 7, the pattern information 710 is a pattern representing a frequency band corresponding to running, the pattern information 720 is a pattern representing a frequency band corresponding to a step-up, and the pattern information 730 is a pattern representing a frequency band And the pattern information 740 may be a pattern indicating a frequency band representing a fall. The fall determining unit 340 may detect the frequencies and amplitudes extracted from the Fourier transformed 75 Fourier information, and store the preliminarily stored and stored values by experiment. Based on the pattern information 710, 720, 730, and 740, it is possible to determine again whether the movement of the pedestrian corresponds to fall, jumping, running, or stair climbing.

For example, when pattern matching is performed based on the FFT-converted values and the pattern information 710, 720, 730, and 740 in a state where the SVM value is larger than the threshold value 1.7g and is determined to fall first, If the FFT values according to the movement of the pedestrian match the pattern information 710, the fall determining unit 340 can determine that the pedestrian is running. In other words, although it is decided to fall first, it can be finally decided by running, not falling, using pattern matching at the time of re-judgment. Thus, it is possible to more accurately determine whether or not a fall occurs by classifying the running, running, and stair climb similar to a fall by judging whether or not the fall is secondary by using pattern matching.

Likewise, if the FFT values corresponding to the movement of the pedestrian match the pattern information 720, the fall determining unit 340 finally determines that the movement of the pedestrian is in the step-up state, and determines that the FFT values according to the movement of the pedestrian are pattern information 730), it can be finally determined that the movement of the pedestrian is in a state of a jump. For example, the fall determining unit 340 may extract FFT values corresponding to a fall of 3 Hz or less, which represent falls, with respect to FFT values according to the motion of a pedestrian. If the number of extracted FFT values is equal to or greater than the predetermined reference number, the fall determining unit 340 can finally determine a fall. That is, if the FFT values corresponding to 3 Hz or less are present in excess of the reference number, it can be finally determined as a fall.

At this time, if the FFT values according to the movement of the pedestrian match the pattern information 740, the fall determining unit 340 can finally determine that a fall has occurred to the pedestrian. As a result, the information receiving unit 310 can operate in conjunction with the server 130 and the cloud push server (i.e., the cloud server 140) managing fall determination systems held by the respective pedestrians, It is possible to transmit information informing the fall of the pedestrian to the registered guardian terminal, emergency terminal, and emergency terminal.

8 is a view for explaining an operation of providing information informing occurrence of a fall of a pedestrian in cooperation with a cloud server in an embodiment of the present invention.

FIG. 8 shows a case in which the occurrence of a fall of a pedestrian is informed to a guardian, an emergency institution, a paramedic, etc. by using a conventional push message function.

For example, a Google Push message (G.C.M.) or the like may be used as a push message. Google Cloud Message (G.C.M) can refer to Google's Push Message Service, which allows developers to send data from their servers to their Android apps. If it is determined by the pedestrian that the fall has occurred in both the first based on the SVM value and the second based on the FFT value, the information receiving unit 310 may transmit a message informing the guardian, the emergency institution, have. In order to utilize the push message, the server 801 may request and receive a project ID and an API key from the cloud push server 802. The server 801 may store and manage the assigned project ID, the API key, and the identification information of the fall determination system 803 in association with each other.

For example, in a case where a fall occurs in the arrow ① and information for informing the fall is received from the fall determination system 803, the server 801 calculates the fall determination information based on the identification information of the fall determination system 803, The telephone number of the emergency unit, and the telephone number of the emergency unit from a database (not shown). Here, the database may be included in the server 801 or may exist separately from the server 801 to exchange information through the network.

In arrow (2), the server 802 receives a message including a telephone number of the extracted guardian, emergency agency, and emergency unit, API key information to be used for notifying fall, and identification information of the fall determination system 803 , Pedestrian ID, phone number, etc.) to the cloud push server 802. [

Then, the push server 802 can transmit a push message to the guardian's terminal based on the terminal telephone number of the guardian, the identification information of the fall determination system, and API key information to inform the pedestrian of the occurrence of a fall to the terminal of the guardian . For example, a push message informing '010-111-1234 (a pedestrian's phone number) has fallen' can be transmitted to the guardian's terminal. At the same time, a push message informing '010-111-1234 (a pedestrian's telephone number has occurred)' can be transmitted to the terminal of the emergency institution and the terminal of the emergency unit. As described above, by using the push message to inform the user of the fall, the accident can receive appropriate feedback from the emergency institution or the guardian, such as emergency measures, hospital transportation due to falls, and the like.

As described above, whether or not a fall is primarily determined based on the SVM value and the threshold value based on the inertia information, and when the fall is determined, it is determined whether or not the fall is secondarily detected through pattern matching based on the FFT value. The accuracy of judgment can be improved. In other words, in the case of a behavior in which the SVM value is difficult to distinguish from falls such as running, stair climbing, and jumping, the SVM used in determining a certain frequency band occurring only during running, stair climbing, The accuracy of the fall determination can be improved by judging again whether the fall determined first is a true fall or whether it is a jump, a step up, or a run. For example, if the SVM value is larger than 1.7 g and the transformed FFT values have a component less than 3 Hz, the fall can be finally determined.

The method according to an embodiment may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions to be recorded on the medium may be those specially designed and configured for the embodiments or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. For example, it is to be understood that the techniques described may be performed in a different order than the described methods, and / or that components of the described systems, structures, devices, circuits, Lt; / RTI > or equivalents, even if it is replaced or replaced.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (10)

Receiving inertia information according to a movement of a pedestrian from a wearable device worn by a pedestrian;
Calculating a signal vector magnitude (SVM) value based on the inertia information;
Performing a Fourier transform on the SVM values for a predetermined period of time; And
Determining whether a fall of the pedestrian is based on pattern information based on the Fourier-transformed information and pattern information indicating a motion of a predetermined pedestrian;
Lt; / RTI >
The inertia information is separately processed into a main thread for calculating the SVM value using the Async Task and a work thread for performing the Fourier transform,
The step of calculating the SVM value comprises:
Receiving the inertia information from the wearable device at a predetermined reference time and temporarily storing the inertia information in a buffer; And
Converting the inertia information temporarily stored in the buffer into an SVM value and temporarily storing the SVM value in the buffer
Lt; / RTI >
Wherein performing the Fourier transform comprises:
Adding SVM values corresponding to a predetermined sampling size of SVM values temporarily stored in the buffer to an array list; And
Performing Fourier transform on the SVM values added to the array list
And a fall determination method. The method according to claim 1,
Wherein the pattern information indicating the movement of the pedestrian,
Including a plurality of patterns classified into different specific frequency bands according to the movement of a pedestrian
And a fall determination method. The method according to claim 1,
Wherein performing the Fourier transform comprises:
Determining whether the pedestrian falls or not based on the calculated SVM value and a predetermined threshold value; And
Performing Fourier transform on the SVM values for the predetermined time as the pedestrian is primarily determined to have fallen,
And a fall determination method. The method of claim 3,
The step of determining whether the pedestrian has fallen comprises:
And re-determining whether or not the pedestrian primarily determined based on the pattern information is fallen
And a fall determination method. delete The method according to claim 1,
When it is determined that the pedestrian has fallen, at least one of the guardian terminal, the emergency terminal, and the emergency terminal, previously registered through the cloud push server, is provided with information informing the fall of the pedestrian
And a fall determination method. The method according to claim 1,
If the pedestrian is determined to have fallen, a push message informing the fall of the pedestrian is transmitted to at least one of the pre-registered guardian terminal, the emergency terminal, and the emergency agent terminal
And a fall determination method. The method according to claim 1,
The wearable device,
And an inertia measurement sensor which is manufactured as a wrist band type and is worn by the pedestrian and measures a movement of the pedestrian by the movement of the pedestrian
And a fall determination method. An information receiving unit for receiving inertia information according to a movement of a pedestrian from a wearable device worn by a pedestrian;
A calculation unit for calculating a signal vector magnitude (SVM) value based on the inertia information;
A Fourier transform unit for performing a Fourier transform on the SVM values for a predetermined period of time; And
Based on the pattern information based on the Fourier-transformed information and the pattern information indicating the movement of the pedestrian determined,
Lt; / RTI >
The inertia information is separately processed into a main thread for calculating the SVM value using the Async Task and a work thread for performing the Fourier transform,
The calculation unit may calculate,
Receiving the inertia information from the wearable device at a predefined reference time, temporarily storing the inertia information in the buffer, converting the inertia information temporarily stored in the buffer into an SVM value, and temporarily storing the SVM value in the buffer,
Wherein the Fourier transform unit comprises:
Adding SVM values corresponding to a predetermined sampling size of SVM values temporarily stored in the buffer to an array list and performing Fourier transform on SVM values added to the array list
And a fall determination system. 10. The method of claim 9,
Wherein the pattern information indicating the movement of the pedestrian,
Including a plurality of patterns classified into different specific frequency bands according to the movement of a pedestrian
And a fall determination system.

Images