Motion correction and training method, device and system for home fitness

Abstract

The invention relates to an action correction and training method, device and system for home fitness. The method comprises the following steps: dividing actions of a user into a correct action category and a wrong action category according to a preset user action category; storing the correct action category and the wrong action category as known classification variables; receiving user action information; storing the received user action information as unknown classification variables; calculating the distance between the unknown classification variable and the known classification variable, and sorting according to a distance increasing order; selecting K points with the minimum distance from the unknown classification variable, and determining the occurrence frequency of the category where the first K points are located; taking the category with the highest occurrence frequency of the first K points as a predicted action category of an unknown classification variable; outputting a predicted action category of an unknown classification variable; and performing classified statistics on each group of user action information trained by the user, and determining the number of action categories. Action correction of the user during home exercise is conveniently achieved, and the exercise effect is improved.

Description

Technical Field

[0001] The invention relates to an intelligent fitness device, and in particular to a method, device and system for correcting and training movements for home fitness. Background Art

[0002] In modern life, people are under great pressure in life, lack of exercise and fitness, and their physical fitness is generally getting worse and worse. Long-term intake of fast food, high-sugar, and high-calorie foods leads to obesity and increased body fat rate, which endangers people's health. Therefore, fitness has become an indispensable part. Fitness is a kind of sports, such as various freehand aerobics, rhythmic gymnastics, body gymnastics, and various self-resistance movements. Gymnastics and yoga can enhance strength, flexibility, increase endurance, improve coordination, and control the ability of various parts of the body, thereby making the body strong. With the acceleration of the pace of life and the development of a large number of online fitness videos and fitness apps, fitness at home has become the choice of more and more people. Home fitness is particularly important for maintaining physical and mental health, but when exercising at home, it is easy to have non-standard movements because there is no professional coach to correct the movements. Long-term non-standard movements are prone to solidified wrong movements, which affect the exercise effect and physical health. Summary of the invention

[0003] Based on this, it is necessary to provide a movement correction and training method, device and system for home fitness, in view of the fact that when doing home fitness, it is easy to have non-standard movements due to the lack of professional coaches to correct the movements. Long-term non-standard movements can easily form solidified wrong movements, affecting the exercise effect and physical health problems.

[0004] The present invention provides a method for correcting and training movements for home fitness, comprising the following steps:

[0005] Classifying the user's actions into a correct action category and an incorrect action category according to preset user action categories;

[0006] Store the correct action categories and incorrect action categories as known categorical variables;

[0007] Receiving user motion information, the user motion information including acceleration information and its changes, and / or distance information and its changes, and / or photosensitivity information and its changes;

[0008] The received user action information is stored as an unknown categorical variable;

[0009] Calculate the distance between the unknown categorical variable and the known categorical variable, and sort them in increasing order of distance;

[0010] Select the K points with the smallest distance to the unknown classification variable and determine the number of occurrences of the categories of the first K points;

[0011] The category with the highest number of occurrences in the first K points is used as the predicted action category of the unknown categorical variable;

[0012] Output the predicted action category of the unknown categorical variable;

[0013] Each group of user action information trained by the user is classified and counted, and the number of correct action categories and / or the number of incorrect action categories are output.

[0014] In one embodiment, the method further includes the following steps: when the predicted action category of the unknown classification variable is an incorrect action category, outputting prompt information.

[0015] In one embodiment, the distance calculation formula between the unknown classification variable and the known classification variable is as follows:

[0016]

[0017] Among them, d ij It represents the distance between the unknown categorical variable i and the known categorical variable j, q represents the order, q=1 or 2 or +∞, and k represents the number of samples selected with the minimum distance to the unknown categorical variable.

[0018] The present invention provides a home fitness movement correction and training device, comprising:

[0019] A preset action category module, used to classify the user's actions into a correct action category and an incorrect action category according to preset user action categories;

[0020] A known categorical variable module, used for storing correct action categories and incorrect action categories as known categorical variables;

[0021] A user motion information module, used to receive user motion information, including acceleration information and its changes, and / or distance information and its changes, and / or photosensitivity information and its changes;

[0022] An unknown categorical variable module, used for storing received user action information as an unknown categorical variable;

[0023] The distance sorting module is used to calculate the distance between the unknown categorical variables and the known categorical variables and sort them in ascending order of distance;

[0024] The sample selection module is used to select the K points with the smallest distance to the unknown classification variable and determine the number of occurrences of the category of the first K points;

[0025] The module for predicting action categories is used to take the category with the highest number of occurrences of the first K points as the predicted action category of the unknown classification variable;

[0026] A predicted action category output module is used to output the predicted action category of the unknown classification variable;

[0027] The classification statistics output module is used to classify and count each group of user action information trained by the user, and output the number of correct action categories and / or the number of incorrect action categories.

[0028] In one embodiment, a prompt module is further included, wherein the prompt module is used to output prompt information when the action category prediction module predicts an incorrect action category.

[0029] In one embodiment, the prompt module is a vibration prompt module or a voice prompt module.

[0030] The present invention provides a home fitness movement correction and training system, comprising:

[0031] A sensor for detecting user motion information, including an acceleration sensor, a photosensor and an ultrasonic sensor, wherein the photosensor and the ultrasonic sensor are both installed on the fitness equipment, and the acceleration sensor is worn on the user;

[0032] A control mainboard, used to receive data collected by the sensor, identify and judge the correct category and the wrong category of the user's action, and output the number of correct categories and / or the number of wrong categories;

[0033] The display screen is connected to the control mainboard and is used to display the categories and numbers of user actions.

[0034] In one embodiment, a prompt device is further included, and when the control mainboard determines that the user action is of an incorrect category, the prompt device outputs a prompt message.

[0035] In one embodiment, the prompt device is a vibration prompt device, which is worn on the user, and the vibration prompt device is connected to the control mainboard via a wireless network.

[0036] In one embodiment, the prompt device is a voice prompt device, and the voice prompt device is connected to the control mainboard.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. By pre-storing correct action categories and common incorrect action categories, and then attributing the training actions of the user during use to a category to which the majority of the k most adjacent samples in the feature space belong, it is possible to determine whether the user's exercise actions are correct, and output correct results or incorrect results. At the same time, the number of correct actions and incorrect actions are counted separately, which facilitates the user's action correction during home exercise and improves the exercise effect.

[0039] 2. Strong targeting: By entering common error action sets in advance, the detection system can be made more targeted; Personalization: Suitable for each user's training habits, using machine learning functions to create a more personalized training set for each user; Compatibility: The integrated design makes it very convenient to place the wiring of various sensors, display screens, switches, etc. inside various auxiliary training devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1 A flow chart of a method for correcting and training movements for home fitness in one embodiment;

[0042] Figure 2 A schematic diagram of a distance calculation algorithm for a method for correcting and training movements for home fitness in one embodiment;

[0043] Figure 3 A schematic diagram of a home fitness movement correction and training system in one embodiment;

[0044] Figure 4 A schematic diagram of a movement correction and training device for home fitness in one embodiment;

[0045] Figure 5 A flow chart of the training part of a method for correcting and training movements for home fitness in one embodiment;

[0046] Figure 6 The present invention is a flowchart of the testing part of a method for correcting and training movements for home fitness in one embodiment.

[0047] Reference numerals:

[0048] 300, sensor; 301, acceleration sensor; 302, photosensor; 303, ultrasonic sensor; 310, control mainboard; 320, display screen; 330, prompt device; 400, prompt module; 410, preset action category module; 420, known classification variable module; 430, user action information module; 440, unknown classification variable module; 450, distance sorting module; 460, sample selection module; 470, predicted action category module; 480, predicted action category output module; 490, classification statistics output module. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] It should be noted that when a component is referred to as being "fixed to" or "disposed on" another component, it may be directly on the other component or there may be a central component. When a component is considered to be "connected to" another component, it may be directly connected to the other component or there may be a central component at the same time. The terms "vertical", "horizontal", "upper", "lower", "left", "right" and similar expressions used in the specification of the present invention are for illustrative purposes only and do not represent the only implementation method.

[0051] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0052] In the present invention, unless otherwise clearly specified and limited, a first feature being “above” or “below” a second feature may mean that the first feature is directly in contact with the second feature, or the first feature and the second feature are indirectly in contact through an intermediate medium. Moreover, a first feature being “above”, “above” or “above” a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being “below”, “below” or “below” a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0053] Unless otherwise defined, all technical and scientific terms used in the specification of the present invention have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used in the specification of the present invention includes any and all combinations of one or more related listed items.

[0054] like Figure 1 As shown, in one embodiment, the present invention provides a method for correcting and training movements for home fitness, comprising the following steps:

[0055] S110: Classifying the user's actions into a correct action category and an incorrect action category according to preset user action categories;

[0056] S120: storing the correct action category and the incorrect action category as known classification variables;

[0057] S130: receiving user motion information, where the user motion information includes acceleration information and its changes, and / or distance information and its changes, and / or photosensitivity information and its changes;

[0058] S140: storing the received user action information as an unknown classification variable;

[0059] S150: Calculate the distance between the unknown categorical variable and the known categorical variable, and sort them in ascending order of distance;

[0060] S160: Select K points with the smallest distance from the unknown classification variable, and determine the number of occurrences of the categories to which the first K points belong;

[0061] S170: taking the category with the highest number of occurrences of the first K points as the predicted action category of the unknown classification variable;

[0062] S180: Output the predicted action category of the unknown classification variable;

[0063] S190: Classify and count each group of user action information trained by the user, and output the number of correct action categories and / or the number of incorrect action categories.

[0064] By pre-storing correct action categories and common incorrect action categories, and then attributing the training actions during user use to a category to which the majority of the k most adjacent samples in the feature space belong, it is possible to determine whether the user's exercise actions are correct, and output correct or incorrect results. At the same time, the number of correct actions and incorrect actions are counted separately, which facilitates the user's action correction during home exercise and improves the exercise effect.

[0065] In this embodiment, the following steps are also included: when the predicted action category of the unknown classification variable is an incorrect action category, a prompt message is output to conveniently warn the user of the standardization of the action.

[0066] In this embodiment, the distance calculation formula between the unknown classification variable and the known classification variable is as follows:

[0067]

[0068] Among them, d ijrepresents the distance between the unknown categorical variable i and the known categorical variable j, q represents the order, q=1 or 2 or +∞, k represents the number of samples selected with the minimum distance to the unknown categorical variable, k=1,2,3,…p.

[0069] See also Figure 2 , the five-pointed star represents the unknown classification variable, the triangle represents the correct action category, and the square represents the wrong action category. When p=3, it means that there are three samples closest to the unknown classification variable, of which there are two correct action categories and one wrong action category, so the unknown classification variable is the correct action category; when p=6, it means that there are six samples closest to the unknown classification variable, of which there are two correct action categories and four wrong action categories, so the unknown classification variable is the wrong action category.

[0070] like Figure 4 As shown, in one embodiment, the present invention provides a motion correction and training device for home fitness, including a preset action category module 410, a known classification variable module 420, a user action information module 430, an unknown classification variable module 440, a distance sorting module 450, a sample selection module 460, a predicted action category module 470, a predicted action category output module 480, and a classification statistics output module 490.

[0071] The preset action category module 410 is used to classify the user's actions into a correct action category and an incorrect action category according to preset user action categories.

[0072] The known classification variable module 420 is used to store the correct action category and the incorrect action category as known classification variables.

[0073] The user motion information module 430 is used to receive user motion information, and the user motion information module 430 includes acceleration information and its changes, and / or distance information and its changes, and / or photosensitivity information and its changes.

[0074] The unknown categorical variable module 440 is used to store the received user action information as an unknown categorical variable.

[0075] The distance sorting module 450 is used to calculate the distance between the unknown categorical variables and the known categorical variables, and sort them in ascending order of distance.

[0076] The sample selection module 460 is used to select K points with the smallest distance from the unknown classification variable and determine the number of occurrences of the categories to which the first K points belong.

[0077] The predicted action category module 470 is used to use the category with the highest number of occurrences of the first K points as the predicted action category of the unknown classification variable.

[0078] The predicted action category output module 480 is used to output the predicted action category of the unknown classification variable.

[0079] The classification statistics output module 490 is used to classify and count each group of user action information in user training, and output the number of correct action categories and / or the number of incorrect action categories.

[0080] In this embodiment, a prompt module 400 is further included, and the prompt module 400 is used to output prompt information when the predicted action category module is an incorrect action category.

[0081] In this embodiment, the prompt module 400 is a vibration prompt module or a voice prompt module.

[0082] like Figure 3 As shown, in one embodiment, the present invention provides a home fitness movement correction and training system, including a sensor 300, a control mainboard 310, a display screen 320 and a prompt device 330.

[0083] The sensor 300 is used to detect user motion information. The sensor 300 includes an acceleration sensor 301, a photosensor 302 and an ultrasonic sensor 303. The photosensor 302 and the ultrasonic sensor 303 are both installed on the fitness equipment. The acceleration sensor 301 is worn on the user. The acceleration sensor 301 is connected to the control motherboard 310 via a wireless network.

[0084] The control motherboard 310 is used to receive data collected by the sensor 300, and identify and determine the correct category and incorrect category of the user's action, and output the number of correct categories and / or the number of incorrect categories. The control motherboard 310 uses the Arduino 101 development board equipped with a machine learning Curie module.

[0085] The display screen 320 is connected to the control mainboard 310 , and the display screen 320 is used to display the category and number of user actions.

[0086] When the control mainboard 310 determines that the user action is of an incorrect type, the prompt device 330 outputs prompt information.

[0087] In one embodiment, the prompt device 330 is a vibration prompt device, which is worn on the user, and is connected to the control mainboard via a wireless network.

[0088] In one embodiment, the prompt device 330 is a voice prompt device, and the voice prompt device is connected to the control mainboard.

[0089] The home fitness movement correction and training system includes two modes: a training learning mode and a judgment test mode.

[0090] See also Figure 5 , the training learning mode includes the following steps:

[0091] Step 1: Initialize the settings first, display the correct training group to start, and the user presses the button to start training. If the button is not pressed, remind the user to press the button. If the button has been pressed, output "i = 0";

[0092] Step 2: Display "i+1", then display "3, 2, 1 start", the sensor collects the number of samples, the collection stops, and outputs "i=i+1";

[0093] Step 3: When i is less than the number of trainings, return to step 2. When i is equal to the number of trainings, the training of this group ends.

[0094] See also Figure 6 , determining the test mode includes the following steps:

[0095] Step 1: Collect data first, record the data in neurons, classify the collected data, and store the category number in the category (data classification) variable;

[0096] Step 2: When the category variable is "1" and the data measured by the ultrasonic sensor 303 and the photosensitive sensor 302 are within the specified range, "correct" is displayed and the total number of correct categories is output; when the category variable is not "1" and the category variable is "32767", "unrecognizable" is displayed and the total number of unrecognizable categories is output; when the category variable is not "1" and the category variable is not "32767", "error" is displayed and the total number of error categories is output.

[0097] Application scenarios:

[0098] When doing push-ups on a yoga mat, two photosensitive sensors 302 and two ultrasonic sensors 303 are symmetrically embedded in the yoga mat, and two acceleration sensors 301 are worn on the user's arms respectively. When a vibration prompt device is selected, the vibration prompt device can be worn on the user. When a voice prompt device is selected, it is only necessary to connect the voice prompt device to the control motherboard 310 through a wireless network, or the voice prompt device and the control motherboard 310 can be integrated into one.

[0099] The acceleration sensor 301 collects motion data, and after filtering, deduplication, compression, and packaging, sends it to the Curie module for learning, identifying and judging whether the user's shoulder movements are standard, whether there is shrugging, excessive range of motion of the shoulder blades, or internal rotation of the shoulders. The auxiliary judgment device uses the ultrasonic sensor 303 to judge whether the user has problems with sticking out the buttocks or sagging the waist, and uses two photosensitive sensors 302 to identify the placement direction of the user's palms to avoid sports injuries caused by excessive internal rotation of the palms.

[0100] The whole system is divided into two stages: training and learning, and testing. In the training stage, the user first performs standard push-ups under the guidance of others. After recognition and learning, the Curie module will remember this action. The user then presses a button to end the training of this state, and performs incorrect push-ups and trains again. After this process is completed, the 128 neurons of Arduino101 store the data of correct and incorrect actions, and each piece of data has a label of the category to which it belongs. The above process is a more ideal situation. In actual operation, considering that it is unlikely that the user will make incorrect actions or standard actions continuously, it can also be classified after performing push-ups.

[0101] Next, in the test phase, the user will do a push-up alone, and the Curie module on Arduino101 can classify the user's actions as correct or incorrect. This judgment result will be combined with the result of the auxiliary judgment device, and finally the display screen 320 will display whether the user's action is standard. After this process matures, the functions of the system will be further refined, and the two categories of correct and incorrect can be refined into multiple categories such as "correct", "shoulder internal rotation", "palm internal rotation", "excessive scapular movement", "shrugging", "collapsed waist", "buttocks", etc., to provide more detailed guidance for users' exercise.

[0102] In addition to the recognition and judgment functions, the system also has a push-up counting function, which can count the number of correct and incorrect push-ups respectively. Finally, the user can control the system and obtain feedback information through a display screen 320 and two buttons.

[0103] The home fitness movement correction and training system is based on the phenomenon that there are often wrong movements in push-up exercises, and attempts to design a system for movement recognition and error correction using machine learning. The main functional modules in the system have machine learning functions, which can classify and identify several common wrong movements that are difficult to judge with the naked eye, and achieve a more targeted training effect. The other sensors 300 in the system are auxiliary judgment modules, which identify some wrong movements that are easier to judge and do not require machine learning functions to classify, such as wrong hand shapes, collapsed waists and buttocks, etc. When in use, the system can achieve the effect of judging the right and wrong of each push-up movement of the user during training and classifying the wrong movements by learning the data of the user's correct movements and different wrong movements, so as to achieve higher targeting. In addition, the system also has an automatic counting function, which can give the number of correct and wrong movements respectively. All information is presented to the user through a display screen 320, which is more centralized and concise.

[0104] The present invention is not limited to application on yoga mats, but can be used on other fitness equipment. The specific application mode on other fitness equipment is the same as the application mode on yoga mats, which will not be repeated here.

[0105] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0106] The above-described embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the attached claims.

Claims

1. A method of movement correction and training for home fitness, It is characterized in that include: Classifying the user's actions into a correct action category and an incorrect action category according to preset user action categories; Store the correct action categories and incorrect action categories as known categorical variables; Receiving user motion information, the user motion information including acceleration information and its changes, and / or distance information and its changes, and / or photosensitivity information and its changes; The received user action information is stored as an unknown categorical variable; Calculate the distance between the unknown categorical variable and the known categorical variable, and sort them in increasing order of distance; Select the K points with the smallest distance to the unknown classification variable and determine the number of occurrences of the categories of the first K points; The category with the highest number of occurrences in the first K points is used as the predicted action category of the unknown categorical variable; Output the predicted action category of the unknown categorical variable; Classify and count each group of user action information trained by the user, and output the number of correct action categories and / or the number of incorrect action categories.

2. The method for correcting and training movements for home fitness according to claim 1, It is characterized in that The method also includes the following steps: when the predicted action category of the unknown classification variable is an incorrect action category, outputting a prompt message.

3. The method for correcting and training movements for home fitness according to claim 1 or 2, It is characterized in that The distance calculation formula between the unknown classification variable and the known classification variable is as follows: Among them, d ij It represents the distance between the unknown categorical variable i and the known categorical variable j, q represents the order, q=1 or 2 or +∞, and k represents the number of samples selected with the minimum distance to the unknown categorical variable.

4. A movement correction and training device for home fitness, It is characterized in that include: A preset action category module, used to classify the user's actions into a correct action category and an incorrect action category according to preset user action categories; A known categorical variable module, used for storing correct action categories and incorrect action categories as known categorical variables; A user motion information module, used to receive user motion information, including acceleration information and its changes, and / or distance information and its changes, and / or photosensitivity information and its changes; An unknown categorical variable module, used for storing received user action information as an unknown categorical variable; The distance sorting module is used to calculate the distance between the unknown categorical variables and the known categorical variables and sort them in ascending order of distance; The sample selection module is used to select the K points with the smallest distance to the unknown classification variable and determine the number of occurrences of the category of the first K points; The module for predicting action categories is used to take the category with the highest number of occurrences of the first K points as the predicted action category of the unknown classification variable; A predicted action category output module is used to output the predicted action category of the unknown classification variable; The classification statistics output module is used to classify and count each group of user action information trained by the user, and output the number of correct action categories and / or the number of incorrect action categories.

5. The movement correction and training device for home fitness according to claim 4, It is characterized in that It also includes a prompt module, which is used to output prompt information when the action category prediction module is an incorrect action category.

6. The movement correction and training device for home fitness according to claim 5, It is characterized in that The prompt module is a vibration prompt module or a voice prompt module.

7. A movement correction and training system for home fitness, It is characterized in that include: A sensor for detecting user motion information, including an acceleration sensor, a photosensor and an ultrasonic sensor, wherein the photosensor and the ultrasonic sensor are both installed on the fitness equipment, and the acceleration sensor is worn on the user; A control mainboard, used to receive data collected by the sensor, identify and judge the correct category and the wrong category of the user's action, and output the number of correct categories and / or the number of wrong categories; The display screen is connected to the control mainboard and is used to display the categories and numbers of user actions.

8. The home fitness movement correction and training system according to claim 7, It is characterized in that It also includes a prompt device, which outputs prompt information when the control mainboard determines that the user action is of an incorrect category.

9. The home fitness movement correction and training system according to claim 8, It is characterized in that The prompting device is a vibration prompting device, which is worn on the user, and is connected to the control mainboard via a wireless network.

10. The home fitness movement correction and training system according to claim 8, It is characterized in that The prompt device is a voice prompt device, and the voice prompt device is connected to the control mainboard.

Images