Analysis apparatus and analysis method

Abstract

To provide an analysis device which can present feedback information for controlling the quality of a ball.SOLUTION: An analysis device comprises: a position information acquisition unit which acquires position information of a body portion of a subject in a pitching motion; a parameter acquisition unit which acquires a parameter about a ball pitched by the subject and a parameter about a plurality of segments of the subject; an angle calculation unit which calculates angles of a plurality of joints of the subject in the pitching motion; a torque calculation unit which calculates torque acting on the joint of each segment in the pitching motion; a pattern calculation unit which calculates a cooperation pattern between the joints of the subject in the pitching motion on the basis of the angle of each joint in time-series and each torque in time-series; and an output control unit which outputs feedback information for the pitching motion of the subject on the basis of the similarity between the calculated cooperation pattern and one or more reference patterns associated with the quality of the ball.SELECTED DRAWING: Figure 16

Description

[Technical Field]

[0001] This disclosure relates to an analysis device and an analysis method. [Background technology]

[0002] In recent years, the existence of muscle synergy has been proposed as a mechanism to simplify the vast amount of information associated with movement. Muscle synergy is thought to be a mechanism that simplifies the redundant degrees of freedom handled by the central nervous system by integrating the movement information of muscles with similar functions and intervening between the central nervous system and each muscle. In response to this, attempts have been made to extract inter-joint coordination patterns, called muscle synergy or dynamic synergy, using mathematical methods such as principal component analysis, non-negative matrix factorization, and singular value decomposition. For example, Non-Patent Literature 1 describes the progress of inter-joint coordination pattern analysis using singular value decomposition. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Yuichiro Hayashi, Nobuyoshi Tsujiuchi, Shota Nakamura, Yuta Makino, Mayu Asano, Yasushi Matsuda, Yotaro Tsuchiya, Proposal of a quantitative analysis method for lower limb joint rotational movement including joint moment and angle, Transactions of the Japan Society of Mechanical Engineers, Vol. 82, No. 834 (2016), pp. 1-14. [Overview of the project] [Problems that the invention aims to solve]

[0004] Mathematically extracted synergies have been shown to be related to the activity of spinal interneurons using spinal animal models, suggesting that it may be possible to indirectly investigate spinal neural circuits. On the other hand, in pitching motions that involve high-speed acceleration of the extremities, it is unclear how the pitcher's central nervous system controls the ball's characteristics (e.g., spin rate, ball speed). Therefore, while ball characteristics are an effective parameter for pitchers to strike out batters, there has been a challenge in that specific methods for improving how to control the body to change ball characteristics (e.g., increase or decrease spin rate) remain unknown.

[0005] One objective of this disclosure is to provide an analysis device and method capable of providing feedback information for controlling the ball's trajectory by analyzing a subject's pitching motion. [Means for solving the problem]

[0006] An analysis device according to one embodiment includes: a position information acquisition unit that acquires position information of the subject's body parts during a throwing motion; a parameter acquisition unit that acquires parameters relating to the ball thrown by the subject and parameters relating to multiple segments of the subject; an angle calculation unit that calculates the angles of multiple joints of the subject during a throwing motion based on the position information of the body parts; a torque calculation unit that calculates the torque acting on the joints of each segment during a throwing motion based on the equations of motion relating to the ball thrown by the subject and the subject's fingers, the position information of the body parts, parameters relating to the ball thrown by the subject and parameters relating to multiple segments; a pattern calculation unit that calculates the coordination pattern between the subject's joints during a throwing motion based on the time-series angles of each joint and the time-series torques; and an output control unit that outputs feedback information for the subject's throwing motion based on the similarity between the calculated coordination pattern and one or more reference patterns associated with the ball's trajectory.

[0007] Preferably, the pattern calculation unit calculates left and right singular vectors by performing singular value decomposition on an observation matrix including the angles of each joint in time series and each torque in time series, and calculates, as a cooperative pattern, a time pattern indicating a matrix in which the left singular vectors are arranged in the column direction and a spatial pattern indicating a matrix in which the right singular vectors are arranged in the column direction.

[0008] Preferably, the analysis device further includes a similarity calculation unit that calculates the similarity between the calculated cooperative pattern and each of one or more reference patterns. Each of the one or more reference patterns is associated with a pitching motion type having a correlation with the quality of the ball. The output control unit outputs, as feedback information, the pitching motion type associated with the reference pattern corresponding to the highest similarity among the calculated similarities.

[0009] Preferably, the output control unit further outputs the calculated cooperative pattern. Preferably, the output control unit further outputs information indicating the joint with the greatest contribution based on the calculated cooperative pattern.

[0010] Preferably, the similarity calculation unit further calculates the similarity between the cooperative pattern of each of a plurality of subjects including the subject and the cooperative pattern of another subject, and classifies the plurality of subjects into a plurality of groups based on each calculated similarity. [[ID=A]] [[ID=B]]

[0011] Preferably, the output control unit outputs information indicating the group to which each of the plurality of subjects belongs.

[0012] Preferably, the parameters regarding the ball thrown by the subject include the mass of the ball and the acceleration of the ball. The parameters regarding the plurality of segments include the mass, acceleration, angular velocity, angular acceleration, and moment of inertia of each segment, the vector from the proximal joint to the distal joint of each segment, the vector from the proximal joint of each segment to the center of mass position of the segment, and the vector from the center of mass position of the finger segment of the subject to the action points of the ball and the finger.

[0013] Preferably, the multiple segments include an upper arm segment, a forearm segment, a palmar segment, and a finger segment. The joints of the upper arm segment, forearm segment, palmar segment, and finger segment are the shoulder joint, elbow joint, wrist joint, and MP (metacarpophalangeal) joint, respectively.

[0014] An analysis method according to another embodiment includes the steps of: acquiring positional information of a subject's body parts during a throwing motion; acquiring parameters relating to the ball thrown by the subject and parameters relating to multiple segments of the subject; calculating the angles of multiple joints of the subject during a throwing motion based on the positional information of the body parts; calculating the torque acting on the joints of each segment during a throwing motion based on the equations of motion relating to the ball thrown by the subject and the subject's fingers, the positional information of the body parts, parameters relating to the ball thrown by the subject and parameters relating to multiple segments; calculating the coordination pattern between the subject's joints during a throwing motion based on the time-series angles of each joint and the time-series torques; and outputting feedback information for the subject's throwing motion based on the similarity between the calculated coordination pattern and one or more reference patterns associated with the ball's trajectory. [Effects of the Invention]

[0015] According to this disclosure, by analyzing the subject's pitching motion, it becomes possible to provide feedback information for controlling the ball's trajectory. [Brief explanation of the drawing]

[0016] [Figure 1] This is a diagram illustrating the overall configuration of the analysis system. [Figure 2] This is a block diagram showing the hardware configuration of the analysis device. [Figure 3] This is a block diagram showing the hardware configuration of a sensor device. [Figure 4] This is a diagram illustrating a rigid body model according to this embodiment. [Figure 5] This is a flowchart illustrating an example of how the analysis system works. [Figure 6] This shows the joint angle and joint torque of the MP joint. [Figure 7] This shows the joint angle and joint torque of the wrist joint. [Figure 8] This shows the joint angle and joint torque of the elbow joint. [Figure 9] This shows the joint angle and joint torque of the shoulder joint. [Figure 10] This graph shows the contribution rate of each mode. [Figure 11] This figure shows the pitching motion type and inter-joint coordination patterns in the first group. [Figure 12] This figure shows the pitching motion type and inter-joint coordination patterns in the second group. [Figure 13] This figure shows the pitching motion type and inter-joint coordination patterns in the third group. [Figure 14] This figure shows the relationship between each subject's SPV and throwing motion type. [Figure 15] This diagram shows the relative relationships of each subject. [Figure 16] This is a block diagram showing an example of the functional configuration of an analysis device. [Figure 17] This figure shows the relationship between each subject's ball speed and pitching motion type. [Modes for carrying out the invention]

[0017] Embodiments of the present invention will be described below with reference to the drawings. In the following description, identical parts are denoted by the same reference numerals. Their names and functions are also the same. Therefore, detailed descriptions of them will not be repeated.

[0018] <Overall Structure> Figure 1 is a diagram illustrating the overall configuration of the analysis system 1000. Referring to Figure 1, the analysis system 1000 is a system for analyzing the inter-joint coordination in the pitching motion of subject 5. In this embodiment, subject 5 throws a ball 2 in which a sensor device 20 is built in. However, the ball 2 may be a general baseball without a built-in sensor device.

[0019] The analysis system 1000 includes an analysis device 10, a ball 2 with built-in sensor equipment 20, and a plurality of cameras 30. In this embodiment, the three-dimensional position (e.g., three-dimensional coordinates) of the subject 5 and the ball 2 is acquired using an optical motion capture system. Reflective markers are attached to the subject 5's body and the ball 2, and the three-dimensional coordinates of the image signals of the reflective markers are acquired using an optical three-dimensional motion analysis system consisting of a plurality of cameras 30 and the analysis device 10. The cameras 30 are, for example, infrared cameras. The reflective markers are attached, for example, to the surface of the ball 2, the fingers, back of the hands, forearms, upper arms, torso, etc. of the subject 5.

[0020] The above configuration is not limited to any system configuration that acquires the three-dimensional positions of subject 5 and ball 2. For example, a motion capture system may be configured using an analysis device 10 with a camera. Alternatively, a motion capture system using other methods such as inertial sensors, mechanical, magnetic, or video sensors may be employed.

[0021] The analysis device 10 consists of a laptop PC (Personal Computer). However, the analysis device 10 can be implemented as any device of any type. For example, the analysis device 10 may be a smartphone, tablet device, desktop PC, etc.

[0022] The analysis device 10 communicates with the sensor device 20 using a wireless communication method. For example, BLE (Bluetooth® low energy) is used as the wireless communication method. However, the analysis device 10 may also use other wireless communication methods such as Bluetooth® or Wi-Fi (local area network). In addition, the analysis device 10 is connected to the camera 30 using a wired communication method, but it may also be configured to be connected to the camera 30 using a wireless communication method.

[0023] The sensor device 20 detects acceleration, angular velocity, and geomagnetic field in the sensor coordinate system (i.e., the local coordinate system). Specifically, the sensor device 20 includes an acceleration sensor, an angular velocity sensor, and a geomagnetic field sensor. The acceleration sensor detects acceleration in the direction of three mutually orthogonal axes (x, y, and z axes). The angular velocity sensor detects angular velocity around the three axes (x, y, and z axes). The geomagnetic field sensor detects geomagnetic data indicating the magnetic field (magnetic flux density) in the direction of the three axes (x, y, and z axes).

[0024] The analysis device 10 receives time-series data acquired by the acceleration sensor, angular velocity sensor, and geomagnetic sensor. For example, the analysis device 10 uses a known ball analysis application program to analyze the acceleration data in the three axes and the geomagnetic data in the three axes to calculate the velocity and rotation speed of the ball 2. The "rotation speed" is the number of rotations per unit time of the ball 2 immediately after release.

[0025] The analysis device 10 analyzes the three-dimensional positional information of the subject 5 obtained by the camera 30 to identify the coordination pattern between the joints of the subject 5 during the pitching motion and outputs feedback information for the pitching motion.

[0026] <Hardware Configuration> (Analysis device 10) Figure 2 is a block diagram showing the hardware configuration of the analysis device 10. Referring to Figure 2, the analysis device 10 includes, as its main components, a CPU (Central Processing Unit) 102, memory 104, input device 106, display 108, input / output interface (I / F) 110, and communication interface (I / F) 112. Each of these components is connected to the others via a bus.

[0027] The CPU 102 controls the operation of each part of the analysis device 10 by reading and executing a program stored in memory 104. More specifically, the CPU 102 implements the processing of the analysis device 10, which will be described later, by executing the program.

[0028] Memory 104 is implemented using RAM (Random Access Memory), ROM (Read-Only Memory), a hard disk, etc. Memory 104 stores programs executed by CPU 102, or data used by CPU 102.

[0029] The input device 106 receives operational input for the analysis device 10. The input device 106 can be implemented as, for example, a keyboard, buttons, or a mouse. Alternatively, the input device 106 may be implemented as a touch panel.

[0030] The display 108 displays images, text, and other information on the screen based on signals from the CPU 102.

[0031] The input / output interface 110 communicates signals with the camera 30. Typically, the input / output interface 110 communicates with the camera 30 using a wired communication method such as USB (Universal Serial Bus).

[0032] The communication interface (I / F) 112 allows for the exchange of various data between the analysis device 10 and the sensor device 20. Wireless communication methods such as Bluetooth® and Wi-Fi (Local Area Network) can be used as the communication method. Wired communication methods using USB or similar technologies may also be used.

[0033] (Sensor device 20) Figure 3 is a block diagram showing the hardware configuration of the sensor device 20. Referring to Figure 3, the sensor device 20 includes, as its main components, a CPU 202 for executing various processes, a memory 204 for storing programs and data executed by the CPU 202, an acceleration sensor 205, an angular velocity sensor 206, a geomagnetic sensor 208, a communication interface (I / F) 210 for communicating with the analysis device 10, and a battery 212 for supplying power to the various components of the sensor device 20.

[0034] <Rigid body model> Figure 4 is a diagram illustrating the rigid body model according to this embodiment. Referring to Figure 4, in this embodiment, a two-segment model is employed in which the finger segments (index finger, middle finger, and ring finger) and the palm segment are considered as two rigid bodies in order to calculate the finger torque during the pitching motion. The finger portion is defined as one link from the metacarpophalangeal joint (MP joint) of the middle finger to the fingertip. In addition, in the actual pitching motion, the ball moves while rolling along the fingers just before release. Therefore, the movement of the point of application of the force acting between the fingers and the ball is taken into consideration.

[0035] The finger segment is defined by the line segment connecting the fingertip and the MP joint, and the palm segment is defined by the line segment connecting the MP joint and the wrist joint. Similarly, the forearm segment is defined by the line segment connecting the wrist joint and the elbow joint (not shown), and the upper arm segment (not shown) is defined by the line segment connecting the elbow joint and the shoulder joint (not shown). The mass, center of mass position, moment of inertia, and rotational semicircular ratio of the palm, fingers, forearm, and upper arm are estimated using known methods based on the subject's weight and height, etc. The center position of the ball is estimated, for example, from the coordinates of multiple markers attached to the ball.

[0036] <Example of operation> Figure 5 is a flowchart illustrating an example of the operation of the analysis system 1000. Typically, each step shown in Figure 5 is performed by the CPU 102 of the analysis device 10.

[0037] The analysis device 10 acquires time-series data of three-dimensional positional information (e.g., three-dimensional positional coordinates) of each body part (e.g., fingers, back of hand, forearm, upper arm, trunk, etc.) of the subject 5 performing the throwing motion (step S10). For example, the analysis device 10 acquires (calculates) three-dimensional positional information of reflective markers attached to each part of the subject 5 by combining the two-dimensional positional information of each reflective marker acquired by each camera 30 with information on the positional relationship of each camera 30.

[0038] The analysis device 10 acquires parameters for each segment of subject 5 (e.g., finger segment, palm segment, forearm segment, upper arm segment) (step S12). The parameters for each segment include the segment's mass, acceleration, angular velocity, and moment of inertia. Typically, the mass and moment of inertia of each segment are estimated from subject 5's weight and height, etc. The acceleration and angular velocity of each segment are calculated based on time-series three-dimensional positional information.

[0039] The analysis device 10 uses the acquired time-series 3D position information to calculate multiple joint angles of the subject 5 in a time series (step S14). The multiple joint angles include MP joint angles, wrist joint angles, elbow joint angles, and shoulder joint angles.

[0040] The analysis device 10 calculates the joint torque acting on the proximal joint of each segment during the pitching motion based on a predetermined equation of motion, three-dimensional position information, and parameters for each segment (step S16). Details of the joint torque calculation method will be described later.

[0041] The analysis device 10 calculates the inter-joint coordination pattern of subject 5 during the throwing motion based on the time-series joint angles and the time-series torques of each joint (step S18). Details of the calculation method for the inter-joint coordination pattern will be described later.

[0042] The analysis device 10 calculates the similarity between the calculated inter-joint coordination pattern and each of the one or more reference patterns associated with the ball rotation speed (step S20). Details of the similarity calculation method will be described later.

[0043] The analysis device 10 identifies the pitching motion type of subject 5 based on the calculated similarity (step S22). Specifically, the analysis device 10 identifies the pitching motion type associated with the reference pattern corresponding to the highest similarity among the similarities as the pitching motion type of subject 5.

[0044] The analysis device 10 displays feedback information for the pitching motion of the subject being tested (step S24). Specifically, the analysis device 10 displays the identified pitching motion type as the feedback information on the display 108.

[0045] <Method for calculating joint torque> The method for calculating joint torque in step S16 of Figure 5 will be explained below.

[0046] The analysis device 10 calculates the MP joint torque of the fingers, the wrist joint torque, the elbow joint torque, and the shoulder joint torque by inverse power calculation.

[0047] The following equations (1) to (3) hold true from the equations of motion relating ball 2 and the fingers of subject 5.

[0048]

number

[0049] These equations yield equations (4), (5), (6), and (7), from which the MP joint torque T3, wrist joint torque T2, elbow joint torque T1, and shoulder joint torque T0 can be calculated.

[0050]

number

[0051] In equations (1) to (7) above, the subscripts are numbered so that the segment and its proximal joint have the same number. For segments, i=0,1,2,3 correspond to the humeral segment, forearm segment, palmar segment, and finger segment, respectively. For joints, i=0,1,2,3 correspond to the shoulder joint, elbow joint, wrist joint, and MP joint, respectively. That is, the proximal joints of the humeral segment, forearm segment, palmar segment, and finger segment are the shoulder joint, elbow joint, wrist joint, and MP joint, respectively.

[0052] I i The inertia tensor of a segment is shown. For example, I3 represents the inertia tensor of a finger segment.

[0053] a gi This indicates the acceleration vector at the center of mass of the segment. For example, a g3 This shows the acceleration vector at the center of mass of the finger segment. ball ω represents the acceleration vector at the center of the ball. iand ω i · indicates the angular velocity and angular acceleration of each segment respectively. For example, ω3 and ω3· indicate the angular velocity and angular acceleration of the finger segment respectively. Note that "ω·" means a character with a dot on top of the ω character. Note that the angular velocity and angular acceleration of the segment are calculated by the first derivative and the second derivative of the joint angle of the proximal joint of the segment respectively. For example, the angular velocity and angular acceleration of the finger segment are calculated by the first derivative and the second derivative of the joint angle of the MP joint.

[0054] L gi indicates the vector from the proximal joint to the center-of-mass position of the segment. For example, L g3 indicates the vector from the MP joint to the center-of-mass position of the finger segment. L i indicates the vector from the proximal joint to the distal joint. For example, L3 indicates the vector from the MP joint to the wrist joint. l indicates the vector from the center-of-mass position of the finger segment to the action point of the ball and the finger.

[0055] F ball indicates the force acting on the ball. F3 indicates the force acting on the finger. m ball indicates the mass of the ball. m i indicates the mass of the segment. For example, m3 indicates the mass of the finger segment. T i indicates the torque acting on the joint. For example, T3 indicates the MP joint torque acting on the MP joint.

[0056] Note that I i is obtained from the moment of inertia of the segment. a gi a ball ω i and ω i · are calculated from the time-series 3D position information. L gi L i l are calculated from the 3D position information and the known center-of-mass ratio. F ball is calculated from the acceleration of the ball in the stationary coordinate system.

[0057] <Method for calculating inter-joint coordination patterns> The calculation method for the inter-joint coordination pattern in step S18 of Figure 5, and the calculation method for the similarity corresponding to step S20, will be explained.

[0058] To quantitatively evaluate the inter-joint coordination of the upper limbs during human throwing motion, an evaluation method that performs singular value decomposition including joint angles and joint torques is applied to throwing data. Joint torque is a representative physical quantity in terms of dynamics that is considered to contribute significantly to the rotational movement of the upper limb joints.

[0059] Using a time-series data pattern in which MP joint angle, wrist joint angle, elbow joint angle, shoulder joint angle, MP joint torque, wrist joint torque, elbow joint torque, and shoulder joint torque are arranged in order, the joint angle θ- j (t i ) and joint torque M- j (t i The matrix is ​​represented by the matrix shown in equation (8), where i=1,…,m and j=1,…,p. m and p are the number of data points. In this embodiment, p is 4. Note that "θ-" means the letter θ with a line drawn through it. The same applies to "M-" and "R-".

[0060]

number

[0061] The matrix shown in equation (8) is defined as the observation matrix. In the observation matrix, the rows represent each joint angle and each joint torque, and the columns represent the time series. Next, singular value decomposition is performed on the observation matrix. When the joint angles and joint torques are expanded using mutually orthogonal basis vectors with singular value decomposition, the observation matrix is ​​expressed by equation (9).

[0062]

number

[0063] Here, the singular value λ jThis corresponds to the contribution of each orthonormal basis vector, which is designated as mode Md1 and mode Md2 in descending order of singularity. Also, v j (t) shows the activity pattern of each base, z j T (θ-,M-) represents the coordination pattern of physical quantities in each basis. In this embodiment, v j (t) is also called the time pattern, z j T (θ-,M-) is also referred to as the spatial pattern. In this embodiment, the inter-joint coordination pattern includes both a temporal pattern and a spatial pattern.

[0064] The j-th singular value λ for all observed modes j The contribution rate γ j This is expressed by equation (10).

[0065]

number

[0066] For example, the smallest j with a cumulative contribution rate exceeding 90% is defined as the number of modes. The similarity between the inter-joint coordination pattern of subject 5 and other inter-joint coordination patterns is calculated using cosine similarity. For example, the spatial pattern vector is z a , the spatial pattern vector z b Therefore, the similarity s between these spatial patterns ab This is calculated by formula (11).

[0067]

number

[0068] Furthermore, the similarity between time patterns is expressed in equation (11) by the spatial pattern vector z a time pattern vector v a Replace with the spatial pattern vector z b time pattern vector v b This is equivalent to replacing it with [the other term].

[0069] <Examples> (Joint angle, joint torque) Here, we calculate the joint angles and joint torques for each of the seven subjects, perform singular value decomposition on the observation matrix of these time-series data, and explain the results of calculating the temporal and spatial patterns, respectively.

[0070] Figures 6-9 show the average joint angle and torque data for seven subjects. Specifically, Figure 6 shows the joint angle and torque of the MP joint. Figure 7 shows the joint angle and torque of the wrist joint. Figure 8 shows the joint angle and torque of the elbow joint. Figure 9 shows the joint angle and torque of the shoulder joint. In each figure, the time change of joint angle and torque from the moment of stepping foot contact (approximately 110 ms before the ball release) to the moment of release is shown. The moment of ball release is defined as the moment when the distance between the center of the ball and the fingertip marker exceeds the sum of the radius of the ball, the radius of the fingertip marker, and the thickness of the fingers.

[0071] The movements of the MP joint and wrist joint, as well as the changes in MP joint and wrist joint torque, as shown in Figures 6 and 7, were similar to those of previous studies. Furthermore, the maximum values ​​of MP joint flexion torque and wrist joint palmar flexion torque were 14.1 ± 4.1 N·m and 27.5 ± 9.5 N·m, respectively, which are similar to the results of previous studies, where the MP joint flexion torque was approximately 15 N·m and the wrist joint palmar flexion torque was approximately 20 N·m.

[0072] Referring to Figure 8, the elbow joint underwent flexion to extension, supination to pronation, and pronation torque was generated from supination torque to release. These results were similar to those found in previous studies.

[0073] The shoulder joint movement and changes in shoulder joint torque shown in Figure 9 were similar to those reported in previous studies. Furthermore, the maximum values ​​of shoulder joint abduction torque and horizontal adduction torque were 58.9 ± 21.7 N·m and 57.4 ± 55.6 N·m, respectively, which were similar to those reported in previous studies, where they were approximately 50 N·m each.

[0074] Furthermore, the elbow joint torque transitioned from flexion torque to extension torque during release. This differed from the results of previous studies, where flexion torque was generated from extension torque to release. This is likely because previous studies used a model that treated the palm, fingers, and ball as a single segment, unlike the model used in this embodiment.

[0075] The overall similarities described above are considered to support the validity of the joint angles and joint torques obtained in this study.

[0076] (Number of modes) For each of the seven subjects, the contribution rate of each orthonormal basis vector was calculated from the singular values ​​using equation (10), and its mean and standard deviation were determined.

[0077] Figure 10 is a graph showing the contribution rate of each mode. Referring to Figure 10, the contribution rate of each mode indicates the proportion of each mode to the overall motion. Therefore, the spatial pattern in each mode can represent the overall motion to that extent. The average number of modes was 1.6 ± 0.5. Therefore, focusing on mode Md1, which represents a particularly major feature, the following analysis was performed.

[0078] (Relationship between throwing motion type and interjoint coordination pattern) Singular value decomposition was performed on the observation matrix of time-series data for each subject's joint angles and joint torques to calculate temporal and spatial patterns, respectively. Focusing on each subject's mode Md1, the similarity between subjects was evaluated using cosine similarity. In this example, subjects with an absolute value of 0.5 or higher in cosine similarity were considered to belong to the same group. If a subject was classified into multiple groups, the group with the higher average similarity within that group was designated as the subject's group. As a result, the seven subjects could be classified into three groups.

[0079] Figures 11 to 13 show the relationship between pitching motion types and inter-joint coordination patterns (spatial and temporal patterns) in three groups. Specifically, Figure 11 shows the pitching motion types and inter-joint coordination patterns in the first group. Figure 12 shows the pitching motion types and inter-joint coordination patterns in the second group. Figure 13 shows the pitching motion types and inter-joint coordination patterns in the third group.

[0080] Referring to Figure 11, in the first group to which subjects A and B were classified, focusing on the spatial pattern reveals that the contribution of shoulder joint internal rotation torque is significant. Therefore, the throwing motion type of the first group will also be referred to as the "shoulder internal rotation torque type."

[0081] Referring to Figure 12, in the second group to which subjects C and D were classified, focusing on the spatial pattern reveals that elbow extension torque makes a significant contribution. Therefore, the throwing motion type of the second group will also be referred to as the "elbow extension torque type."

[0082] Referring to Figure 13, in the second group to which subjects E, F, and G were classified, focusing on the spatial pattern reveals that the contribution of shoulder joint horizontal adduction torque is significant. Therefore, the throwing motion type of the third group is also referred to as the "shoulder horizontal adduction torque type."

[0083] Furthermore, Figure 14 shows the results focusing on the relationship between the pitching motion types of the 1st to 3rd groups and each subject's SPV (spin per velocity). SPV is the value obtained by dividing the ball rotation speed by the ball velocity. The ball rotation speed and ball velocity are calculated from acceleration data and geomagnetic data acquired by the sensor device 20 of ball 2.

[0084] Figure 14 shows the relationship between each subject's SPV and pitching motion type. Referring to Figure 14, we can see that subject A, a shoulder internal rotation torque type belonging to Group 1, is the pitcher with the highest SPV among all 7 subjects, and subject B, of the same type, is the pitcher with the second highest SPV. Furthermore, subject C, a elbow extension torque type belonging to Group 2, is the pitcher with the third highest SPV, and subject D, of the same type, is the pitcher with the fourth highest SPV. In addition, subjects E, F, and G, of the shoulder horizontal adduction torque type belonging to Group 3, are the pitchers with the fifth, sixth, and seventh highest SPVs, respectively.

[0085] This indicates a correlation between SPV and pitching motion type. Specifically, it is understood that pitchers with high SPV belong to the shoulder internal rotation torque type, while pitchers with low SPV belong to the shoulder horizontal adduction type. This suggests that pitchers with different spin rates may have different posterior neural activity.

[0086] The results above suggest that acquiring a movement pattern that enhances the contribution of shoulder joint internal rotation torque is necessary to increase rotational speed. On the other hand, acquiring a movement pattern that enhances the contribution of shoulder joint horizontal adduction torque is considered necessary to decrease rotational speed.

[0087] Furthermore, when the similarity between each subject was visualized using multidimensional scaling, the results shown in Figure 15 were obtained. Multidimensional scaling is a method that visualizes the relative relationships between objects by replacing the strength of the relationship or similarity between them with the distance between points on a map.

[0088] Figure 15 shows the relative relationships of each subject. Referring to Figure 15, it can be seen that some subjects with shoulder internal rotation torque type and some subjects with shoulder horizontal adduction type are highly similar.

[0089] <Functional Configuration> Figure 16 is a block diagram showing an example of the functional configuration of the analysis device 10. Referring to Figure 16, the analysis device 10 includes a position information acquisition unit 250, a parameter acquisition unit 252, a joint angle calculation unit 254, a joint torque calculation unit 256, a pattern calculation unit 258, a similarity calculation unit 260, and an output control unit 262. Typically, these are implemented by the CPU 102 of the analysis device 10. Note that some or all of these functional configurations may be implemented in hardware.

[0090] The position information acquisition unit 250 acquires position information of the subject's body parts (e.g., fingers, back of hand, forearm, upper arm, torso) during the throwing motion. Specifically, the position information acquisition unit 250 acquires (calculates) the three-dimensional position information of each reflective marker based on the two-dimensional position information of each reflective marker attached to the subject's body parts acquired by each camera 30 and the positional relationship information of each camera 30. The position information acquisition unit 250 may also be configured to acquire the three-dimensional position information of each body part of the subject's body using other methods (e.g., inertial sensor type, mechanical type, magnetic type, video type, etc.).

[0091] The parameter acquisition unit 252 acquires parameters related to the ball 2 thrown by the subject 5 and parameters related to multiple segments of the subject 5 (e.g., upper arm segment, forearm segment, palm segment, and finger segment). The parameters related to the ball 2 include the mass of the ball 2 (e.g., m ball ) and the acceleration of ball 2 (for example, a ball ) and include.

[0092] The parameters for multiple segments are the mass of each segment (e.g., m i ), acceleration (for example, a gi ), angular velocity (for example, ω i), angular acceleration (for example, ω i •) and moment of inertia (e.g., I i ) and the vector from the proximal joint to the distal joint of each segment (for example, L i ) and the vector from the proximal joint of each segment to the center of mass of that segment (for example, L gi The acceleration, angular velocity, and angular acceleration of each segment, as well as the vector from the proximal joint to the distal joint of the segment and the vector from the proximal joint of each segment to the center of mass of that segment, are calculated based on time-series 3D positional information. The mass and moment of inertia of each segment are estimated from the weight of subject 5, etc.

[0093] The joint angle calculation unit 254 calculates the angles of multiple joints (e.g., MP joints, wrist joints, elbow joints, and shoulder joints) of the subject 5 during the throwing motion based on the positional information of the body parts.

[0094] The joint torque calculation unit 256 calculates the torque acting on the proximal joints of each segment (e.g., MP joint, wrist joint, elbow joint, and shoulder joint) during the throwing motion, based on the equations of motion for the ball 2 thrown by the subject 5 and the subject 5's fingers (e.g., equations (1) to (3)), the position information of the body parts, and the parameters for each segment. Specifically, it calculates the MP joint torque T3, wrist joint torque T2, elbow joint torque T1, and shoulder joint torque T0 shown by equations (4) to (7).

[0095] The pattern calculation unit 258 calculates the inter-joint coordination pattern of subject 5 during the throwing motion based on the time-series joint angles and time-series joint torques. Specifically, the pattern calculation unit 258 calculates left and right singular vectors by performing singular value decomposition on an observation matrix (e.g., equation (8)) that includes the time-series joint angles and time-series joint torques. The pattern calculation unit 258 then calculates a matrix (e.g., v) obtained by arranging the left singular vectors in the column direction. j (t)) represents a time pattern and a matrix obtained by arranging the right singular vectors in the column direction (for example, z jT The spatial pattern representing (θ-,M-)) is calculated as the inter-joint coordination pattern.

[0096] The similarity calculation unit 260 calculates the similarity between the calculated inter-joint coordination pattern and each of the one or more reference patterns. Specifically, the similarity calculation unit 260 calculates the similarity using cosine similarity as shown in equation (11). Each of the one or more reference patterns is associated with a pitching motion type that correlates with ball rotation speed. For example, the first reference pattern is an inter-joint coordination pattern (time pattern and spatial pattern) of the shoulder internal rotation torque type that tends to have a "high" ball rotation speed. The second reference pattern is an inter-joint coordination pattern of the elbow extension torque type that tends to have a "normal" ball rotation speed. The third reference pattern is an inter-joint coordination pattern of the shoulder horizontal adduction torque type that tends to have a "low" ball rotation speed.

[0097] In other cases, the similarity calculation unit 260 calculates the similarity between the inter-joint coordination pattern of each of the multiple subjects and the inter-joint coordination pattern of another subject. Based on the calculated similarity, the similarity calculation unit 260 may classify the multiple subjects into multiple groups. For example, the similarity calculation unit 260 classifies subjects with an absolute similarity of 0.5 or higher into the same group. If a subject is classified into multiple groups, the similarity calculation unit 260 classifies the subject into the group with the highest average similarity among the multiple groups.

[0098] The output control unit 262 outputs feedback information for the subject 5's pitching motion based on the similarity between the calculated inter-joint coordination pattern and one or more reference patterns associated with ball rotation speed. For example, the first reference pattern is an inter-joint coordination pattern associated with a "high" ball rotation speed, the second reference pattern is an inter-joint coordination pattern associated with a "normal" ball rotation speed, and the third reference pattern is an inter-joint coordination pattern associated with a "low" ball rotation speed.

[0099] Specifically, the output control unit 262 outputs the pitching motion type associated with the reference pattern corresponding to the highest similarity among the calculated similarities as feedback information. This allows subject 5 to receive feedback on whether their pitching motion type belongs to a pitching motion type with a high ball spin rate or a pitching motion type with a high ball spin rate. Therefore, subject 5 (or their coach) can determine what kind of pitching form they should aim for in order to increase (or decrease) the ball spin rate.

[0100] In other situations, the output control unit 262 may further output the calculated inter-joint coordination pattern. This allows subject 5 to determine which joint torques contribute most significantly to their throwing motion.

[0101] Furthermore, in other situations, the output control unit 262 outputs information indicating the joint that makes the greatest contribution based on the calculated inter-joint coordination pattern. Specifically, the output control unit 262 outputs information indicating the joint corresponding to the spatial pattern with the largest absolute value. For example, in the example in Figure 11, the largest spatial pattern is "shoulder joint internal rotation torque," so the output control unit 262 outputs information indicating "shoulder joint."

[0102] In other contexts, the output control unit 262 outputs information indicating the group to which each of the multiple subjects belongs. This allows subject 5 to determine whether other subjects have a similar pitching motion (i.e., belong to the same group) or a different pitching motion (i.e., belong to a different group). For example, if subject 5 has a lower ball rotation rate compared to other subjects, it can determine whether subject 5 should improve their pitching motion to increase the ball rotation rate, or whether they should improve something else.

[0103] <Advantages> According to this embodiment, subject 5 receives feedback on which of several pitching motion types associated with ball spin rate their pitching motion belongs to. Therefore, subject 5 (or their coach) can determine what kind of pitching motion they should aim for in order to control the ball spin rate (for example, by raising or lowering it).

[0104] <Other embodiments> (1) In the above-described embodiment, a configuration focusing on ball spin rate as a ball quality was explained, but a configuration focusing on ball speed as a ball quality is also possible.

[0105] Figure 17 shows the relationship between each subject's ball speed and pitching motion type. Referring to Figure 17, it can be seen that subject D, who belongs to Group 2 and has an elbow extension torque type, is the pitcher with the highest ball speed among all 7 subjects, and subject C, who also belongs to the same type, is the pitcher with the second highest ball speed. It can also be seen that the ball speeds of subjects A and B, who belong to Group 1 and have a shoulder internal rotation torque type, and subjects E, F, and G, who belong to Group 3 and have a shoulder horizontal adduction torque type, are roughly the same.

[0106] For example, the elbow extension torque type interjoint coordination pattern (e.g., the second reference pattern) is a pattern that tends to result in a "high" ball speed. The shoulder internal rotation torque type and shoulder horizontal adduction torque type interjoint coordination patterns (e.g., the first and third reference patterns) are patterns that tend to result in a "normal" ball speed.

[0107] Therefore, Subject 5 receives feedback on which of the multiple pitching motion types associated with ball speed their pitching motion belongs to. Thus, Subject 5 (or their coach) can determine what pitching motion they should aim for in order to control the ball speed (e.g., raise or lower it).

[0108] (2) It is also possible to provide a program that makes the computer function and performs the control described in the above-described embodiment. Such a program can be recorded on a non-temporary computer-readable recording medium such as a flexible disk, CD-ROM (Compact Disk Read Only Memory), ROM, RAM, and memory card attached to the computer and provided as a program product. Alternatively, the program can be recorded on a recording medium such as a hard disk built into the computer and provided. Furthermore, the program can be provided by downloading it over a network.

[0109] (3) The configurations described above as embodiments are examples of the configuration of the present invention, and can be combined with other known technologies, and can be modified, such as by omitting parts, without departing from the spirit of the present invention.

[0110] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of the present invention is indicated by the claims rather than by the foregoing description, and all modifications within the meaning and scope equivalent to the claims are intended to be included. [Explanation of Symbols]

[0111] 2 balls, 5 subjects, 10 analysis device, 20 sensor equipment, 30 cameras, 104, 204 memory, 106 input device, 108 display, 110 input / output interface, 205 acceleration sensor, 206 angular velocity sensor, 208 geomagnetic sensor, 212 storage battery, 250 position information acquisition unit, 252 parameter acquisition unit, 254 joint angle calculation unit, 256 joint torque calculation unit, 258 pattern calculation unit, 260 similarity calculation unit, 262 output control unit, 1000 analysis system.

Claims

1. A position information acquisition unit that acquires position information of the subject's body parts during the throwing motion, A parameter acquisition unit that acquires parameters related to the ball thrown by the subject and parameters related to multiple segments of the subject, An angle calculation unit that calculates the angles of multiple joints of the subject during a throwing motion based on the positional information of the body parts, A torque calculation unit calculates the torque acting on each joint of a segment during a throwing motion, based on the equation of motion for the ball thrown by the subject and the subject's fingers, the position information of the body parts, parameters related to the ball thrown by the subject, and parameters related to the plurality of segments. A pattern calculation unit calculates the coordination pattern between the joints of the subject during a throwing motion based on the time-series angles of each of the aforementioned joints and the time-series torques of each of the aforementioned joints, The system includes an output control unit that outputs feedback information for the subject's pitching motion based on the similarity between the calculated coordination pattern and one or more reference patterns associated with the ball's trajectory, The pattern calculation unit, A left singular vector and a right singular vector are calculated by performing singular value decomposition on an observation matrix that includes the time-series angles of each joint and the time-series torques. An analysis device that calculates a time pattern representing a matrix obtained by arranging the left singular vectors in the column direction, and a spatial pattern representing a matrix obtained by arranging the right singular vectors in the column direction, as the cooperative pattern.

2. The system further includes a similarity calculation unit that calculates the similarity between the calculated coordination pattern and each of the one or more reference patterns, Each of the one or more reference patterns is associated with a pitching motion type that correlates with the ball's trajectory. The analysis apparatus according to claim 1, wherein the output control unit outputs the pitching motion type associated with the reference pattern corresponding to the highest similarity among the calculated similarities as feedback information.

3. The analysis apparatus according to claim 2, wherein the output control unit further outputs the calculated coordination pattern.

4. The analysis apparatus according to claim 2, wherein the output control unit further outputs information indicating the joint that makes the greatest contribution based on the calculated coordination pattern.

5. The similarity calculation unit, For each of the subjects, including the aforementioned subject and other subjects, the degree of similarity between the cooperation pattern of that subject and the cooperation pattern of another subject is further calculated. The analysis apparatus according to claim 2, which classifies the plurality of subjects into a plurality of groups based on the calculated similarity scores.

6. The analysis apparatus according to claim 5, wherein the output control unit outputs information indicating the group to which each of the plurality of subjects belongs.

7. The parameters relating to the ball thrown by the subject include the mass of the ball and the acceleration of the ball. The analysis apparatus according to claim 1 or 2, wherein the parameters relating to the plurality of segments include the mass, acceleration, angular velocity, angular acceleration and moment of inertia of each segment, a vector from the proximal joint to the distal joint of each segment, a vector from the proximal joint to the center of mass of each segment, and a vector from the center of mass of the subject's finger segment to the point of application of the ball and finger.

8. The plurality of segments include an upper arm segment, a forearm segment, a palmar segment, and a finger segment. The analysis apparatus according to claim 1 or 2, wherein the joints of the upper arm segment, the forearm segment, the palmar segment, and the finger segment are the shoulder joint, the elbow joint, the wrist joint, and the MP (Metacarpophalangeal) joint, respectively.

9. A step of acquiring positional information of the subject's body parts during the throwing motion, The steps include obtaining parameters relating to the ball thrown by the subject and parameters relating to multiple segments of the subject, A step of calculating the angles of multiple joints of the subject during the throwing motion based on the positional information of the body parts, A step of calculating the torque acting on each joint of the segment during a throwing motion, based on the equation of motion for the ball thrown by the subject and the subject's fingers, the position information of the body parts, parameters related to the ball thrown by the subject, and parameters related to the plurality of segments. A step of calculating the coordination pattern between the joints of the subject during a throwing motion based on the time-series angles of each of the aforementioned joints and the time-series torques of each of the aforementioned joints, The process includes the step of outputting feedback information for the subject's pitching motion based on the similarity between the calculated coordination pattern and one or more reference patterns associated with the ball's trajectory, The step of calculating the aforementioned coordination pattern is: The left singular vector and the right singular vector are calculated by performing singular value decomposition on an observation matrix that includes the time-series angles of each joint and the time-series torques. An analysis method comprising calculating a time pattern representing a matrix obtained by arranging the left singular vectors in the column direction, and a spatial pattern representing a matrix obtained by arranging the right singular vectors in the column direction, as the cooperative pattern.

Images