Gait analysis device, gait analysis method, and computer program product

Abstract

A gait analysis device, a gait analysis method, and a gait analysis program. In the gait analysis device, a step frequency change with respect to a running speed is acquired with respect to running of a subject. A step length change with respect to a running speed is acquired with respect to running of the subject. A determination section (80) calculates a principal component score from the slope of the step frequency change and the slope of the step length change of the subject, based on a principal component analysis model that is generated in advance based on measured values of a plurality of runners, and determines which of a plurality of gait types including a stride type and a step frequency type the running of the subject conforms to, based on the calculated principal component score.

Description

TECHNICAL FIELD

[0001] The present application relates to a technique for analyzing a running method of a marathon runner, and particularly relates to a running method analysis device, a running method analysis method, and a running method analysis program. BACKGROUND

[0002] In long-distance running such as a marathon, as a running method of a runner, there are known a "cadence running method" and a "stride running method". In general, the former is a running method characterized by a relatively large number of steps per unit time (also referred to as "cadence" or "pace") and a relatively short step length (distance of one step, also referred to as "stride"), and the latter is a running method characterized by a relatively small number of steps per unit time and a relatively long step length, but there is no clear definition.

[0003] In recent years, running shoes suitable for the cadence running method and running shoes suitable for the stride running method have also been developed. Therefore, a runner can sometimes select a more suitable shoe by understanding whether his or her running method is of the cadence type or the stride type (for example, refer to Patent Literature 1).

[0004] PRIOR ART DOCUMENTS

[0005] PATENT LITERATURE

[0006] Patent Literature 1: Japanese Patent No. 4856427 SUMMARY OF THE INVENTION

[0007] PROBLEMS TO BE SOLVED BY THE INVENTION

[0008] However, in the past, except for cases in which the tendency of the step frequency or the step length is particularly remarkable, there was no clear criterion for determining which of the cadence type and the stride type a running method of a runner conforms to, and it was only possible to rely on subjective judgment.

[0009] In this case, the present inventors have analyzed running records of a large number of runners, and as a result, have found a method for determining the two based on the tendency of the running method according to an objective criterion.

[0010] The present application was completed in view of this problem, and aims to provide a technique for analyzing a running method of a runner.

[0011] MEANS OF SOLVING THE PROBLEM

[0012] To solve the problem, a gait analysis device according to an aspect of the present application includes: a step frequency acquisition section that acquires a slope of a change in a step frequency change with respect to a change in a running speed with respect to a running of a subject; a step length acquisition section that acquires a slope of a change in a step length change with respect to a change in the running speed with respect to the running of the subject; a determination section that calculates a principal component score from the slope of the step frequency change and the slope of the step length change of the subject according to a principal component analysis model that is generated in advance on the basis of measured values of a plurality of runners, and determines which one of a plurality of gait types including a stride type and a step type the running of the subject conforms to according to the calculated principal component score; and a result output section that outputs a result of the determination.

[0013] Another aspect of the present application is a gait analysis method. The method includes: a process of acquiring a slope of a change in a step frequency change with respect to a change in a running speed with respect to a running of a subject; a process of acquiring a slope of a change in a step length change with respect to a change in the running speed with respect to the running of the subject; a process of calculating a principal component score from the slope of the step frequency change and the slope of the step length change of the subject according to a principal component analysis model that is generated in advance on the basis of measured values of a plurality of runners; a process of determining which one of a plurality of gait types including a stride type and a step type the running of the subject conforms to according to the calculated principal component score; and a process of outputting a result of the determination.

[0014] Further, any combination of the above-described elements or replacement of elements or expressions of the present application with each other between a method, a device, a program, a storage medium in which a program is stored, a system, and the like is also effective as an aspect of the present application.

[0015] Effects of the Invention

[0016] According to the present application, a gait of a runner can be analyzed simply to obtain information that is useful for a user. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 is a diagram showing a basic structure of a gait analysis system.

[0018] Figure 2 is a diagram comparing a step length change and a step frequency change between a stride type and a step type in the same running speed range.

[0019] Figure 3 is a diagram illustrating a relationship between a slope of a step frequency change and a slope of a step length change at a plurality of running speeds.

[0020] Figure 4 is a functional block diagram showing a basic structure of a user terminal.

[0021] Figure 5is a functional block diagram showing the basic structure of the running style analysis server.

[0022] Figure 6 is a graph showing the distribution of the principal components calculated by principal component analysis.

[0023] Figure 7 is a graph showing the distribution of the first principal component calculated by principal component analysis.

[0024] Figure 8 is a flowchart showing the basic processing in the running style analysis server.

[0025] Figure 9 is a graph showing the relationship between the distribution of the first principal component calculated by principal component analysis and the range of the running style type.

[0026] [Explanation of symbols]

[0027] 10: user terminal

[0028] 16: wearable device

[0029] 20: running style analysis server

[0030] 30: running style analysis system

[0031] 74: step frequency acquisition section

[0032] 75: step length acquisition section

[0033] 76: data analysis section

[0034] 77: principal component analysis section

[0035] 78: average value calculation section

[0036] 80: determination section

[0037] 82: model storage section

[0038] 83: score calculation section

[0039] 84: determination processing section

[0040] 90: output section

[0041] 92: result output section

[0042] 94: recommendation output section DETAILED DESCRIPTION

[0043] Hereinafter, the present application will be described based on suitable embodiments with reference to the respective drawings. In the embodiments, variations, the same or equivalent constituent elements are denoted by the same symbols, and repeated description is appropriately omitted. In the respective drawings, parts that are not important in explaining the embodiments are omitted.

[0044] Here, the "running method analysis device" mentioned in the technical solution can be realized by a combination of a server program executed on a network server or in the cloud and a server, or by a combination of a program executed on an information terminal such as a smartphone or tablet or a personal computer and the device. Alternatively, it can also be realized by a combination of a program executed on a wearable device including various built-in sensors and the wearable device. In the following embodiments, examples of a running method analysis device realized by a combination of a server program and a server and a running method analysis system including a user's terminal or wearable device are described.

[0045] (First Embodiment)

[0046] In this embodiment, it is assumed that a user who is a runner himself / herself performs analysis of a running method by wearing a running shoe (hereinafter referred to as "shoe") for a step frequency type running method or a shoe for a step length type running method. First, the user wears various wearable devices while running, acquires information required for analysis through various sensors, and transmits the information to the user's terminal. Then, the user transmits the information from the terminal to a server and obtains the analysis result from the server.

[0047] Figure 1 A basic structure of a running method analysis system is shown. The running method analysis system 30 includes, for example, a user terminal 10, a wearable device 16, and a running method analysis server 20. The user wears the wearable device 16 such as a running watch 12 or a motion sensor 14 on the arm or waist, performs running, and acquires various detection data through the running watch 12 or the motion sensor 14. The running watch 12 or the motion sensor 14 includes a sensor such as a positioning module or a 9-axis motion sensor. The running speed is acquired based on the relationship between the time information and the position information detected by the positioning module, and the step frequency is acquired based on the information detected by the 9-axis motion sensor. In addition, the step length is acquired based on the running distance measured by the positioning module and the step frequency (step length = running distance ÷ step frequency). The running log acquired by the user terminal 10 is transmitted to the running method analysis server 20 via a network 18, the running method is analyzed by the running method analysis server 20, and it is determined which of a plurality of running methods including a step frequency type and a step length type is appropriate.

[0048] In a modification, instead of the wearable device 16, a positioning module or a motion sensor built into a smartphone serving as the user terminal 10 can be used. In other modifications, as data representing the running state of the examinee, a pattern in which running speed and step frequency are acquired from an image captured by a high-speed camera by a technique such as motion capture or ground reaction force detection using a force plate can be provided. In this case, the running state data for the examinee can be acquired by an operator other than the user (for example, a clerk of a store) operating the user terminal 10, and the running analysis server 20 can be caused to perform running analysis.

[0049] The information on "step frequency" is a value in units of steps per second (Hz) or steps per minute (spm), and in the case where a runner who runs a full course of a marathon in 3 hours and 30 minutes runs at a race pace, the steps per minute are usually limited to a range of 175 spm to 205 spm on average. In addition, the information on "step length" is an average step length (m) obtained by dividing the running distance for one minute by the steps per minute. Furthermore, the data on step frequency and step length as the analysis target in the present embodiment is not limited to the running speed of an elite runner who runs a full course of a marathon in 3 hours, and can be data at a running speed corresponding to a full course time of more than 3 hours, for example, a full course time of 4 hours or less, as long as the relationship between the slope of the change in step frequency and the slope of the change in step length under a plurality of running speeds described later is detected.

[0050] Figure 2 is a graph in which the change in step length and the change in step frequency are compared between the stride type and the pace type in the same running speed range. Figure 2 (a) of FIG. 1 is a scatter plot illustrating the relationship between running speed and step length and the relationship between running speed and step frequency of a runner of the stride type (personal best time for a full course of a marathon: 2 hours, 36 minutes, and 7 seconds). Figure 2 (b) of FIG. 1 is a scatter plot illustrating the relationship between running speed and step length and the relationship between running speed and step frequency of a runner of the pace type (personal best time for a full course of a marathon: 2 hours, 40 minutes, and 0 seconds). The horizontal axis is running speed [m / s], and the vertical axis is step length [m] or step frequency (steps per minute) [spm]. In this graph, step length (black circle marks) and step frequency (white circle marks) when running in a range (4.0 m / s to 6.0 m / s) including 4.17 m / s (4-minute pace for 1 km) to about 5.56 m / s (3-minute pace for 1 km) are plotted.

[0051] In Figure 2In the case of the stride type runner shown in (a), the stride length increases significantly over a wide range of approximately 0.55 m from approximately 1.45 m to approximately 2 m, in a manner proportional to the increase in running speed. This indicates that the slope of the regression line representing the increase in stride length relative to the increase in running speed, i.e., the change in stride length 110, is relatively large. In particular, relative to the increase in running speed from 4.17 m / s (4 minutes per kilometer pace) to approximately 5.56 m / s (3 minutes per kilometer pace), the stride length increases from 1.49 m to 1.88 m, reaching +0.39 m (approximately 26%).

[0052] In contrast, the stride frequency (steps per minute) of stride-type runners gradually increases within a small range of 169 spm to 183 spm in a manner proportional to the increase in running speed. However, the slope of the regression line representing the increase in stride frequency relative to the increase in running speed, i.e., the stride frequency change 111, is small and almost flat. In particular, relative to the increase in running speed from 4.17 m / s (4 minutes per kilometer pace) to approximately 5.56 m / s (3 minutes per kilometer pace), the stride frequency (steps per minute) increases from 169 spm to 177 spm, a mere +8 spm (approximately 5%).

[0053] exist Figure 2 In the case of the cadence type runner shown in (b), the stride length increases by about 0.3 m in a range from about 1.45 m to about 1.75 m in a manner proportional to the increase in running speed. This means that the slope of the regression line representing the increase in stride length relative to the increase in running speed, i.e., the change in stride length 112, is smaller than that of the stride length type. In particular, relative to the increase in running speed from 4.17 m / s (4 minutes per kilometer pace) to about 5.56 m / s (3 minutes per kilometer pace), the stride length increases from 1.48 m to 1.68 m, only +0.2 m (about 14%).

[0054] In contrast, the stride frequency (steps per minute) of cadence-type runners increases significantly over a wide range from 170 spm to 225 spm, proportional to the increase in running speed. This means the slope of the regression line representing the increase in stride frequency relative to the increase in running speed, i.e., the change in stride frequency 113, is larger than that of stride length-type runners. Specifically, relative to the increase in running speed from 4.17 m / s (4 minutes per kilometer pace) to approximately 5.56 m / s (3 minutes per kilometer pace), the stride frequency (steps per minute) increases from 170 spm to 198 spm, a gain of +28 spm (approximately 16%).

[0055] Figure 3is a graph illustrating the relationship of the slope of the step frequency change to the slope of the step length change with respect to the change in running speed of a plurality of runners. If the slope of the step frequency (the number of steps per second) change is taken as the horizontal axis and the slope of the step length change is taken as the vertical axis, it is seen that there is a negative correlation distributed in the area 101 that is downwardly inclined to the right as shown in the graph. That is, it is seen that the greater the slope of the step frequency change, the smaller the slope of the step length change, and the smaller the slope of the step frequency change, the greater the slope of the step length change. If a regression analysis is performed on the relationship of the slope of the step frequency change to the slope of the step length change, a regression line 100 having a negative slope as shown in the graph is obtained.

[0056] Figure 4 is a functional block diagram showing the basic structure of the user terminal. In this drawing, a block diagram focusing on functions is depicted, and these function blocks can be implemented in various forms by hardware, software, or a combination of these. The user terminal 10 can be an information terminal such as a smartphone or a tablet terminal, or a device such as a personal computer. The user terminal 10 includes at least each function of the running log recording section 50, the display section 52, the data processing section 54, the operation processing section 56, and the data communication section 58. As hardware, the user terminal 10 is constituted by, for example, a central processing unit (CPU), a read only memory (ROM), a random access memory (RAM), a touch screen, a communication module, and the like. Further, as a modification example, each function of the user terminal 10 shown in this drawing can also be built into the form of the wearable device 16 as an integrated device.

[0057] The running log recording section 50 acquires various detection data from the wearable device 16 via the communication module such as a close proximity wireless communication, and records it in the form of a running log. The detection data acquired from the wearable device 16 includes, for example, position information received from a satellite positioning system such as a global positioning system (GPS) and information indicating the date and time of its acquisition, information of the step frequency (per unit time, for example, the number of steps per minute). The running log recording section 50 records information of the running time, the running distance, the running speed for each prescribed distance or each prescribed time, the step frequency, and the like in the form of a running log in a prescribed storage area based on the detection data acquired from the wearable device 16.

[0058] The operation processing section 56 accepts operation input for the user's instruction. The display section 52 displays the running log recorded by the running log recording section 50 in the screen based on the user's instruction via the operation processing section 56. The data processing section 54 selects data of the step frequency at various running speeds from the running log recorded by the running log recording section 50 based on the user's instruction via the operation processing section 56, and transmits the selected data to the running style analysis server 20 in the form of the subject's running log via the data communication section 58. Further, in a modification, the data processing section 54 can also be configured in such a manner that the entire running log is transmitted to the running style analysis server 20 via the data communication section 58, and the required data is selected on the side of the running style analysis server 20. In addition, data of the step length can also be generated by dividing the running distance by the step frequency, and included in the running log.

[0059] Figure 5 is a functional block diagram showing the basic structure of the running style analysis server. In the present drawing, a block diagram focusing on functions is depicted, and these functional blocks can be realized in various forms by hardware, software, or a combination of these. The running style analysis server 20 can be, for example, a server computer in hardware. The running style analysis server 20 has at least each function of a data reception section 70, a data accumulation section 72, a step frequency acquisition section 74, a step length acquisition section 75, a data analysis section 76, a determination section 80, and an output section 90. The running style analysis server 20 is configured by, for example, a CPU, a ROM, a RAM, a communication module, and the like as hardware.

[0060] The data reception section 70 receives data of the running speed and the step frequency contained in the subject's running log from the user terminal 10 and saves it in the data accumulation section 72. In the data accumulation section 72, a data group of the step frequency and the step length of the running log of a plurality of runners based on past measurements is accumulated. Principal component analysis is performed on the data group accumulated in the data accumulation section 72, and stored in the determination section 80 as a principal component analysis model. The principal component analysis model is used to determine whether the running style of the runner is the step frequency type or the step length type based on a newly acquired running log. Further, Figure 2 The "step frequency" mentioned in the above means mainly the number of steps per minute (spm), but the value of the "step frequency" used as a calculation target of the principal component analysis can also use the number of steps per second (Hz) obtained by dividing the number of steps per minute by 60 in the calculation. In the calculation of the principal component analysis, either one of the number of steps per second (Hz) and the number of steps per minute (spm) can also be used as the value of the step frequency. However, in order to avoid the mixture of these multiple units in the calculation, one of them is used as a uniform reference, and becomes the target of the principal component analysis.

[0061] The data analysis section 76 includes a principal component analysis section 77 and an average value calculation section 78. The principal component analysis section 77 performs principal component analysis on the data sets accumulated in the data accumulation section 72, and causes the determination section 80 to store the generated principal component analysis model. That is, the principal component analysis section 77 calculates the slope of the stride frequency change and the slope of the stride length change at various running speeds from the data sets based on the stride frequency and the stride length of the running logs of the plurality of runners, and performs principal component analysis with the slope of the stride frequency change as the first observation variable and the slope of the stride length change as the second observation variable. The average value calculation section 78 calculates the average value of the slope of the stride frequency change and the average value of the slope of the stride length change from the data sets based on the stride frequency and the stride length of the running logs of the plurality of runners, and stores them in the determination section 80.

[0062] Figure 6 The graph shows the distribution of the principal components calculated by the principal component analysis. The lower section of the graph is a scatter plot that takes the first principal component PC1 calculated by the principal component analysis based on the data sets of the plurality of runners as the horizontal axis and the second principal component PC2 calculated by the principal component analysis as the vertical axis. The first principal component PC1 is normalized by dividing by the maximum value of its absolute value to a value in the range of -1.0 to 1.0. As in the lower section, the plurality of data points 142 are distributed in the region 140 that is long in the horizontal axis direction. That is, the first principal component PC1 is distributed in a relatively large range of -0.75 to 0.85 across the central value 0.0 shown by the first dotted line 114, and the second principal component PC2 is distributed in a relatively small range of -0.1 to 0.1 across the central value 0.0 shown by the second dotted line 116. The larger the first principal component on the negative side (left direction of the graph), the stronger the tendency toward the stride type, and the larger the first principal component on the positive side (right direction of the graph), the stronger the tendency toward the step frequency type. The bar graph of the upper section of the graph shows that the distribution of the runners on the negative side, that is, the stride type, is concentrated in a relatively small range of -0.4 to 0.0, and the distribution of the runners on the positive side, that is, the step frequency type, is dispersed in a relatively large range of 0.0 to 0.6.

[0063] Returning to Figure 5 , the determination section 80 includes a model storage section 82, a score calculation section 83, and a determination processing section 84. The model storage section 82 stores the principal component analysis model. The principal component analysis model is generated in the form of an equation that calculates the principal component score based on the principal component load calculated by the principal component analysis section 77 and the average value of the slope of the stride frequency change and the average value of the slope of the stride length change calculated by the average value calculation section 78. Furthermore, the principal component score is normalized to a value in the range of -1.0 to 1.0 by dividing by the maximum value of its absolute value. The equation of the principal component analysis model stored in the model storage section 82 is represented by the following equation.

[0064] [Equation 1]

[0065]

[0066] Equation 1 represents the slope SF of the step frequency change obtained when the new object is used as the judgment object. Slope The slope SL of the change in step length Slope Multiply the matrix by the rotation matrix of the principal component loadings generated in advance through principal component analysis, and subtract the slope SF of the step frequency change. Slope The average value and the slope SL of the step length change Slope The matrix of average values ​​yields the first principal component score. PC1 Second principal component score PC2 The matrix. The score calculation unit 83 can be based on the principal component analysis model of formula 1 stored in the model storage unit 82, and calculate the score based on the newly obtained slope SF of the step frequency change. Slope The slope SL of the change in step length Slope Calculate the first principal component score. PC1 Second principal component score PC2 .

[0067] Here, the contribution rate of the first principal component PC1 is 98.2%, and the contribution rate of the second principal component PC2 is 1.8%. Thus, it can be seen that the contribution rate of the first principal component PC1 is overwhelmingly higher than that of the second principal component PC2. The first principal component PC1 alone can be used to explain the type of relationship between the slope of the change in stride frequency and the slope of the change in stride length, that is, to determine which type of runner the runner belongs to.

[0068] Figure 7 This represents the distribution of the first principal component obtained through principal component analysis. The lower part of the figure is a scatter plot with the first principal component PC1, obtained through principal component analysis based on multiple runner datasets, as the horizontal axis. In this scatter plot, a two-dimensional distribution is represented with the second principal component PC2 as the vertical axis. Figure 6 Unlike the scatter plot, only a one-dimensional distribution along the horizontal axis is plotted. As shown in the figure, the dispersion is only observed in the distribution of the first principal component PC1, as shown in the bar chart above. In the first principal component PC1, the larger the value on the negative side (to the left of the figure), the stronger the tendency towards stride length, and the larger the value on the positive side (to the right of the figure), the stronger the tendency towards stride frequency.

[0069] Thus, runner type can be sufficiently determined using only the first principal component PC1, therefore it is possible to... Figure 7 By performing dimensionality compression in this way, the score calculation part (83) does not need to calculate the second principal component score. PC2 However, only the score of the first principal component is calculated. PC1 The decision processing unit 84 only bases its decision on the first principal component score. PC1determines the runner type. Alternatively, the first principal component score Score PC1 determines the runner type. Alternatively, the first principal component score Score PC2 determines the runner type. Alternatively, the first principal component score Score PC1 determines the runner type. Alternatively, the first principal component score Score PC1 determines the runner type. Alternatively, the first principal component score Score PC1 determines the runner type. Alternatively, the first principal component score Score

[0070] Returning to Figure 5 The process of determining the runner type of a subject based on data of the subject newly acquired as a determination target will be described. The step frequency acquisition section 74 acquires data of step frequency at various running speeds from the running log of the subject saved in the data accumulation section 72 with respect to the running of the subject. The step frequency acquisition section 74 derives a regression formula by regression analysis with the step frequency as a target variable and the running speed as an explanatory variable, and calculates the slope of the change in step frequency based on the regression formula. The step frequency acquisition section 74 performs regression analysis on data of at least two points as data indicating the relationship between the running speed and the step frequency, but since the more data to be analyzed, the smaller the error of the regression formula and the higher the accuracy, it is desirable to analyze data of three or more points.

[0071] The step length acquisition section 75 acquires data of step length at various running speeds from the running log of the subject saved in the data accumulation section 72 with respect to the running of the subject. In the case where data of step length is not included in the running log, the step length is calculated by dividing the running distance by the step frequency. The step length acquisition section 75 derives a regression formula by regression analysis with the step length as a target variable and the running speed as an explanatory variable, and calculates the slope of the change in step length based on the regression formula. The step length acquisition section 75 performs regression analysis on data of at least two points as data indicating the relationship between the running speed and the step length, but since the more data to be analyzed, the smaller the error of the regression formula and the higher the accuracy, it is desirable to analyze data of three or more points.

[0072] The determination unit 80, based on the principal component analysis model stored in the model storage unit 82, determines which of several running styles, including stride length and cadence, the subject's running conforms to. In this embodiment, running styles are categorized into cadence and stride length types to determine which type is correct. More specifically, the scoring unit 83 calculates the principal component score based on the slope of the subject's stride frequency and stride length changes using the principal component analysis model. Furthermore, the determination processing unit 84 determines whether the subject's running conforms to either stride length or cadence based on the principal component score.

[0073] The judgment processing unit 84 uses the average of the score range as a benchmark and determines which of several running styles, including stride length type and cadence type, the subject's running conforms to by comparing the principal component scores with the average. The score calculation unit 83 sets the average of the score range of principal component scores calculated by the principal component analysis model stored in the model storage unit 82 as the benchmark value for distinguishing between cadence type and stride length type. The average of the score range is, for example,... Figure 7 The first dashed line 114 indicates 0.0. The judgment processing unit 84 determines the subject's first principal component score. PC1 If the score is above 0.0, it is considered to conform to the step frequency type, and the score is based on the subject's first principal component score. PC1 If the score is below 0.0, it is determined to conform to the stride type. The judgment processing unit 84 determines the subject's first principal component score. PC1 If the value is the same as 0.0, it can be determined that it conforms to both the step frequency type and the step length type.

[0074] Thus, as long as the slopes of the changes in the subject's stride frequency and stride length are obtained, it is easy and accurate to determine whether the subject is a cadence-type or stride-length-type runner. Furthermore, by pre-storing the principal component loads (e.g., values ​​of a 2×2 matrix) and averages obtained from the running logs of multiple runners through principal component analysis, it is possible to objectively determine whether the subject conforms to a cadence-type or stride-length-type runner through a simple calculation as shown in Equation 1, with a light processing load. In this sense, it is sufficient to calculate whether the subject conforms to a cadence-type or stride-length-type runner, even without the need for calculation by the running analysis server 20, on the user terminal 10 or wearable device 16.

[0075] In addition, in the case where the running style is determined based on the distribution or relative value of the numerical value of the principal component score, unlike the case where the determination is made by whether the measured value of the step frequency or step length exceeds a predetermined reference value, it is not necessary to prepare a reference value as an absolute value in advance. Therefore, for example, even if the running speed at which the characteristics of the step frequency or step length are clearly exhibited is not limited to a high speed of a competition pace of a marathon race of 3 hours or less, there is no case where an objective reference value as an absolute value cannot be set, and the running style can be determined for runners or measured values of a wide range of running speeds.

[0076] The output section 90 includes a result output section 92, a recommendation output section 94, and a data transmission section 96. The result output section 92 outputs the determination result of the determination section 80 to the user terminal 10 via the data transmission section 96. That is, the result output section 92 displays the determination result on the screen of the user terminal 10 by transmitting the determination result of which of the running styles of the subject, the step frequency type or the step length type, to the user terminal 10.

[0077] The recommendation output section 94 determines the shoe to be recommended from among a plurality of shoes including a shoe suitable for a runner of the step frequency type and a shoe suitable for a runner of the step length type, based on the result of the determination made by the determination processing section 84. The recommendation output section 94 generates information introducing the shoe to be recommended and outputs it. In this way, as long as the slope of the change in the step frequency and the slope of the change in the step length of the running of the subject can be obtained, it is possible to determine which of the shoe suitable for a runner of the step frequency type and the shoe suitable for a runner of the step length type should be recommended with ease and with good accuracy.

[0078] Figure 8 is a flowchart showing the basic processing in the running style analysis server. The data reception section 70 acquires the running log of the subject (S10), the step frequency acquisition section 74 acquires the slope of the change in the step frequency at a plurality of running speeds from the running log of the subject (S12), the step length acquisition section 75 acquires the slope of the change in the step length at a plurality of running speeds from the running log of the subject (S14), and the score calculation section 83 calculates the principal component score based on the principal component analysis model from the slope of the change in the step frequency and the slope of the change in the step length of the subject (S16). The determination processing section 84 determines which of the step length type and the step frequency type the running of the subject conforms to by comparing the principal component score of the subject with the average value (S18), and the determination recommendation output section 94 determines which of a plurality of shoes including a shoe suitable for a runner of the step frequency type and a shoe suitable for a runner of the step length type should be recommended (S20). The result output section 92 outputs the result of the determination made by the determination processing section 84 to the user terminal 10 (S22), the recommendation output section 94 generates the recommendation information of the shoe (S24), and outputs it to the user terminal 10 (S26).

[0079] (Second Embodiment)

[0080] In the present embodiment, unlike the first embodiment which determines which of the two runner types of the step frequency type and the step length type and which of the shoes corresponds thereto, three runner types of the step frequency type, the step length type, and an intermediate type corresponding to an intermediate of them are classified, and which of the three runner types and which of the shoes corresponds thereto is determined. Hereinafter, the differences from the first embodiment will be described, and the commonalities will be omitted.

[0081] For example, which of the three runner types of the step frequency type, the step length type, and the intermediate type corresponding to an intermediate of them corresponds thereto is determined as follows. That is, the determination processing section 84 determines that the step frequency type corresponds thereto in a case where the principal component score of the subject is in a prescribed first reference range which is higher than the average, determines that the step length type corresponds thereto in a case where the principal component score is in a prescribed second reference range which is lower than the average, and determines that the intermediate type corresponds thereto in a case where the principal component score is in a prescribed third reference range which is lower than the first reference range and higher than the second reference range.

[0082] Figure 9 The relationship between the distribution of the first principal component obtained by the principal component analysis and the range of the runner type is shown. In the present embodiment, density estimation based on a mixture Gaussian model which is a mixture of a plurality of Gaussian distributions is used. That is, assuming that the distribution of the principal component score is a mixture of three Gaussian distributions of the step frequency type, the step length type, and an intermediate type of them, which of the runner types corresponds thereto is determined according to which of one or more of the three Gaussian distributions the principal component score of the subject corresponds to, and which of the shoes is recommended. The initial conditions of the mixture Gaussian model are such that, on the premise of the classification into the three runner types, for example, the vertex of the first Gaussian distribution is set to the average of the principal component scores corresponding to the step length type (negative side) in the Figure 7 , the vertex of the second Gaussian distribution is set to the average of the principal component scores of the entire (for example, 0.0), and the vertex of the third Gaussian distribution is set to the average of the principal component scores corresponding to the step frequency type (positive side) in the Figure 7 . Each Gaussian distribution is set to a standard deviation of 1.0. The result of the density estimation based on the initial conditions is that the step length type Gaussian distribution 120 having the value shown by the third dotted line 118 as the vertex, the intermediate type Gaussian distribution 122 having the value shown by the first dotted line 114 as the vertex, and the step frequency type Gaussian distribution 124 having the value shown by the fourth dotted line 119 as the vertex are obtained.

[0083] Here, as the classification method of the running style, it is possible to consider a case where the ranges of the three running styles are set so as not to overlap each other and a case where the ranges of the three running styles are set so as to overlap each other. In the case where the ranges of the three running styles are set so as not to overlap each other, as illustrated, the first range 130 of the value or less than that indicated by the third broken line 118 is set as the stride type, the range of the value from the third broken line 118 to the fourth broken line 119, that is, the second range 131 is set as the intermediate type, and the third range 132 of the value or more than that indicated by the fourth broken line 119 is set as the cadence type.

[0084] In the case where the ranges of the three running styles are set so as to overlap each other, as illustrated, the fourth range 133 of the value or less than that indicated by the first broken line 114 is set as the stride type, the range of the value from the third broken line 118 to the fourth broken line 119, that is, the second range 131 is set as the intermediate type, and the fifth range 134 of the value or more than that indicated by the first broken line 114 is set as the cadence type. In this case, when the principal component score is included in the range of the value from the third broken line 118 to the first broken line 114, the determination processing section 84 can determine that it conforms to both the stride type and the intermediate type, or can determine that it is the intermediate type close to the stride type. When the principal component score is included in the range of the value from the first broken line 114 to the fourth broken line 119, the determination processing section 84 can determine that it conforms to both the cadence type and the intermediate type, or can determine that it is the intermediate type close to the cadence type. In the case where the intermediate type close to the stride type or the intermediate type close to the cadence type is distinguished and determined, in total, it is possible to classify substantially four running styles including the stride type and the cadence type.

[0085] The recommendation output section 94 determines the shoe to be recommended from a plurality of shoes including a shoe suitable for a runner of the cadence type, a shoe suitable for a runner of the stride type, and a shoe suitable for a runner of the intermediate type in accordance with the determination result of the running style by the determination processing section 84. However, in the case where the ranges of the three running styles are set so as not to overlap each other as described above, the recommendation output section 94 classifies the plurality of shoes into a shoe suitable for the cadence type, a shoe suitable for the stride type, and a shoe suitable for the intermediate type and stores them.

[0086] On the other hand, in a case where the ranges for the three running style types overlap each other, the recommendation output section 94 can recommend the shoe of the step frequency type in a case where it is determined to be the step frequency type, the shoe of the step length type in a case where it is determined to be the step length type, and both the shoe of the step frequency type and the shoe of the step length type in a case where it is determined to be the intermediate type, on the basis of the shoe being classified as both the shoe suitable for the step frequency type and the shoe suitable for the step length type. Alternatively, the shoe of the step frequency type in a case where it is determined to be the step frequency type, the shoe of the step length type in a case where it is determined to be the step length type, the shoe of the step frequency type in a case where it is determined to be both the step frequency type and the intermediate type, and the shoe of the step length type in a case where it is determined to be both the step length type and the intermediate type can be recommended on the basis of the shoe being classified as the shoe of the step frequency type, the shoe of the step length type, and the shoe of the intermediate type.

[0087] Further, the classification method of the running style type and the classification method of the shoe do not necessarily coincide with each other. For example, it can be determined which running style type is appropriate as the running style type by classifying it as both the step frequency type and the step length type, on the other hand, as to the shoe, it can be determined which shoe classification is appropriate by classifying it as the step frequency type, the step length type, and the intermediate type. In addition, in a case where the ranges for the three running style types overlap each other, the shoe of the step frequency type in a case where it is determined to be the step frequency type, the shoe of the step length type in a case where it is determined to be the step length type, and both the shoe of the step frequency type and the shoe of the step length type in a case where it is determined to be the intermediate type can be recommended on the basis of the shoe being classified as both the shoe suitable for the step frequency type and the shoe suitable for the step length type. Figure 9 In the example shown, the method of estimating a plurality of Gaussian distributions by density estimation based on a Gaussian mixture model, and determining the running style type and the recommendation of the shoe in accordance with which Gaussian distribution the principal component score conforms to is explained, and the range of the principal component score in which each running style type is set is based on the design idea of each shoe, and does not necessarily have to be set only by a statistical method. It is also possible to slightly adjust the range on the basis of the numerical range (for example, each Gaussian distribution as shown in FIG. 6) obtained by a statistical method, and thereby set a more appropriate range as the range of each running style type. Figure 9 In the example shown, the method of estimating a plurality of Gaussian distributions by density estimation based on a Gaussian mixture model, and determining the running style type and the recommendation of the shoe in accordance with which Gaussian distribution the principal component score conforms to is explained, and the range of the principal component score in which each running style type is set is based on the design idea of each shoe, and does not necessarily have to be set only by a statistical method. It is also possible to slightly adjust the range on the basis of the numerical range (for example, each Gaussian distribution as shown in FIG. 6) obtained by a statistical method, and thereby set a more appropriate range as the range of each running style type.

[0088] The present application has been explained on the basis of the embodiments. Those skilled in the art will understand that the embodiments are illustrative, and various modifications of each constituent element or the combination of each process can exist, and such modifications are also within the scope of the present application. Hereinafter, the modifications will be explained.

[0089] In the embodiment, an example of performing running analysis in the form of a running analysis system 30 including the user terminal 10 and the running analysis server 20 is explained. In the modification, each function for running analysis can be realized by a form executed on a device directly operated by the user such as a smartphone or a tablet, a personal computer, and the like, rather than by a form executed on the running analysis server 20.

[0090] Further, if the embodiment is generalized, the following forms can be obtained.

[0091] [Form 1]

[0092] A running method analysis device characterized by comprising:

[0093] a step frequency acquisition section that acquires a slope of a change in step frequency with respect to a change in running speed with respect to running by the subject;

[0094] a step length acquisition section that acquires a slope of a change in step length with respect to a change in running speed with respect to running by the subject;

[0095] a determination section that calculates a principal component score from the slope of the change in step frequency and the slope of the change in step length of the subject based on a principal component analysis model that is generated in advance based on measured values of a plurality of runners, and determines which of a plurality of running method types including a stride type and a step frequency type the running by the subject conforms to based on the calculated principal component score; and

[0096] a result output section that outputs a result of the determination.

[0097] [Mode 2]

[0098] The running method analysis device according to Mode 1, characterized in that the determination section stores, as the principal component analysis model, an algorithm for calculating a principal component score based on a principal component load amount obtained by performing principal component analysis on a data set of the slope of the change in step frequency and the slope of the change in step length with respect to a change in running speed in measured values of a plurality of runners in advance.

[0099] [Mode 3]

[0100] The running method analysis device according to Mode 2, characterized in that the determination section stores, in advance, respective averages of the slope of the change in step frequency and the slope of the change in step length with respect to a change in running speed in measured values of a plurality of runners, and stores, as the principal component analysis model, an algorithm for calculating the principal component score by multiplying data acquired by the step frequency acquisition section and the step length acquisition section by the principal component load amount to find a difference from the averages.

[0101] [Mode 4]

[0102] The running method analysis device according to any one of Modes 1 to 3, characterized in that the determination section determines which of a plurality of running method types including a stride type and a step frequency type the running by the subject conforms to by comparing the calculated principal component score with an average of a range of scores that can be calculated as a principal component score by the principal component analysis model, based on the average as a reference.

[0103] [Mode 5]

[0104] The running method analysis device according to Aspect 4, wherein the determination unit determines that the runner is of the step frequency type when the principal component score is equal to or higher than the average value, and determines that the runner is of the step length type when the principal component score is lower than the average value.

[0105] [Aspect 6]

[0106] The running method analysis device according to Aspect 4, wherein the determination unit determines that the runner is of the step frequency type when the principal component score is within a prescribed first reference range that is higher than the average value, determines that the runner is of the step length type when the principal component score is within a prescribed second reference range that is lower than the average value, and determines that the runner is of the intermediate type when the principal component score is within a prescribed third reference range that is lower than the first reference range and higher than the second reference range.

[0107] [Aspect 7]

[0108] The running method analysis device according to any one of Aspects 1 to 6, further comprising a recommendation output unit that outputs information recommending at least one of a plurality of shoes including a shoe suitable for a runner of the step frequency type and a shoe suitable for a runner of the step length type, based on a result of the determination.

[0109] [Aspect 8]

[0110] The running method analysis device according to any one of Aspects 1 to 7, further comprising a recommendation output unit that outputs information recommending one or more shoes from a plurality of shoes including a shoe suitable for a runner of the step frequency type, a shoe suitable for a runner of the step length type, and a shoe suitable for a runner of both the step frequency type and the step length type, based on which of one or more Gaussian distributions the calculated principal component score belongs to, by mixing a plurality of Gaussian distributions each having a respective principal component score as a peak into a mixture Gaussian model.

[0111] [Aspect 9]

[0112] A running method analysis method, comprising:

[0113] a process of acquiring a slope of a change in step frequency with respect to a change in running speed with respect to running by a subject;

[0114] a process of acquiring a slope of a change in step length with respect to a change in running speed with respect to running by the subject;

[0115] a process of calculating a principal component score from the slope of the change in step frequency and the slope of the change in step length by a principal component analysis model that is generated in advance based on measured values of a plurality of runners.

[0116] a process of determining which one of a plurality of running pattern types including a stride pattern type and a cadence pattern type the running of the subject conforms to, based on the calculated principal component score; and

[0117] a process of outputting the result of the determination.

[0118] [Configuration 10]

[0119] A running pattern analysis program characterized by causing a computer to realize the following functions:

[0120] a function of acquiring a slope of a change in a step frequency with respect to a running speed with respect to the running of the subject;

[0121] a function of acquiring a slope of a change in a step length with respect to a running speed with respect to the running of the subject;

[0122] a function of calculating a principal component score based on the slope of the change in the step frequency and the slope of the change in the step length of the subject, based on a principal component analysis model that is generated in advance based on measured values of a plurality of runners, and determining which one of a plurality of running pattern types including a stride pattern type and a cadence pattern type the running of the subject conforms to, based on the calculated principal component score; and

[0123] a function of outputting the result of the determination.

[0124] Industrial applicability

[0125] The present application relates to a technology of analyzing a running pattern of a marathon runner.

Claims

1. A running method analysis device, characterized by, including: a step frequency acquisition section that acquires a slope of a change in a step frequency with respect to a change in a running speed with respect to running of a subject; a step length acquisition section that acquires a slope of a change in a step length with respect to a change in a running speed with respect to running of the subject; a determination section that calculates a principal component score from the slope of the change in the step frequency and the slope of the change in the step length with respect to running of the subject, based on a principal component analysis model generated in advance from measured values of a plurality of runners measured in the past, the measured values of the plurality of runners including step frequency data and step length data at a plurality of running speeds, and determines that the running of the subject conforms to one of a plurality of running style types including a pace type running style and a step frequency type running style, based on the calculated principal component score; and a result output section that outputs a result of the determination, wherein the determination section stores, as the principal component analysis model, a model generated using principal component load amounts and the measured values of the plurality of runners measured in the past, the principal component load amounts being obtained by performing principal component analysis in advance on a data set of the slope of the change in the step frequency and the slope of the change in the step length with respect to a change in a running speed in the measured values of the plurality of runners.

2. The running analysis apparatus according to claim 1, characterized in that, The determination section stores, as the principal component analysis model, each average of the slope of the change in the step frequency and the slope of the change in the step length with respect to a change in a running speed in the measured values of a plurality of runners in advance, and stores, as the principal component analysis model, an arithmetic formula for calculating the principal component score by multiplying data acquired by the step frequency acquisition section and the step length acquisition section by the principal component load amounts to find a difference from the average.

3. The running analysis apparatus according to claim 1, characterized in that, The determination section determines that the running of the subject conforms to one of the plurality of running style types by comparing the calculated principal component score with an average of a range of scores that can be calculated as the principal component score by the principal component analysis model.

4. The running analysis apparatus according to claim 3, characterized in that, The determination section determines that the running of the subject conforms to the step frequency type running style when the principal component score is equal to or higher than the average, and determines that the running of the subject conforms to the pace type running style when the principal component score is lower than the average.

5. The running analysis apparatus according to claim 3, characterized in that, The plurality of running style types further include an intermediate type running style, The determination section determines that the running of the subject conforms to the step frequency type running style when the principal component score is in a prescribed first reference range that is higher than the average, determines that the running of the subject conforms to the pace type running style when the principal component score is in a prescribed second reference range that is lower than the average, and determines that the running of the subject conforms to the intermediate type running style when the principal component score is in a prescribed third reference range that is lower than the first reference range and higher than the second reference range.

6. The running analysis apparatus of claim 1, wherein further including: a recommendation output section that recommends at least one of a plurality of shoes based on the result of the determination and outputs information, the at least one of the plurality of shoes including at least one of a shoe for a runner of the step frequency type running style and a shoe for a runner of the pace type running style.

7. The running analysis apparatus of claim 1, wherein Further comprising: a recommendation output section that, based on a mixture Gaussian model that is formed by mixing a plurality of Gaussian distributions whose vertices are the respective principal component scores, outputs information that recommends one or more shoes from among a plurality of shoes, according to which of the plurality of Gaussian distributions the calculated principal component score belongs, wherein the plurality of shoes include a shoe for a runner who is suitable for the stride type running, a shoe for a runner who is suitable for the pace type running, and a shoe for a runner who is suitable for both the stride type running and the pace type running.

8. A running method analysis method characterized by, Further comprising: a process of acquiring a slope of a change in step frequency with respect to a change in running speed with respect to running by the subject; a process of acquiring a slope of a change in step length with respect to a change in running speed with respect to running by the subject; a process of calculating a principal component score from the slope of the change in step frequency and the slope of the change in step length of the subject, based on a principal component analysis model that is generated in advance from measured values of a plurality of runners measured in the past, wherein the measured values of the plurality of runners include step frequency data and step length data at a plurality of running speeds; a process of determining that the running by the subject conforms to one of a plurality of running type, from the calculated principal component score; and a process of outputting a result of the determination, wherein the principal component analysis model is a model that is generated using principal component load amounts and the measured values of the plurality of runners measured in the past, and the principal component load amounts are obtained by performing principal component analysis in advance on a data set of the slope of the change in step frequency and the slope of the change in step length with respect to a change in running speed in the measured values of the plurality of runners.

9. A computer program product comprising a run-time analysis program, characterized in that, The running type analysis program causes a computer to function as: a function of acquiring a slope of a change in step frequency with respect to a change in running speed with respect to running by the subject; a function of acquiring a slope of a change in step length with respect to a change in running speed with respect to running by the subject; a function of calculating a principal component score from the slope of the change in step frequency and the slope of the change in step length of the subject, based on a principal component analysis model that is generated in advance from measured values of a plurality of runners measured in the past, wherein the measured values of the plurality of runners include step frequency data and step length data at a plurality of running speeds, and the plurality of running types include a stride type running and a pace type running; and a function of outputting a result of the determination, wherein the principal component analysis model is a model that is generated using principal component load amounts and the measured values of the plurality of runners measured in the past, and the principal component load amounts are obtained by performing principal component analysis in advance on a data set of the slope of the change in step frequency and the slope of the change in step length with respect to a change in running speed in the measured values of the plurality of runners.

Images