KR20190104704A - a terminal for detecting fall - Google Patents

Abstract

Disclosed is a terminal for detecting falling to determine whether a user gets hurt from a fall. According to one embodiment of the present invention, the terminal for detecting falling comprises: a sensor unit receiving an acceleration signal for determining a falling accident; and a control unit determining a falling event by applying a first algorithm related to a rule-based approach and a second algorithm related to a pattern recognition method to the acceleration signal received from the sensor unit.

Description

Terminal for detecting fall

The present invention relates to a terminal for detecting a fall. More particularly, the present invention relates to a terminal including an application for falling detection.

Exponentially falling falls are recognized as the leading cause of physical, mental and economic problems in the growing elderly population worldwide. According to the WHO statistics, 28% to 35% of senior citizens over the age of 64 suffer from falls each year, and the percentage increases from 32% to 42% for people over 70 years old. These falls account for 90% of hip and wrist fractures and 60% of head injuries. In addition to these injuries, long-term lying down from falls can have serious consequences such as dehydration, hypothermia and death.

Moreover, frequent falls in older people can cause fear of falling, which can adversely affect living an independent and socially active life. Therefore, there is a need for immediate support to minimize the serious health complications caused by lying down for a long time during falls, and the development of an automated and reliable rapid fall detection system is essential.

There is conventionally an algorithm that uses a three-axis accelerator to recognize a fall. This means that the movement of the body caused by natural walking in daily behavior can be expressed as the acceleration values of the X, Y, and Z axis of the body. Through observation and analysis of the acceleration signal, it is possible to identify the ideal body movement such as a fall. Start with the idea that it is.

However, the conventional method only describes the general method and notification function related to general fall-related signal detection, and lacks a specific fall classification value or a specific methodology for securing performance accuracy. There is a problem in terms of usability.

Republic of Korea Patent Publication 1009884590000

S. Abbate, M. Avvenuti, F. Bonatesta, G. Cola, P. Corsini, and A. Vecchio, "A smartphone-based fall detection system," Pervasive and Mobile Computing, vol. 8, no. 6, pp. 883 ?? 899, 2012. H. Kerdegari, S. Mokaram, K. Samsudin, and A. R. Ramli, "A pervasive neural network based fall detection system on smart phone," Journal of Ambient Intelligence and Smart Environments, vol. 7, no. 2, pp. 221 ?? 230, 2015.

We propose a detector that accurately and quickly determines the actual fall event from the acceleration signal input through the acceleration sensor.

The fall detector according to an embodiment of the present invention includes a sensor unit for receiving an acceleration signal for determining a fall event and a first algorithm for a rule-based approach to an acceleration signal received from the sensor unit, and a second algorithm for a pattern recognition method. It includes a control unit for determining the fall event by applying.

The fall detector according to an exemplary embodiment may accurately and quickly determine the actual fall event from the acceleration signal input through the acceleration sensor.

1 schematically illustrates a fall detection and emergency alert system.
2 illustrates an example of a user interface of an application for a fall detector according to an exemplary embodiment.
3 is a block diagram for explaining a fall detector according to the present invention.
4 is a flowchart illustrating an operation of a fall detector according to an exemplary embodiment.
5 shows the magnitude of the acceleration sensing received by the controller 180.
6 shows a comparison of a fall event detection result through an algorithm according to the present invention and a conventional method.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, but are not limited or limited by the embodiments of the present invention. In describing the present invention, a detailed description of known functions or configurations may be omitted to clarify the gist of the present invention.

Terms including ordinal numbers such as first and second may be used to describe various components, but the components are not limited by the terms. The terms are used only for the purpose of distinguishing one component from another.

When a component is referred to as being "connected" or "connected" to another component, it may be directly connected to or connected to that other component, but it may be understood that other components may be present in between. Should be. On the other hand, when a component is said to be "directly connected" or "directly connected" to another component, it should be understood that there is no other component in between.

Singular expressions include plural expressions unless the context clearly indicates otherwise.

In this application, the terms "comprises" or "having" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

Hereinafter, a terminal (hereinafter, referred to as a "falling detector") for improving fall detection of the above-described prior art will be described. In a specific embodiment, the fall detector may be a smartphone. Application included in the fall detector according to an embodiment of the present invention proposes a specific algorithm for analyzing a signal detected from the user of everyday life.

1 schematically illustrates a fall detection and emergency alert system.

The fall detector according to an embodiment of the present invention uses a two-step algorithm to analyze a user's movement. As illustrated in FIG. 1, upon detecting a fall, the fall detector vibrates and displays a warning cancel screen. In detail, an application in the fall detector may determine a fall situation according to an algorithm according to an embodiment of the present disclosure. In this case, the application may vibrate the fall detector and display a related message on the display.

If the user does not cancel the displayed warning within a specified time (eg 30 seconds), the fall detector activates a sound alarm and delivers a message with location information via the designated emergency contact.

2 illustrates an example of a user interface of an application for a fall detector according to an exemplary embodiment.

2A illustrates an example of a main screen of an application.

2 (b) shows an example of a setting screen of an application.

2 (c) shows an example of a warning cancellation message according to a fall detection of an application. As described above with reference to FIG. 1, when the user does not cancel the cancellation message through a touch or the like within a predetermined time, the application may determine that an emergency has occurred to the user.

2D illustrates an example of a feedback screen of an application.

In one embodiment, the mail screen of the application may include 1) start / stop of the fall detection process, 2) application configuration settings, and 3) summary display of important settings.

According to an embodiment of the present disclosure, the setting screen of the application may include setting functions related to personal information, a location of a moving fall detector, a fall detection service priority, and a fall event notification. The fall alert notification setting function may include an alarm sound and duration, a countdown timer for canceling the alert, and an emergency contact item.

3 is a block diagram for explaining a fall detector according to the present invention.

The fall detector 100 includes a wireless communication unit 110, an input unit 120, a sensing unit 140, an output unit 150, an interface unit 160, a memory 170, a controller 180, and a power supply unit 190. ) May be included. The components shown in FIG. 3 are not essential to implementing the fall detector, so the fall detector described herein may have more or fewer components than those listed above.

More specifically, the wireless communication unit 110 of the components, between the fall detector 100 and the wireless communication system, between the fall detector 100 and another fall detector 100, or the fall detector 100 and the external server It may include one or more modules that enable wireless communication therebetween. In addition, the wireless communication unit 110 may include one or more modules connecting the fall detector 100 to one or more networks.

The wireless communication unit 110 may include at least one of the broadcast receiving module 111, the mobile communication module 112, the wireless internet module 113, the short range communication module 114, and the location information module 115. .

The input unit 120 may include a camera 121 or an image input unit for inputting an image signal, a microphone 122 for inputting an audio signal, an audio input unit, or a user input unit 123 for receiving information from a user. , Touch keys, mechanical keys, and the like. The voice data or the image data collected by the input unit 120 may be analyzed and processed as a control command of the user.

The sensing unit 140 may include one or more sensors for sensing at least one of information in the fall detector, surrounding environment information surrounding the fall detector, and user information. For example, the sensing unit 140 may include a proximity sensor 141, an illumination sensor 142, an illumination sensor, a touch sensor, an acceleration sensor, a magnetic sensor, and gravity. G-sensor, Gyroscope Sensor, Motion Sensor, RGB Sensor, Infrared Sensor, Infrared Sensor, Finger Scan Sensor, Ultrasonic Sensor Optical sensors (e.g. cameras 121), microphones (see 122), battery gauges, environmental sensors (e.g. barometers, hygrometers, thermometers, radiation detection sensors, Thermal sensors, gas sensors, etc.), chemical sensors (eg, electronic noses, healthcare sensors, biometric sensors, etc.). On the other hand, the fall detector disclosed herein may utilize a combination of information sensed by at least two or more of these sensors.

The output unit 150 is used to generate an output related to sight, hearing, or tactile sense, and includes at least one of a display unit 151, an audio output unit 152, a hap tip module 153, and an optical output unit 154. can do. The display unit 151 forms a layer structure with or is integrally formed with the touch sensor, thereby implementing a touch screen. The touch screen may serve as a user input unit 123 that provides an input interface between the fall detector 100 and the user, and may provide an output interface between the fall detector 100 and the user.

The interface unit 160 serves as a path to various types of external devices connected to the fall detector 100. The interface unit 160 connects a device equipped with a wired / wireless headset port, an external charger port, a wired / wireless data port, a memory card port, and an identification module. It may include at least one of a port, an audio input / output (I / O) port, a video input / output (I / O) port, and an earphone port. In the fall detector 100, in response to an external device being connected to the interface unit 160, appropriate control associated with the connected external device may be performed.

In addition, the memory 170 stores data supporting various functions of the fall detector 100. The memory 170 may store a plurality of application programs or applications that are driven by the fall detector 100, data for operating the fall detector 100, and instructions. At least some of these applications may be downloaded from an external server via wireless communication. In addition, at least some of these application programs may exist on the fall detector 100 from the time of shipment for basic functions of the fall detector 100 (for example, a call forwarding, a calling function, a message receiving, and a calling function). Meanwhile, the application program may be stored in the memory 170 and installed on the fall detector 100 to be driven by the controller 180 to perform an operation (or function) of the fall detector.

In addition to the operations related to the application program, the controller 180 typically controls the overall operation of the fall detector 100. The controller 180 may provide or process information or a function appropriate to a user by processing signals, data, information, and the like, which are input or output through the above-described components, or by driving an application program stored in the memory 170.

In addition, the controller 180 may control at least some of the components described with reference to FIG. 1A in order to drive an application program stored in the memory 170. In addition, the controller 180 may operate by combining at least two or more of the components included in the fall detector 100 to drive the application program.

The power supply unit 190 receives power from an external power source and an internal power source under the control of the controller 180 to supply power to each component included in the fall detector 100. The power supply unit 190 includes a battery, which may be a built-in battery or a replaceable battery.

At least some of the components may operate in cooperation with each other in order to implement an operation, a control, or a control method of the fall detector according to various embodiments described below. In addition, the operation, control, or control method of the fall detector may be implemented on the fall detector by driving at least one application program stored in the memory 170.

4 is a flowchart illustrating an operation of a fall detector according to an exemplary embodiment.

The fall detector operating method according to an exemplary embodiment of the present invention includes a two-stage fall detection method. In addition, in another embodiment, a total of three steps may be included by using a first step. In a preferred embodiment, the two-step drop detection method may include a rule-based approach based on a threshold determination method (TBM) and a pattern recognition method based on a multiple kernel (Multiple Kernel Learning Support Vector Machine (MKL-SVM)). have. And the third step added may be a recovery step.

The falling detector receives an acceleration sensing value related to the falling through the sensor (S101). In a specific embodiment, the controller of the fall detector (hereinafter, referred to as “control unit”) 180 may continuously receive the acceleration sensing value through the sensor unit 140. The input sensing value may include a peak value that may be determined as a fall-like event.

The magnitude of the acceleration vector used by the controller 180 to determine the fall similar event is expressed by Equation 1.

Here, A _x , A _y and A _z represent acceleration signals for the x, y and z axes of the sensor, respectively.

Meanwhile, FIG. 5 illustrates an acceleration sensing magnitude that the controller 180 receives. As illustrated in FIG. 5, the controller 180 may receive an acceleration sensing value that changes with time. In addition, the sensed acceleration magnitude may include a peak value that can be viewed as a fall-like event as shown in FIG. 5. Return to Figure 4 again.

Steps S103 to S109 to be described below may be classified as first algorithms. In a specific embodiment, the first algorithm may be referred to as a threshold determination method.

The controller 180 determines whether an acceleration peak value due to impact exceeds a threshold value (S103). In one embodiment, if the acceleration peak value exceeds the threshold, the controller 180 proceeds to the next processing step. In another embodiment, when the acceleration peak value does not exceed the threshold, another acceleration peak value is input. At this time, in a specific embodiment, the threshold value may be 3g. Where 1 g represents the acceleration of gravity on the ground and 3 g is three times the acceleration of gravity.

As described above, when the acceleration peak value exceeds the threshold value, the controller 180 determines whether the acceleration peak value due to post impact exceeds the threshold value (S105).

Falls are a single event that occurs instantaneously and fall like events do not characterize repeatability. Therefore, in an embodiment of the present disclosure, the controller 180 determines the event as a fall-like event when the acceleration peak identified through the acceleration sensor exceeds a predetermined threshold value and there is no additional peak for a certain period. In a specific embodiment, if the acceleration vector magnitude (AVM) exceeds the threshold and there are no additional peaks above the threshold for a period of time, the fall detector determines the acceleration peak as a fall like event. In this case, the time interval for determining the repeatability of the drop-like event may be 3.5 seconds.

The controller 180 determines whether the inactivity test value is less than a specific value when there is no acceleration pit exceeding a threshold value after the acceleration peak received in step S105 (S107). In an embodiment, when the inactivity test value is less than the specific value, the controller 180 determines whether the posture test value exceeds the specific value (S109).

Using an inactivity test and a posture test, the controller 180 effectively discards the fall-like activities of daily living, which are the majority of the acceleration peaks, and falls like events. Can be determined. Here, the fall-like general activity refers to an event having a low probability of being an actual fall among the events identified through the acceleration sensor, and a fall-like event refers to an event that is most likely to be an actual fall among the events identified through the acceleration sensor. If the probability that the event is a fall is more than a predetermined probability through the following additional algorithm, the fall detector determines the event as a fall.

In a specific embodiment, when it is determined that the acceleration peak received through the steps S103 to S105 is related to the fall-like event, the controller 180 determines whether the user's movement is detected for a predetermined time. If the user is inactive for the predetermined time, the controller 180 uses an average absolute acceleration magnitude variation (AAMV). In more detail, the average absolute acceleration magnitude change is expressed by Equation 2 below.

At this time, if there is little or no user movement after detecting the acceleration peak due to the collision, the average absolute acceleration magnitude change value is calculated to be close to zero. In a specific embodiment, the value close to zero may be less than 0.6.

The controller 180 determines a change in posture of the user before and after the acceleration peak is detected. In this case, the controller 180 uses the angle value for determining the posture change of the user. If the measured degree of change in the tilt angle exceeds a specific value, the controller 180 may assume that a fall occurs and the user's posture rapidly changes, with a high probability. Hereinafter, the tilt angle and the tilt angle variation used to determine the posture change will be described.

Tilt angle: This feature roughly indicates the direction of the fall detector with the body of the user holding the fall detector. The x and y axes of the fall detector sensor basically capture the standing position. If the user's posture is from the standing position to the lying position, the tilt angle is calculated as a value between 0 and 90 degrees. The tilt angle is expressed as in Equation 3.

Here, A _lp [i] and AVM _ip [i] represent low-pass filtered axis acceleration (x, y or z axis) and combined acceleration, respectively. Equation 4 below is used to approximate the gravity component and suppress the movement signal.

Tilt Angle Variation (TAV): The tilt angle value may change for the same posture depending on the alignment difference between the user's vertical axis and the fall detector axis. In order to overcome this issue, instead of using the tilt angle directly, the present invention may use a change in tilt angle during a fall like event. Accordingly, the tilt angle variation value may be defined as an absolute value of a difference between an average of tilt angle values during a specific time interval before an event occurs and an average of tilt angle values during a specific time interval after an event is issued. In a specific embodiment, the specific time interval is 0.5s before the event occurs. 1s, 0.5s after occurrence ?? It can be 1s.

Through the above-described steps, the sensed acceleration peak may be estimated as the actual fall. However, there may be a fall like event that is not discarded by the first algorithm, the determination of which is performed by a second algorithm using a pattern recognition method.

The controller 180 extracts the feature points from the event determined as the fall-like event by the first algorithm (S111). At this time, the features extracted by the controller 180 are as shown in Table 1. Specifically, features may include posture related (19-21), energy related (15-18), statistically related (9-14), and criteria based (1-8).

feature No. Feature name feature No. Feature name One AAMV-avg, absolute accel. Mag. Variation 9-10 Mean and median 2 IDI-impact durationindex 11 Range 3 MPI-maximum peak index 12 Variance 4 MVI-minimum valley index 13 Standard deviation 5 PDI-peak durationindex 14 RMS-root mean square 6 ARI-activity ratio index 15-17 SAA _{X / Y / Z} -summed axis acceleration 7 FFI-free fall index 18 SMA-summed magnitude area 8 SCI-step count index 19-21 TAV _{X / Y / Z} -tilt angle variation

Here, summed axis acceleration (SAA) represents the activity intensity for a particular axis. Then, the summed axis acceleration is defined as in Equation 5.

And the summed magnitude area SMA represents the sum of the absolute values of the x, y and z axis accelerations. The sum of the size areas is defined as in Equation 6.

The controller 180 determines whether the sensed acceleration signal event is an actual fall event by using the extracted feature set as an input to the running model (S113).

In a specific embodiment, the controller 180 uses a multi-kernel learning support vector machine (MKL-SVM) to recognize fall patterns and fall-like general activities.

Multi-kernel learning involves learning both the coefficients a _i ^* , b ^* and the kernel weight (d _m ) in a single optimization problem. The decision function for the kernel algorithm is expressed as in Equation 7.

Here, a _i ^* and b ^* represent optimized support vectors and bias coefficients using training data {x _i , y _i } ^l _{i = 1} , respectively. As shown in Equation 8, the kernel K (x, x _i ) is basically a convex combination of basis kernels.

Here, M represents the total number of kernels and K _m represents the basis kernels.

When it is determined that the acceleration signal received through the above-described second algorithm is an actual fall event, the controller determines whether the user recovers (S115). In the case of a minor fall, the user may not be at great risk, in which case no special measures may be necessary. Therefore, the fall detector according to an embodiment of the present invention sells the user's recovery rather than unconditionally performing the related process based on the detected fall event.

In a specific embodiment, the controller 180 checks the activity. The controller 180 determines whether the acceleration vector magnitude exceeds a specific value. If the magnitude of the acceleration vector does not exceed a specific value, the controller 180 may determine the user's state as unconscious or impaired. For example, the specific value may be 1.5 g.

In addition, the controller 180 may detect a change in posture and determine whether the user recovers. When the tilt angle change value related to the posture change is less than the specific value, the controller 180 may determine that the user has recovered. For example, the specific value may be 20 degrees.

When the controller 180 determines that the user's state is not recovered from the fall through the above-described steps, the controller 180 performs a fall related process (S117).

In a specific embodiment, the controller 180 may notify the user or people around the current emergency situation through the output unit 150. According to an embodiment, the display unit 151 may display a warning cancel message. The warning cancel message is the message described with reference to FIG. 2. In another embodiment, the sound output unit 152 may output a warning sound. In another embodiment, the haptic module 153 may vibrate the fall detector. In another embodiment, the light output unit 154 may output light to notify the surroundings of an emergency situation.

In another embodiment, the controller 180 may transmit an emergency situation through the wireless communication unit 110. In one embodiment, when a user fails to cancel the above-mentioned warning cancellation message despite a predetermined time elapsed, the wireless communication unit 110 may transmit a rescue message using a stored emergency contact.

6 shows a comparison of a fall event detection result through an algorithm according to the present invention and a conventional method.

As shown in FIG. 6, the comparison of Comparative Invention 1 (Non-Patent Document 1) and Comparative Invention 2 (Non-Patent Document 2) with the present invention showed that the present invention showed superior performance over the comparative inventions. Can be.

The algorithm described above may be implemented in an application. In this case, the fall detector may include an application that implements the above-described algorithm as a terminal.

The present invention has been described above with reference to preferred embodiments thereof, which are merely examples and are not intended to limit the present invention, and those skilled in the art do not depart from the essential characteristics of the present invention. It will be appreciated that various modifications and applications are not possible that are not illustrated above. For example, each component specifically shown in the embodiment of the present invention can be modified. And differences relating to such modifications and applications will have to be construed as being included in the scope of the invention defined in the appended claims.

Claims (19)

A sensor unit configured to receive an acceleration signal for determining a fall event; And
And a controller configured to determine a fall event by applying a first algorithm of a rule-based approach and a second algorithm of a pattern recognition method to the acceleration signal received from the sensor unit.
Fall detector The method of claim 1,
The first algorithm may include comparing a peak value included in an input acceleration signal with a specific threshold value to determine whether the peak value exceeds a specific threshold value;
If the peak value exceeds a certain threshold value, determining if there is an additional peak value exceeding the threshold value for a period of time; And
If there is no additional peak value for a period of time, determining an event corresponding to the peak value as a fall like event;
Fall detector. The method of claim 2,
The first algorithm may further include determining whether the fall detector user is inactive and whether the posture changes when there is no additional peak value for a predetermined time.
Fall detector. The method of claim 3, wherein
Determining whether or not the fall detector user is inactive,
Determining whether the absolute value of acceleration magnitude change is less than a specific value;
Determining whether or not the posture changes,
Determining whether the tilt angle variance absolute value exceeds a specific value
Fall detector. The method of claim 1,
The second algorithm,
Extracting a feature from the acceleration signal; And
Determining whether an event corresponding to a peak value included in an acceleration signal is an actual fall event by using the extracted feature as an input value of multi-kernel learning;
Fall detector. The method of claim 1,
If the controller determines that a fall event has occurred from the acceleration signal based on the first algorithm and the second algorithm, the controller further performs a third algorithm for determining whether the user recovers.
Fall detector. The method of claim 6,
The third algorithm,
Determining whether the absolute value of the magnitude of acceleration variation exceeds a specific value and determining whether the absolute value of the tilt angle variation is less than a specific value.
Fall detector. The method of claim 6,
The control unit
If it is determined that the user has not recovered through the third algorithm,
To carry out a fall-related process
Fall detector. The method of claim 8,
A display unit for outputting a warning cancellation message in accordance with the fall-related process,
If the control unit does not receive an input for canceling the warning cancellation message within a predetermined time, the control unit transmits a rescue message through the wireless communication unit.
Fall detector. Receiving an acceleration signal for determining a fall event;
Determining a fall event by applying a first algorithm of a rule-based approach and a second algorithm of a pattern recognition method to the received acceleration signal.
How to operate a fall detector. The method of claim 10,
The first algorithm may include comparing a peak value included in an input acceleration signal with a specific threshold value to determine whether the peak value exceeds a specific threshold value;
If the peak value exceeds a certain threshold value, determining if there is an additional peak value exceeding the threshold value for a period of time; And
If there is no additional peak value for a period of time, determining an event corresponding to the peak value as a fall like event;
How to operate a fall detector. The method of claim 11,
The first algorithm may further include determining whether the fall detector user is inactive and whether the posture changes when there is no additional peak value for a predetermined time.
How to operate a fall detector. The method of claim 12,
Determining whether or not the fall detector user is inactive,
Determining whether the absolute value of acceleration magnitude change is less than a specific value;
Determining whether or not the posture changes,
Determining whether the tilt angle variance absolute value exceeds a specific value
How to operate a fall detector. The method of claim 10,
The second algorithm,
Extracting a feature from the acceleration signal; And
Determining whether an event corresponding to a peak value included in an acceleration signal is an actual fall event by using the extracted feature as an input value of multi-kernel learning;
How to operate a fall detector. The method of claim 10,
If the controller determines that a fall event has occurred from the acceleration signal based on the first algorithm and the second algorithm, the controller further performs a third algorithm for determining whether the user recovers.
How to operate a fall detector. The method of claim 15,
The third algorithm,
Determining whether the absolute value of the magnitude of acceleration variation exceeds a specific value and determining whether the absolute value of the tilt angle variation is less than a specific value.
How to operate a fall detector. The method of claim 15,
The control unit
If it is determined that the user has not recovered through the third algorithm,
To carry out a fall-related process
How to operate a fall detector. The method of claim 17,
A display unit for outputting a warning cancellation message in accordance with the fall-related process,
If the control unit does not receive an input for canceling the warning cancellation message within a predetermined time, the control unit transmits a rescue message through the wireless communication unit.
How to operate a fall detector. A program comprising the algorithm of claim 10.

Images