WO2018030743A1

WO2018030743A1 - Method and apparatus for recognizing exercise

Info

Publication number: WO2018030743A1
Application number: PCT/KR2017/008534
Authority: WO
Inventors: 정주호; 정창근
Original assignee: 주식회사 비플렉스
Priority date: 2016-08-09
Filing date: 2017-08-08
Publication date: 2018-02-15
Also published as: WO2018030734A1; CN109310913A; CN109310913B; JP2019531772A; JP6765505B2; CN109414608A; CN109414608B

Abstract

Disclosed are a method for recognizing exercise and an apparatus for same, comprising: a sensor signal-collecting unit worn on the head of a user and including a three-axial direction acceleration sensor, the three-axial directions including upward and downward, left and right, and forward and backward directions, and a location measuring sensor for measuring the location of the user; and an exercise posture-deriving unit for deriving a walking or running exercise state value including the acceleration, velocity, and location of the center of mass of the user, by receiving signals from the sensor signal-collecting unit and using a three-axial direction acceleration signal and location signal, and deriving the walking or running posture by analyzing the walking or running exercise state value.

Description

Exercise recognition method and device

The present invention relates to an exercise recognition method and apparatus, and more particularly, to an exercise recognition method and apparatus for collecting and analyzing walking and driving exercise data of a user.

In general, it has been pointed out that the amount of exercise in daily life of modern people is quite insufficient to maintain proper physical health. Accordingly, there is a growing interest in systematic exercise methods that effectively promote health. Specifically, exercises such as exercises to quickly and systematically train the body, correction of postures to promote health in the long term, and exercises suitable for elderly people whose physical ability has deteriorated with a long lifespan Interest in sports is increasing. As one of the exercise methods to meet such various needs, there is a walking exercise that anyone can easily do.

Most people walk in unconsciously accustomed postures because anyone who walks without physical problems can walk. However, because the human body is not perfectly symmetrical, in most cases the walking takes place in an unbalanced and incorrect position. Continued walking in this wrong posture can skew muscles and skeletons, and can also cause a variety of systemic pain. In the general public, such a poor walking posture causes a problem of deteriorating physical health, and especially in a growing child or an elderly person whose physical ability is deteriorated, body distortion and deterioration of health are more severe. On the other hand, professionals such as athletes and dancers need more physical skills than the general public, there is a problem such as generating a limit to improve the physical ability.

As such, correct walking posture is important from the general public to professionals, and thus, various studies on how to effectively correct the walking posture have been conducted.

According to the prior art, a pressure sensor mounted on a shoe or a footrest is mostly used to detect walking. However, in the case of recognizing and analyzing the walking posture according to the prior art, there is a risk that the pressure sensor may be damaged in the process of being continuously pressurized for a long period of time. Can cause. In addition, the size of a person's foot varies greatly depending on the individual, and accordingly, shoes having a pressure sensor must be produced in various sizes, and thus, there is a problem in that production efficiency and economy are considerably inferior. In addition, there is a problem that the family can not share a single walking posture correction device and must be purchased separately according to the size of the foot to increase the economic burden.

As such, in the technology of recognizing, detecting, and analyzing walking in order to correct a walking posture, there is an increasing demand for a technology capable of effectively detecting, detecting, and analyzing walking in a manner other than a pressure sensor.

The present invention has been made in view of the above-described technical problem, the object of the present invention is to substantially compensate for the various problems caused by the limitations and disadvantages in the prior art, to be worn on the user's body The present invention provides a motion recognition method and apparatus for walking and driving monitoring, which can be formed to recognize, detect, and analyze walking effectively and accurately using only a sensor measuring a dynamic physical quantity of a user, such as an acceleration sensor and a position sensor. And a computer readable recording medium having recorded thereon a program for executing the method.

According to an embodiment of the present invention, the motion recognition method includes a three-axis acceleration sensor including up, down, left and right, front and rear, and a position measuring sensor for measuring a user position and using a motion recognition device worn on a user's head. A motion recognition method comprising: a data collection and motion recognition step of determining whether to walk and drive by collecting and analyzing three-axis acceleration; And an acceleration-based motion state value derivation step of calculating the motion state values of the user's center of mass using the collected three-axis accelerations.

According to an embodiment of the present invention, the data collection and motion recognition step may include: an up and down acceleration collection step of collecting up and down acceleration; A peak detection step of detecting peaks of the collected up-down acceleration; A motion detecting step of determining whether the vertical acceleration peak value is equal to or greater than a predetermined threshold value; A three-axis acceleration collection step of collecting three-axis acceleration when the vertical acceleration peak value is equal to or greater than a predetermined threshold value; A Fourier transform step of Fourier transforming the three-axis acceleration to derive a frequency response graph; And an exercise type determination step of determining whether the walking and driving movements or other movements are compared by comparing the frequency response graph with a predetermined frequency response modification or size reference.

According to an embodiment of the present invention, before the peak detection step or the Fourier transform step, the motion recognition method, the noise removal step of removing noise by passing the vertical acceleration or triaxial acceleration through a predetermined band pass filter It further includes.

According to one embodiment of the present invention, the step of determining the exercise type, if the following equation is satisfied, determine the user's exercise as walking and driving state, otherwise it is determined as other exercise.

M _{z, p} / M _{z, other} > c _z and M _{x, p} / M _{x, other} > c _x

(From here,

a _z : acceleration in the vertical direction,

f _p : frequency with the maximum magnitude in the Fourier transform result of the up-down acceleration a _z ,

M _{z, p} : Vertical acceleration in a Fourier transform result of a _z a f _p to the center frequency and the total energy of the frequency components in the vertical direction based on the band with a bandwidth of less than 1Hz,

M _{z, other} : Sum of energy of remaining frequency components excluding the vertical reference band from the Fourier transform result of vertical acceleration a _z ,

c _z : a predetermined vertical reference threshold,

a _x : the horizontal acceleration,

M _{x, p} : The Fourier transform results in the lateral direction acceleration a _x in f _p / 2 of the center frequency and the total energy of the frequency components in the left and right directions based on the band with a bandwidth of less than 1Hz,

M _{x, other} : Sum of energy of remaining frequency components excluding the left and right reference bands from the Fourier transform result of the left and right acceleration ax,

c _x : predetermined left and right reference threshold)

According to an embodiment of the present invention, the step of determining the exercise type, when the user exercise is walking and driving state, if there is a section satisfying the following equation in the Fourier transform result of the up and down acceleration, the user movement is determined as the driving state Otherwise, it is determined as a walking state.

a _z <k

(From here,

k: stance determination constant)

According to an embodiment of the present invention, the acceleration-based movement state value deriving step may include: a mass center acceleration derivation step of deriving an acceleration of the user's center of mass by multiplying each value of the three-axis acceleration by a predetermined gain value; And a center of mass velocity and position derivation step of deriving the velocity and position of the center of mass of the user using pre-measured user key information, user position information, and center of mass acceleration.

In addition, according to an embodiment of the present invention, the exercise recognition apparatus includes a three-axis acceleration sensor including up, down, left and right, front and rear, and a position measuring sensor for measuring the user position and collecting the sensor signal worn on the user's head part; Receiving a signal from the sensor signal collection unit, using the three-axis acceleration and position signals to derive a walking or driving motion state value including the acceleration, speed, position of the user's center of mass, the walking or driving motion state value It includes an exercise posture derivation unit to derive a walking or driving posture by analyzing the.

According to an embodiment of the present invention, the sensor signal collector further includes a three-axis angular velocity sensor.

According to an embodiment of the present invention, the position measuring sensor measures a user position.

According to an embodiment of the present invention, the exercise recognition apparatus further includes an exercise correction generator configured to generate posture correction information by comparing the walking posture derived by the exercise posture deriving unit and the reference posture.

According to an embodiment of the present invention, the exercise recognition device, the correction information output unit for converting the posture correction information generated by the exercise correction generating unit as information that can be recognized by the user, including sound, illustration, image, and output It includes more.

According to an embodiment of the present invention, the exercise recognition apparatus is configured to transmit and store the walking posture derived by the exercise posture derivation unit to an external database and accumulate it.

According to an embodiment of the present invention, the exercise recognition apparatus, the signal transmission between the sensor signal collection unit and the exercise posture deriving unit is made by at least one wireless communication selected from Bluetooth, Wi-Fi, NFC.

In addition, according to an embodiment of the present invention, the pressure center path-based exercise posture deriving method includes a three-axis acceleration sensor including up, down, left, and right, and a position measuring sensor for measuring a user's position. An exercise pose derivation method using a worn exercise pose derivation apparatus, comprising: an acceleration vector at a center of mass position using motion state values of a user's center of mass calculated using three-axis acceleration collected by the exercise pose derivation apparatus Estimating a pressure center path by projecting onto the ground in a direction; An exercise type determination step of determining whether the user is walking or traveling from the pattern of the up-down acceleration graph; And a motion posture derivation step of deriving posture information including stride length, interpolation, interpolation angle, and asymmetry based on the estimated pressure center path and the three-axis acceleration.

According to an embodiment of the present invention, the step of estimating the pressure center path may include a direction of the pressure center as a ratio of the left and right acceleration to the sum of the vertical acceleration and the gravity acceleration and the forward and backward acceleration to the sum of the vertical acceleration and the gravity acceleration. Determining a pressure center direction; Assuming that the center of mass is located at a height determined by multiplying a predetermined user's key information by a predetermined inference constant, the pressure center projecting the pressure center position by projecting onto the ground in the direction determined in the pressure center direction determining step Location inference step.

According to an embodiment of the present invention, the pressure center path estimating step may further include a pressure center position correcting step that is corrected by a value obtained by multiplying a pressure center position inferred in the pressure center position inference step by a predetermined front-rear and left-right correction constant. Include.

According to an embodiment of the present invention, the step of deriving the exercise posture may include defining a local maximum as an intermediate support point in the vertical acceleration measured in the time domain when the user's exercise is running, and in the time domain when the user's exercise is walking. A middle support point determining step of defining a local minimum as an intermediate support point in the vertical acceleration measured in the step; If the user's motion is driving, the section that is represented by constant value in the up-down acceleration measured in the time domain is determined as the floating section, and the remaining section is determined as the one-foot support section. And a section classification determination step of determining a section in which the peak value is formed in the vertical acceleration as the double support section and determining the remaining section as the one support section.

According to an embodiment of the present invention, the exercise posture deriving step may include: calculating an average speed by measuring user location information at predetermined time intervals; Calculating a walking frequency by measuring the number of intermediate supporting points during the time interval; And calculating the stride length by dividing the average speed by the walking frequency to calculate the stride length in the front-rear direction.

According to one embodiment of the present invention, the step of deriving the exercise posture, calculates the interpolation in the left and right directions using the pressure center position value corresponding to the intermediate support point.

According to an embodiment of the present invention, the step of deriving the exercise posture, calculates a complementary angle using a pressure center position value corresponding to a start point of the one foot support section and a pressure center position value corresponding to an end point of the one foot support section. do.

According to an embodiment of the present invention, the step of deriving the exercise posture may include: identifying a foot supported by a sign of left and right acceleration measured in a time domain; Comparing the peak value, the valley value, and the difference value between the two accelerations measured in the time domain, the left and right asymmetry and repeatability are calculated.

In addition, according to an embodiment of the present invention, the pressure center path-based exercise posture deriving device includes a three-axis acceleration sensor including up, down, left, and right, and a position measuring sensor for measuring a user position. A sensor signal collector to be worn; Receiving a signal from the sensor signal collection unit, using the three-axis acceleration and position signals to derive a walking or driving motion state value including the acceleration, speed, position of the user's center of mass, the walking or driving motion state value From the pressure center path is estimated from, and analyzing the pressure center path includes a movement posture derivation unit for deriving a walking or driving posture.

According to the present invention, by using the characteristic analysis algorithm of the present invention, such as measuring acceleration, position, etc. in the user's body (e.g., head or waist), and converting it into a center of mass motion state value, it is effective and accurate. Recognize, detect, and analyze walking.

In addition, according to the present invention, by using the characteristic analysis algorithm of the present invention, such as measuring the acceleration, position, etc. in the user's body, converting it to the center of mass motion state value, and estimating the pressure center path therefrom, It can recognize, detect and analyze walking in an accurate and accurate way. In particular, the present invention uses the acceleration measured at the position most similar to the movement of the center of mass of the user's body (specifically, the acceleration in the left and right directions is measured at the head side, and the acceleration and the position in the front and rear directions are measured at the waist side). Measurement, and the acceleration in the up and down direction is made to measure at the head side or the waist side), more accurate acceleration and position can be measured.

In addition, according to the present invention, there is an advantage in that only a sensor for measuring a dynamic physical quantity of a user, such as an acceleration sensor and a position sensor, can be used in terms of device configuration. That is, by using a pressure sensor that recognizes walking by being pressed by the user's foot, there are various problems such as device durability and life deterioration problem, and production and use of a separate device according to the user's body size. However, in the case of the present invention, since the technical configuration itself of arranging the pressure sensor which is the cause of this problem on the foot part is completely excluded, various problems as described above are fundamentally eliminated. In addition, it is also possible to obtain effects such as improved user convenience and economic efficiency in each user or producer.

1 illustrates use states of the exercise posture deriving apparatus according to an embodiment of the present invention.

Figure 2 shows a schematic diagram of the exercise posture derivation apparatus according to an embodiment of the present invention.

3 is a flowchart illustrating a method of deriving an exercise posture according to an exemplary embodiment of the present invention.

4 shows a relationship diagram between a center of mass and a center of pressure in accordance with one embodiment of the present invention.

5 illustrates pressure center direction determination and position inference according to an embodiment of the present invention.

6 illustrates an example of an estimated pressure center path in accordance with an embodiment of the present invention.

7 illustrates an example of an up and down acceleration graph with respect to time during walking and driving according to an exemplary embodiment of the present invention.

8 illustrates an example of an acceleration signal measurement result according to an embodiment of the present invention.

9 is a flowchart illustrating a motion recognition method according to another embodiment of the present invention.

10 is a detailed flowchart of a data collection and exercise recognition step according to another embodiment of the present invention.

11 shows an example of an acceleration signal measurement result according to another embodiment of the present invention.

12 is a detailed flowchart of an acceleration-based exercise state value derivation step according to another embodiment of the present invention.

13 is a view illustrating usage states of the exercise posture deriving apparatus according to another embodiment of the present invention.

14 is a schematic diagram of an apparatus for deriving an exercise posture according to another embodiment of the present invention.

15 shows a schematic diagram of an apparatus for quantifying injury risk according to another embodiment of the present invention.

16 is a flowchart of a method of quantifying an injury risk according to another embodiment of the present invention.

FIG. 17 is a graph illustrating acceleration in up and down directions according to another embodiment of the present invention. FIG.

18 is a view illustrating an inclination on a vertical acceleration graph when driving in accordance with another embodiment of the present invention.

19 illustrates an impact amount on an up and down acceleration graph while driving according to another embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Like reference numerals in the drawings denote like elements, and the size of each element in the drawings may be exaggerated for clarity.

The exercise posture deriving apparatus 100 according to an exemplary embodiment of the present invention is worn on a user's head as shown in FIG. 1. The exercise posture deriving apparatus 100 according to the present exemplary embodiment may be formed in a band or a hat worn on a head as shown in FIG. 1, and may be made in a form of being plugged into an ear like an earphone by making it more compact. It will be apparent to those skilled in the art that the present invention may be in other forms.

The exercise posture deriving apparatus 100 according to the present exemplary embodiment includes a sensor signal collecting unit 110 and an exercise posture deriving unit 121 as illustrated in FIG. 2. The exercise posture deriving apparatus 100 according to the present exemplary embodiment further includes an exercise correction generator 122 and a correction information output unit 130.

The sensor signal collection unit 110 includes a three-axis acceleration sensor 111 including up, down, left, right, front and rear, and a position measuring sensor 112 for measuring a user position. The three-axis acceleration sensor 111 may be generally employed by selecting an appropriate one from among sensors used to measure acceleration in the three-axis direction, such as a form in which a gyroscope is embedded. Position measuring sensor 112 is for measuring the absolute position of the user, for example, it can be made to measure the user's position using a GPS signal, or the high-precision satellite navigation technology that is more accurate than the recent GPS has been developed Sensors to which this technique is applied may also be used. In addition, the sensor signal collector 110 may further include a three-axis angular velocity sensor 113 as shown in FIG. 2 to increase the accuracy in the motion recognition and analysis process which will be described in more detail below.

The sensor signal collection unit 110 according to the present embodiment is worn on the user's head as shown in FIG. 1 to measure the dynamic physical quantity of the user such as acceleration, speed, and position. Conventionally, a pressure sensor provided in a shoe and a footrest that is directly pressed by a foot for monitoring a gait is used. Accordingly, there is a problem that damage to the sensor occurs quickly, the device durability and life is shortened. In addition, there is a problem such as deterioration of walking recognition and analysis accuracy due to damage to the device during use, convenience and economical degradation due to frequent device replacement. In addition, when such a device is provided in a shoe, a separate device is required for each user according to the size of the user's foot, which increases the user's convenience and economical deterioration. There was a problem.

However, the present embodiment completely escapes the concept of using the pressure pressed by the foot in gait recognition, and measures the dynamic physical quantity of the user such as acceleration, speed and position measured in the user's head as shown in FIG. The characteristic analysis algorithm of the present invention to be described is applied to realize recognition, detection and analysis of gait. As such, the present embodiment differs from the prior art in the measurement position and measurement physical quantity. At this time, the root cause of the various problems pointed out in the prior art comes from the technical configuration of 'positioning the pressure sensor on the foot', according to the present invention, only the configuration alone eliminates the various problems as described above. Can be.

The exercise posture deriving unit 121 receives a signal from the sensor signal collecting unit 110 and uses a three-axis acceleration and position signal to include a walking or driving motion state value including acceleration, speed, and position of the user's center of mass. Derivation, and analyzes the walking or driving motion state value to derive a walking or driving posture. Specifically, the exercise posture deriving unit 121 estimates a pressure center path from the walking or driving motion state value, and analyzes the pressure center path to derive a walking or driving posture. Since the analysis algorithm of the present invention used by the exercise posture derivation unit 121 will be described in more detail later, the description is omitted here.

Meanwhile, the exercise posture deriving unit 121 may be formed in an integrated circuit form that may perform various calculations, and may be formed on the sensor signal collecting unit 110 and one substrate, or may be formed in a form of a separate computer. . In this case, when the exercise posture derivation unit 121 is formed separately from the sensor signal collection unit 110, as illustrated in FIG. 2 for transmitting a signal between the sensor signal collection unit 110 and the exercise posture derivation unit 121. The communication unit 114 may be provided as well. The communication unit 114 may be made by wire or wireless communication. The wireless communication may use Bluetooth, Wi-Fi and NFC technologies, but it is apparent to those skilled in the art that other wireless communication technologies may be used.

The exercise correction generator 122 compares the walking posture derived by the exercise posture deriving unit 121 and the reference posture to generate posture correction information. As described above, the exercise posture deriving unit 121 derives a walking or driving posture of the user based on the signal collected by the sensor signal collection unit 110. From this, the stride, which is one of the elements of the walking posture, can be obtained. In this case, the exercise correction generator 122 includes the optimal key stride relationship data for each walking and traveling speed, and compares the walking posture information of the corresponding user with an excessively wider stride than the height of the corresponding user. It is easy to determine whether it is narrow and to calculate the amount of stride correction that should be reduced or increased if it is outside the optimum range.

The correction information output unit 130 converts and outputs the posture correction information generated by the exercise correction generator 122 as information recognizable by a user including sound, illustration, and image. For example, when it is necessary to reduce the stride length by calculating the stride length correction amount, a voice such as "Reduce stride length" is output through a speaker provided on the exercise posture derivation apparatus 100, or a warning sound is sounded. The user may be not aware of the optimal stride length and may be encouraged to change the walking posture. Alternatively, the present invention may be realized in various forms, such as connected to a smartphone, a computer, a dedicated display, or the like, so that accurate calibration information may be output as an illustration or an image.

In addition, the exercise posture deriving apparatus 100 may transmit and store the walking posture derived by the exercise posture deriving unit 121 to an external database 140 to accumulate the posture. A user who needs such walking or driving motion analysis may be a general person who performs daily walking or jogging to promote health, or an expert who trains to improve physical ability. Naturally, it is desirable to be able to see the change. In addition, when the exercise analysis data is accumulated and stored in a large amount as described above, such data may be used as big data and used for various statistics or analysis, and thus various applications are possible.

In the exercise posture deriving method according to an exemplary embodiment of the present invention, the exercise posture deriving apparatus 100 performs an analysis such as detecting a user's exercise and determining whether the user is walking or driving. At this time, as described above, the analysis algorithm used in the present invention uses a dynamic physical quantity measured in the user's head, the exercise pose derivation apparatus 100 is a three-axis acceleration including at least up, down, left, right, front and rear What is necessary is to include the sensor 111 and the position measurement sensor 112 for measuring a user position, and the analysis algorithm to be described below may be performed in the exercise pose derivation unit 121. In addition, the exercise posture deriving apparatus 100 may further include various additional components described above for improving the function of the apparatus.

3 is a flowchart illustrating a method of deriving an exercise posture according to an exemplary embodiment of the present invention. The exercise posture deriving method according to the present embodiment includes a pressure center path estimating step, an exercise type determination step, and an exercise posture derivation step, as shown. Each step will be described in more detail below.

In the step of estimating the pressure center path, the mass is obtained by using the motion state values of the user's center of mass calculated using the triaxial accelerations a _x , a _y , and a _z collected by the exercise posture deriving apparatus 100. The pressure center path is estimated by projecting onto the ground in the direction of the acceleration vector from the center position.

In the exercise posture deriving apparatus 100, the three-axis acceleration sensors 111 may be collected by using the three-axis acceleration sensor 111, and may be integrated by using the three-axis acceleration sensor 111, or collected using the position measurement sensor 112 per hour. The location information can be used to find both speed and location. On the other hand, in general, when analyzing the movement of a certain object is analyzed based on the movement of the center of mass of the object, since the exercise posture deriving apparatus 100 is provided in the user's head, the values measured here It is desirable to convert the exercise state value to an analysis. The conversion of the value measured at the user's head position to the value at the user's center of mass can be easily derived by appropriately multiplying a gain value previously obtained using body information such as user key information.

By deriving the motion state values of the center of mass (per hour acceleration / velocity / position for each direction, frequency analysis, etc.), the pressure center path can be estimated therefrom. The human body behaves by using the reaction pressure applied to the supporting foot during walking or driving. The sum of these reaction pressures is called the ground reaction force (GRF), and the center of this pressure is called the center of pressure (COP). It is found that the ground reaction generated at this time has a characteristic from the center of pressure to the center of mass (COM).

In the present invention, this biomechanical property is used inversely, and the pressure center is inferred by projecting onto the ground in the vector direction of the force measured at the center of mass.

The direction of pressure center refers to the direction from the center of mass toward the pressure center. The step of estimating the pressure center path may be configured to first determine the pressure center direction, and then project in this direction to infer the pressure center position. More specifically, first, in the pressure center direction determination step, as shown in FIG. 5, the ratio of the left and right accelerations (a _x ) to the sum of the up and down accelerations (a _z ) and the gravitational acceleration (g) and the up and down accelerations are shown. The direction of the pressure center is determined by the ratio of the forward and backward acceleration (a _y ) to the sum of (a _z ) and gravity acceleration (g). When the direction of the pressure center is determined as described above, in the next pressure center position inference step, it is assumed that the center of mass is located at a height determined by multiplying a predetermined user key information by a predetermined inference constant. Projection to the ground in the direction determined in the direction determination step infers the pressure center position. The analogy constant here refers to the height of the center of mass according to the height of the user. In general, it is well known that a child's center of mass is proportionally higher than an adult's center of mass, and that a male's center of mass is proportionally higher than a woman's center of mass. For example, it is known that the center of mass of an adult male is on average 55.27% of the height, in which case the analogy constant is 0.5527. Thus, for example, by inputting the child / adult and the male / female classification information when the user key information is input, an appropriate analogy constant can be selected and used for calculation.

In order to further increase the accuracy of the pressure center position thus obtained, a pressure center position correction step may be further performed, wherein the pressure center position inferred in the pressure center position inference step is corrected by a value obtained by multiplying predetermined constants for forward and backward and left and right directions. . Here, the front-rear and left-right direction correction constants are constants capable of statistically matching the pressure center position obtained by the projection method as described above with the actual front-rear and left-right pressure centers.

In the movement type determination step, it is determined whether the vehicle is walking or traveling from the pattern of the up-down acceleration a _z graph.

6 is an example of a landing place pattern obtained as an estimated pressure center path. As shown, it can be seen that the left and right feet are alternately supporting the ground.

On the other hand, the distinction between walking and driving is that one or both feet always touch the ground in walking, while one or both feet always float from the ground in walking.

7 illustrates an example of an up and down acceleration graph with respect to time during walking and driving. In the case of the graph shown in FIG. 7 (A), it can be seen that a peak occurs when both feet touch the ground. In the case of the graph shown in FIG. 7 (B), the moment when both feet float from the ground It can be seen that a constant value section in which the vertical acceleration a _z becomes the minimum value exists. In this case, when the walking and the driving are different from each other, the patterns of the up and down acceleration (a _z ) graphs may be different from each other to determine whether the current movement of the user is walking or driving.

In the step of deriving the movement posture, the posture information including the stride length, interpolation, interpolation angle, and left and right asymmetry is derived based on the estimated pressure center path and the three-axis acceleration (a _x , a _y , a _z ). 6 will be described in more detail with reference to the example of the pressure center path of FIG. 6 and the example of the up and down acceleration during walking or driving of FIG. 7.

First, it has been described above that the appearance of the user's exercise is slightly different when walking and driving, and of course, there are also common items. As described above, in walking, one or both feet always touch the ground, and in driving, one or both feet always float from the ground. That is, there is a section in which only one foot is supported in common when walking and driving. In consideration of these points, the exercise posture derivation step includes a step of determining an intermediate support point for first determining an intermediate support point, and a step classification determination step for determining a bipedal support section, a single support section, and an airborne section, Basic information for deriving a posture while distinguishing walking and driving is formed.

First, the walking motion is described as follows. At the moment when the heel of one foot hits the ground, the toe of the other foot does not fall off the ground, ie the feet are supported. In this state, only one foot supports the ground while the other foot falls off the ground, and the other foot moves forward while slicing the air and the human body also moves forward. At the moment when the heel of the other foot touches the ground, the foot of one foot is not separated from the ground, that is, the state where both feet are supported is reestablished. In this process, while the human body is moving forward with only one foot, the person's head does not shake significantly in the vertical direction (a local minimum is formed at the vertical acceleration (a _z )). At the moment, the largest shakes in the vertical direction (a peak value is formed at the vertical acceleration a _z ).

That is, the walking motion can be divided into sections in which both feet are on the ground, sections in which only one foot is on the ground, and shaking in the vertical direction is the least while only one foot is on the ground. This aspect of the motion is well illustrated in FIG. 7A and as shown in this example, in the mid-point determination step, the local at the vertical acceleration a _z measured in the time domain when the user's motion is walking. Define the minimum as the middle support point. Also, in the section classification determining step, when the user's motion is walking, the section in which the peak value is formed in the vertical acceleration (a _z ) measured in the time domain is determined as the double support section, and the remaining section is determined as the one support section.

Next, the driving movement is solved and described as follows. Start with the moment one foot forwards to the ground (the other foot floats in the air at this moment). In this state, one foot spurts the ground and the human body moves forward with both feet floating in the air, and with both feet stir in the air, the front and rear are changed and the other foot comes forward. As soon as the other foot comes out to touch the ground, the moment of spurting the ground is reestablished. In this process, the person's head shakes most in the vertical direction at the moment of spurting the ground with one foot (the local maximum is formed at the acceleration (a _z ) in the vertical direction), while in the air while moving in the air almost in the vertical direction. It is not shaken (a constant value is formed at the vertical acceleration a _z ).

That is, the driving motion may be divided into sections in which both feet are in the air and sections in which only one foot is on the ground, and the shaking in the vertical direction is the least while both feet are in the air. This aspect of the exercise is well illustrated in FIG. 7B and as shown in this example, in the intermediate support point determination step, the local at the vertical acceleration a _z measured in the time domain when the user's movement is running. Define the maximum as the middle support point. Also, in the section classification determining step, when the user's movement is driving, the section that is represented by the constant value in the vertical acceleration a _z measured in the time domain is determined as the floating section, and the remaining section is determined as the one foot support section. Herein, the constant value appearing in the floating section is a predetermined value of the signal level level when the accelerometer does not apply any external force other than gravity, and may be appropriately determined as a value close to approximately zero. In other words, the constant value is a reference value for determining the current stance. In this sense, the constant value may be called a stance phase constant. If the interval is large, it is determined as one foot support interval.

Thus, when the basic information for deriving the attitude is derived, it is possible to derive a walking or driving posture such as stride length, interpolation, interpolation, left and right asymmetry.

Stride length: First, the average speed is calculated by measuring user location information at predetermined time intervals. Next, the walking frequency is calculated by measuring the number of intermediate supporting points during the time interval. Finally, the user's stride length can be accurately calculated by dividing the average speed by the walking frequency.

Interpolation: Interpolation in the left and right directions can be calculated using the pressure center position value corresponding to the intermediate support point. That is, by applying the time value corresponding to the intermediate support point shown in FIG. 7 (A) or (B) to the pressure center position value shown in FIG. 6, when the pressure center position corresponding to this time value is found, the left foot depresses the ground. The position and the position of the right foot are out of the ground, and the distance between them can be measured to accurately calculate the user's interpolation.

Compensation angle: The complementary angle may be calculated using a pressure center position value corresponding to a start point of the one foot support section and a pressure center position value corresponding to an end point of the one foot support section. To solve the problem, the heel of the foot support at the beginning of the foot, and the foot of the foot at the end of the foot to the ground.

In other words, to obtain the angle between the pressure center positions as described above, to obtain the angle of the heel position and the toe position of the moment when the ground is taken, that is, the angle of view, that is, the user's angle of view can be calculated accurately in this way. .

Left and right asymmetry: First, identify the foot supporting the sign of the left and right acceleration (a _x ) measured in the time domain. Next, by comparing the difference between the peak value, golgap and two of the vertical acceleration (a _z) measured in the time domain. In other words, by comparing the peak value, the valley value, and the like when the left foot is supported with the right foot, the left and right asymmetry of the user can be accurately calculated. It is also possible to calculate the repeatability of walking or driving in the same way.

FIG. 8 is an example of the acceleration signal measurement result, and it can be seen that the left and right asymmetry of the vertical acceleration a _z is strongly shown in the bottom graph of FIG. 8.

By deriving a walking or driving posture such as stride length, interpolation, interpolation, left and right asymmetry, and the like, the user can monitor in real time whether the user is walking or driving in the correct posture. Of course, at this time, the stride, interpolation, interpolation, and left and right asymmetry values corresponding to the optimal posture may be stored in advance, and the calibration amount may be calculated by comparing the current posture values currently being monitored. By notifying or storing the information to the user in real time so that the user can check later, the user can effectively correct his posture by walking or driving to a more correct posture.

9 is a flowchart illustrating a motion recognition method for walking and driving monitoring according to another embodiment of the present invention.

The motion recognition method according to the present embodiment is largely composed of two steps, that is, a data collection and motion recognition step of determining whether to walk and drive by collecting and analyzing three-axis accelerations (a _x , a _y , a _z ); And an acceleration-based motion state value derivation step of calculating the motion state values of the user's center of mass using the collected three-axis accelerations (a _x , a _y , a _z ). Hereinafter, the detailed steps of each step will be described in more detail.

As shown in FIG. 10, the data collection and motion recognition step includes a vertical acceleration collection step, a peak detection step, a motion detection step, a triaxial acceleration collection step, a Fourier transform step, and a motion shape determination step. The data collection and exercise recognition step recognizes whether the exercise is taking place in the user, and if the exercise corresponds to walking or driving.

As shown in FIG. 10, data variables to be collected are initially initialized to prepare for performing exercise recognition.

In the up-and-down acceleration collection step, instead of collecting all of the three-axis acceleration (a _x , a _y , a _z ), the up-down acceleration a _z is collected once. The collected up-and-down acceleration a _z may be used as it is, but it is more preferable to pass a noise removing step of removing noise by passing through a predetermined band pass filter. At this time, the band pass filter may be formed, for example, 0.1 to 5 Hz corresponding to the walking or running frequency of a general person, but it is apparent to those skilled in the art that this range may be appropriately changed.

In the peak detection step, the peaks of the vertical acceleration a _z thus collected are detected. In the motion detection step, the motion occurs by determining whether the vertical acceleration a _z peak value is greater than or equal to a predetermined threshold. Judge whether or not burned. If it is determined that the exercise did not occur in the motion detection step, the process returns to the initial preparation step again, the variable is initialized.

In other words, in performing the analysis in the motion recognition apparatus 100, when it is made to always collect the three-axis acceleration (a _x , a _y , a _z ), unnecessary calculation load is generated when the motion is not performed, thereby consuming power. Problems such as fever may occur. On the other hand, the most significant difference in the case of the user sitting or moving backwards, or walking or running, is the degree that the user shakes up and down, that is, the vertical acceleration (a _z ). Therefore, by first collecting the vertical acceleration (a _z ) of the user wearing the motion recognition device 100, if the value is above a certain threshold value is determined that the user is walking or running to start the full-scale motion detection from then on Therefore, the unnecessary computational load problem described above can be prevented.

In the triaxial acceleration collection step, as described above, if the vertical acceleration _az peak value is equal to or greater than a predetermined threshold value, the triaxial acceleration a _x , a _y , a _z are collected. Here, likewise, the collected three-axis accelerations (a _x , a _y , a _z ) may be used as they are, but it is more preferable to pass a noise removing step of removing noise by passing through a predetermined band pass filter. The band pass filter at this time may be formed in the same manner as the band pass filter previously used to remove the vertical acceleration a _z noise, or may be appropriately changed and set.

In the Fourier transform step, the frequency response graph is derived by Fourier transforming the three-axis accelerations (a _x , a _y , a _z ), and the frequency response graph is determined based on a predetermined frequency response modification or magnitude in the motion type determination step. In comparison with the step of determining whether the walking and driving movement or other movement. If it is determined that the exercise corresponding to the walking and driving movement has not occurred in the exercise type determination step, the process returns to the initial preparation stage again and the parameter initialization is performed. Otherwise, if it is determined that the walking and driving movement has occurred, the acceleration-based movement state value is derived. Steps will be performed.

As such, in the exercise type determination step, it is determined whether the exercise of the user is walking and driving.

11 illustrates an example of an acceleration signal measurement result when walking according to another exemplary embodiment of the present invention. On the left side of FIG. 11 is a time domain acceleration graph in which the horizontal axis is time / vertical axis (x), front and rear (y), and vertical axis (z), respectively, in three-axis accelerations (a _x , a _y , a _z ). On the right side, there is shown a frequency response graph derived through the Fourier transform step in which the horizontal axis represents the frequency and the vertical axis represents the magnitude as described above.

Of course, during walking or running, periodically shakes up and down, front and rear, left and right direction, that is, a periodic signal is generated as shown on the left side of FIG. At this time, since the left and right feet are alternately walking or running, the frequency of the periodic signal in the left and right directions has a value 1/2 of the frequency of the periodic signal in the up and down directions and the front and rear directions. It's easy to see in the graph on the right. On the other hand, one foot or both feet are always in contact with the ground when walking, while one or both feet are floating from the ground when driving. That is, when the walking or running is made, large shaking in the head must occur periodically.

In consideration of the above-described points, the exercise type determination step may be performed to determine the user's movement as a walking and driving state if the following equation is satisfied, and to determine other movements. In other words, when the following relation is easily solved and explained, it is determined that walking or driving is performed when the degree of periodic shaking in the vertical direction and the left and right directions is greater than a certain level.

M _{z, p} / M _{z, other} > c _z and M _{x, p} / M _{x, other} > c _x

(From here,

a _z : acceleration in the vertical direction,

c _z : a predetermined vertical reference threshold,

a _x : the horizontal acceleration,

M _{x, p:} the Fourier transform results in the lateral direction acceleration a _x in f _p / 2 of the center frequency and the total energy of the frequency components in the left and right directions based on the band with a bandwidth of less than 1Hz,

M _{x, other} : Sum of energy of remaining frequency components excluding the left and right reference bands from the Fourier transform result of left and right acceleration a _x ,

c _x : predetermined left and right reference threshold)

When it is determined that the user exercise is walking and driving by satisfying the above equation, it is now necessary to determine whether the exercise is walking or driving. At this time, as described above, one foot or both feet are always in contact with the ground during walking, and one or both feet are floating from the ground during driving. Here, at the time when both feet are lifted up from the ground while driving, there is no external force exerted on the user in the up and down direction because both feet are flying in the air, so the up and down acceleration (a _z ) is the minimum constant value at this time. Will be formed.

In consideration of the points described above, the step of determining the movement type determines that the user's motion is a driving state if there is a section satisfying the following equation in the Fourier transform result of the up-down acceleration a _z , otherwise walking It is made to judge the state.

a _z <k

(From here,

k: stance determination constant)

Here, the stance phase constant is a predetermined value of the signal level level when the accelerometer does not apply any external force other than gravity, and may be appropriately determined as a value close to approximately zero.

When the user's walking or driving motion is detected through the data collection and motion recognition step including the detailed steps as described above, an acceleration-based exercise state value derivation step is performed using the collected variables.

As shown, the acceleration-based motion state value derivation step includes a mass center acceleration derivation step, a center of mass velocity and a location derivation step.

In the derivation of the center of gravity acceleration, the acceleration of the user center of mass is derived by multiplying each of the three-axis accelerations a _x , a _y , a _z by a predetermined gain value. In general, when analyzing the motion of an object, it is analyzed based on the motion of the center of mass of the object. Since all the variables used in the analysis are measured in the user's head, it is referred to as the state of motion of the center of mass. To convert. This gain value may be represented by a constant vector γ, and may be previously obtained using body information such as user key information.

In the center of mass velocity and position derivation step, the speed and position of the center of mass of the user are derived using previously measured user key information, user position information, and center of mass acceleration. That is, it is possible to obtain the velocity and position of the center of mass (integrated by the integral constant value) by integrating the center of mass acceleration obtained as described above, or the velocity of the center of mass using the user position information measured in time by the position measuring sensor. You can also get the location. There is an error as much as the integral constant between these two calculated values. By comparing them properly, the velocity and position of the center of mass can be accurately calculated.

As described above, according to the present invention, it is possible to accurately determine whether the user is walking or driving by using the acceleration, the position, and the like measured at the user's head. You can see exactly how you are moving (ie how the acceleration, velocity, and position of the center of mass are represented). Therefore, based on this, it is possible to derive various elements of walking or driving posture and to use it for correcting posture.

As shown in FIG. 13, the exercise posture deriving apparatus 1300 according to the present exemplary embodiment is configured to be divided into a user's body, more specifically, a head and a waist. That is, in the exercise posture deriving apparatus 1300 according to the present embodiment, as shown in the schematic diagram of FIG. 13, the head sensor signal collecting unit 1310H worn on the head is formed to be inserted into the ear like an earphone, and the waist The waist sensor signal collector 1310W worn on the side may be formed to be plugged into a belt. Of course, the present invention is not limited thereto, and for example, the head sensor signal collecting unit 1310H may be variously modified, such as a headband shape, a glasses shape, a shape to be attached to a separate hat, a helmet shape, and the like. Of course.

Exercise posture deriving apparatus 1300 according to another embodiment of the present invention, as shown in Figure 14, the head sensor signal collector 1310H, the waist sensor signal collector 1310W and exercise posture derivation unit 1442. The exercise posture deriving apparatus 1300 further includes an exercise correction generator 1422 and a correction information output unit 1430.

The head sensor signal collection unit 1310H includes a head side triaxial acceleration sensor 1411H including up, down, left and right, front and rear, and the waist side sensor signal collection unit 1310W is configured to be vertical, horizontal, horizontal, And a position measuring sensor 1412W for measuring a user's position and a waist side three-axis acceleration sensor 1411W including front and rear.

The head side and waist side three-

axis acceleration sensors

1411H and 1411W may be appropriately selected and used among sensors generally used for measuring acceleration in three axes, such as a form in which a gyroscope is incorporated. The position measuring sensor 1412W is for measuring an absolute position of a user. For example, the position measuring sensor 1412W may be configured to measure a user's position using a GPS signal. Sensors to which this technique is applied may also be used. In addition, the head and waist

sensor signal collectors

1310H and 1310W are three-axis angular velocity sensors 1412H as shown in FIG. 14 to increase the accuracy in the motion recognition and analysis process, which will be described in more detail below. ) May be further included.

In the exercise posture deriving apparatus 1300 according to the present exemplary embodiment, the head and waist sensor

signal collecting units

1310H and 1310W are worn on the head and waist of the user as shown in FIG. Measure the user's dynamic physical quantity. In particular, in the present embodiment, when measuring the dynamic physical quantity of the user for the posture derivation, it is made to use the value measured at the position that appears most similar to the movement of the center of mass of the user's body. For example, in the present invention, the acceleration in the left and right directions is measured at the head side, the acceleration and the position in the front and rear directions are measured at the waist side, and the acceleration in the vertical direction is measured at the head side or the waist side. More specifically, in the case of the up-down acceleration, the up-down acceleration is quite accurate even when measured from the head side or the waist side, so that the value measured at either the head side or the waist side can be selectively used. Alternatively, the average value of the measured values on both sides may be used.

Of course, in general, the acceleration sensor is made to measure the acceleration in the three axis direction, that is, up and down, left and right, front and rear, and thus collected in the head sensor signal collector 1310H alone or the waist sensor signal collector 1310W alone. Of course, it is also possible to perform various calculations to be described later using the up, down, left, and right directions. However, when walking and driving are performed, the left and right movements in the head and the left and right movements of the center of mass of the user's body appear more similar, and the forward and backward movements in the waist and the front and rear movements of the center of mass of the user's body are more similar. Appears. In addition, the vertical movement is similar in the head, waist and center of mass. On the other hand, in the motion recognition method of the present invention to be described later, the motion recognition or posture derivation is made using the dynamic physical quantity at the center of mass of the user's body. Putting these things together, the left and right accelerations are measured on the head side, the front and back accelerations are measured on the waist side, and the head or waist side is appropriately selected to measure the vertical acceleration or both sides. By calculating the acceleration in the vertical direction as a value obtained by the average value, it is possible to obtain the effect that the final motion recognition, posture derivation, and the like are much more accurate.

The exercise posture deriving unit 1421 receives a signal from the head and waist sensor

signal collecting units

1310H and 1310W, and uses the 3-axis acceleration and position signals to determine the acceleration, velocity, and position of the user's center of mass. Deriving a walking or driving exercise state value, and analyzing the walking or driving exercise state value serves to derive a walking or driving posture. Specifically, the exercise posture deriving unit 1421 estimates a pressure center path from the walking or driving motion state value, and derives a walking or driving posture by analyzing the pressure center path. The analysis algorithm of the present invention used by the exercise posture deriving unit 1421 will be described in more detail later, and thus description thereof is omitted.

Meanwhile, the exercise posture deriving unit 1421 may be formed in an integrated circuit that may perform various calculations, and may be formed on one substrate as one body with the waist sensor signal collecting unit 1310W, or may be a separate computer. It may also be in the form. In addition, the head and waist sensor

signal collecting unit

1310H and 1310W may be provided with a head communication unit 1413H and a waist communication unit 1413W, respectively, for signal transmission with the exercise posture deriving unit 1421. . When the waist sensor signal collecting unit 1310W is integrally formed with the exercise posture deriving unit 1421, the waist side communication unit 1310W may be directly connected to the exercise posture deriving unit 1421 to transmit a signal, or The waist communication unit 1413W may serve to receive a signal transmitted from the head communication unit 1413H and transmit the signal to the exercise posture deriving unit 1421. Head and waist communication unit (1413H) (1413W) may be made of a wire, or may be made to transmit a signal using at least one wireless communication selected from Bluetooth, Wi-Fi, NFC to increase user convenience. .

The exercise correction generating unit 1422 generates the posture correction information by comparing the walking posture derived by the exercise posture deriving unit 1421 and the reference posture. The exercise posture deriving unit 1421 derives a walking or driving posture of the user based on the signals collected by the head and waist sensor

signal collecting units

1310H and 1310W as described above. The driving direction, the speed, and the like of the user can be derived while driving, and the stride length, which is one of the elements of the walking posture, can be obtained from this. In this case, the exercise correction generator 1422 embeds optimal key stride relationship data for each walking and traveling speed, and compares the walking posture information of the corresponding user with an excessively wide stride length compared to the height of the corresponding user. It is easy to determine whether it is narrow and to calculate the amount of stride correction that should be reduced or increased if it is out of the optimum range.

The correction information output unit 1430 converts and outputs the posture correction information generated by the exercise correction generator 1422 as information recognizable by a user including sound, illustration, and image. For example, if it is necessary to reduce the stride length by calculating the stride length correction amount, a voice such as "Reduce stride length" is output through a speaker provided on the exercise posture derivation device 1300, or a warning sound sounds. The user may be not aware of the optimal stride length and may be encouraged to change the walking posture. In particular, in this configuration, since the calibration information output unit 1430 is integrally formed with the head sensor signal collection unit 1310H, it is preferable that the calibration information output unit 1430 is disposed close to the user's information collection organ, that is, the eye, the ear, and the like, to transmit the information. Do. That is, as a specific example, when the head sensor signal collection unit 1310H is formed in the form of earphone plug into the ear as shown in the example shown in Figure 13, and the output information type is a voice, sound signal, the efficiency of calibration information transmission is maximized Can be. Alternatively, the present invention may be realized in various forms, such as connected to a smartphone, a computer, a dedicated display, or the like, so that accurate calibration information may be output as an illustration or an image.

In addition, the exercise posture deriving apparatus 1300 may be configured to transmit and store the walking posture derived by the exercise posture deriving unit 1421 to an external database 1440 to accumulate the posture. A user who needs such walking or driving motion analysis may be a general person who performs daily walking or jogging to promote health, or an expert who trains to improve physical ability. Naturally, it is desirable to be able to see the change. In addition, when the exercise analysis data is accumulated and stored in a large amount as described above, such data may be used as big data and used for various statistics or analysis, and thus various applications are possible.

In the exercise posture derivation apparatus 1300 according to the present exemplary embodiment as described above, specific and comprehensive examples of the movement information recognition through the exercise posture derivation and the output of the correction information are as follows. First, the left and right acceleration is collected at the head side, the front and rear acceleration and the position are collected at the waist side, and the up and down acceleration is collected at the head side or the waist side, similar to the movement of the center of mass of the user's body as described above. .

The waist sensor signal collector 1310W may be integrally formed with the exercise pose derivation unit 1421, and thus, the physical quantities collected by the head sensor signal collector 1310H may be derived through the head communication unit 1413H. It is transmitted to the part 1421 side. At this time, the waist side communication unit 1413W provided in the waist sensor signal collecting unit 1310W receives the signal and transmits the signal to the exercise posture deriving unit 1421.

Using the physical quantities such as the acceleration and the position collected in this way, the exercise posture deriving unit 1421 derives the walking and driving postures of the user. The exercise correction generator 1422 generates the posture correction information by comparing the derived actual priest with the ideal reference posture. The exercise posture deriving unit 1421 and the exercise correction generating unit 1422 may also be integrally formed, that is, they are integrally formed with the waist sensor signal collecting unit 1310W.

In order for the posture correction information thus generated to be effectively delivered to the user, it is preferable that the information is transmitted from the head side close to the eyes, ears, and the like, which are the user's information collection institutions. As described above, when the head sensor signal collection unit 1310H and the calibration information output unit 1430 are integrally formed, the generated posture correction information sequentially processes the waist communication unit 1413W and the head communication unit 1413H. The calibration information output unit 1430 is delivered to the calibration information output unit 1430 to effectively transmit the calibration information in the form of delivering a voice message to the user's ear.

According to another exemplary embodiment of the present invention, a method for deriving an exercise posture may be performed by using the exercise posture derivation apparatus 1300 to detect a user's exercise and determine whether the user is walking or driving. At this time, as described above, the analysis algorithm used in the present invention uses a dynamic physical quantity measured in the user's head and waist bar, the exercise posture derivation apparatus 1300 is at least a head including a top, bottom, left and right, front and rear The three-axis direction acceleration sensor 1411H side, the three-axis direction acceleration sensor 1411W, including the up, down, left, right, front and rear, and a position measuring sensor 1412W for measuring the user's position, the exercise posture In the deriving unit 1421, an analysis algorithm to be described below may be performed. In addition, the exercise posture deriving apparatus 1300 may further include various additional components described above for improving the function of the apparatus.

Injury risk quantification device 1500 according to the present embodiment is a device for informing the user of the risk of injury that may occur during walking or driving. More specifically, it is as follows. When walking or driving, it is well known that a variety of reasons, such as a bad posture or a hard ground, can cause a strain on the ankle, knee, waist, etc., which may lead to injury. In order to prevent such a risk, conventionally, only a measure such as wearing a functional sneaker such as shock absorbing, etc., has not been an accurate indicator to know how much a risk of injury actually occurs. In the present embodiment, the risk of injury is quantified as a determination index, and when the risk of injury rises above a certain level, the user is notified of the degree of danger as an alarm. This allows the user to take appropriate measures to stop walking or driving properly, correct posture, change shoes, change walking or driving course before injury, and ultimately occur during walking or driving. This can greatly reduce the risk of injury.

The injury risk quantification device 1500 according to the present embodiment includes a sensor signal collection unit 1510, a controller 1520, and an alarm unit 1530. The injury risk quantification device 1500 may further include a database 1540.

The sensor signal collector 1510 includes an acceleration sensor 1511 and is worn on the upper body except the user's arm. The sensor signal collection unit 1510 may be a single dog or a plurality of dogs. The sensor signal collector 1510 may be formed in two and may be worn on each of the user's head and waist. In this case, the sensor signal collector that is worn on the user's head may be the head sensor signal collector 1510H and the user's waist. The sensor signal collector worn on the side may be divided into the waist sensor signal collector 1510W. As a specific example of the wearing state, the head sensor signal collection unit 1510H worn on the head side is made of an ear plug like earphone, and the waist sensor signal collector 1510W worn on the waist side is plugged into a belt. It may be made in the form. Of course, the present invention is not limited thereto, and for example, the head sensor signal collecting unit 1510H may be variously modified, such as a headband shape, a glasses shape, a form attached to a separate hat, a helmet shape, and the like. Of course. In addition, although not shown in the drawing, the sensor signal collector 1510 may be worn anywhere except the user's arm, for example, to be worn or inserted into the breast pocket of the garment to be worn on the chest, Various modifications can be made, such as a form worn using a separate vest or harness.

The sensor signal collector 1510 includes an acceleration sensor 1511 as described above. The acceleration sensor 1511 may generally employ and select an appropriate one from among sensors used to measure acceleration in the three axis direction, such as a form in which a gyroscope is embedded. On the other hand, the sensor signal collector 1510 may be provided with a controller 1520 that performs a role such as performing a calculation and controlling the acceleration data signal collected by the acceleration sensor 1511 directly. Alternatively, the controller 1520 may be implemented in various ways, such as to be implemented in the form of an app on a conventionally used smartphone. That is, when the control unit 1520 is implemented as a separate device from the sensor signal collection unit 1510, the sensor signal so that the acceleration data signal collected by the acceleration sensor 1511 can be smoothly transmitted to the control unit 1520. The collection unit 1510 may further include a communication unit 1512. Such signal transmission may be made by wired communication through wiring, or may be made by wireless communication such as Bluetooth, Wi-Fi, NFC, etc., and may be adopted by selecting an appropriate form according to required conditions or required performance. .

Although it will be described in more detail in the description of the injury risk quantification method according to the present embodiment, the present embodiment uses the vertical acceleration in determining the risk of injury.

In addition, the present embodiment uses the vertical acceleration to quantify the risk of injury. In general, when traveling, the left and right movements in the head and the left and right movements of the center of mass of the user's body appear more similarly, and the back and forth movements in the waist and the forward and backward movements of the center of mass of the user's body are relatively similar. appear. In addition, the vertical movement is similar in both the upper body and the center of mass, including the head to the waist. However, the arm part of the upper body, except for the movement of the center of mass in addition to the movement in the forward and backward direction, the arm is excluded. In consideration of this point, the acceleration in the vertical direction may be measured in any of the upper body except the arm. Incidentally, the acceleration in the up and down direction is quite accurate even when measured in the upper body except the arm, so that the value measured at either the head side or the waist side may be selectively used, or the measured value at both sides may be The average value may be appropriately selected.

The control unit 1520 receives the signal from the sensor signal collection unit 1510, derives at least one injury risk determination index calculated based on the vertical acceleration (a _z ), and using the injury risk determination index It determines and controls the occurrence of alarm. More specifically, the control unit 1520 derives at least one is a risk judgment indicators to the portion, the average slope, the selection of the maximum slope, the maximum impact force, the impulse of the vertical acceleration (a _z) in the vertical direction acceleration (a _z) Through this, the risk of injury is quantified and the degree of danger is determined. Derivation of the injury risk determination index performed by the controller 1520 will be described in more detail with reference to the method of quantifying injury risk according to the present embodiment.

The actual implementation form of the controller 1520 may be formed in various ways depending on the need or purpose. That is, the controller 1520 may be formed on one substrate as an integrated circuit form capable of performing various calculations and integrally with the sensor signal collector 1510, or a separate dedicated device (that is, for quantifying injury risk) It may be made in the form of an independent device (manufactured only as a) or a separate computer, etc.), or may be implemented in the form of an app on a smart phone that is being used as described above. As described above, when the controller 1520 is integrally formed with the sensor signal collector 1510, the controller 1520 may be configured to receive a signal directly from the acceleration sensor 1511. On the other hand, if the control unit 1520 is formed independently of the sensor signal collection unit 1510, such as in the form of a separate device or a smart phone app is made to receive a signal from the acceleration sensor 1511 by wire or wireless communication Can be.

The alarm unit 1530 receives an alarm generation control signal from the control unit 1520, and serves to alert the user of a risk of injury. The controller 1520 derives at least one injury risk determination index calculated based on the up and down acceleration a _z and uses the same to determine whether an alarm has occurred. In this case, the user is notified of the danger by controlling the alarm to be generated.

The alarm unit 1530 outputs an alarm signal as information that can be recognized by a user including sound, illustration, and image. For example, when the alarm unit 1530 is in the form of a speaker that outputs sound, a warning sound may be sounded if the risk of injury is greater than a reference level. Alternatively, when the device according to the present embodiment is applied to augmented reality glasses such as Google Glass, the alarm unit 1530 outputs a red warning figure or an image flickering on the augmented reality glasses, or "the risk of injury is several percent. It is also possible to output a message such as "." Alternatively, the alarm unit 1530 may be implemented as a thermoelectric element and may be in direct or indirect contact with the user's skin. If the risk of injury is greater than or equal to the standard, the alarm unit 1530 may alert the user by being cold or hot. As another example, when the user is visually impaired, the alarm unit 1530 may be configured in a form that can be changed by braille to be recognized by the sense of touch. As such, the alarm unit may be formed in any form as long as the alarm unit can output the alarm signal as information that can be recognized by the user.

In addition, the injury risk quantification device 1500 may be configured to accumulate and store injury risk data including an injury risk alarm occurrence time point and an injury risk determination index value at the corresponding time point to an external database 1540. A user who needs such walking or driving motion analysis may be a general person who performs daily walking or jogging to promote health, or an expert who trains to improve physical ability. Naturally, it is desirable to be able to see the change. In addition, when the exercise analysis data is accumulated and stored in a large amount as described above, such data may be used as big data and used for various statistics or analysis, and thus various applications are possible.

The injury risk quantification method according to the present embodiment includes an acceleration sensor 1511 as described above, and is measured by using at least one sensor signal collector 1510 worn on the upper body except the user's arm. Injury risk determination indicators are used to quantify the risk of injury using directional acceleration (a _z ). To this end, the injury risk quantification method according to the present embodiment includes a data collection step, a determination index derivation step, an injury risk determination step, and an injury risk warning step. In addition, the method may further include a noise removal step to increase the accuracy of derivation of the injury risk determination index. Referring to each step shown in Figure 16 in more detail as follows.

In the data collection step, the up-down acceleration a _z measured by the sensor signal collector 1510 is collected. The collected up-and-down acceleration a _z may be used as it is, but it is more preferable to pass a noise removing step of removing noise by passing through a predetermined band pass filter. In this case, the band pass filter may be formed at, for example, 0.1 to 5 Hz, which corresponds to a walking or running frequency of a general person, but, of course, the range may be appropriately determined.

In the determination index deriving step, at least one injury risk determination index calculated based on the vertical acceleration a _z is derived. At this time the injury risk judgment indicators may be a vertical direction acceleration (a _z) the average slope, maximum slope of the vertical acceleration (a _z), the maximum impact force, at least one selected from the amount of impact. Each judgment indicator will be described in more detail later.

In the injury risk determination step, it is determined whether the injury risk determination index is greater than a predetermined criterion. At this time, the injury risk determination index may be a plurality as described above, may generate an alarm when any one of the multiple determination index is above the reference, may generate an alarm when all above the reference, or appropriately prioritized You can also rank and raise alarms in stages. In the injury risk determination step, if the injury risk determination index is smaller than a predetermined criterion, the alarm returns to the data collection step without generating an alarm. In addition, in the injury risk determination step, it is preferable to make the determination using the injury risk determination index calculated by collecting data of at least two cycles with respect to the up-down acceleration (a _z ) data represented by the periodic signal. .

In the injury risk warning step, when at least one of the injury risk determination indexes is greater than a predetermined criterion, the user is alerted of the injury risk. As described above, the warning form of the injury risk can be in various forms such as sound, illustration, image, and the like, and the user is actively alerted to the action to reduce the risk of injury (end of exercise, posture correction, shoe replacement, course change). Etc.) can ultimately greatly reduce the risk of injury.

Hereinafter will be described in more detail the process of deriving several examples and each of the injury risk determination indicators used in the present invention.

As shown, the up-and-down acceleration a _z appears in a periodic form with respect to time (which is natural because walking or driving itself is a periodic movement). The driving motion is described as follows. Start with the moment one foot forwards to the ground (the other foot floats in the air at this moment). In this state, one foot spurts the ground and the body of the person moves forward with both feet floating in the air, and with both feet stir up in the air, the other foot changes forward and the other foot comes forward. As soon as the other foot comes out to touch the ground, the moment of spurting the ground is reestablished. In this process, the person's head shakes most in the vertical direction at the moment of spurting the ground with one foot (the local maximum is formed at the acceleration (a _z ) in the vertical direction), while in the air while moving in the air almost in the vertical direction. It is not shaken (a constant value is formed at the vertical acceleration a _z ).

The moment the foot hits the ground, the most impact is applied to the joint, and this impact is shown in the form of the first peak in the vertical acceleration graph as shown in FIG. 17. The risk of injury will vary depending on the degree of impact at this time, and in the present invention, it was used as the basis of quantified judgment by indexing it. As such a surface is determined, in the present invention, as described above, it uses the maximum slope, the maximum impact force, the impulse of the vertical direction acceleration average slope, vertical acceleration (a _z) of (a _z).

18 is a view illustrating an inclination on a vertical acceleration graph when driving in accordance with another embodiment of the present invention. This will be described the process of deriving the average slope and the maximum slope of the vertical acceleration (a _z ).

First, when the injury risk determination index is selected as the average inclination value of the vertical acceleration (a _z ), the injury risk determination index is calculated using the following equation.

(Where, a _z : up and down acceleration, mean: average value calculation function, i: index number, t _i : i-th time, t _i _-1 : i-1 th time, t _c : shock start time, t _m : Impact end time)

The impact start time actually means the moment the foot lands on the ground. This may be determined as a time when the vertical acceleration a _z rises upward through a predetermined reference value (for example, 0.3 m / s ² ) near zero at a value of zero or less. Herein, the specific value of the reference value for determining the impact start time may be appropriately determined from a value of 0.5 m / s ² or less as in the above-described example. The impact end time is the point at which the first peak value appears and can be easily seen intuitively on the graph. The index i is an index of times digitized by dividing the time from the impact start time to the impact end time by n, and n may be appropriately determined as necessary.

The average slope value is thus the average value of the n slope values obtained at each interval when n is divided from the impact start time to the impact end time. FIG. 18 illustrates a vertical acceleration a _z graph in one period, and the average slope value as described above may be obtained in one such period. On the other hand, as shown in FIG. 17, while driving, the graph of the form as shown in FIG. 18 is repeatedly repeated, and the average slope value as described above may be obtained for each period (that is, for each step). At this time, in the determination index deriving step, an average vertical loading rate calculated as a product of the user mass m and the average slope may be further calculated.

On the other hand, when the injury risk determination index is selected as the maximum inclination value of the vertical acceleration (a _z ), the injury risk determination index is calculated using the following equation.

i = 1, 2,... , n,

t ₀ = t _c , t _n = t _m

(Where, a _z : up and down acceleration, max: maximum calculation function, i: index number, t _i : i time, t _i _-1 : i-1 time, t _c : shock start time, t _m : Impact end time)

The maximum slope is the maximum of n slope values obtained between the shock start time and the impact end time in one period (one step) as described in the description of the average slope. At this time, in the determination index derivation step, the instantaneous vertical loading rate calculated as the product of the user mass m and the maximum slope may be further calculated.

19 illustrates an impact amount on an up and down acceleration graph while driving according to another embodiment of the present invention. This section describes the process of deriving the maximum impact force and the amount of impact.

First, when the injury risk determination index is selected as the maximum impact force value, the injury risk determination index is calculated using the following equation.

(Here, _az : acceleration in the up and down direction, m: user mass, t _m : end time of impact)

As described above, since the impact end time is the time at which the first peak value appears, the time at which the maximum impact force is naturally shown is the impact end time. In FIG. 19, the first peak (1st peak) of the up-down acceleration a _z is indicated, and the value multiplied by the user's mass (m) is the maximum impact force value.

On the other hand, when the injury risk determination index is selected as the impact amount value, the injury risk determination index is calculated using the following equation.

(Here, a _z : up and down acceleration, m: user mass, t _c : shock start time, t _m : shock end time)

In FIG. 19, the up-down acceleration a _z graph area between the impact start time and the impact end time is shown, and the value obtained by multiplying the area by the user's mass (m) is the impact amount value.

Although the preferred embodiment of the present invention has been described in detail above, the scope of the present invention is not limited thereto, and various modifications and equivalent other embodiments are possible. Therefore, the true technical protection scope of the present invention will be defined by the appended claims.

For example, an apparatus according to an exemplary embodiment of the present invention may include a bus coupled to units of each of the apparatus as shown, at least one processor coupled to the bus, and a command, received And a memory coupled to the bus for storing messages or generated messages, and coupled to at least one processor for performing instructions as described above.

In addition, the system according to the present invention can be embodied as computer readable codes on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored. The computer-readable recording medium may include a magnetic storage medium (eg, ROM, floppy disk, hard disk, etc.) and an optical reading medium (eg, CD-ROM, DVD, etc.). The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

Claims

In the motion recognition method using a motion recognition device made of a three-axis acceleration sensor including the up, down, left, right, front and rear and a position measuring sensor for measuring the user position and worn on the user's head,

A data collection and motion recognition step of determining whether to walk and drive by collecting and analyzing three-axis acceleration;

And an acceleration-based exercise state value deriving step of calculating the exercise state values of the user's center of mass using the collected three-axis accelerations.
The method of claim 1, wherein the data collection and exercise recognition step,

A vertical acceleration collection step of collecting vertical acceleration;

A peak detection step of detecting peaks of the collected up-down acceleration;

A motion detecting step of determining whether the vertical acceleration peak value is equal to or greater than a predetermined threshold value;

A three-axis acceleration collection step of collecting three-axis acceleration when the vertical acceleration peak value is equal to or greater than a predetermined threshold value;

A Fourier transform step of Fourier transforming the three-axis acceleration to derive a frequency response graph;

And an exercise type determination step of comparing the frequency response graph with a predetermined frequency response modifying or magnitude criterion to determine whether it is a walking and driving exercise or other exercise.
The method of claim 2, wherein the exercise recognition method comprises:

Before the peak detection step or the Fourier transform step,

And a noise removing step of removing the noise by passing the vertical acceleration or the three-axis acceleration through a predetermined band pass filter.
The method of claim 2, wherein the determining of the type of exercise comprises:

If the following equation is satisfied, the user's motion is judged to be a walking and running state, and if not, it is determined as other motion.

M z, p / M z, other > c z and M x, p / M x, other > c x

(From here,

a z : acceleration in the vertical direction,

f p : frequency with the maximum magnitude in the Fourier transform result of the up-down acceleration a z ,

M z, p : Vertical acceleration in a Fourier transform result of a z a f p to the center frequency and the total energy of the frequency components in the vertical direction based on the band with a bandwidth of less than 1Hz,

M z, other : Sum of energy of remaining frequency components excluding the vertical reference band from the Fourier transform result of vertical acceleration a z ,

c z : a predetermined vertical reference threshold,

a x : the horizontal acceleration,

M x, p : The Fourier transform results in the lateral direction acceleration a x in f p / 2 of the center frequency and the total energy of the frequency components in the left and right directions based on the band with a bandwidth of less than 1Hz,

M x, other : Sum of energy of remaining frequency components excluding the left and right reference bands from the Fourier transform result of left and right acceleration a x ,

c x : predetermined left and right reference threshold)
The method of claim 4, wherein the determining of the type of exercise comprises:

When the user's exercise is walking and driving,

And a user's motion is judged to be a driving state when there is a section satisfying the following equation in the Fourier transform result of the up-down acceleration.

a z <k

(From here,

k: stance determination constant)
The method of claim 1, wherein the acceleration-based exercise state value deriving step comprises:

A center of mass acceleration derivation step of deriving an acceleration of the center of mass of the user by multiplying each value of the triaxial accelerations by a predetermined gain value;

And a center of mass velocity and a position derivation step of deriving a velocity and a position of the center of mass of the user using previously measured user key information, user position information, and center of mass acceleration.
A sensor signal collector including a three-axis acceleration sensor including up, down, left, and right, and a position measuring sensor measuring a user position and worn on a user's head;

Receiving a signal from the sensor signal collection unit, using the three-axis acceleration and position signals to derive a walking or driving motion state value including the acceleration, speed, position of the user's center of mass, the walking or driving motion state value Exercise recognition device comprising a movement posture derivation unit for deriving a walking or driving posture by analyzing the.
The method of claim 7, wherein the sensor signal collector,

Motion recognition device further comprises a three-axis angular velocity sensor.
The method of claim 7, wherein the position measuring sensor,

Exercise recognition device, characterized in that for measuring the user position.
The apparatus of claim 7, wherein the exercise recognition device comprises:

And a workout correction generator configured to compare the walking posture derived by the exercise posture deriving unit and the reference posture to generate posture correction information.
The apparatus of claim 10, wherein the exercise recognition device comprises:

And a correction information output unit which converts and outputs posture correction information generated by the exercise correction generation unit as information recognizable by a user including sound, illustration, and image, and outputs the correction information.
The apparatus of claim 7, wherein the exercise recognition device comprises:

The exercise recognizing apparatus, characterized in that the walking posture derived by the exercise posture deriving unit is stored in a cumulative transmission to an external database.
The apparatus of claim 7, wherein the exercise recognition device comprises:

The signal transmission between the sensor signal collector and the exercise posture derivation unit,

Exercise recognition device characterized in that made by at least one wireless communication selected from Bluetooth, Wi-Fi, NFC.
In the exercise posture deriving method using the exercise posture deriving device which comprises a three-axis acceleration sensor including the up, down, left, right, front and rear and a position measuring sensor for measuring the user position, which is worn on the user's head,

The pressure center path is projected on the ground in the direction of acceleration vector from the center of mass position and the pressure center path is calculated using the motion state values of the user's center of mass calculated using the 3-axis acceleration collected by the exercise attitude deriving device. Estimating step;

An exercise type determination step of determining whether the user is walking or traveling from the pattern of the up-down acceleration graph;

And a movement attitude derivation step of deriving attitude information including stride length, interpolation, interpolation, and left and right asymmetry based on the estimated pressure center path estimate and three-axis acceleration.
The method of claim 14, wherein the step of estimating the pressure center path,

A pressure center direction determining step of determining the direction of the pressure center by the ratio of the left and right accelerations to the sum of the vertical accelerations and the gravity accelerations, and the ratio of the front and rear accelerations to the sum of the vertical accelerations and the gravity accelerations;

Assuming that the center of mass is located at a height determined by multiplying a predetermined user's key information by a predetermined inference constant, the pressure center projecting the pressure center position by projecting onto the ground in the direction determined in the pressure center direction determining step Pressure center path-based exercise posture deriving method comprising the step of inferring position.
The method of claim 15, wherein the step of estimating the pressure center path,

And a pressure center position-based exercise posture derivation step of correcting the pressure center position inferred in the pressure center position inference step by a value obtained by multiplying a predetermined front-rear and left-right direction correction constants.
The method of claim 14, wherein the deriving the exercise posture,

If the user's motion is driving, the local maximum is defined as the intermediate support point in the vertical acceleration measured in the time domain. If the user's motion is walking, the local minimum is defined as the intermediate support point in the vertical acceleration measured in the time domain. Intermediate support point determination step;

If the user's motion is driving, the section that is represented by constant value in the up-down acceleration measured in the time domain is determined as the floating section, and the remaining section is determined as the one-foot support section. A pressure center path-based exercise posture derivation method comprising determining a section in which a peak value is formed in the vertical acceleration as a double support section, and determining a remaining section as a single support section.
The method of claim 17, wherein the deriving the exercise posture,

Calculating an average speed by measuring user location information at predetermined time intervals;

Calculating a walking frequency by measuring the number of intermediate supporting points during the time interval;

And calculating the stride length by dividing the average speed by the walking frequency to calculate the stride length in the front-rear direction.
The method of claim 17, wherein the deriving the exercise posture,

Pressure center path-based exercise posture derivation method, characterized in that for calculating the interpolation in the left and right directions using the pressure center position value corresponding to the intermediate support point.
The method of claim 17, wherein the deriving the exercise posture,

A pressure center path-based exercise posture derivation method comprising calculating a complementary angle using a pressure center position value corresponding to a start point of the one foot support section and a pressure center position value corresponding to an end point of the one foot support section.
The method of claim 17, wherein the deriving the exercise posture,

Identifying a foot supported by the sign of the left-right acceleration measured in the time domain;

Comprising the step of comparing the peak value, the bone value and the difference value between the two accelerations measured in the time domain, the pressure center path-based exercise attitude derivation method characterized in that it calculates the left and right asymmetry and repeatability.
A sensor signal collector including a three-axis acceleration sensor including up, down, left, and right, and a position measuring sensor measuring a user position and worn on a user's head;

Receiving a signal from the sensor signal collection unit, using the three-axis acceleration and position signals to derive a walking or driving motion state value including the acceleration, speed, position of the user's center of mass, the walking or driving motion state value A pressure center path-based exercise posture deriving device comprising: an exercise posture deriving unit configured to estimate a pressure center path from the head, and to analyze the pressure center path to derive a walking or driving posture.