An intelligent wooden man robot capable of measuring strength and a motion analysis system

Abstract

The present invention relates to an intelligent wooden man robot and motion analysis system capable of measuring skill. The system analyzes the weights of the hitting points in each area of ​​the wooden man according to the distance from the key hitting point; uses a pressure sensor to obtain the force data of hitting the wooden man; monitors the hitting points and force of the wooden man, and calculates the hitting skill value in combination with the weights of the hitting points in each area of ​​the wooden man; uses a K-means cluster analysis method to obtain the skill level of different hitting skill values, and emits lights of different brightness at the hitting points according to different skill levels; extracts the image feature data of the hitting action video stream, compares it with the standard action image feature data, and evaluates the trainee's action in combination with the corresponding skill level; and emits lights of different colors according to the evaluation results of the hitting action. The present invention solves the problem that it is difficult to interact with lights according to the accuracy of different hitting parts of the wooden man in the prior art to accurately evaluate the trainee's skill and action.

Description

Technical Field

[0001] The invention belongs to the technical field of intelligent wooden man piles, and relates to an intelligent wooden man pile robot capable of measuring skill and a motion analysis system. Background Art

[0002] The wooden man is a traditional Chinese boxing equipment, usually used to train martial arts skills. It is a vertically placed round log with fixed protrusions on the top and sides to simulate a real enemy. During training, it is attacked and hit from different directions to improve martial arts skills. The shape and structure of the wooden man can be adjusted to suit different training needs. Modern technology has introduced intelligent elements into wooden man training, such as robots that can simulate real-life attacks and hits to help practitioners better improve their martial arts skills.

[0003] At present, most intelligent wooden dummy interacts with trainees through lights, such as testing skills and analyzing movements. In the prior art, when testing skills, the light interaction for different hitting parts of the wooden dummy is the same, that is, the light only changes in a fixed way considering the hitting force. However, in reality, the accuracy of the hitting part of the wooden dummy is closely related to the accuracy of the movement. It is difficult for the prior art to interact with lights according to the accuracy of the hitting parts of the wooden dummy to accurately evaluate the trainee's skills and movements. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention provides an intelligent wooden man robot capable of measuring skill and a motion analysis system.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] In a first aspect, the present application provides an intelligent wooden man robot capable of measuring skill, comprising a striking point analysis module, a strength acquisition module, a skill analysis module and a light interaction module, wherein the striking point analysis module, the strength acquisition module, the skill analysis module and the light interaction module are communicatively connected, wherein:

[0007] The hitting point analysis module is used to identify the key hitting points of the wooden dummy and analyze the weights of the hitting points in each area of ​​the wooden dummy according to the distance from the key hitting points;

[0008] The force acquisition module uses a pressure sensor to acquire force data of hitting the wooden man;

[0009] The skill analysis module is used to monitor and analyze the hitting points and strength of the wooden man, and calculate the hitting skill value based on the weights of the hitting points in each area of ​​the wooden man;

[0010] The lighting interaction module is used to input the striking skill value into the skill level analysis model, output the corresponding skill level, and emit lights of different brightness at the striking point according to different skill levels; the skill level analysis model, the skill level is obtained by dividing the large-scale striking skill value data set in the historical period through the K-means clustering method.

[0011] Furthermore, the weights of the hitting points in each area of ​​the wooden dummy are analyzed and calculated using the following formula:

[0012]

[0013] In the formula, ω represents the weight of a certain hitting point; D i represents the distance between the hitting point and the i-th key hitting point; N represents the number of key hitting points near the hitting point.

[0014] Furthermore, the pressure sensor includes a pile head pressure sensor, a pile chest pressure sensor, a pile belly pressure sensor, a pile hand pressure sensor and a pile foot pressure sensor; the pile head pressure sensor is used to measure the force of hitting the pile head; the pile chest pressure sensor is used to measure the force of hitting the pile chest; the pile belly pressure sensor is used to measure the force of hitting the pile belly; the pile hand pressure sensor is used to measure the force of hitting the pile hand; the pile foot pressure sensor is used to measure the force of hitting the pile foot.

[0015] Furthermore, the striking power value is calculated by the formula:

[0016] P = ω × F,

[0017] Where P is the striking power value of a certain striking point; ω is the weight of the striking point; and F is the force applied to the striking point.

[0018] Furthermore, the K-means clustering method comprises the following steps:

[0019] S21. Data collection: Collect a large-scale dataset of skill strike values ​​from historical periods;

[0020] S22, normalization processing: performing normalization processing on the skill hitting value data set;

[0021] S23, initializing cluster centers: randomly selecting several data objects in the data space as initial cluster centers;

[0022] S24, initializing clusters: calculating the Euclidean distances between all data objects and the initial cluster centers, and dividing each data object into the initial cluster centers with the smallest Euclidean distances to form initial clusters;

[0023] S25, updating the cluster center: calculating the average value of the data objects of each initial cluster, and taking the average value as the new cluster center;

[0024] S26. Determine the final cluster: repeatedly calculate the Euclidean distance between all data objects and the new cluster center, divide each data object into the new cluster center with the smallest Euclidean distance to form a new cluster; calculate the average value of each new cluster as the new cluster center again, until the new cluster center no longer changes, determine the final cluster, and determine the skill level corresponding to each final cluster according to the data characteristics of the final cluster.

[0025] Furthermore, in step S22, the normalization process is calculated using the following formula:

[0026]

[0027] Where Y i is the normalized i-th striking power value data sample; K i is the i-th striking power value data sample; K max Represents the maximum value of all striking power value data samples; K min Indicates the minimum value of all striking power data samples.

[0028] Furthermore, in step S24, the Euclidean distance is calculated by the formula:

[0029]

[0030] In the formula, d(x,C i ) represents the Euclidean distance function; x is the normalized data object; C i represents the i-th cluster center; n is the number of samples in the data set; x j Represents the data object of the jth variable in the data set; C ij represents the i-th cluster center of the j-th variable.

[0031] In a second aspect, the present application provides a motion analysis system for an intelligent wooden man robot capable of measuring skill, which is applied to an intelligent wooden man robot capable of measuring skill as described above, and includes a motion acquisition module and a motion evaluation module, wherein the motion acquisition module and the motion evaluation module are communicatively connected, wherein:

[0032] The action acquisition module is used to acquire a video stream image of a trainee's striking action, wherein the striking action includes a striking point and a striking posture;

[0033] The action evaluation module is used to extract the image feature data of the striking action video stream, compare it with the standard action image feature data, and evaluate the trainer's action in combination with the corresponding skill level, specifically:

[0034] Obtaining feature similarity between striking action video stream images and standard action images;

[0035] The feature similarity and skill level are used to fuse data and obtain motion vector data;

[0036] According to the action vector data, a comprehensive score of the striking action is obtained.

[0037] Furthermore, the action evaluation module also sends instructions to the light interaction module to emit lights of different colors according to the comprehensive score of the striking action.

[0038] Beneficial effects of the present invention:

[0039] According to the distance from the key striking point, the weight of the striking points in each area of ​​the wooden man is analyzed; the force data of the striking on the wooden man is obtained by using a pressure sensor; the striking points and force of the wooden man are monitored, and the striking skill value is calculated in combination with the weights of the striking points in each area of ​​the wooden man; the skill level of different striking skill values ​​is obtained by using the K-means clustering analysis method, and lights of different brightness are emitted at the striking points according to different skill levels; the image feature data of the striking action video stream is extracted, compared with the standard action image feature data, and the trainee's action is evaluated in combination with the corresponding skill level; different colors of lights are emitted according to the striking action evaluation results. The present invention solves the problem that it is difficult to interact with lights according to the accuracy of different striking parts of the wooden man to accurately evaluate the trainee's skill and action in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0041] Figure 1 The figure is a structural diagram of an intelligent wooden man robot capable of measuring power in the present invention.

[0042] Figure 2 The present invention is a structural diagram of an intelligent wooden man robot motion analysis system capable of measuring skill. DETAILED DESCRIPTION

[0043] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0044] See also Figure 1-2In the first aspect, the present application provides an intelligent wooden man robot capable of measuring skill, comprising a striking point analysis module, a strength acquisition module, a skill analysis module and a light interaction module, wherein the striking point analysis module, the strength acquisition module, the skill analysis module and the light interaction module are communicatively connected, wherein:

[0045] The hitting point analysis module is used to identify the key hitting points of the wooden dummy and analyze the weights of the hitting points in each area of ​​the wooden dummy according to the distance from the key hitting points;

[0046] In this embodiment, the key striking points refer to the most critical points on the surface of the wooden man that need to be struck. These points are usually the most important parts in martial arts moves, such as the head, chest, abdomen, waist, shoulders, arms, legs, etc. In actual training, the trainee needs to hit these key striking points as accurately as possible to improve the effect of the practice and improve the skill level. By identifying these key striking points, the distance between the trainee's striking position and the key striking points can be recorded in real time, and feedback can be provided on the accuracy of their striking.

[0047] The weight of the hitting points in each area of ​​the wooden dummy can be evaluated by calculating the distance from the key hitting point. Generally, the closer the hitting point is to the key hitting point, the higher its weight is, because these parts are more important for the execution of skills and the effect of attack. When performing weight analysis, a mathematical model or algorithm can be used to quantify the importance of the hitting point. For example, the inverse of the distance can be used as a weight factor, and the closer the hitting point is to the key hitting point, the higher the weight is, and the farther the hitting point is, the lower the weight is.

[0048] Furthermore, the weights of the hitting points in each area of ​​the wooden dummy are analyzed and calculated using the following formula:

[0049]

[0050] In the formula, ω represents the weight of a certain hitting point; D i represents the distance between the hitting point and the i-th key hitting point; N represents the number of key hitting points near the hitting point.

[0051] It should be noted that the number of N usually depends on the number of key striking points within 10 cm from the striking point.

[0052] The force acquisition module uses a pressure sensor to acquire force data of hitting the wooden man;

[0053] In this embodiment, the pressure sensor can be installed in the key parts of the wooden man, such as the head, chest, abdomen and other positions, to detect the force applied by the strike. When the trainee strikes, the pressure sensor can sense and record the pressure or force applied by the strike. Through the data collected by the sensor, information about the force of the strike can be obtained, including the intensity of the strike, the size of the force, etc. These force data can be used to analyze and identify the strike effect of the trainee, such as determining whether the force of the strike reaches the expected level, the improvement of the striking technique, etc. At the same time, by comparing with the preset striking force, real-time feedback can be provided to help the trainee make adjustments and improvements to improve the quality and accuracy of the strike.

[0054] Furthermore, the pressure sensor includes a pile head pressure sensor, a pile chest pressure sensor, a pile belly pressure sensor, a pile hand pressure sensor and a pile foot pressure sensor; the pile head pressure sensor is used to measure the force of hitting the pile head; the pile chest pressure sensor is used to measure the force of hitting the pile chest; the pile belly pressure sensor is used to measure the force of hitting the pile belly; the pile hand pressure sensor is used to measure the force of hitting the pile hand; the pile foot pressure sensor is used to measure the force of hitting the pile foot.

[0055] The skill analysis module is used to monitor and analyze the hitting points and strength of the wooden man, and calculate the hitting skill value based on the weights of the hitting points in each area of ​​the wooden man;

[0056] In this embodiment, visual sensors or other methods can be used to monitor and analyze the hitting points of the wooden man. By identifying the key hitting points of the wooden man and recording the positions of the hitting points, the hitting accuracy of the trainee can be known. The hitting points of key parts and their vicinity will be given higher weights because they are more important for skill execution and attack effects. Combining the strength data and weights of the hitting points, the hitting skill value can be calculated. The strength of the hitting points is multiplied by the corresponding weights, and finally a comprehensive skill value is obtained. By calculating the skill value, the trainee can obtain a comprehensive evaluation of the hitting effect and ability. This evaluation can help the trainee understand his or her hitting skills and strength performance, and make necessary adjustments and improvements in training to improve the skill level and hitting quality.

[0057] Furthermore, the striking power value is calculated by the formula:

[0058] P = ω × F,

[0059] Where P is the striking power value of a certain striking point; ω is the weight of the striking point; and F is the force applied to the striking point.

[0060] The lighting interaction module is used to input the striking skill value into the skill level analysis model, output the corresponding skill level, and emit lights of different brightness at the striking point according to different skill levels;

[0061] In this embodiment, the light interaction module is used to interact with the trainer and record the trainer's skill of each strike in real time. By analyzing the skill level of different striking skill values, the light interaction module can emit lights of different brightness. Generally, the higher the skill level, the higher the brightness of the light emitted. This visual interactive effect can help the trainer understand his or her striking situation in real time to quickly make corrections and adjustments. However, the current classification of the level of striking skill values ​​is not clear, and people's thresholds for dividing the upper and lower levels are mostly based on experience or subjective judgment, resulting in inaccurate division results, thereby causing the problem of reduced interactive effect. To this end, this embodiment collects a large-scale striking skill value data set in historical periods, and uses the K-means clustering method to divide the data set to obtain a skill level with a clear data boundary (threshold). K-means clustering is a common unsupervised learning algorithm for dividing data points into different clusters. By calculating the similarity between data points, the K-means algorithm can classify data points into the closest cluster. This method helps to discover hidden patterns and structures in the data, thereby realizing cluster analysis and classification of data. The K-means algorithm is used to divide the striking skill value dataset into different clusters, each cluster represents the striking skills with similar characteristics, which can reflect.

[0062] Therefore, the skill level analysis model, the skill level is obtained by dividing the large-scale striking skill value data set in the historical period through the K-means clustering method.

[0063] After the skill level is determined by the K-means clustering method, the real-time acquired striking skill value data is input into the skill level analysis model to output the corresponding skill level.

[0064] Furthermore, the K-means clustering method comprises the following steps:

[0065] S21. Data collection: Collect a large-scale dataset of skill strike values ​​from historical periods;

[0066] S22, normalization processing: performing normalization processing on the skill hitting value data set;

[0067] In this embodiment, the normalization process can make the clustering algorithm more stable and converge faster, which is conducive to obtaining more accurate and reliable clustering results.

[0068] S23, initializing cluster centers: randomly selecting several data objects in the data space as initial cluster centers;

[0069] In this embodiment, when initializing the cluster centers, a common practice is to randomly select several data objects in the data space as the initial cluster centers. The reason for this is that by randomly selecting the initial cluster centers, it can be ensured that the algorithm will not be affected by the position of a specific initial cluster center, thereby avoiding falling into a local optimal solution. By randomly selecting data objects in the data space as the initial cluster centers, the algorithm can have a certain degree of randomness, which helps to explore different aspects of the data, and can increase the robustness and reliability of the algorithm, and improve the quality of the final clustering results. Randomly selecting data objects as initial cluster centers is a common and effective initialization method.

[0070] S24, initializing clusters: calculating the Euclidean distances between all data objects and the initial cluster centers, and dividing each data object into the initial cluster centers with the smallest Euclidean distances to form initial clusters;

[0071] S25, updating the cluster center: calculating the average value of the data objects of each initial cluster, and taking the average value as the new cluster center;

[0072] S26. Determine the final cluster: repeatedly calculate the Euclidean distance between all data objects and the new cluster center, divide each data object into the new cluster center with the smallest Euclidean distance to form a new cluster; calculate the average value of each new cluster as the new cluster center again, until the new cluster center no longer changes, determine the final cluster, and determine the skill level corresponding to each final cluster according to the data characteristics of the final cluster.

[0073] Furthermore, in step S22, the normalization process is calculated using the following formula:

[0074]

[0075] Where Y i is the normalized i-th striking power value data sample; K i is the i-th striking power value data sample; K max Represents the maximum value of all striking power value data samples; K min Indicates the minimum value of all striking power data samples.

[0076] Furthermore, in step S24, the Euclidean distance is calculated by the formula:

[0077]

[0078] In the formula, d(x,C i) represents the Euclidean distance function; x is the normalized data object; C i represents the i-th cluster center; n is the number of samples in the data set; x j Represents the data object of the jth variable in the data set; C ij represents the i-th cluster center of the j-th variable.

[0079] In a second aspect, the present application provides a motion analysis system for an intelligent wooden man robot capable of measuring skill, which is applied to an intelligent wooden man robot capable of measuring skill as described above, and includes a motion acquisition module and a motion evaluation module, wherein the motion acquisition module and the motion evaluation module are communicatively connected, wherein:

[0080] The action acquisition module is used to acquire a video stream image of a trainee's striking action, wherein the striking action includes a striking point and a striking posture;

[0081] The action evaluation module is used to extract the image feature data of the striking action video stream, compare it with the standard action image feature data, and evaluate the trainer's action in combination with the corresponding skill level, specifically:

[0082] Obtaining feature similarity between striking action video stream images and standard action images;

[0083] The feature similarity and skill level are used to fuse data and obtain motion vector data;

[0084] According to the action vector data, a comprehensive score of the striking action is obtained.

[0085] In this embodiment, the extraction and comparison of striking action image feature data, and the evaluation of the trainee's action in combination with the skill level can be achieved through the following steps:

[0086] 1) Data collection: Collect a set of image data of standard movements as reference. These standard movements can be examples of movements performed by experts or experienced trainers;

[0087] 2) Feature extraction: Use computer vision technology to extract feature data from the trainer's striking action images. Common feature extraction methods include traditional feature extraction algorithms based on manual design (such as directional gradient histograms, local binary patterns, etc.), and feature extraction methods based on deep learning (such as feature extraction of convolutional neural networks);

[0088] 3) Feature comparison: Compare the trainee's striking action image feature data with the standard action image feature data. Various similarity measurement methods (such as Euclidean distance, cosine similarity, etc.) can be used to measure the similarity between the two.

[0089] 4) Action evaluation: Combined with the skill level, the trainee's action is evaluated according to the feature comparison results. Even if the trainee's action is consistent with the standard action, the strength of the strike is also within the scope of action analysis. Therefore, the skill level is monitored to evaluate the trainee's action, and the fusion data of the feature similarity between the skill level and the action image can be used as the result of the action evaluation.

[0090] Furthermore, the action evaluation module also sends instructions to the light interaction module to emit lights of different colors according to the comprehensive score of the striking action.

[0091] Beneficial effects of the present invention:

[0092] According to the distance from the key striking point, the weight of the striking points in each area of ​​the wooden man is analyzed; the force data of the striking on the wooden man is obtained by using a pressure sensor; the striking points and force of the wooden man are monitored, and the striking skill value is calculated in combination with the weights of the striking points in each area of ​​the wooden man; the skill level of different striking skill values ​​is obtained by using the K-means clustering analysis method, and lights of different brightness are emitted at the striking points according to different skill levels; the image feature data of the striking action video stream is extracted, compared with the standard action image feature data, and the trainee's action is evaluated in combination with the corresponding skill level; different colors of lights are emitted according to the striking action evaluation results. The present invention solves the problem that it is difficult to interact with lights according to the accuracy of different striking parts of the wooden man to accurately evaluate the trainee's skill and action in the prior art.

[0093] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. An intelligent wooden man robot capable of measuring skill, characterized in that: It includes a striking point analysis module, a force acquisition module, a skill analysis module and a light interaction module, wherein the striking point analysis module, the force acquisition module, the skill analysis module and the light interaction module are communicatively connected, wherein: The hitting point analysis module is used to identify the key hitting points of the wooden dummy and analyze the weights of the hitting points in each area of ​​the wooden dummy according to the distance from the key hitting points; The force acquisition module uses a pressure sensor to acquire force data of hitting the wooden man; The skill analysis module is used to monitor and analyze the hitting points and strength of the wooden man, and calculate the hitting skill value based on the weights of the hitting points in each area of ​​the wooden man; The lighting interaction module is used to input the striking skill value into the skill level analysis model, output the corresponding skill level, and emit lights of different brightness at the striking point according to different skill levels; the skill level analysis model, the skill level is obtained by dividing the large-scale striking skill value data set in the historical period through the K-means clustering method.

2. The intelligent wooden man robot capable of measuring power according to claim 1, characterized in that: The weights of the hitting points in each area of ​​the wooden dummy are calculated using the following formula: ， In the formula, ω Indicates the weight of a certain hitting point; D i Indicates that the hitting point is i The distance between the key striking points; N Indicates the number of key hitting points near the hitting point.

3. The intelligent wooden man robot capable of measuring power according to claim 1, characterized in that: The pressure sensor includes a pile head pressure sensor, a pile chest pressure sensor, a pile belly pressure sensor, a pile hand pressure sensor and a pile foot pressure sensor; the pile head pressure sensor is used to measure the force of hitting the pile head; the pile chest pressure sensor is used to measure the force of hitting the pile chest; the pile belly pressure sensor is used to measure the force of hitting the pile belly; the pile hand pressure sensor is used to measure the force of hitting the pile hand; the pile foot pressure sensor is used to measure the force of hitting the pile foot.

4. The intelligent wooden man robot capable of measuring power according to claim 1, characterized in that: The striking power value is calculated by the formula: ， In the formula, P It is the striking power value of a certain striking point; ω is the weight of the hitting point; F is the force applied to the striking point.

5. The intelligent wooden man robot capable of measuring power according to claim 1, characterized in that: The K-means clustering method comprises the following steps: S21. Data collection: Collect a large-scale dataset of skill strike values ​​from historical periods; S22, normalization processing: performing normalization processing on the skill hitting value data set; S23, initializing cluster centers: randomly selecting several data objects in the data space as initial cluster centers; S24, initializing clusters: calculating the Euclidean distances between all data objects and the initial cluster centers, and dividing each data object into the initial cluster centers with the smallest Euclidean distances to form initial clusters; S25, updating the cluster center: calculating the average value of the data objects of each initial cluster, and taking the average value as the new cluster center; S26. Determine the final cluster: repeatedly calculate the Euclidean distance between all data objects and the new cluster center, divide each data object into the new cluster center with the smallest Euclidean distance to form a new cluster; calculate the average value of each new cluster as the new cluster center again, until the new cluster center no longer changes, determine the final cluster, and determine the skill level corresponding to each final cluster according to the data characteristics of the final cluster.

6. The intelligent wooden man robot capable of measuring power according to claim 5, characterized in that: In step S22, the normalization process is performed using the following calculation formula: ， In the formula, Y i After normalization, i samples of striking power value data; K i For the i samples of striking power value data; K max Indicates the maximum value of all striking power value data samples; K min Indicates the minimum value of all striking power data samples.

7. The intelligent wooden man robot capable of measuring power according to claim 5, characterized in that: In step S24, the Euclidean distance is calculated using the following formula: ， In the formula, d ( Y i , C i ) represents the Euclidean distance function; Y i is the normalized data object; C i Indicates i Cluster centers; n is the number of samples in the data set; Y ij Indicates the data set j A data object with a variable; C ij Indicates j The variable i Cluster centers.

8. A motion analysis system for an intelligent wooden man robot capable of measuring skill, applied to an intelligent wooden man robot capable of measuring skill as claimed in any one of claims 1 to 7, characterized in that: It includes an action acquisition module and an action evaluation module, the action acquisition module and the action evaluation module are communicatively connected, wherein: The action acquisition module is used to acquire a video stream image of a trainee's striking action, wherein the striking action includes a striking point and a striking posture; The action evaluation module is used to extract the image feature data of the striking action video stream, compare it with the standard action image feature data, and evaluate the trainer's action in combination with the corresponding skill level, specifically: Obtaining feature similarity between striking action video stream images and standard action images; The feature similarity and skill level are used to fuse data and obtain motion vector data; According to the action vector data, a comprehensive score of the striking action is obtained.

9. The intelligent wooden man robot motion analysis system capable of measuring skill according to claim 8, characterized in that: The action evaluation module also sends instructions to the light interaction module according to the comprehensive score of the striking action to emit lights of different colors.

Images