Stability evaluation method and system for tennis shot action considering various ball paths

Abstract

The present application relates to a kind of tennis shot action stability evaluation method and system considering various ball routes, method includes: step 1, the relevant data of the shot action of trainer is collected, and the cumulative hand contribution speed-time sequence is calculated;Step 2, based on dynamic time warping theory, the shot action stability evaluation score is calculated;Step 3, construct ordinary Kriging model;Step 4, select semi-variogram function fitting;Step 5, the correlation model of shot action stability evaluation score and training data is evaluated.The present application method captures the posture data generated in the shot action of trainer by inertial motion capture system, and it is converted into cumulative hand contribution speed, to quantify the speed generated in hand by each joint of human body, and based on dynamic time warping theory, the shot action stability of trainer is evaluated.Compared with the traditional experience-based evaluation method, the present application method is more objective and accurate.

Description

Technical Field

[0001] This invention belongs to the field of sports training technology and relates to a method for evaluating tennis movements. Specifically, it relates to an evaluation method for assessing the stability of a trainee's return movements with various ball trajectories. Background Technology

[0002] Tennis enjoys widespread influence and popularity worldwide. In tennis, the shot, as the most core technical element, is a crucial indicator of an athlete's skill level, and the consistency of the shot is a key factor. The varied trajectories in real matches demand quick reactions from athletes. From straight shots to cross-court shots, from baseline deep shots to drop shots at the net, different shot trajectories require different shot movements and strategies. Therefore, mastering various shot techniques and consistently performing well in multiple shots is crucial for tennis skill and an important part of technical training. However, in current training, coaches and athletes often rely on experience to judge the consistency of an athlete's shot, which is often not objective enough. Furthermore, traditional training methods mostly provide feedback on shots at a single target point, making it difficult to assess the consistency of an athlete's shot across different shot trajectories. Therefore, a more scientific, comprehensive, and personalized evaluation method is needed to help trainees accurately assess the consistency of their shot across diverse shot trajectories, understand their current training progress and shortcomings, and conduct more targeted training. Summary of the Invention

[0003] The technical problem this invention aims to solve is the difficulty in comprehensively evaluating the stability of tennis hitting actions. It provides a method for evaluating the stability of tennis hitting actions that considers diverse ball trajectories. The method quantifies the velocities generated by various joints in the hand by quantifying the cumulative hand contribution velocity. Based on dynamic time curvature theory, it assesses the stability of the trainee's hitting action. Finally, it uses a Kriging model for spatial modeling to predict the trainee's hitting performance under different ball trajectories. This helps trainees efficiently and accurately assess their control over different ball trajectories.

[0004] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0005] A method for evaluating the stability of tennis shot considering diverse ball trajectories includes the following steps:

[0006] Step 1: Collect relevant data on the trainee's hitting action; the data includes the landing speed, the distance between the hitting point and the trainee's body, the rotational angular velocity of various parts of the body, and the swing speed; and calculate the cumulative hand contribution speed-time series.

[0007] Step 2: Calculate the stability evaluation score of the hitting action based on the dynamic time curvature theory;

[0008] Step 3, construct a standard kriging model;

[0009] Step 4: Select a semi-variogram function for fitting;

[0010] Step 5: Evaluate the correlation model between the stability evaluation score of the hitting action and the training data.

[0011] Furthermore, the specific steps of step 1 are as follows:

[0012] Step 101: Collect landing speed and the distance between the hitting point and the trainee After performing N shots, record the landing speed of the ball for each shot. Distance between the hitting point and the trainee The measured data are divided according to the landing velocity. Arrange the data from smallest to largest and divide them into C speed intervals. Then, rank the data within each speed interval according to the distance between the hitting point and the trainee. Arrange the data in ascending order and divide them into D speed intervals, ensuring that each interval contains M data points; in summary, There are a total of Group data. , () indicates the range of the trainee's landing speed and the range of the front-to-back distance between the hitting point and the trainee's body. Training data during grouping;

[0013] Step 102: Collect the rotational angular velocity of each part; specifically, the lower limb driving the hip ( ) with angular velocity Rotation, where angular velocity Depend on , , Components, chest ( Drive the shoulder at angular velocity Rotation, shoulder ( Drive the upper arm at an angular velocity Rotation, elbow ( Drive the forearm at an angular velocity Rotation, hand ( The swing speed is .

[0014] Step 103: Calculate the contribution rate of each part to the hand. The calculation formula is:

[0015]

[0016] in, Represents joint The velocity contributed by the hand at time t This represents the spatial position of joint k at time t;

[0017] At that time, the hip area, At that time, it was the chest. At that time, the shoulder area At the elbow, At that time, it was the hand;

[0018] Step 104: Calculate the cumulative hand contribution speed The calculation method involves summing the contribution speeds of each part of the hand; the calculation formula is:

[0019]

[0020] in, This represents the cumulative hand contribution velocity at joint k at time t;

[0021] This represents the spatial position of joint k at time t;

[0022] Step 105: Record the cumulative hand contribution speed-time series; correlate the cumulative hand contribution speed with the acquisition time, that is, the horizontal axis is the time axis and the vertical axis is the cumulative hand contribution speed value, and record it as the cumulative hand contribution speed-time series (hereinafter referred to as speed-time series).

[0023] Furthermore, the specific steps of step 2 are as follows:

[0024] Step 201: Process the velocity-time series data obtained in Step 1. Each velocity-time series includes data from four locations: hip, chest, shoulder, and elbow. The velocity-time series data for the second stroke is represented as follows:

[0025]

[0026] in, Represents the joint area. At that time, the hip area, At that time, it was the chest. At times, the shoulder At the elbow, Indicates the first During the second strike, the joint area velocity-time series, Indicates joint location exist The cumulative hand contribution speed at any given moment.

[0027] Step 202: Arbitrarily select two velocity time series and combine them, and use the Dynamic Time Warping (DTW) algorithm to calculate each combined part. The optimal path distance is recorded as ;in, Indicates the first During the second strike, the joint area velocity-time series, Indicates the first During the second strike, the joint area The velocity-time series. In summary, there are a total of... A combination.

[0028] Step 203: Calculate the DTW values ​​of the four joints in M ​​strokes using the DTW values ​​of a single combination of individual parts. The average optimal path distance under each speed-time combination is calculated, and the beginner's path distance is obtained. Weighted stability score of the second shot The specific calculation formula is as follows:

[0029]

[0030] in, The weighted stability score; Assigning a weight to joint k in the analysis of hitting stability, and setting... =0.42、 =0.20, =0.27, =0.11; The score is used to evaluate the stability of the ball striking at joint k. For a certain interval, the first Optimal path distance under the combination of group speed and time series.

[0031] Step 204: Calculate the stability score of the hitting action. ; The training data is The stability evaluation score at that time. The calculation formula is:

[0032] .

[0033] Furthermore, the following constraints must be met when using the DTW algorithm to find a path:

[0034] (1) Boundary conditions: At the beginning, the starting point must be The destination must be Where a and b are velocity-time series. , The length of the time series;

[0035] (2) Continuity: After leaving the starting point, it is necessary to find the next data combination in the time series corresponding to the current data combination. Jumping on the time axis is not allowed.

[0036] (3) Monotonicity: When searching for the next data combination, it is not possible to go back on the timeline;

[0037] The optimal path distance between two velocity-time series is calculated as follows:

[0038]

[0039] in, , representing a velocity-time series of length a. and a velocity-time series of length b The optimal path distance between them; Representing velocity-time series The a-th element With velocity-time series The b-th element The distance between them.

[0040] Furthermore, in step 3, kriging is a statistical technique that calculates the value at a point within a random field based on observations from locations near that point; the interpolation formula for ordinary kriging is:

[0041]

[0042] in, Indicates a random point within the field. The predicted value, from point Observations of n nearby points and relative points Weighting coefficients The linear combination determines it.

[0043] Furthermore, step 3 is detailed below:

[0044] Step 301: Set the assumptions. The premise for applying the standard Kriging method is that the hitter's hitting motion is consistent and uniform. The trainer uses... Stability score for the action during conditional shot With the same expectation, it can be concluded that the motion stability evaluation score of a random point in the field is composed of the expected value of the motion stability evaluation score and the random deviation value under that condition. The motion stability evaluation score of a random point in the field is:

[0045]

[0046] in, represent Action stability evaluation score, Represents the expected value. , Represents any The random bias value under the given conditions; let the variance of the motion stability evaluation score be... ,have

[0047]

[0048] Therefore, the variance of random bias is also a constant. The variogram of ordinary kriging It is known, specifically: ;

[0049] express The variability function;

[0050] The variogram is a statistic that describes the spatial correlation of random fields and random processes. It is defined as the variance of the difference between two points in space.

[0051] Step 302: Set unbiased estimation conditions;

[0052] The weights in ordinary kriging must satisfy the requirement of unbiased estimation, which requires that the expected value of the estimated random bias is 0. Mathematically, this can be expressed as:

[0053] Note: The principle of the Kriging model is that the expected value of the estimated random bias is zero. The formula below is the mathematical expression for achieving a zero expected value for the estimated random bias. From this, we can deduce... This is used to solve the weight matrix in step 305.

[0054]

[0055] because ,Right now ,

[0056] At the same time, for any They all In conclusion, , ;

[0057] Step 303: Set the semivariance;

[0058] The semivariance of the motion stability evaluation score The equivalent form is ;

[0059] Step 304: Optimize the objective function;

[0060] Set Kriging error This represents the uncertainty or error in the predicted values ​​of the stability evaluation scores for each action; [the text abruptly ends here, likely due to an incomplete sentence or missing information.] The smallest group Solve using the Lagrange multiplier method;

[0061] Step 305: Construct the Lagrange multiplier.

[0062] In order to make Minimize, construct the Lagrange multiplier:

[0063]

[0064] in, For Lagrange functions, These are constraints. It is a Lagrange multiplier;

[0065] Will right Find the partial derivative and set its value to 0, that is ;

[0066] Solving this equation will ultimately yield the following matrix equation:

[0067] How is this matrix expression solved? Used for final calculation of predicted values .

[0068]

[0069] vector ,in, These are the weights used in the final calculation of the predicted value.

[0070] Ordinary Kriging model refers to The formula above shows the weight matrix.

[0071] Furthermore, the Kriging method relies on a semivariogram. The semivariogram can fit the semivariogram of all data pairs based on distance, and has three characteristic parameters: sill, range, and nugget value. Essentially, the semivariogram is a multi-parameter, nonlinear optimization method.

[0072] Step 4 is detailed below:

[0073] Step 401: Select the semivariogram; for hitting training, the difference in motion stability between two sets of training data with close spatial correlation is often small. This is specifically reflected in the semivariogram. The difference is small. However, for two sets of training data with relatively low spatial correlation, the difference in motion stability tends to increase. This means that the semivariance... As the spatial distance between training data increases, the corresponding motion stability evaluation scores no longer correlate. Their variances begin to flatten, and the sample values ​​become uncorrelated. Common semivariograms include spherical models, Gaussian models, linear models, damped sine models, and exponential models.

[0074] Step 402: Substitute the semi-mutation function into In the next step, we solve the matrix equation in step 305 to obtain... Then, according to the formula Calculate the stability score of the predicted hitting action. ;

[0075] Step 403: Plot the predicted stability evaluation score of the hitting action on a surface to obtain the correlation model between the stability evaluation score of the hitting action and the training data.

[0076] Furthermore, step 5 specifically involves the following steps:

[0077] Step 501: Select the evaluation model. Compare the predicted stability evaluation scores of each action in Step 403 with the measured values ​​of the stability evaluation scores of the predicted points. Evaluate the correlation model between the stability evaluation scores of the hitting action in Step 403 and the training data by establishing the mean square error (MSE) plot, root mean square error (RMSE) plot, and mean absolute error (MAE) plot to ensure that the correlation model between the stability evaluation scores of the hitting action and the training data has high accuracy.

[0078] Step 502: Construct an error map; construct a mean square error map and observe the areas with collapses within the mean square error map. The range allows trainees to hit the ball more times and gain experience. By using training data within a certain range, higher evaluation accuracy can be obtained.

[0079] The formula for calculating the mean squared error (MSE) is as follows: .

[0080] A stability evaluation system for tennis shot considering diverse ball trajectories includes:

[0081] The data acquisition module is used to collect relevant data on the trainee's hitting action and calculate the cumulative hand contribution velocity-time series.

[0082] The calculation module is used to calculate the stability evaluation score of the hitting action based on the dynamic time bending theory;

[0083] Modules for building standard kriging models;

[0084] The "Select Fit" module is used to choose the semivariogram function to fit.

[0085] The evaluation module is used to evaluate the correlation model between the stability evaluation score of the hitting action and the training data.

[0086] An apparatus comprising one or more processors;

[0087] Memory, used to store one or more programs;

[0088] When the one or more programs are executed by the one or more processors, the one or more processors implement the stability evaluation method for tennis shot actions that takes into account various ball trajectories as described above.

[0089] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the stability evaluation method for tennis shot actions considering diverse ball trajectories as described above.

[0090] Compared with the prior art, the beneficial effects of the present invention are:

[0091] 1. The method described in this invention captures the posture data generated during the trainee's hitting motion using an inertial motion capture system, and converts it into cumulative hand contribution velocity to quantify the velocity generated by each joint in the hand. Based on dynamic time curvature theory, the stability of the trainee's hitting motion is evaluated. Compared to traditional experience-based evaluation methods, the method described in this invention is more objective and accurate.

[0092] 2. The method described in this invention comprehensively reflects the trainee's technical level by evaluating the stability of their hitting motion under various ball trajectories. Furthermore, because the data originates from the trainee's posture data generated during actual hitting situations, this method can accurately identify areas for improvement compared to traditional, less personalized training methods.

[0093] 3. The method described in this invention uses a Kriging model for spatial correlation modeling, which can predict the stability performance of trainees under different ball trajectory conditions. Compared with traditional methods that judge hitting stability based on the trainee's past performance, this prediction function based on spatial correlation can help trainees understand their shortcomings in advance, thereby developing more targeted training plans and conducting training more efficiently. Attached Figure Description

[0094] To better describe and explain the technical solution of the invention, and to clearly demonstrate the content and process of the invention, the invention will be further described below with reference to the accompanying drawings:

[0095] Figure 1This is a flowchart of the stability evaluation method for tennis hitting actions considering various ball trajectories described in this invention;

[0096] Figure 2 The curve represents the function curve of the Gaussian model. Detailed Implementation

[0097] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the present invention will be described in detail below with reference to the accompanying drawings.

[0098] Step 1: Collect relevant data on the trainee's hitting motion. This includes the landing velocity, the distance between the hitting point and the trainee's body, the angular velocities of various body parts, and the swing speed, and calculate the velocity-time series. The specific steps are as follows:

[0099] Step 101: Use a high-speed camera to collect the landing velocity. and the distance between the hitting point and the trainee After 500 shots, the landing speed of each shot was recorded. Distance between the hitting point and the trainee The measured data were divided according to the landing velocity. Arrange the data from smallest to largest and divide them into 10 speed ranges. Then, analyze the data measured within each speed range according to the distance between the hitting point and the trainee. Arrange the data in ascending order and divide them into 10 equal speed intervals, ensuring that each interval contains 5 data points. , () indicates the range of the trainee's landing speed and the range of the front-to-back distance between the hitting point and the trainee's body. Training data for each group.

[0100] Step 102: Use inertial sensors to collect the rotational angular velocity and linear velocity of various parts. Set in the tennis hitting motion, the lower limbs drive the hips ( ) with angular velocity Rotation, where angular velocity Depend on , , Components, chest ( Drive the shoulder at angular velocity Rotation, shoulder ( Drive the upper arm at an angular velocity Rotation, elbow ( Drive the forearm at an angular velocity Rotation, hand ( The swing speed is .

[0101] Step 103: To more clearly identify the velocity generated by a specific body part, the velocity generated by each body part on the arm is processed independently using the hand contribution velocity method, and the contribution velocity of each part to the hand is calculated. The calculation formula is:

[0102]

[0103] in, This represents the hand's contribution velocity at joint k at time t. This represents the spatial position of joint k at time t.

[0104] Step 104: To better represent the kinematic chain, add the hand contribution velocity at joint k to the hand contribution velocity of the preceding joint to obtain the cumulative hand contribution velocity. Calculate the cumulative hand contribution velocity. The calculation method involves summing the contribution speeds of each part of the hand. The calculation formula is:

[0105]

[0106] in, This represents the cumulative hand contribution velocity at joint k at time t.

[0107] Step 105: Correspond the cumulative hand contribution speed with the collection time, that is, the horizontal axis is the time axis and the vertical axis is the cumulative hand contribution speed value, and record it as the cumulative hand contribution speed-time series (hereinafter referred to as speed-time series).

[0108] Step 2: Calculate the stability evaluation score of the hitting action based on the dynamic time curvature theory; specific steps:

[0109] Step 201: Process the data based on the velocity-time series obtained in Step 1. Collect data on the same hitting type from beginners. After each velocity-time series, the data is processed. Since each velocity-time series includes data from four joints: hip, chest, shoulder, and elbow, the... The velocity-time series data during the next stroke is represented as follows:

[0110]

[0111] Where k represents a joint, with the hip being 1, the chest 2, the shoulder 3, and the elbow 4. Indicates the first During the second strike, the joint area velocity-time series, Indicates joint location exist The cumulative hand contribution speed at any given moment.

[0112] Step 202: Arbitrarily select two velocity time series and combine them, and use the DTW algorithm to calculate each combined part. The optimal path distance is recorded as . Indicates the first During the second strike, the joint area velocity-time series, Indicates the first During the second strike, the joint area The velocity-time series. Since 5 velocity-time series were collected, there are 5(5-1) / 2=10 velocity-time series combinations. The following constraints need to be satisfied when using the DTW algorithm to find a path:

[0113] 1. Boundary conditions: At the beginning, the starting point must be... The destination must be Where a and b are velocity-time series. , The length of the time series.

[0114] 2. Continuity: After leaving the starting point, it is necessary to find the next data combination in the time series corresponding to the current data combination. Jumping on the time axis is not allowed.

[0115] 3. Monotonicity: When searching for the next data combination, you cannot go back on the timeline.

[0116] The optimal path distance between two velocity-time series is calculated as follows:

[0117]

[0118] in, , representing a velocity-time series of length a. and a velocity-time series of length b The optimal path distance between them; Representing velocity-time series The a-th element With velocity-time series The b-th element The distance between them.

[0119] Step 203: After traversing the optimal path distance of joint k under all velocity-time series combinations, calculate the average optimal path distance of the four joints in 5 shots under 10 velocity-time combinations using the DTW value of a single combination for a single joint. This allows for the calculation of the average optimal path distance between the 10 velocity-time series, further leading to the weighted stability score for the beginner in 5 shots. The specific calculation formula is as follows:

[0120]

[0121] in, The weighted stability score; Assigning a weight to joint k in the analysis of hitting stability, and setting... =0.42、 =0.20, =0.27, =0.11; The score is used to evaluate the stability of the ball striking at joint k. For the mth Group velocity - optimal path distance for time series.

[0122] Step 204: Obtain the stability score for the hitting action. , The training data is The stability evaluation score at that time. The calculation formula is:

[0123]

[0124] Step 3: Construct a standard Kriging model.

[0125] Kriging is a statistical technique that, within a random field, calculates the value at a given point based on observations from its vicinity. The interpolation formula for ordinary kriging is:

[0126]

[0127] in, Indicates a random point within the field. The predicted value, from point Observations of n nearby points and relative points Weighting coefficients The linear combination determines it.

[0128] Step 301: Set the assumptions. The premise for applying the standard Kriging method is that the hitter's hitting motion is consistent and uniform. The trainer uses... Stability score for the action during conditional shot Since they share the same expectation, it can be concluded that the motion stability evaluation score at a random point within the field is composed of the expected value of the motion stability evaluation score and the random deviation value under that condition. Therefore, the motion stability evaluation score at a random point within the field is:

[0129]

[0130] in, represent Action stability evaluation score, Represents the expected value. , Represents any The random bias value under the given conditions. Let the variance of the motion stability evaluation score be... ,have

[0131] ,

[0132] Therefore, the variance of random bias is also a constant. .

[0133] The variogram is a statistic describing the spatial correlation of random fields and random processes; it is defined as the variance of the difference between two points in space. The variogram of ordinary kriging... It is known, specifically:

[0134] Step 302: Set unbiased estimation conditions.

[0135] The weights in ordinary kriging must satisfy the requirement of unbiased estimation, that is, the expected value of the estimated random bias must be 0, which can be expressed mathematically as:

[0136]

[0137] because ,Right now ,

[0138] At the same time, for any They all In conclusion, it can be deduced that , .

[0139] Step 303: Set the semivariance.

[0140] The semivariance of the motion stability evaluation score The equivalent form is .

[0141] Step 304: Optimize the objective function.

[0142] Set Kriging error This represents the uncertainty or error in the predicted values ​​of the stability evaluation scores for each action. Because

[0143]

[0144]

[0145] Then there is

[0146] From step 303,

[0147]

[0148] thereby

[0149]

[0150] Next, we will search for... The smallest group The solution is obtained by applying the Lagrange multiplier method.

[0151] Step 305: Construct the Lagrange multiplier.

[0152] In order to make Minimize, construct the Lagrange multiplier:

[0153]

[0154] in, For Lagrange functions, These are constraints. It is a Lagrange multiplier.

[0155] Will right Find the partial derivative and set its value to 0, that is ;

[0156] Solving this equation will ultimately yield the following matrix equation:

[0157]

[0158] vector ,in, These are the weights used in the final calculation of the predicted value.

[0159] Step 4: Select a semi-variogram function for fitting.

[0160] The Kriging method relies on a semivariogram. The semivariogram can fit the semivariogram of all data pairs based on distance, and it has three characteristic parameters: sill, range, and nugget value. Essentially, the semivariogram is a multi-parameter, nonlinear optimization method.

[0161] Step 401: Select the semivariogram. For hitting training, the difference in motion stability between two sets of training data with close spatial correlation is often small. This is specifically reflected in the semivariogram. The difference is small. However, for two sets of training data with relatively low spatial correlation, the difference in motion stability tends to increase. This means that the semivariance... As the spatial distance between training data increases, the corresponding motion stability scores no longer correlate. Their variances begin to flatten, and the sample values ​​become uncorrelated. Common semivariograms include spherical models, Gaussian models, linear models, damped sine models, and exponential models. Since the Gaussian function uses a normal probability distribution curve, it gradually rises on the y-axis. Motion stability indices can be considered relatively similar over short distances. Therefore, in this method, a Gaussian model is preferred to fit the relationship between the semivariogram and the spatial correlation between the two sets of training data.

[0162] The curve of the Gaussian model is as follows Figure 2 As shown, gold nuggets The sill represents the value of y-truncation of the function corresponding to the Gaussian model. The value representing the first flattening of the Gaussian model; range This represents the distance at which the Gaussian model first flattens out. As the spatial correlation between the two sets of training data decreases, the semivariance increases, reaching the point of variation. At this point, there is no spatial autocorrelation between sample locations, and the system begins to reach a level close to flat.

[0163] The function of the Gaussian model is:

[0164]

[0165] Where h is the spatial step size. The semivariance with distance h The value of the semivariogram is denoted as . The smaller the value of the semivariogram, the stronger the spatial correlation, and vice versa.

[0166] Step 402: Substitute the semi-mutation function. Substitute the semi-mutation function into... In the next step, we solve the matrix equation in step 305 to obtain... Then, according to the formula Calculate the stability score of the predicted hitting action. .

[0167] Step 403: Plot the predicted stability evaluation score of the hitting action on a surface to obtain the correlation model between the stability evaluation score of the hitting action and the training data.

[0168] Step 5: Evaluate the correlation model between the stability evaluation score of the hitting action and the training data.

[0169] Step 501: Select the evaluation model. Compare the predicted stability evaluation scores of each action in Step 403 with the measured stability evaluation scores of the predicted points. Evaluate the correlation model between the stability evaluation scores of the hitting action in Step 403 and the training data by establishing the mean square error (MSE) plot, root mean square error (RMSE) plot, and mean absolute error (MAE) plot to ensure that the correlation model between the stability evaluation scores of the hitting action and the training data has high accuracy.

[0170] The formula for calculating the mean squared error (MSE) is:

[0171]

[0172] The formula for calculating the root mean square error (RMSE) is:

[0173]

[0174] The formula for calculating the mean absolute error (MAE) is:

[0175]

[0176] The number of data sets.

[0177] Since the mean squared error (MSE) represents the squared difference between the predicted and actual motion stability measurements, a larger error value has a greater impact on the goodness of fit, which helps to more intuitively observe the prediction error of the motion stability model. Therefore, in this method, the mean squared error (MSE) is preferred for evaluating the motion stability model.

[0178] Step 502: Construct an error map.

[0179] Construct a mean squared error plot and examine the areas that collapse within it. The range allows trainees to hit the ball more times and gain experience. By using training data within a certain range, higher evaluation accuracy can be obtained.

[0180] The present invention provides another embodiment of a tennis shot stability evaluation system considering diverse ball trajectories, comprising:

[0181] The data acquisition module is used to collect relevant data on the trainee's hitting action and calculate the cumulative hand contribution velocity time series.

[0182] The calculation module is used to calculate the stability evaluation score of the hitting action based on the dynamic time bending theory;

[0183] Modules for building standard kriging models;

[0184] The "Select Fit" module is used to choose the semivariogram function to fit.

[0185] The evaluation module is used to evaluate the correlation model between the stability evaluation score of the hitting action and the training data.

[0186] Based on the above-mentioned method for evaluating the stability of tennis shot considering diverse ball trajectories, this invention provides another device. The device includes, but is not limited to, one or more processors and a memory.

[0187] Memory, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions corresponding to the tennis shot stability evaluation method considering diverse ball trajectories in this embodiment of the invention. The processor executes the software programs, instructions, and modules stored in the memory to perform various functional applications and data processing of the vehicle, thereby implementing the aforementioned tennis shot stability evaluation method considering diverse ball trajectories.

[0188] The memory may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the terminal. Furthermore, the memory may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.

[0189] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for evaluating the stability of tennis hitting actions considering various ball trajectories. This method for evaluating the stability of tennis hitting actions considering various ball trajectories includes the following steps:

[0190] Step 1: Collect relevant data on the trainee's hitting action; the data includes the landing speed, the distance between the hitting point and the trainee's body, the rotational angular velocity of various parts of the body, and the swing speed; and calculate the cumulative hand contribution speed time series.

[0191] Step 2: Calculate the stability evaluation score of the hitting action based on the dynamic time curvature theory;

[0192] Step 3, construct a standard kriging model;

[0193] Step 4: Select a semi-variogram function for fitting;

[0194] Step 5: Evaluate the correlation model between the stability evaluation score of the hitting action and the training data.

[0195] The computer-readable storage medium provided by the present invention has computer-executable instructions that are not limited to the method operations described above, but can also perform related operations in the tennis hitting action stability evaluation method considering various ball trajectories provided in any embodiment of the present invention.

[0196] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0197] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line DSL) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk, SSD), etc.

[0198] In the above embodiments, the various units and modules are divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0199] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be included within the scope of protection of the present invention. Furthermore, all content not described in detail in this specification is prior art known to those skilled in the art.

Claims

1. A method for evaluating the stability of tennis hitting motions considering diverse ball trajectories, characterized in that, Includes the following steps: Step 1: Collect relevant data on the trainee's hitting action; the data includes the landing speed, the distance between the hitting point and the trainee's body, the rotational angular velocity of various parts of the body, and the swing speed; and calculate the cumulative hand contribution speed-time series. Step 2: Calculate the stability evaluation score of the hitting action based on the dynamic time bending theory; Step 2 is detailed below: Step 201: Each velocity-time series includes data from four locations: hip, chest, shoulder, and elbow. The velocity-time series data for the second stroke is represented as follows: in, Represents the joint area. At that time, the hip area, At that time, it was the chest. At times, shoulder At the elbow, Indicates the first During the second strike, the joint area velocity-time series, Indicates joint location exist The cumulative hand contribution speed at any given moment; Step 202: Arbitrarily select two velocity time series and combine them, and use the Dynamic Time Warp (DTW) algorithm to calculate each combined part. The optimal path distance is recorded as ; Indicates the first During the second strike, the joint area velocity-time series, Indicates the first During the second strike, the joint area Velocity-time series; Step 203: Calculate the DTW values ​​of the four joints in M ​​strokes using the DTW values ​​of a single combination of individual parts. The average optimal path distance under each speed-time combination is calculated, and the beginner's path distance is obtained. Weighted stability score of the second shot The specific calculation formula is as follows: in, The weighted stability score; Assigning a weight to joint k in the analysis of hitting stability, and setting... =0.42, =0.20, =0.27, =0.11; The score is used to evaluate the stability of the ball striking at joint k. For a certain interval, the first Optimal path distance under a combination of group speed and time series; Step 204: Calculate the stability score of the hitting action. ; The training data is Stability evaluation score at that time; The calculation formula is: Step 3, construct a standard kriging model; Step 3 is detailed below: Step 301: Set the assumptions; the application of the ordinary Kriging method is based on the premise that the batter's hitting motion is consistent, and the trainer uses... Stability score for the action during a conditional shot With the same expectation, it can be concluded that the motion stability evaluation score of a random point in the field is composed of the expected value of the motion stability evaluation score and the random deviation value under that condition. The motion stability evaluation score of a random point in the field is: in, represent Action stability evaluation score, Represents the expected value. , Represents any The random bias value under the given conditions; let the variance of the motion stability evaluation score be... ,have Variation function of ordinary kriging It is known, specifically: express The variability function; Step 302: Set unbiased estimation conditions; The weights of ordinary kriging must satisfy the requirement of unbiased estimation, which requires that the expected value of the estimated random bias is 0, expressed as: because ,Right now , At the same time, for any They all In conclusion, , ; Step 303: Set the semivariance; semivariance of motion stability evaluation score The equivalent form is ; Step 304: Optimize the objective function; Set Kriging error This represents the uncertainty or error in the predicted values ​​of the stability evaluation scores for each action; [the text abruptly ends here, likely due to an incomplete sentence or missing information.] The smallest group Solve using the Lagrange multiplier method; Step 305: Construct the Lagrange multiplier; In order to make Minimize, construct the Lagrange multiplier: in, For Lagrange functions, These are constraints. It is a Lagrange multiplier; Will right Find the partial derivative and set its value to 0, that is ; Solving this equation will ultimately yield the following matrix equation: vector ,in, These are the weights used in the final calculation of the predicted value; Step 4: Select a semi-variogram function for fitting; Step 4 is detailed below: Step 401: Select the semivariogram; the semivariogram includes the spherical model, Gaussian model, linear model, damped sine model, and exponential model; Step 402: Substitute the semi-mutation function into In the next step, we solve the matrix equation in step 305 to obtain... Then, according to the formula Calculate the stability score of the predicted hitting action. ; Step 403: Plot the predicted stability evaluation score of the hitting action on a surface to obtain the correlation model between the stability evaluation score of the hitting action and the training data; Step 5: Evaluate the correlation model between the stability evaluation score of the hitting action and the training data; Step 5 consists of the following steps: Step 501: Select the evaluation model; compare the predicted stability evaluation scores of each action in Step 403 with the measured values ​​of the stability evaluation scores of the predicted points, and evaluate the correlation model between the stability evaluation scores of the hitting action in Step 403 and the training data by establishing the mean square error (MSE) plot, root mean square error (RMSE) plot, and mean absolute error (MAE) plot to ensure that the correlation model between the stability evaluation scores of the hitting action and the training data has high accuracy. Step 502: Construct an error map; construct a mean square error map and observe the areas with collapses within the mean square error map. The range allows trainees to hit the ball more times and gain experience. Training data within the specified range; The formula for calculating the mean squared error (MSE) is: .

2. The method for evaluating the stability of tennis hitting motion considering diverse ball trajectories according to claim 1, characterized in that, Step 1: The specific steps are as follows: Step 101: Collect landing speed and the distance between the hitting point and the trainee After performing N shots, record the landing speed of the ball for each shot. Distance between the hitting point and the trainee The measured data are divided according to the landing velocity. Arrange the data from smallest to largest and divide them into C speed intervals. Then, rank the data within each speed interval according to the distance between the hitting point and the trainee. Arrange the data in ascending order and divide them into D speed intervals, ensuring that each interval contains M data points; , () indicates the range of the trainee's landing speed and the range of the front-to-back distance between the hitting point and the trainee's body. Training data during grouping; Step 102: Collect the rotational angular velocity of each part; specifically, the lower limbs drive the hips with an angular velocity... Rotation, where angular velocity Depend on , , Component composition, chest drives shoulders with angular velocity Rotation, shoulder drives upper arm at angular velocity Rotation, elbow drives forearm at angular velocity The rotation speed of the hand swing is ; Step 103: Calculate the contribution rate of each part to the hand. The calculation formula is: in, Represents joint The velocity contributed by the hand at time t This represents the spatial position of joint k at time t; At that time, the hip area, At that time, it was the chest. At times, the shoulder At the elbow, At that time, it was the hand; Step 104: Calculate the cumulative hand contribution speed The calculation method involves summing the contribution speeds of each part of the hand; the calculation formula is: in, This represents the cumulative hand contribution velocity at joint k at time t; This represents the spatial position of joint k at time t; Step 105: Record the cumulative hand contribution rate - time series; correlate the cumulative hand contribution rate with the acquisition time, that is, the horizontal axis is the time axis and the vertical axis is the cumulative hand contribution rate value, and record it as the cumulative hand contribution rate - time series.

3. The method for evaluating the stability of tennis hitting motion considering diverse ball trajectories according to claim 1, characterized in that, When using the DTW algorithm to find a path, the following constraints must be met: (1) Boundary conditions: At the beginning, the starting point must be The destination must be Where a and b are velocity-time series. , The length of the time series; (2) Continuity: After leaving the starting point, it is necessary to find the next data combination in the time series corresponding to the current data combination. Jumping on the time axis is not allowed. (3) Monotonicity: When searching for the next data combination, it is not possible to go back on the time axis; The optimal path distance between two velocity-time series is calculated as follows: in, , representing a velocity-time series of length a. and a velocity-time series of length b The optimal path distance between them; Representing velocity-time series The a-th element With velocity-time series The b-th element The distance between them.

4. The method for evaluating the stability of tennis hitting motion considering diverse ball trajectories according to claim 1, characterized in that: In step 3, Kriging is a statistical technique that calculates the value at a point within a random field based on observations from locations near that point. The interpolation formula for ordinary Kriging is: in, Indicates a random point within the field. The predicted value, from point Observations of n nearby points and relative points Weighting coefficients The linear combination determines this.

5. The method for evaluating the stability of tennis hitting motion considering diverse ball trajectories according to claim 4, characterized in that: The semivariogram fits the semivariogram of all data pairs based on the distance and has three characteristic parameters: sill, range, and nugget value.

6. The evaluation system for a method of evaluating the stability of tennis hitting action considering diverse ball trajectories according to any one of claims 1 to 5, characterized in that, include: The data acquisition module is used to collect relevant data on the trainee's hitting action and calculate the cumulative hand contribution velocity-time series. The calculation module is used to calculate the stability evaluation score of the hitting action based on the dynamic time curvature theory; Modules for building standard kriging models; The "Select Fit" module is used to choose the semivariogram function to fit. The evaluation module is used to evaluate the correlation model between the stability evaluation score of the hitting action and the training data.

7. An apparatus, characterized in that: Includes one or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by the processor, it implements the method as described in any one of claims 1-5.

Images