JP2017124086A - Training classification system, training classification method and training classification server - Google Patents

Abstract

PROBLEM TO BE SOLVED: To divide a period in which trainings are performed into periods in which a different kind of a training is performed, when plural kinds of trainings are continuously performed in a sport, and classify similar trainings for every kind.SOLUTION: A communication terminal 30 of a training classification system 1 transfers sensor data which is measured by a sensor device 10 to a training classification server 20. The training classification server 20 analyzes the acquired sensor data and calculates an exercise feature amount, then detects a section which matches a feature of the exercise feature amount in an interval of the sport, as an interval, then divides whole training section into respective training sections, then classifies training items of the training section using the exercise feature amount of the divided training section.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a technique for dividing and classifying different types of training by time based on information obtained by measuring human movement using data measured by a sensor device that can be worn on a human body.

A system that records data such as exercise intensity (exercise amount per unit time) and speed in sports competitiveness improvement and education, records and evaluates athletes' movements during competition and training, and supports guidance and conditioning To be done. In this conventional system, the viewpoint of evaluation differs depending on the type of training. Therefore, when performing multiple types of training consecutively on a day, the recorded data is divided in advance for each training and classified before evaluation. There was a need.
For example, Patent Document 1 discloses a technique for detecting points that are stationary during a series of exercises using a detector mounted on the body of a plurality of users.

JP 2012-65943 A

  However, with the technique described in Patent Document 1, it is only possible to detect the time point when each of a plurality of users has completely stopped and compare the stationary timing, and cannot capture the characteristics of intervals in various sports. . In other words, during multiple types of training, athletes (users) are not necessarily in a completely stopped state, spending an interval period with less movement than during training, and are waiting for the next training In some cases.

  In view of such a background, the present invention has been made, and the present invention divides the time of different types of training when performing a plurality of trainings in sports continuously, and performs similar training for each type. It is an object to provide a training classification system, a training classification method, and a training classification server that can be classified into two categories.

  In the training classification system of the present invention, the communication terminal includes sensor data transfer means for transferring the sensor data measured by the sensor device to the training classification server, and the training classification server analyzes the acquired sensor data to obtain the motion feature amount. Calculating and detecting the section that matches the feature of the motor feature in the sport interval as an interval, dividing the entire training section into individual training sections, and using the motor feature of the divided training section The training classification means for classifying the training items in the training section is provided.

  According to the present invention, a training classification system, a training classification method, and a training that divide the time of different types of training and classify similar trainings by type when performing a plurality of trainings in sports continuously. A classification server can be provided.

It is a figure showing the whole training classification system composition concerning this embodiment. It is a flowchart which shows the flow of the training classification process which the training classification server which concerns on this embodiment performs. It is a figure which shows the data structural example of the sensor data which concern on this embodiment. It is a figure which shows the data structural example of the exercise | movement feature-value data which concern on this embodiment. It is a figure for demonstrating the example of calculation of the inclination of the arm stored in the movement feature-value data based on this embodiment. It is a flowchart which shows the flow of the interval detection process which the training classification | category means based on this embodiment performs. It is a figure which shows the example which read the time series data of the exercise intensity about the several users who performed the same training. It is a figure which shows the time series data of the average value about the exercise intensity of the multiple users who performed the same training. It is a figure which shows the example which extracted the area of the interval from the time series data of the average value of the exercise intensity of a plurality of users. It is a figure for demonstrating the method of clustering the divided training item. It is a figure which shows the example which clustered the divided training item shown in FIG. 9 with the exercise | movement feature-value. It is a figure which shows the example which reclassified the training item of the same kind which is not continuous about the result (FIG. 11) which clustered the divided training item by the exercise | movement feature-value. It is a figure which shows the example which put together the continuous training item of the same kind about the result (FIG. 12) which reclassified the training item of the same type which is not continuous. It is a figure showing an example of data composition of training classification data concerning this embodiment. It is a flowchart which shows the flow of the performance evaluation process which the training classification server which concerns on this embodiment performs. It is a figure which shows the data structural example of the performance data which concerns on this embodiment. It is a figure which shows the example of the performance evaluation figure which the analysis result display means of the communication terminal concerning this embodiment displayed performance evaluation information on the display part.

  Next, the training classification system 1, the training classification method, and the training classification server 20 in a mode for carrying out the present invention (hereinafter referred to as “this embodiment”) will be described.

<Overall configuration of training classification system>
First, the overall configuration of the training classification system 1 according to the present embodiment will be described.
FIG. 1 is a diagram showing an overall configuration of a training classification system 1 according to the present embodiment.
As shown in FIG. 1, the training classification system 1 includes a sensor device 10 (wearable terminal) worn by a user who exercises on an arm, a communication terminal 30 that transmits and receives data to and from the sensor device 10, and a communication The training classification server 20 is configured to be connected to and communicated with the terminal 30 via a router (not shown) or the network 5.
In this training classification system 1, the sensor device 10 measures the exercise performed by the user during training as sensor data and transmits it to the communication terminal 30. The communication terminal 30 transmits the sensor data received from the sensor device 10 to the training classification server 20. The training classification server 20 divides a plurality of continuous trainings performed by the user by time by analyzing sensor data, and classifies the divided trainings by type. Further, the training classification system 1 is characterized in that it extracts only data of training sections classified into the same type, and can compare performance such as exercise intensity for each time series or each user.

  The sensor device 10 may be one (one user continuously uses one sensor device 10) when used in sports individual competitions (land, swimming, etc.), or team sports (soccer, rugby, etc.) ), The sensor device 10 may be attached to each of a plurality of users. In the following description, an example will be described in which the communication terminal 30 receives information from a plurality of sensor devices 10 attached to each user in team sports and analyzes the information in the training classification server 20.

<Sensor device>
First, the configuration of the sensor device 10 according to the present embodiment will be described.
As shown in FIG. 1, the sensor device 10 includes, as a microcomputer, for example, an MCU (Micro Controller Unit) 11, a communication unit 12, an input unit 13, a display unit 14, a flash memory 15, and an acceleration sensor 16. It is comprised including.

  The sensor device 10 is a device that is mounted on a user's arm or the like and measures sensor data. For example, the MCU 11 measures the acceleration sensor 16 at a timing that is defined in advance by the firmware about 10 to 1000 times per second. Read the value (sensor data). Then, the MCU 11 outputs and stores the measurement value (sensor data) to the flash memory 15 at a predefined timing, and further reads out the stored measurement value (sensor data) by the communication unit 12 to communicate with the communication terminal 30. It is characterized by transmitting to.

The MCU 11 is a microprocessor that controls the entire sensor device 10, and the control content is defined in advance by firmware.
The acceleration sensor 16 is a sensor that measures movement (acceleration). In addition to the acceleration sensor 16, the sensor device 10 according to the present embodiment includes an angular velocity (gyro) sensor, a geomagnetic sensor, and the like, and measures the state of movement to measure more detailed movement. Good.

  The communication unit 12 has a wireless communication function suitable for driving with a small battery such as a button battery, and transmits data (for example, sensor data) output from the MCU 11 to the communication unit 12 to the communication terminal 30 or the like.

The input unit 13 can receive input information from a push button (not shown) and the like, can send a signal to the MCU 11 at an arbitrary timing, and can activate processing defined in advance by firmware that controls the MCU 11. For example, the time when the push button is depressed can be recorded in the flash memory 15 via the MCU 11, and the remaining battery level and the current time can be displayed on the display unit 14.
The display unit 14 is, for example, a liquid crystal display or an organic EL display, and displays characters and graphics.

  The flash memory 15 is a non-volatile memory, and temporarily stores acceleration data (sensor data) detected by the acceleration sensor 16. When the data (sensor data) measured by the acceleration sensor 16 is stored in the flash memory 15, the date and time when the data is measured when the data is referred to later by adding the time information held by the MCU 11. And data measured by other sensor devices 10 can be analyzed synchronously.

<Communication terminal>
Next, the configuration of the communication terminal 30 according to the present embodiment will be described.
The communication terminal 30 is a computer having a wireless communication function, an arithmetic function for executing an application, a function for displaying received data for showing to a user, and the like, and is configured by, for example, a tablet terminal or a PC (Personal Computer). . The communication terminal 30 receives data (sensor data) measured by the sensor device 10 by a dedicated application that cooperates with the sensor device 10, and sends the data to the training classification server 20 via a router (not shown) and the network 5. Send. The communication terminal 30 is characterized in that the training classification server 20 receives and displays the result of analyzing the data (sensor data) and presents it to the user.
As shown in FIG. 1, the communication terminal 30 includes a CPU (Central Processing Unit) 31, a communication unit 32, an input unit 33, a display unit 34, a RAM (Random Access Memory) 35, and an application 36. Consists of including.

The CPU 31 is a central processing unit that controls the entire communication terminal 30, and the CPU 31 executes processing by expanding the application 36 in the RAM 35.
The communication unit 32 includes a communication interface (not shown) that transmits and receives information via a communication line. The communication unit 32 transmits and receives information between the sensor device 10 and the training classification server 20.

The input unit 33 includes an input unit (input interface) with a keyboard, a touch panel, or the like (not shown), and acquires input information from a user or the like.
The display unit 34 is, for example, a liquid crystal display or an organic EL display, and displays characters and graphics.
The RAM 35 temporarily reads an application (program) when the CPU 31 performs arithmetic processing. In addition, the RAM 35 stores sensor data acquired from the sensor device 10 and analysis results received from the training classification server 20 (such as “performance evaluation information” described later).

  The application 36 is a program describing processing contents executed by the communication terminal 30 and includes a sensor data transfer unit 361, an information input unit 362, and an analysis result display unit 363. The application 36 is stored in storage means (not shown) (flash memory, hard disk, etc.).

The sensor data transfer unit 361 executes processing for receiving the sensor data received from the sensor device 10 and transmitting it to the training classification server 20 via the communication unit 32.
The information input unit 362 acquires information input from the keyboard or the touch panel via the input unit 33. For example, the information input unit 362 accepts input of each training name of training classification data 252 (see FIG. 14) described later, “performance evaluation request information” described later, and the like.
The analysis result display unit 363 acquires the analysis result (training classification data 252 (see FIG. 14) and performance evaluation information) of the sensor information acquired from the sensor device 10 from the training classification server 20 and displays it on the display unit 34. For example, the analysis result display unit 363 displays an analysis result such as a performance evaluation diagram 300 (see FIG. 17) described later on the display unit 34 based on the performance evaluation information.

<Training classification server>
Next, the configuration of the training classification server 20 according to the present embodiment will be described.
The training classification server 20 acquires and analyzes sensor data measured by the sensor device 10 to divide a plurality of continuous trainings performed by the user by time, and classifies the divided trainings by type. Further, the training classification server 20 extracts data of training sections classified into the same type, and analyzes results (performance evaluation information) for comparing performance based on exercise feature values such as exercise intensity for each time series or each user. ) And transmitted to the communication terminal 30.
As shown in FIG. 1, the training classification server 20 includes a CPU 21, a communication unit 22, a RAM 23, a program 24, and a database (DB) 25.

The CPU 21 is a central processing unit that controls the entire training classification server 20, and the CPU 21 executes processing by expanding each means of the program 24 in the RAM 23.
The communication unit 22 includes a communication interface (not shown) that transmits and receives information via a communication line. The communication unit 22 transmits and receives information to and from the communication terminal 30 via the network 5.
The RAM 23 reads the program 24 when the CPU 21 performs arithmetic processing. The RAM 23 temporarily stores sensor data and the like measured by the sensor device 10 and acquired from the communication terminal 30.

  The program 24 describes the processing contents executed by the training classification server 20 and includes a training classification unit 241, a performance evaluation unit 242, and a visualization unit 243. The program 24 is stored in a storage means (not shown) such as a hard disk.

  The training classification unit 241 generates movement feature amount data 251 (see FIG. 4 described later) by analyzing data (sensor data 250 in FIG. 3 described later) measured by the sensor device 10. Then, the training classification unit 241 generates training classification data 252 (see FIG. 14 described later) by dividing and classifying a series of training performed by the user. The details of the processing executed by the training classification unit 241 will be described with reference to FIG.

  The performance evaluation unit 242 refers to the data to be evaluated (exercise feature quantity data 251 in FIG. 4, training classification data 252 in FIG. 14, etc.) for each divided and classified training, and the same type of training at different dates and times. The comparison data (performance data 253 in FIG. 16 described later) is generated.

  The visualization unit 243 generates data for displaying a graph or the like suitable for understanding by the user, using the training comparison data (performance data 253) generated by the performance evaluation unit 242. Specifically, the visualization unit 243 generates performance evaluation information for constructing the performance evaluation diagram shown in FIG. 17 displayed on the display unit 34 by the communication terminal 30, and transmits the performance evaluation information to the communication terminal 30 via the communication unit 22. Send.

  The database (DB) 25 is configured by storage means such as a hard disk, and stores sensor data 250, exercise feature value data 251, training classification data 252, performance data 253, and the like.

The sensor data 250 stores data measured by the sensor device 10. Details of the sensor data 250 will be described later (see FIG. 3).
The exercise feature value data 251 stores values (for example, exercise intensity, walking pitch, speed, arm angle, etc.) obtained by analyzing exercise features in the training section based on the sensor data 250. Details of the motion feature data 251 will be described later (see FIG. 4).

The training classification data 252 stores a result obtained by dividing a plurality of continuous trainings based on the exercise feature value data 251 and classifying them for the same type. Details of the training classification data 252 will be described later (see FIG. 14).
The performance data 253 stores the result of analyzing an index suitable for evaluation of the divided and classified training (for example, an average exercise intensity, a maximum exercise intensity, an average speed, a maximum speed, etc. as an index indicating the state of exercise). Is done. Details of the performance data 253 will be described later (see FIG. 16).

≪Process flow≫
Next, processing executed by the training classification system 1 will be described. First, the process (training classification process) executed by the training classification server 20 will be described in detail.
FIG. 2 is a flowchart showing the flow of training classification processing executed by the training classification server 20 according to the present embodiment.

  As shown in FIG. 2, the training classification unit 241 of the training classification server 20 receives sensor data 250 of a new training section from the communication terminal 30 or receives update data of the sensor data 250, thereby receiving sensor data. 250 is acquired (step S10), stored in the database (DB) 25, and the process is started.

FIG. 3 is a diagram illustrating a data configuration example of the sensor data 250 according to the present embodiment.
As shown in FIG. 3, the sensor data 250 stores acceleration data (sensor data) measured once in the sensor device 10 for each row (one record). Thereby, acceleration data (sensor data) is recorded in time series.
This sensor data 250 includes user ID 2501, date and time 2502, acceleration X [G] 2503 (acceleration in the X-axis direction described later), acceleration Y [G] 2504 (acceleration in the Y-axis direction described later), and acceleration Z [G]. Data for each data item 2505 (acceleration in the Z-axis direction described later) is stored.

The user ID 2501 is a unique identifier for identifying the user wearing the sensor device 10.
The date and time 2502 indicates the date and time when the sensor device 10 measured the acceleration value by the acceleration sensor 16.
An acceleration X2503, an acceleration Y2504, and an acceleration Z2505 indicate the values of the triaxial acceleration measured by the acceleration sensor 16 of the sensor device 10. By storing the acceleration in units such as [G], which is a unit of gravity, even if the specifications such as the resolution of the acceleration sensor used are different, they can be stored in a comparable format.

Returning to FIG. 2, the training classification unit 241 acquires data of the entire training section to be analyzed from the sensor data 250 in the database (DB) 25 (step S <b> 11).
For example, the training classification server 20 acquires information (start date / time, end date / time) specified by the user via the input unit 33 in the communication terminal 30 as the information on the entire training section to be analyzed. The setting information (start date / time, end date / time) of the entire training section to be analyzed may be stored in advance in storage means (not shown) of the training classification server 20, and the information may be stored in advance. It may be used.

  Subsequently, the training classification unit 241 calculates an exercise feature amount using the acquired sensor data 250 of the entire training section (step S12), and stores it in the database (DB) 25 as the exercise feature amount data 251. The exercise feature amount is calculated from data obtained by measuring the user's movement, and is, for example, exercise intensity proportional to calorie consumption, walking pitch indicating the pace of walking or running, speed, arm inclination, and the like.

FIG. 4 is a diagram illustrating a data configuration example of the motion feature amount data 251 according to the present embodiment.
As shown in FIG. 4, the movement feature value data 251 stores one data of unit time calculated based on the sensor data 250 (FIG. 3) for each row (1 record). Thereby, movement feature-value data are recorded in time series. Since the data measured by the sensor device 10 in milliseconds is finer than the time unit that can be recognized by a person, the data is grouped in units of seconds that the person can easily recognize in the motion feature amount.
This motion feature quantity data 251 includes user ID 2511, date and time 2512, exercise intensity [METs] 2513, walking pitch [steps / second] 2514, speed [m / second] 2515, slope θ [rad] 2516, slope φ [rad. Data is stored for each data item 2517. In addition, when calculating these values in unit time (for example, 1 second), it memorize | stores as an average value in unit time.

The user ID 2511 is a unique identifier for identifying the user, and is the same information as the user ID 2501 of the sensor data 250 (FIG. 3).
The date and time 2512 indicates the date and time when the sensor data 250 corresponding to the motion feature amount was measured.

  The exercise intensity [METs] 2513 indicates a typical characteristic of human exercise, and indicates the strength of the human motion (the amount of exercise per unit time) proportional to the calorie consumption. When the unit is METs, it can be converted into calories consumed by calculating by multiplying the weight of the user. When using this exercise intensity (calorie consumption) as an exercise feature amount, the weight of each user is associated with the user ID 2511 and stored in advance.

The walking pitch [steps / second] 2514 is a value indicating the pace of a person's running or walking, and indicates how many steps per second or how many people walked.
The speed [m / sec] 2515 is a value indicating the moving speed of the person, and can be estimated using data such as the exercise intensity, the walking pitch, and the height of the user. When this speed is used as an exercise feature, the height of each user is associated with the user ID 2511 and stored in advance.

  The inclination θ [rad] 2516 and the inclination φ [rad] 2517 are values indicating the state of the inclination of the arm of the sensor device 10 attached to the user's arm or the like by an angle. Hereinafter, this will be described in detail with reference to FIG.

FIG. 5 is a diagram for explaining an example of calculating the arm inclination stored in the motion feature data 251 according to the present embodiment. FIG. 5A is a front view of the sensor device 10. FIG. 5B is a side view of the sensor device 10. FIG. 5C shows a calculation formula for calculating the inclination θ and the inclination φ. Here, it is assumed that the sensor device 10 is mounted on the back side of the hand on the left arm (wrist) in the same manner as a general wristwatch.
As shown in FIGS. 5A and 5B, when the triaxial acceleration is set in the X axis, Y axis, and Z axis directions for the acceleration sensor 16, the acceleration (x (t) , Y (t), z (t)) are measured by the acceleration sensor 16. Then, as shown in FIG. 5C, using the measurement value of the acceleration sensor 16, the inclination θ (t) that is the angle at which the arm is raised with respect to the ground is calculated by the equation (1). Further, an inclination φ (t) that is an angle by which the arm is rotated (twisted) is calculated by the equation (2).

Returning to FIG. 2, the training classification unit 241 uses the time series data of the exercise feature amount shown in the exercise feature amount data 251 to take a break between trainings performed continuously by the user or to a different type of training. An interval is detected by extracting a feature of an interval (time between successive trainings) indicating a time during which the transition or preparation is performed (interval detection processing: step S13).
The characteristics of this interval are, for example, that there is no running for a certain period of time in training based on running such as soccer or athletics, and that walking or standing still. Also, in competitions that mainly use the upper body, the arm is lowered for a certain time from the tilt of the arm. The training classification unit 241 detects the interval by extracting a section that matches the feature of the motion feature value at the interval in the target sport.

Here, the interval detection process executed by the training classification unit 241 will be described in detail.
FIG. 6 is a flowchart showing a flow of interval detection processing executed by the training classification unit 241 according to this embodiment. In FIG. 6, a team competition training such as a searcher will be described as an example.

First, the training classification unit 241 refers to the exercise feature value data 251 (FIG. 4), and reads out exercise feature values such as exercise intensity and walking pitch of continuous training sections to be analyzed (step S131).
FIG. 7 shows an example in which time-series data of exercise intensity is read as the exercise feature amount. Here, in FIGS. 7A to 7D, time series data (U ₁ , U ₂ , U ₃ , U ₄ ) of exercise intensity when a series of soccer training is executed by a team (four users). ).

Returning to FIG. 6, the training classification unit 241 then adds the numerical values of all target users (all the teams) to generate average data (step S <b> 132). In team competitions, the features of intervals and movements of users belonging to the team are similar, so that the features can be easily captured by summing them up.
FIG. 8 shows time-series data of average values of exercise intensity of a plurality of users. Here, the training classification means 241 adds and averages the time series data (U ₁ , U ₂ , U ₃ , U ₄ ) of the exercise intensity of the four users shown in FIGS. 7 (a) to 7 (d). The time series data (U _A ) of the average value of exercise intensity is shown.

Returning to FIG. 6, the training classification means 241 calculates a moving average with a time width close to the interval time of the game, such as 1 minute (step S133).
By calculating this moving average, the increase / decrease in exercise intensity is further smoothed, and the characteristics of exercise intensity can be easily grasped.

Subsequently, the training classification means 241 is based on information such as the exercise intensity calculated by the moving average and the walking pitch and speed calculated in the same manner, a section that matches the feature of the exercise feature amount in the interval, for example, a window such as one minute. A section indicating that the exercise intensity data is not in the running state between the widths is detected as an interval (step S134). Then, the training classification unit 241 ends the interval detection process.
FIG. 9 shows an example in which intervals are extracted from time-series data of average values of exercise intensity of a plurality of users. Here, from the time-series data of the average value of exercise intensity shown in FIG. 8, exercise for a predetermined time, for example, about 1 minute in the example of soccer training, the exercise level is less than the running level (the exercise intensity is less than the predetermined value). The result of detecting the interval (I ₁ , I ₂ , I ₃ , I ₄ , I ₅ , I ₆ ) is shown. As shown in FIG. 9, each training section can be divided and divided by intervals.

  Returning to FIG. 2, in step S <b> 14, the training classification unit 241 determines the individual continuous training sections, that is, any one of a plurality of training items, according to the intervals detected in step S <b> 13. Divide as a section to be associated (training section).

  Subsequently, the training classification unit 241 reads out the exercise feature amount of the divided training section from the exercise feature amount data 251 (see FIG. 4). The training classification unit 241 then clusters the training items corresponding to each of the plurality of training sections divided using the motion feature quantities of the plurality of axes using a known algorithm, and collects the training items having similar features. (Step S15).

FIG. 10 is a diagram for explaining a technique for clustering the training items corresponding to the divided training sections according to the present embodiment. Here, as shown in FIG. 9, a technique is shown in which training items corresponding to training sections divided and divided by intervals are clustered based on the corresponding motion feature data.
Each training item is arranged at a corresponding position on a space of multiple axis (N-axis) motion feature values (a, b, c: for example, exercise intensity, walking pitch, speed), and training items that are close to each other are the same. A known algorithm is used that is optimized to be combined into training categories (A, B, C, D).

11, training the classification unit 241, the time series data U _A of the average value of motion intensity of a plurality of users in FIG. 9 shows the results of the training items with similar characteristics were grouped into the same type. Here, the clustering method illustrated in FIG. 10 is applied to the training items in each training section divided by intervals (I ₁ , I ₂ , I ₃ , I ₄ , I ₅ , I ₆ ).

As a result, as illustrated in FIG. 11, the training item in the training section t ₀ is “A”, the training item in the training section t ₁ is “B”, the training item in the training section t ₂ is “C”, The training item for training section t ₃ is “C”, the training item for training section t ₄ is “B”, the training item for training section t ₅ is “B”, and the training item for training section t ₆ is “D”. And

  Returning to FIG. 2, in step S <b> 16, the training classification unit 241 reclassifies training items of the same type that are not consecutive in time series among the training items classified by clustering as different types of training.

FIG. 12 shows the result of the training classification means 241 reclassifying the same type of training items that are not continuous in time series among the classified training items (A, B, C, D) shown in FIG. Yes.
In FIG. 12, the training item “B” in the training sections t ₁₄ and t ₁₅ that is not continuous with the training item “B” in the training section t ₁ is reclassified as the training item “E”.

  Returning to FIG. 2, in step S <b> 17, the training classification unit 241 collects a plurality of training items continuous in time series into one training item. In the training items collected at this time, information indicating that the same training has been repeatedly performed for the number of the training items collected is stored in training classification data 252 (see FIG. 14) described later.

FIG. 13 shows an example in which the training classification means 241 combines consecutive training items of the same type from the reclassified training items (A, B, C, D, E) shown in FIG. Show. The combined training items “C” and “E” indicate that the same type of training is repeated twice.
Note that the same result can be obtained even if steps S16 and S17 are processed in the reverse order.

  Returning to FIG. 2, in step S <b> 18, the training classification unit 241 stores the training item information divided for each type in the training classification data 252.

FIG. 14 is a diagram illustrating a data configuration example of the training classification data 252 according to the present embodiment.
As shown in FIG. 14, the training classification data 252 stores one training item divided and classified for each row (one record). The training classification data 252 stores data for each data item of training ID 2521, team ID 2522, start date 2523, end date 2524, number of sets 2525, implementation user ID 2526, input name 2527, and input tag 2528.

The training ID 2521 is a unique identifier for identifying each training item.
The team ID 2522 is an identifier unique to the team when training is performed on a team basis mainly in the team competition. This team includes a plurality of users who were trained at the same time. When training is performed by one person rather than by a plurality of persons, the team and the individual can be the same (that is, a team composed of one person).

The start date and time 2523 and the end date and time 2524 indicate the time at which each training item was performed, and indicate the time delimited by the interval.
The number of sets 2525 is stored as the number of sets that are repeatedly performed when the training classification means 241 summarizes training items that are separated by intervals but continuously performed.

  The implementation user ID 2526 is an identifier indicating a user who belongs to the team indicated by the team ID 2522 and trained at the same time, and stores a plurality of user IDs when configured by a plurality of people.

  The input name 2527 is a name given so that each user can easily recognize each training item later (details will be described later).

In the input tag 2528, information that cannot be detected by the sensor device 10 when a training item is performed, for example, subjective information of a user or an instructor, environmental information such as weather, and the like are stored after training (details). Is described later).
A plurality of input tags 2528 can be input to one training item, and may be text information or the like, and as shown in FIG. 14, a unique number (identifier) is attached to each content as tag information. Also good.

  In this way, the training classification unit 241 of the training classification server 20 can divide and classify a series of training performed by the user by analyzing the sensor data 250 measured by the sensor device 10.

  Next, the training classification unit 241 of the training classification server 20 acquires the information of the input name 2527 and the input tag 2528 by transmitting the training classification data 252 stored in the database (DB) 25 in step S18 to the communication terminal 30. Then, the training classification data 252 is updated (step S19). The information of the input name 2527 and the input tag 2528 is information that cannot be detected by the sensor device 10, and may be referred to as training related information below.

  Specifically, the training classification unit 241 transmits training classification data 252 newly stored in the database (DB) 25 to the communication terminal 30. The analysis result display unit 363 of the communication terminal 30 displays the received training classification data 252 on the display unit 34. Then, the user or a leader (such as a coach) who directs the training refers to the displayed training classification data 252 (specifically, the start date / time 2523, the end date / time 2524, the set number 2525, the execution user ID 2526, etc.). Thus, the training actually performed is identified, and the information of the input name 2527 is input by the input unit 33 of the communication terminal 30 as the name of the training. Similarly, a user, a coach, or the like inputs subjective information, environment information, or the like as information of the input tag 2528 through the input unit 33 of the communication terminal 30. The information (training related information) of the input name 2527 and the input tag 2528 thus input is transmitted to the training classification server 20 by the information input unit 362 of the communication terminal 30 via the communication unit 32. The training classification unit 241 of the training classification server 20 updates the training classification data 252 using the received information (training related information) of the input name 2527 and the input tag 2528.

  For example, as shown in FIG. 14, in the field of the input name 2527 of the training classification data 252, “up”, “pass basic practice”, “4 to 2 pass practice”, etc. are stored as the name of training. . As shown in FIG. 14, the input tag 2528 may be a number (“11”, “22”, “25”) as an identifier uniquely set by the user or the like, and the weather and temperature at the time of training, the training place Alternatively, the names of players determined to be in good or bad condition during the training may be stored.

  Note that the training classification means 241 obtains the training related information such as the input name 2527 and the input tag 2528 from the communication terminal 30, after the training classification server 20 receives the request from the communication terminal 30, the training classification server 20 receives the training classification data 252. May be transmitted. In addition, for the training items classified as the same type of training in the clustering of the training items in step S15, the information of the input name 2527 input in the past may be stored as it is.

  By doing so, when displaying the result of the performance evaluation process by the performance evaluation means 242 described later on the communication terminal 30, the training related information such as the input name 2527 and the input tag 2528 which are the names of the training are also combined. By displaying the information, the user can grasp the training items at a glance.

<Performance evaluation process>
Next, the performance evaluation process executed by the performance evaluation unit 242 and the visualization unit 243 of the training classification server 20 according to the present embodiment will be described.
FIG. 15 is a flowchart showing the flow of performance evaluation processing executed by the training classification server 20 according to the present embodiment. In this performance evaluation process, the training classification means 241 executes the above-described training classification process (see FIG. 2), and information indicating that the exercise feature data 251 and the training classification data 252 are stored in the database (DB) 25. Is executed when the performance evaluation means 242 obtains it.

  First, the performance evaluation unit 242 of the training classification server 20 refers to the exercise feature data 251 and the training classification data 252 in the database (DB) 25, and generates performance data 253 (FIG. 16) indicating the state of exercise (FIG. 16). Step S20).

FIG. 16 is a diagram illustrating a data configuration example of the performance data 253 according to the present embodiment.
As shown in FIG. 16, the performance data 253 stores information indicating the performance of one user for one training item for each row (one record). This performance data 253 is for each data item of user ID 2531, training ID 2532, average exercise intensity [METs] 2533, maximum exercise intensity [METs] 2534, average speed [m / sec] 2535, and maximum speed [m / sec] 2536. Is stored.

The user ID 2531 is a unique identifier for identifying the user wearing the sensor device 10. This user ID 2531 is the same information as the user ID 2501 of the sensor data 250 (see FIG. 3).
The training ID 2532 is a unique identifier for identifying each training item. This training ID 2532 is the same information as the training ID 2521 of the training classification data 252 (see FIG. 14).

Average exercise intensity [METs] 2533 indicates an average value of the exercise intensity of the user during the training section of the training item.
Maximum exercise intensity [METs] 2534 indicates the maximum value of the user's exercise intensity during the training section of the training item.

The average speed [m / sec] 2535 indicates the average value of the moving speed of the user during the training section of the training item.
The maximum speed [m / sec] 2536 indicates the maximum value of the moving speed of the user during the training section of the training item.

  The performance evaluation unit 242 generates data on the items for evaluating performance as described above for a section to be analyzed for predetermined training, and stores the data as performance data 253 in the database (DB) 25.

Returning to FIG. 15, next, the visualization means 243 of the training classification server 20 refers to the training classification data 252 and the performance data 253, and generates performance evaluation information that becomes comparison data of the same type of training at different dates and times ( Step S21).
This performance evaluation information is information for displaying the analysis result of training on the communication terminal 30 as an easy-to-understand graph or the like for the user or his / her instructor (coach or the like).

For example, the visualization means 243 receives the performance evaluation request information requesting the analysis result of the performance evaluation from the communication terminal 30 to start the performance evaluation information generation process or the performance evaluation information of the training classification server 20 in advance. If the set condition is satisfied as a generation process start condition, the performance evaluation information generation process can be started.
The performance evaluation request information received from the communication terminal 30 stores, for example, training ID, user ID, evaluation items (for example, average exercise intensity 2533) to be evaluated, and training period information. These pieces of information are also stored in the performance evaluation information generation conditions set in advance.

  Subsequently, the visualization unit 243 transmits the generated performance evaluation information to the communication terminal 30 (step S22) and causes the display unit 34 of the communication terminal 30 to display the performance evaluation information.

FIG. 17 shows an example of a performance evaluation diagram 300 in which the performance evaluation information generated by the visualization unit 243 of the training classification server 20 according to the present embodiment is displayed on the display unit 34 by the analysis result display unit 363 of the communication terminal 30. .
As shown in FIG. 17, the performance evaluation diagram 300 includes a plurality of periods (first to second) for the same type of training items (here, “4-to-2 pass practice”) divided and classified by the training classification unit 241. In the 4th week), data of evaluation items (average exercise intensity) are read and the transition is visualized. Thereby, the user and its instructor (coach etc.) can grasp | ascertain correctly the change and growth of each user's performance.

  As described above, according to the training classification system 1, the training classification method, and the training classification server 20 according to the present embodiment, when a plurality of trainings in a sport are continuously performed, different types of training times are obtained. While dividing, similar training can be classified by type. In addition, it is possible to extract only data of training sections classified into the same type, and to compare performance such as exercise intensity for each time series or each user.

Furthermore, according to the training classification system 1, the training classification method, and the training classification server 20 according to the present embodiment, the following effects can also be achieved.
The exercise data (sensor data) of a plurality of users measured by the sensor device 10 can be collected together in the training classification server 20, and the training, the user performance during the training, etc. can be collectively stored, analyzed, and visualized. be able to.

  In addition, by the training classification process shown in FIG. 2, it is possible to extract movement features (motion feature amounts) peculiar to various sport events from the sensor data, and to detect intervals using the time series data. As a result, the entire continuous training section is divided into training sections for executing each training item, and training items having similar motion characteristics (motion feature quantities) can be classified and stored together. it can.

  In addition, the interval detection process shown in FIG. 6 makes it possible to accurately calculate the interval by utilizing the fact that the movement characteristics (motion feature quantities) of a plurality of users in time series are similar, particularly in a team competition (such as soccer). Can be detected.

  Further, the sensor data 250 (FIG. 3), which is a simple physical quantity, is converted from a plurality of physical activity feature quantities (for example, exercise intensity, walking pitch, speed, arm, etc.) indicating exercise such as the exercise feature quantity data 251 of FIG. Can be made easier to discriminate the similarity and interval of training.

  Further, by adopting the data structure of the training classification data 252 as shown in FIG. 14, only the sensor data is added to the information on the divided training items obtained by analyzing the sensor data in addition to the team and the implementing user. Input information (name of training, subjective information, environmental information, etc.) that cannot be discriminated can be centrally managed. Thereby, when a user etc. browses this training classification data 252 after training, the contents of training can be easily grasped.

  In addition, with the data structure of the performance data 253 as shown in FIG. 16, various statistics of exercise intensity and speed indicating the user's performance can be recorded for each training item. Therefore, it is possible to read and visualize the optimum statistics (performance data) for each training item.

  In addition, a performance evaluation chart 300 as shown in FIG. 17 is created, and for each training of the same type, different date / time (time-series date / time) and performance data 253 of a plurality of users can be read and displayed. Users' performance can be compared fairly.

1 training classification system 5 network 10 sensor device 11 MCU
12, 22, 32 Communication unit 13, 33 Input unit 14, 34 Display unit 15 Flash memory 16 Acceleration sensor 20 Training classification server 21, 31 CPU
23, 35 RAM
24 programs 25 database (DB)
30 Communication Terminal 36 Application 241 Training Classification Unit 242 Performance Evaluation Unit 243 Visualization Unit 250 Sensor Data 251 Movement Feature Data 252 Training Classification Data 253 Performance Data 300 Performance Evaluation Diagram 361 Sensor Data Transfer Unit 362 Information Input Unit 363 Analysis Result Display Unit

Claims (7)

A sensor device that measures exercise when performing sports training with a sensor, a communication terminal that is connected to one or more of the sensor devices, and training that acquires sensor data measured by the sensor via the communication terminal A classification server,
The communication terminal is
Sensor data transfer means for acquiring the sensor data measured by the sensor device and transferring it to the training classification server,
The training classification server
The sensor data is acquired, and the acquired sensor data is analyzed to calculate an exercise feature amount indicating an exercise feature, which matches the feature of the exercise feature amount of the sport in an interval that is a time between successive training. The section to be detected is detected as the interval, and the entire training section is divided into individual training sections by dividing the section by the interval, and similar to each other using the motion feature values of the plurality of divided training sections. A training classification system comprising: a training classification unit that classifies the training so that training items of a plurality of training sections serving as exercise feature values are combined into one training item. The training classification server
Performance evaluation means for generating performance data indicating the state of exercise in the training section using the exercise feature amount in the training section of the training item for each classified training item;
Visualization means for generating performance evaluation information for comparing the performance data of the generated training item and past performance data of the same training item generated in the past, and transmitting the performance evaluation information to the communication terminal,
The communication terminal is
The training classification system according to claim 1, further comprising analysis result display means for receiving the performance evaluation information and displaying the performance evaluation information on a display device. The training classification means acquires training-related information, which is information that cannot be detected by the sensor device, including the training name of the training item from the communication terminal, stores the training-related information in association with the classified training item,
The training classification system according to claim 2, wherein the visualization means generates performance evaluation information with the training-related information and transmits the performance evaluation information to the communication terminal. The sensor data transfer means of the communication terminal acquires the sensor data measured in the same entire training section from each of the plurality of sensor devices, and transfers the sensor data to the training classification server.
The training classification means of the training classification server analyzes each of a plurality of acquired sensor data to calculate the motion feature amount, generates average value data of the calculated plurality of motion feature amounts, and generates the generated average 4. The section according to claim 1, wherein a section that matches the feature of the movement feature amount at the interval in the sport is detected as the interval using the value data. 5. Training classification system. The exercise feature amount is an exercise intensity indicating an exercise amount per unit time,
The said training classification | category means calculates the said exercise intensity from the said sensor data, and detects the area which corresponds to the characteristic of the exercise intensity in the said interval in the said sport as the said interval. 5. The training classification system according to any one of 4 above. Sensor device
Execute the step of measuring the exercise when doing sports training with the sensor,
The communication terminal
Acquiring sensor data measured by the sensor device and transferring it to a training classification server;
The training classification server is
The sensor data is acquired, and the acquired sensor data is analyzed to calculate an exercise feature amount indicating an exercise feature, which matches the feature of the exercise feature amount of the sport in an interval that is a time between successive training. The section to be detected is detected as the interval, and the entire training section is divided into individual training sections by dividing the section by the interval, and similar to each other using the motion feature values of the plurality of divided training sections. A training classification method of a training classification system, comprising: performing the step of classifying the training so that training items of a plurality of training sections serving as exercise feature values are combined into one training item. Acquire sensor data that measures the exercise during sports training, analyze the acquired sensor data to calculate the exercise feature amount indicating the feature of the exercise, and in the interval that is the time between successive training A section that matches the characteristics of the sport feature quantity of sports is detected as the interval, and the entire training section is divided into individual training sections by dividing the section by the interval, and each of the divided plurality of training sections A training classification server characterized by comprising training classification means for classifying the training so that training items of a plurality of training sections that are similar to the movement feature amount are combined into one training item using an exercise feature amount .