Behavior analysis method, behavior analysis system, and program

Abstract

A behavior analysis method according to the present disclosure involves: collecting detection results of inertial sensors attached to a plurality of players who execute a specific collision sport; calculating, from the detection results of the inertial sensors, accelerations and angular velocities in global coordinates of the plurality of players to which the inertial sensors are attached; and outputting the calculated accelerations and angular velocities in the global coordinates of the plurality of players at a specific timing in order to detect specific behaviors among the plurality of players.

Description

Behavior analysis method, behavior analysis system, and program

[0001] The present disclosure relates to a behavior analysis method, a behavior analysis system, and a program.

[0002] Collision sports such as rugby, American football, and wrestling tend to pose a high risk of injury due to contact between players. Therefore, ensuring safety during collision sports is crucial. For example, the scrum during a rugby match, in which multiple players crowd together and push against each other, is a highly injury-prone action. Specifically, if various factors disrupt the balance of power between the teams pushing against each other during a scrum (known as "scrum collapsing" or simply "collapsing"), players in the front row of the scrum (more specifically, players in the front row) are particularly at risk of concussion and severe head and neck injuries from head-on impact with the ground. They are also at risk of second-impact syndrome following a concussion. Collapsing is considered a foul during competition, and referees generally halt the game if a collapse occurs to prevent injury.

[0003] For example, International Publication No. 2017 / 183495 describes a technology for monitoring the position, movements, etc. of athletes and the like by attaching various sensors to their clothing.

[0004] From the perspective of improving the safety of a sport, it is important to prevent players from performing actions that pose a high risk of injury, such as the scrum collapsing described above. Therefore, if specific high-risk injury behaviors could be accurately detected in advance or at an early stage, it may be possible to prevent injuries or reduce their severity. Furthermore, if the occurrence of specific high-risk injury behaviors could be identified, it would be possible to prevent injuries by practicing to avoid the occurrence of those behaviors. From this perspective, there is a need for a method for collecting data that can identify players' behaviors and providing that data to users.

[0005] In view of the above-mentioned problems, the present disclosure provides a behavior analysis method, a behavior analysis system, and a program for detecting specific behaviors in collision sports.

[0006] A behavior analysis method according to a first aspect of the present disclosure collects detection results from inertial sensors attached to multiple athletes participating in a specific collision sport, calculates the acceleration and angular velocity in global coordinates of the multiple athletes to which the inertial sensors are attached from the detection results of the inertial sensors, and outputs the calculated acceleration and angular velocity in the global coordinates of the multiple athletes at a specific timing in order to detect specific behavior between the multiple athletes.

[0007] This behavior analysis method can output the acceleration and angular velocity of multiple athletes in the global coordinate system, making it possible to identify with high accuracy the acceleration and angular velocity occurring in each athlete. This makes it possible to detect specific behaviors between athletes by checking the output data. It can also contribute to elucidating the mechanisms behind specific behaviors.

[0008] A behavior analysis method according to a second aspect of the present disclosure is the behavior analysis method according to the first aspect of the present disclosure, wherein the specific collision sport includes rugby, and the specific behavior includes behavior in a scrum.

[0009] This behavior analysis method makes it possible to understand the behavior of players during scrums, where there is a high risk of injury.

[0010] A behavior analysis method according to a third aspect of the present disclosure is the behavior analysis method according to the second aspect of the present disclosure, wherein the behavior in the scrum includes a precursor or occurrence of scrum collapse.

[0011] Such a behavior analysis method can be used to detect scrum collapses, which pose a high risk of injury.

[0012] A behavior analysis method according to a fourth aspect of the present disclosure is a behavior analysis method according to the second or third aspect of the present disclosure, which estimates or identifies the player or area that was a precursor to or caused the scrum collapse.

[0013] This type of behavior analysis method can estimate or identify the players or areas that are precursors to or cause scrum collapses, and can be used to prevent injuries or create training menus that reduce the risk of injury.

[0014] A behavior analysis method according to a fifth aspect of the present disclosure is a behavior analysis method according to any one of the second to fourth aspects of the present disclosure, in which the behavior in the scrum is behavior that differs from the comparison target.

[0015] In such a behavior analysis method, by focusing on differences from a specific comparison target, it is possible to grasp a specific behavior with high accuracy.

[0016] A behavior analysis method according to a sixth aspect of the present disclosure is a behavior analysis method according to any one of the second to fifth aspects of the present disclosure, wherein the difference is a difference in at least one of the connection strength between the players forming the scrum, the stress between the players forming the scrum, and the inclination of the scrum.

[0017] In such a behavior analysis method, it is possible to output data that focuses on parameters that have a strong correlation with scram collapsing.

[0018] A behavior analysis method according to a seventh aspect of the present disclosure is a behavior analysis method according to any one of the second to sixth aspects of the present disclosure, in which the analysis results of behavior in the scrum are displayed superimposed on an image showing the players forming the scrum.

[0019] Such a behavior analysis method makes it easier for the user to understand the behavior in a scrum.

[0020] A behavior analysis method according to an eighth aspect of the present disclosure is a behavior analysis method according to any one of the second to seventh aspects of the present disclosure, wherein the inertial sensor is attached to at least a front-row player among the plurality of players playing the rugby.

[0021] In such a behavior analysis method, scram collapse can be detected with high accuracy using a small number of inertial sensors.

[0022] A behavior analysis method according to a ninth aspect of the present disclosure is a behavior analysis method according to any one of the first to eighth aspects of the present disclosure, in which an extended Kalman filter is applied to the detection results of the inertial sensor before they are used to calculate the acceleration and the angular velocity in the global coordinate system.

[0023] In such a behavior analysis method, the application of an extended Kalman filter can eliminate or reduce errors in the measurement results of the inertial sensor.

[0024] A behavior analysis method according to a tenth aspect of the present disclosure is the behavior analysis method according to any one of the first to ninth aspects of the present disclosure, wherein the inertial sensors are attached to the torsos of the multiple players.

[0025] In such a behavior analysis method, noise contained in the detection results of the inertial sensor can be suppressed.

[0026] A behavior analysis method according to an eleventh aspect of the present disclosure is a behavior analysis method according to any one of the first to tenth aspects of the present disclosure, wherein the calculated acceleration and angular velocity in the global coordinates are output in the form of transition data of the acceleration in the horizontal direction converted using a mechanical model, and transition data of the angular velocity in the up and down rotation directions converted using a mechanical model.

[0027] In such a behavior analysis method, by monitoring transition data and evaluating its stability, it becomes easy to estimate or identify the occurrence of collapse and the player or area that caused the collapse.

[0028] A behavior analysis system according to a twelfth aspect of the present disclosure includes inertial sensors attached to a plurality of athletes participating in a specific collision sport, and a computing device that calculates the acceleration and angular velocity in global coordinates of the plurality of athletes to which the inertial sensors are attached from the detection results of the inertial sensors, and outputs the calculated acceleration and angular velocity of the plurality of athletes at a predetermined timing in order to detect specific behavior between the plurality of athletes.

[0029] This behavior analysis system can output the acceleration and angular velocity of multiple athletes in a global coordinate system, making it possible to identify with high accuracy the acceleration and angular velocity occurring in each athlete. This makes it possible to detect specific behaviors between athletes by checking the output data. It can also contribute to elucidating the mechanisms behind specific behaviors.

[0030] A behavior analysis system according to a thirteenth aspect of the present disclosure is the behavior analysis system according to the twelfth aspect of the present disclosure, further comprising an imaging device that images the plurality of athletes participating in the specific collision sport, and the computing device identifies the predetermined timing based on the imaging results of the imaging device, or the imaging results of the imaging device and the calculated acceleration and angular velocity.

[0031] In such a behavior analysis system, the timing of the start of a scrum can be easily identified by using images captured by an imaging device, and the timing at which a particular behavior may occur can be identified.

[0032] A program according to a fourteenth aspect of the present disclosure causes a computer processor to execute processing to collect detection results from inertial sensors attached to multiple athletes participating in a specific collision sport, calculate accelerations and angular velocities in global coordinates of the multiple athletes to which the inertial sensors are attached from the detection results of the inertial sensors, and output the calculated accelerations and angular velocities of the multiple athletes at specific timings in order to detect specific behaviors between the multiple athletes.

[0033] Such a program can output the acceleration and angular velocity of multiple athletes in the global coordinate system, making it possible to identify with high accuracy the acceleration and angular velocity occurring in each athlete. This makes it possible to detect specific behaviors between athletes by checking the output data. It can also contribute to elucidating the mechanisms behind specific behaviors.

[0034] The behavior analysis method, behavior analysis system, and program disclosed herein can detect specific behaviors in collision sports.

[0035] 11 is an explanatory diagram showing an example of a scrum performed by a plurality of athletes as seen from the side. FIG. 12 is an explanatory diagram showing an example of a scrum performed by a plurality of athletes as seen from above. FIG. 13 is a block diagram showing an example of the hardware configuration of a behavior analysis system according to an embodiment of the present disclosure. FIG. 14 is a functional block diagram schematically showing each function of the arithmetic device shown in FIG. 3. FIG. 15 is a graph showing the change over time in acceleration in a local coordinate system acquired by an inertial sensor. FIG. 16 is a graph showing the change over time in angular velocity in a local coordinate system acquired by an inertial sensor. FIG. 17 is a graph showing the change over time in roll angle and pitch angle for identifying the posture of each athlete, as an example of output data. FIG. 18 is a graph showing the change over time in the horizontal component of acceleration of each athlete, as an example of output data. FIG. 19 is a graph showing the change over time in angular velocity around the top and bottom of each athlete, as an example of output data. FIG. 19 is a diagram schematically showing a state in which a plurality of athletes form a scrum using a dynamic model. FIG. 19 is a diagram showing a dynamic model related to horizontal acceleration focusing on athletes in the front row. FIG. 19 is a graph showing the change over time in the elastic coefficient of a spring and the damping coefficient of a damper in the dynamic model of FIG. 14 is a diagram showing a dynamic model relating to angular velocity in the rotational direction focusing on a rider in the front row. FIG. 15 is a graph showing changes over time in the elastic coefficient of a spring and the damping coefficient of a damper in the dynamic model of FIG. 13. FIG. 16 is a flowchart showing an example of a behavior analysis method according to an embodiment of the present disclosure. FIG. 17 is a diagram showing an example of a method for displaying behavior analysis results. FIG. 18 is a diagram showing an example of a method for displaying behavior analysis results.

[0036] This application is based on Japanese Patent Application No. 2024-027508, filed on February 27, 2024, in Japan, the contents of which are incorporated herein by reference. The present disclosure will become more fully understood from the following detailed description. Further scope of application of the present application will become apparent from the following detailed description. However, the detailed description and specific examples are preferred embodiments of the present disclosure and are set forth for illustrative purposes only. From this detailed description, various changes and modifications will be apparent to those skilled in the art within the spirit and scope of the present disclosure. The applicant does not intend to dedicate any of the described embodiments to the public, and the applicants also consider disclosed modifications and alternatives, even if not literally included within the scope of the claims, to be part of the invention under the doctrine of equivalents. Like reference numbers and names in the various drawings indicate like elements.

[0037] Hereinafter, each embodiment for carrying out the present disclosure will be described with reference to the drawings. Note that the scope necessary for the explanation to achieve the object of the present disclosure will be schematically shown below, and the scope necessary for explaining the relevant parts of the present disclosure will be mainly explained, and the parts for which explanation is omitted will be referred to as publicly known technologies. Furthermore, identical or similar reference numerals will be used for identical or corresponding components in the drawings, and duplicate explanations will be omitted. Furthermore, when a plurality of identical or corresponding components are included in the drawings, only some of them may be referenced to make the drawings easier to understand.

[0038] The behavior analysis method, behavior analysis system, and program according to one embodiment described below use rugby as an example of a collision sport played by athletes, and detect behavior in a rugby scrum, particularly behavior related to the occurrence of collapsing, as a specific behavior. Note that collision sports to which the present disclosure is applicable are not limited to rugby. Furthermore, the specific behavior is not particularly limited as long as it is behavior related to a specific movement in a collision sport, and may be behavior related to a movement other than collapsing.

[0039] Figure 1 is an explanatory side view of an example of a scrum performed by multiple players. Figure 2 is an explanatory top view of an example of a scrum performed by multiple players. A scrum in rugby is conducted when play is stopped due to a foul (e.g., a knock-on or a forward throw) during a match. As shown in Figures 1 and 2, a scrum is usually formed by eight forward players from each of two teams, TA and TB. The forward players PA1-PA8, PB1-PB8 are respectively made up of the three front row players (hereinafter simply referred to as the "front row") PA1-PA3, PB1-PB3 who are positioned furthest forward, the two second row players (hereinafter simply referred to as the "second row") PA4-PA5, PB4-PB5 who are positioned behind the three front row players PA1-PA3, PB1-PB3, and the three back row players (hereinafter simply referred to as the "back row") PA6-PA8, PB6-PB8 who are positioned behind the second row players PA4-PA5, PB4-PB5. In the following explanation, the direction indicated by arrow X in Figures 1 and 2 will be taken as the front-to-back direction, and similarly, the direction indicated by arrow Y will be taken as the left-to-right direction, and the direction indicated by arrow Z will be taken as the up-to-down direction.

[0040] Of the players involved in the scrum described above, the six players in the front rows PA1-PA3 and PB1-PB3 are at the highest risk of injury when a collapse occurs. In relation to this, when a collapse occurs, the front rows PA1-PA3 and PB1-PB3 tend to experience the greatest acceleration and other displacements. Therefore, in this embodiment, inertial sensors SA1-SA3 and SB1-SB3 are attached to at least the front rows PA1-PA3 and PB1-PB3 of each team to monitor their behavior. Note that the players to whom inertial sensors are attached are not limited to the front rows PA1-PA3 and PB1-PB3 described above. For example, all players PA1-PA8 and PB1-PB8 involved in the scrum may be attached.

[0041] The inertial sensors SA1-SA3 and SB1-SB3 are each composed of a sensor capable of measuring at least the acceleration and angular velocity of an object to which they are attached. For example, an IMU (Inertial Measurement Unit) sensor can be used as the inertial sensors SA1-SA3 and SB1-SB3. In this embodiment, the inertial sensors SA1-SA3 and SB1-SB3 are illustrated as sensors that detect the acceleration and angular velocity of the object, i.e., the six front row passengers PA1-PA3 and PB1-PB3, but may also include other sensor functions. Specifically, the inertial sensors SA1-SA3 and SB1-SB3 may include a position sensor (e.g., a GPS (Global Positioning System)) that can detect the position of the inertial sensors SA1-SA3 and SB1-SB3. If the inertial sensors SA1 to SA3 and SB1 to SB3 include a GPS, this can be used to identify the positions of multiple athletes during the behavior analysis described below.

[0042] Furthermore, the inertial sensors SA1 to SA3 and SB1 to SB3 can be attached to any position on the athlete's body, but it is preferable to attach them to the athlete's torso, as this allows for accurate measurement of the athlete's movements. Based on this, this embodiment illustrates an example in which the sensors are attached to the athlete's back (particularly the center between the shoulders) via dedicated clothing, such as underwear, to minimize the impact on the competition.

[0043] The behavior analysis method, behavior analysis system, and program according to this embodiment detect a precursor to scrum collapse or its occurrence as a specific behavior by analyzing the detection results of the inertial sensors SA1 to SA3 and SB1 to SB3. Details of the behavior analysis system 1 according to this embodiment will be described below.

[0044] 3 is a block diagram showing an example of the hardware configuration of a behavior analysis system according to an embodiment of the present disclosure. The behavior analysis system 1 according to this embodiment includes at least inertial sensors attached to multiple athletes (specifically, the above-mentioned inertial sensors SA1 to SA3 and SB1 to SB3), and a computing device 10 that uses the detection results of the inertial sensors to output data for detecting specific behaviors between the multiple athletes.

[0045] The computing device 10 may be configured as a well-known computer, as shown in Fig. 3. The computer may include, for example, a CPU (Central Processing Unit) 11 as an example of a processor, a ROM (Read Only Memory) 12 and a RAM (Random Access Memory) 13 as examples of memory, a storage 14, a communication interface 15, and an input / output interface 16. The above-mentioned components may be connected to each other via an internal bus 17 so as to be able to communicate with each other.

[0046] The CPU 11 is a central processing unit and is an example of a processor that can execute various programs and perform arbitrary arithmetic processing. Specifically, the CPU 11 may be capable of reading various programs stored in the ROM 12 and / or the storage 14 and executing the programs using the RAM 13 as a work area. In this embodiment, the CPU 11 controls each component constituting the arithmetic device 10 and performs various arithmetic processing in accordance with the programs.

[0047] The ROM 12 may be capable of storing various programs and various data, and the RAM 13 may be capable of temporarily storing programs or data as a working area.

[0048] The storage 14 may be configured with a recording medium such as a hard disk drive (HDD), a solid state drive (SSD), or a flash memory, and may store various programs including an operating system and various data necessary to operate the computing device 10. In this embodiment, the ROM 12 or the storage 14 may store programs and various data used to analyze specific behaviors, including behaviors when collapse occurs and behaviors that are precursors to the occurrence of collapse.

[0049] The communication interface (I / F) 15 may be an interface capable of wireless communication, which allows the computing device 10 to communicate with a server computer (not shown), various databases, other terminal devices, etc. via a network, etc. This communication interface 15 may use communication standards such as CAN (Controller Area Network), Ethernet (registered trademark), LTE (Long Term Evolution), FDDI (Fiber Distributed Data Interface), or Wi-Fi (registered trademark).

[0050] The input / output interface (I / F) 16 may be an interface for transmitting and receiving data between the components of the behavior analysis system 1, including at least the inertial sensors SA1 to SA3 and SB1 to SB3. In this embodiment, the input / output interface 16 and the inertial sensors SA1 to SA3 and SB1 to SB3 are connected via wireless communication. The components electrically connected to the input / output interface 16 are not limited to the inertial sensors SA1 to SA3 and SB1 to SB3 described above, and various other components can be assumed, and are not particularly limited. Specifically, the components may include a display (not shown), such as an LCD monitor, a user interface including input means such as a keyboard, a touch panel, or a pointing device, or an external memory.

[0051] In addition to the above-described configuration, the behavior analysis system 1 according to the present embodiment may further include a camera 20 as an example of an imaging device capable of capturing images of multiple athletes during a competition. The camera 20 may be, for example, a two-dimensional camera using an imaging element such as a CCD (Charge Coupled Device) sensor or a CMOS (Complementary Metal Oxide Semiconductor) sensor, but is not limited thereto. The number and placement of the cameras 20 are not particularly limited as long as they are capable of capturing images of the athletes, but may be located on the left and right sides of the competition space and above, for example, as shown in FIGS. 1 and 2 .

[0052] Images captured by camera 20 can be used, for example, to detect the start of a scrum, or to assist in determining the occurrence of a scrum collapse (or its precursors) and identifying the player who caused the scrum collapse.

[0053] Fig. 4 is a functional block diagram schematically showing each function of the arithmetic device shown in Fig. 3. As shown in Fig. 4, the arithmetic device 10 functions to include a data acquisition unit 31, a filter unit 32, a coordinate conversion unit 33, an output data generation unit 34, and an output unit 35. Note that Fig. 4 mainly illustrates only functions related to the generation of output data, and omits functions that are not directly related to the generation of output data.

[0054] The data acquisition unit 31 is mainly realized by the input / output interface 16 and acquires necessary information from the inertial sensors SA1-SA3, SB1-SB3 and the camera 20. Specifically, the data acquired by the data acquisition unit 31 may include the acceleration and angular velocity of the front-row players PA1-PA3, PB1-PB3 equipped with the inertial sensors SA1-SA3, SB1-SB3, and images captured by the camera 20. Of these, the images captured by the camera 20 can be used to identify the start or end of a scrum, or, in the event of a collapse, can be used as a supplementary means to identify the player who caused the collapse. Therefore, it is preferable that the captured images be at least temporarily stored in the storage 14 or the like.

[0055] FIG. 5 is a graph showing the change over time in acceleration in the local coordinate system acquired by the inertial sensor. FIG. 6 is a graph showing the change over time in angular velocity in the local coordinate system acquired by the inertial sensor. Note that FIGS. 5 and 6 show the change over time in acceleration and angular velocity for each of the inertial sensors SA1-SA3 and SB1-SB3. The accelerations and angular velocities of the front rows PA1-PA3 and PB1-PB3 acquired by the data acquisition unit 31 are the accelerations and angular velocities in the local coordinate system based on the inertial sensors SA1-SA3 and SB1-SB3, as shown in FIGS. 5 and 6.

[0056] The filter unit 32, which is primarily implemented by the CPU 11, performs measurement value estimation taking into account errors contained in the acceleration and angular velocity in the local coordinate system acquired by the data acquisition unit 31. In this embodiment, an extended Kalman filter is employed as the specific configuration of the filter unit 32. A general Kalman filter refers to an algorithm that estimates the system state from measurement data. The extended Kalman filter used in this embodiment linearizes nonlinear state equations and observation equations with respect to the local coordinate systems of the inertial sensors SA1 to SA3 and SB1 to SB3, creating an algorithm similar to the theory of the Kalman filter. The filter unit 32 predicts the current situation using the linearized coordinate system, corrects the state estimation using the measurement and observation values ​​output by the inertial sensors SA1 to SA3 and SB1 to SB3, taking into account errors between the predicted values ​​and the observed values, and updates the situation while taking into account the true values ​​of the inertial sensors SA1 to SA3 and SB1 to SB3. By repeating such prediction and updating in the filter unit 32, it is possible to accurately estimate the state of the measurement values ​​of the nonlinear inertial sensors SA1 to SA3 and SB1 to SB3.

[0057] The coordinate conversion unit 33 is mainly realized by the CPU 11, and converts the acceleration and angular velocity of the local coordinate system (also called the object coordinate system) detected by the inertial sensors SA1 to SA3 and SB1 to SB3 into the acceleration and angular velocity of the global coordinate system (also called the world coordinate system or the absolute coordinate system). An example of the coordinate conversion in the coordinate conversion unit 33 will be described below.

[0058] First, a rotation matrix R to be used when converting from the local coordinate system to the global coordinate system is identified. The rotation matrix R can be expressed as in the following equation (1) using the attitude angles of each rider, more specifically, the roll angle φ and pitch angle θ. Here, the roll angle φ and pitch angle θ will be explained using one front-row rider PB3 shown in Figures 1 and 2 as an example. The roll angle φ refers to the angle of rotation about the roll axis A1, and the pitch angle θ refers to the angle of rotation about the pitch axis A2.

[0059] The relationship between the acceleration and angular velocity in the local coordinate system and the acceleration and angular velocity in the global coordinate system can be expressed as the following equations (2) and (3) using the rotation matrix R described above. In addition, A in the above formula ｘ , A ｙ , A ｚ , ω ｘ , ω ｙ , ω ｚ are the acceleration and angular velocity in the local coordinate system, and A ｘ０ , A ｙ０ , A ｚ０ , ω ｘ０ , ω ｙ０ , ω ｚ０ are the acceleration and angular velocity in global coordinates.

[0060] In this embodiment, the coordinate conversion unit 33 converts the acceleration and angular velocity values ​​from the local coordinate system to the global coordinate system because the acceleration and angular velocity values ​​in the local coordinate system are not uniquely determined depending on the athlete's posture and angle, which reduces the accuracy of detecting specific behavior.

[0061] The output data generation unit 34 generates, as output data, data usable by a user to detect collapsing, from the accelerations and angular velocities in the global coordinate system converted by the coordinate transformation unit 33 and information about the posture identified via the filter unit 32 (specifically, the roll angle φ and the pitch angle θ). In other words, the output data generation unit 34 generates, as output data, the accelerations and angular velocities in the global coordinate system of multiple athletes at specific timings. Here, the specific timing refers to a timing at which a specific behavior can occur, and in this embodiment, it refers to a timing that includes the period from the start of the scrum formation movement to the end of the scrum. Furthermore, the specific timing can be identified by analyzing image data captured by the camera 20, or, in cases where there are movements defined in relation to specific behaviors, such as in a scrum (specifically, the crouch, bind, and set movements), by analyzing the detection results of the inertial sensors SA1 to SA3 and SB1 to SB3.

[0062] Furthermore, the format of the output data generated by the output data generation unit 34 is not particularly limited as long as it can quantitatively display the athlete's movements. As an example, output data represented by graphs such as those shown in Figures 7 to 9 below at any timing can be generated.

[0063] FIG. 7 is a graph showing, as an example of output data, the changes in roll angle and pitch angle over time for identifying the posture of each athlete. FIG. 8 is a graph showing, as an example of output data, the changes in the horizontal component of acceleration of each athlete over time. FIG. 9 is a graph showing, as an example of output data, the changes in angular velocity around the top and bottom of each athlete over time. In FIGS. 7 to 9, the posture, acceleration, and angular velocity of each of the front rows PA1 to PA3 and PB1 to PB3 are shown in separate graphs. The changes in roll angle and pitch angle over time shown in FIG. 7 can be primarily used to extract characteristics of the athlete's posture.

[0064] Furthermore, the changes in acceleration and angular velocity over time shown in Figures 8 and 9 can be used primarily to extract characteristics of how the front row riders PA1-PA3 and PB1-PB3 moved. As shown in Figures 8 and 9, it is important from the perspective of detecting specific behavior between multiple riders that the output data be in a format that allows the acceleration and angular velocity of several or all of the multiple riders to be identified. Furthermore, because the acceleration and angular velocity shown in Figures 8 and 9 are values ​​shown in the global coordinate system, the values ​​are specific and it is easier to detect specific behavior than with values ​​shown in the local coordinate system.

[0065] When a collapse occurred, several sets of transition data shown in the graphs in Figures 8 and 9 were generated as output data and checked. In all of the results, it was possible to identify that a specific athlete was generating specific accelerations and angular velocities. It was also found that athletes adjacent to the specific athlete were also affected by the aforementioned specific accelerations and angular velocities, and that this influence diminished in proportion to the distance between the specific athlete and the athlete. Therefore, by checking the above-mentioned output data, it can be said that it is possible to detect the occurrence of a collapse and estimate or identify the athlete or area that caused it with high accuracy.

[0066] Instead of or in addition to the output data described above, it is also possible to generate output data that can estimate the interrelationships between athletes using a dynamic model. Below, we will explain output data that uses a dynamic model as another example of output data generated by the output data generation unit 34.

[0067] Figure 10 is a diagram that shows a scrum formed by multiple players, using a dynamic model. The square on the left side of Figure 10 shows the eight forward players of one team, TA, and the square on the right side shows the eight forward players of the other team, TB. When the scrum formed by the forwards of two teams, TA and TB, is represented by a dynamic model, as shown in Figure 10, there is a spring K between the imaginary boundary line between the two teams, TA and TB, and each team, TA and TB. １ , K. ２ and damper C １ , C ２ Based on this, the state of one team TA at the time of the scrum can be expressed as the following equation (4) using a dynamic model. Similarly, the state of the other team TB at the time of the scrum can be expressed as the following equation (5) using a dynamic model. In addition, m in the above formula １ , m ２ is the mass of each team TA, TB, and c １ , c ２ is the damping coefficient of the damper, and k １ , k ２ is the elastic modulus of the spring.

[0068] By combining the above equations (4) and (5), the dynamic model at the time of scrum can be expressed by the following equation (6). In the above formula (6), mass m １ , m ２ Assuming that the same, the dynamic model during scrum can be expressed by equation (7). In the above formula, a is a coefficient related to the damping on the TA side, b is a coefficient related to the spring (stiffness) on the TA side, e is a coefficient related to the damping on the TB side, and f is a coefficient related to the spring (stiffness) on the TB side.

[0069] One of the causes of a collapsing scrum is the instability of certain players, particularly certain front-row players. Therefore, by applying the above-mentioned dynamic model to the acceleration and angular velocity of each player, more specifically, the horizontal acceleration and angular velocity in the vertical rotational direction, and monitoring these trends, it is highly likely that the occurrence of a collapsing or a precursor to the occurrence of a collapsing scrum can be detected. Therefore, an example of output data using a dynamic model focusing on the movements of each of the front-row players PA1-PA3 and PB1-PB3 is described below.

[0070] FIG. 11 is a diagram showing a dynamic model related to horizontal acceleration focusing on the front-row players. In this example, a dynamic model like that shown in FIG. 11 is generated to identify the horizontal behavior of the players. Here, when multiple players form a scrum, as shown in FIG. 11, it can be assumed that a dynamic model similar to the dynamic model shown in FIG. 10 is generated for each of the front-row players positioned substantially opposite each other (more specifically, three in parallel). The regions in which these three dynamic models are generated are designated as first region E1 to third region E3, respectively, and output data is generated that enables monitoring of the time-varying changes in the elastic coefficient of the spring and the damping coefficient of the damper in each dynamic model. If the above-described monitoring is possible, it becomes easier to estimate or identify in which of the first region E1 to third region E3 a behavior that may cause a collapse has occurred.

[0071] Figure 12 is a graph showing the changes over time in the elastic coefficient of the spring and the damping coefficient of the damper in the dynamic model of Figure 11. When the output data generating unit 34 generates transition data configured as a graph such as that shown in Figure 12 as output data, the user can monitor behavior related to horizontal acceleration in multiple regions, which can make it easier to estimate or identify the cause of the collapse.

[0072] Figure 13 is a diagram showing a dynamic model relating to angular velocity in the vertical rotational direction, focusing on a rider in the front row. Of the first to third regions E1 to E3 described in Figure 11, Figure 13 only shows the first region E1. In this example, a dynamic model such as that shown in Figure 13 is created to identify the rider's behavior in the vertical direction. Output data is then generated that makes it possible to monitor the changes over time in the elastic coefficient of the spring and the damping coefficient of the damper of the dynamic model. If the above-mentioned monitoring is possible, it becomes easier to estimate or identify in which region of the first to third regions E1 to E3 the behavior that causes collapse has occurred.

[0073] Fig. 14 is a graph showing the changes over time in the elastic coefficient of the spring and the damping coefficient of the damper in the dynamic model of Fig. 13. When the output data generating unit 34 generates transition data configured as a graph such as that shown in Fig. 14 as output data, the user can monitor behavior related to angular velocity in the upward and downward rotational directions in a plurality of regions, which can make it easier to estimate or identify the cause of the occurrence of collapse.

[0074] In the above example, the elastic coefficient of the spring and the damping coefficient of the damper are estimated from the above equations, and the stability of the data is monitored in the form of transition data configured in graphs such as those shown in Figures 12 and 14. This makes it possible to accurately detect the occurrence of collapse or signs of collapse.

[0075] Returning to the explanation of FIG. 4 , the output unit 35 is mainly realized by the input / output interface 16 or the communication interface 15, and outputs the output data generated by the output data generation unit 34 to a user interface connected to the calculation device 10 or to a server computer (not shown) connectable via a network. The output data output from the output unit 35 can be used to analyze the cause of a collapse when it occurs. The output operation by the output unit 35 is preferably performed in real time in conjunction with the operation of the inertial sensor. If the output data is output in real time, the occurrence of a collapse or a sign of a collapse can be immediately notified, which can contribute to preventing injuries.

[0076] According to the behavior analysis system 1 of this embodiment, which includes the above-mentioned series of configurations, data reflecting the progression of specific behaviors such as scrum collapsing can be output, and by using this data, it is expected that the behavior in question can be detected with high accuracy.

[0077] One way to use the output data described above is to operate the system during a rugby match as an auxiliary tool for refereeing. In this case, the referee or assistant referee checks the output data output from the behavior analysis system 1 in real time to detect the occurrence of a collapse or its precursor during a scrum. This data can then be used to stop the match or to estimate or identify the player or area that caused the collapse. In particular, the player who caused the collapse may not be identified at a glance by the referee alone who is in the playing area, which can lead to the referee being unsure of the correct decision. Therefore, using this system will contribute to the realization of accurate refereeing.

[0078] Furthermore, the output data of the behavior analysis system according to this embodiment can be used in practice to avoid collapsing. By checking the output data output by this system during scrum practice, it is possible to understand what actions of the players will cause collapsing. Therefore, training can be carried out to avoid actions that cause collapsing, which is expected to reduce the risk of injury during practice and matches.

[0079] Next, a behavior analysis method according to this embodiment will be described below with reference to Fig. 15. The behavior analysis method described below will be exemplified as being implemented using the above-described behavior analysis system 1. The behavior analysis method described below may be implemented by executing a program that causes at least one processor (specifically, CPU 11) of a computer that constitutes the behavior analysis system 1 to execute a predetermined operation. This program may be stored in a storage means such as ROM 12 or storage 14, or may be provided in the form of a non-transitory computer-readable recording medium.

[0080] 15 is a flowchart showing an example of a behavior analysis method according to an embodiment of the present disclosure. The behavior analysis method according to this embodiment includes at least a step of collecting detection results from inertial sensors attached to multiple players playing rugby (corresponding to step S01 described below), a step of calculating accelerations and angular velocities in global coordinates of the multiple players attached to the inertial sensors from the detection results of the inertial sensors (corresponding to step S05 described below), and a step of outputting the calculated accelerations and angular velocities in global coordinates of the multiple players at a specific timing in order to detect collapsing as a specific behavior between the multiple players (corresponding to step S07 described below). This will be described in detail below.

[0081] In the behavior analysis method according to this embodiment, when the behavior analysis system 1 starts operating, it first activates the inertial sensors SA1-SA3 and SB1-SB3 attached to the front row athletes PA1-PA3 and PB1-PB3, and starts receiving data in the calculation device 10 (step S01). The behavior analysis system 1 also activates the camera 20 to start capturing images of the competition space including the athletes PA1-PA8 and PB1-PB8, and starts receiving image data in the calculation device 10 (step S02).

[0082] Next, the calculation device 10 filters the data detected by the inertial sensors SA1 to SA3 and SB1 to SB3 received by the data acquisition unit 31, specifically the accelerations and angular velocities in the local coordinate systems of the inertial sensors SA1 to SA3 and SB1 to SB3, in the filter unit 32 (step S03). The filtered accelerations and angular velocities are then converted from the local coordinate system to the global coordinate system in the coordinate conversion unit 33 (step S04).

[0083] Next, the output data generation unit 34 generates output data based on the acceleration, angular velocity, and attitude information after conversion to the global coordinate system (step S05). The output data may be in the form of transition data as shown in Figures 7 to 9. The timing at which the output data is generated includes at least a period during which a specific behavior, such as during a scrum, may occur. The generated output data is then output to a predetermined output destination by the output unit 35 (step S06).

[0084] In the above example, output data is output in real time based on the detection results of the inertial sensors SA1 to SA3 and SB1 to SB3, but the output timing of the output data is not limited to this. For example, output data may be generated and output only when the start of a scrum is identified based on image data captured by the camera 20 or the detection results of the inertial sensors SA1 to SA3 and SB1 to SB3. Alternatively, output of the output data may be executed only when requested by the user.

[0085] As described above, the behavior analysis method according to this embodiment can also output data that reflects the progression of specific behaviors, such as scrum collapsing. This allows users to use the data to accurately detect the occurrence of the specific behaviors. Furthermore, the detection results of specific behaviors, including their precursors, can be used as information to assist referees in refereeing during a match or to determine training strategies that reduce the risk of injury.

[0086] Next, an example of a method for providing the output data obtained by the above-described behavior analysis method to a user, for example, a match referee or a rugby team coach, will be described.

[0087] FIG. 16 is a diagram schematically illustrating an example of a method for displaying behavior analysis results. FIG. 16A illustrates a state in which scrum collapse has not occurred, and FIG. 16B illustrates a state in which scrum collapse has occurred. For example, the analysis results obtained by the behavior analysis method described above can be provided to a specific user by superimposing them on an image of the players forming the scrum. A specific display method is to superimpose the analysis results on an image in which the players forming the scrum are schematically represented by uniform marks M1, as shown in FIGS. 16A and 16B . The analysis results superimposed on the image here represent the scrum's connection strength using shades of color for the uniform marks M1, the strength of the stress between the players forming the scrum using shades of color for a rectangular mark M2 that surrounds the entire scrum, and the left-right and up-down tilt of the scrum by rotating or deforming the rectangular mark M2 into a trapezoid.

[0088] Here, the image showing the players forming a scrum may include an actual image, a schematic drawing, or the like, and the image may be a two-dimensional or three-dimensional image. The displayed image may show the players from above, from the side, from the front, or from the back, or the viewpoint may be adjustable as needed. Furthermore, the analysis results overlaid on the image may be, for example, acceleration, angular velocity, etc. extracted by an inertial sensor, displayed using colors, arrows, image tilt, etc.

[0089] When the scrum is firmly formed, as shown in Figure 16A, the connection strength of the connecting parts between the players is high, the stress between the players forming the scrum varies little from side to side, and the scrum is often almost evenly tilted. On the other hand, when a scrum collapse is occurring or is in the pre-occurrence stage, as shown in Figure 16B, the connection strength of the connecting parts between the players is weak in some places, the stress between the players forming the scrum varies from side to side, or the scrum itself tends to tilt left and right or up and down. Therefore, by adopting the display method shown in Figure 16, a user viewing the display can immediately grasp the occurrence or pre-occurrence of a scrum collapse.

[0090] In this embodiment, the case where the roll angle and pitch angle are focused on as the posture of the athlete is exemplified, but in addition to or instead of these, an analysis may also be performed taking into account the athlete's yaw angle.

[0091] Furthermore, in the above-described embodiment, the behavior analysis system 1 and behavior analysis method are exemplified as systems that output output data that can be used to detect specific behavior. However, the behavior analysis system 1 may also be configured to estimate or identify the player or area that is a precursor to or the cause of a scrum collapse. One example of such an estimation or identification method is a method in which the detection results of the inertial sensors SA1-SA3, SB1-SB3 are converted into the aforementioned transition data, and the transition data is then input into an estimation model that has been trained in advance using a technique such as machine learning, thereby estimating the occurrence of a specific behavior. Note that a neural network model, for example, may be used as the estimation model. Another example of a training method for the estimation model is a method in which multiple data sets are prepared in advance, and the input data is the transition data obtained by converting the detection results of the inertial sensors SA1-SA3, SB1-SB3 when a scrum is formed, and the output data is data indicating whether a collapse has occurred, and supervised learning is performed using the data sets.

[0092] Furthermore, scrum behavior can also be considered as behavior that differs from a comparison target. Examples of comparison targets that can be used include the average of multiple samples of output data during scrums collected in advance, output data when a collapse occurs estimated using machine learning, output data that would be obtained if an ideal scrum were formed that is explicitly input, output data from a scrum where no collapse occurs, and past data from the same team. By analyzing the differences from these comparison targets using the above-described behavior analysis method or behavior analysis system 1, it is possible to obtain data that is useful for predicting or identifying the occurrence of specific behavior such as scrum collapsing.

[0093] In the above embodiments, the term "processor" refers to a processor in a broad sense, and includes general-purpose processors (e.g., CPU: Central Processing Unit, etc.) and dedicated processors (e.g., GPU: Graphics Processing Unit, ASIC: Application Specific Integrated Circuit, FPGA: Field Programmable Gate Array, programmable logic device, etc.).

[0094] Furthermore, the operations of the processors in the above embodiments may be performed not only by a single processor but also by multiple processors located at physically separate locations working together. Furthermore, the order of the operations of the processors is not limited to the order described in the above embodiments and may be changed as appropriate.

[0095] The present disclosure is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit and scope of the present disclosure, all of which are included in the technical concept of the present disclosure.

[0096] All references, including publications, patent applications, and patents, cited in this specification are herein incorporated by reference to the same extent as if each reference was individually and specifically indicated to be incorporated by reference and set forth in its entirety herein.

[0097] The use of nouns and similar referents in connection with the description of this disclosure (particularly in connection with the claims that follow) shall be construed to cover both the singular and the plural, unless otherwise indicated herein or clearly contradicted by context. The words "comprises," "has," "includes," and "comprises" shall be construed as open-ended terms (i.e., meaning "including, but not limited to"), unless otherwise noted. The recitation of numerical ranges herein is merely intended to serve as a shorthand method for referring individually to each value falling within the range, unless otherwise indicated herein, and each value is incorporated into the specification as if it were individually recited herein. All methods described herein can be performed in any suitable order unless otherwise indicated herein or clearly contradicted by context. Any example or exemplary language used herein (e.g., "such as"), unless otherwise claimed, is intended merely to better illustrate the disclosure and does not pose a limitation on the scope of the disclosure. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the present disclosure.

[0098] Preferred embodiments of the disclosure are described herein, including the best mode known to the inventors for carrying out the disclosure. Variations of these preferred embodiments will become apparent to those skilled in the art upon reading the foregoing description. The inventor expects that skilled persons will apply such variations as appropriate, and intends to practice the disclosure otherwise than as specifically described herein. Accordingly, this disclosure includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by applicable law. Moreover, this disclosure includes any combination of the above-described elements in all variations thereof unless otherwise indicated herein or otherwise clearly contradicted by context.

Claims

1. A behavior analysis method comprising: collecting detection results from inertial sensors attached to multiple athletes participating in a specific collision sport; calculating the acceleration and angular velocity in global coordinates of the multiple athletes to which the inertial sensors are attached from the detection results of the inertial sensors; and outputting the calculated acceleration and angular velocity in global coordinates of the multiple athletes at a specific timing in order to detect specific behavior between the multiple athletes.

2. The behavior analysis method according to claim 1, wherein the specific collision sport includes rugby, and the specific behavior includes behavior in a scrum.

3. The behavior analysis method according to claim 2, wherein the behavior in the scrum includes a precursor or occurrence of scrum collapsing.

4. The behavior analysis method according to claim 3, further comprising the step of estimating or identifying the player or area that was a precursor to or a cause of the scrum collapse.

5. The behavior analysis method according to claim 2, wherein the behavior in the scrum is a behavior regarding differences with a comparison target.

6. The behavior analysis method according to claim 5, wherein the difference is a difference in at least one of the strength of connections between players forming a scrum, the stress between players forming a scrum, and the inclination of the scrum.

7. The behavior analysis method according to claim 2, wherein the analysis results of the behavior in the scrum are displayed superimposed on an image showing the players forming the scrum.

8. The behavior analysis method according to claim 2, wherein the inertial sensor is attached to at least a front row player among the plurality of players playing the rugby game.

9. The behavior analysis method according to claim 1, further comprising applying an extended Kalman filter to the detection results of the inertial sensor before they are used to calculate the acceleration and the angular velocity in the global coordinate system.

10. The behavior analysis method according to claim 1, wherein the inertial sensors are attached to the torsos of the plurality of players.

11. A behavior analysis method according to claim 1, wherein the calculated acceleration and angular velocity in the global coordinates are output in the form of transition data of the acceleration in the horizontal direction converted using a dynamic model, and transition data of the angular velocity in the up and down rotation directions converted using a dynamic model.

12. A behavior analysis system comprising: inertial sensors attached to a plurality of athletes participating in a specific collision sport; and a computing device that calculates the acceleration and angular velocity in global coordinates of the plurality of athletes to which the inertial sensors are attached from the detection results of the inertial sensors, and outputs the calculated acceleration and angular velocity of the plurality of athletes at a predetermined timing in order to detect specific behavior between the plurality of athletes.

13. The behavior analysis system of claim 12, further comprising an imaging device that images the plurality of athletes participating in the specific collision sport, and the computing device identifies the predetermined timing based on the imaging results of the imaging device, or the imaging results of the imaging device and the calculated acceleration and angular velocity.

14. A program that causes a computer processor to execute the following process: collect detection results from inertial sensors attached to multiple athletes participating in a specific collision sport; calculate the acceleration and angular velocity in global coordinates of the multiple athletes to which the inertial sensors are attached from the detection results of the inertial sensors; and output the calculated acceleration and angular velocity of the multiple athletes at a specific timing in order to detect specific behavior between the multiple athletes.

Images