Real-time rest decision-making method in sports wireless teaching based on Actor-Critic network
By using the Actor-Critic network in physical education to monitor students' physiological data and movement status in real time, intelligently remind students to rest, solving the problem of lack of real-time security monitoring in the existing technology and improving students' exercise safety.
Patent Information
- Application Number
- CN202311186579.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-14
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2043-09-14
AI Technical Summary
The lack of effective real-time student status monitoring and safety protection in existing physical education will lead to timely detection of safety hazards such as discomfort in exercise intensity or excessive fatigue, which may lead to safety accidents.
Using the real-time rest decision-making method based on the Actor-Critic network, students' physical performance data are collected in advance in physical education and physiological data are collected in real time during exercise. The Actor and Critic networks are used to determine whether they need rest, and remind them through the wireless network. Students stop exercising.
It has achieved intelligent provision of rest suggestions based on the physiological characteristics and exercise status of different students, adaptively guarantees students' exercise safety, and improves safety in physical education teaching.
Smart Images

Figure CN117787731B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of sports teaching safety, and in particular to a real-time rest decision-making method in sports wireless teaching based on an Actor-Critic network. Background Art
[0002] Physical education plays an important role in school education. It is not only about cultivating and improving students' physical fitness, but also a part of their all-round development. In physical education courses, students participate in various sports activities, such as running, jumping, ball games, etc. These activities have a positive impact on students' physical, psychological and social development. However, physical activities are also accompanied by certain risks, such as excessive fatigue, heat stroke, and physical exhaustion, which may affect students' health and safety;
[0003] However, the current safety protection for students in physical education is still relatively lacking, which is manifested in: inappropriate exercise intensity and training plans or lack of real-time monitoring of student status, so that students are often not found until they express discomfort or even safety accidents occur, and safety hazards cannot be discovered in time to remind students to rest in advance;
[0004] The Chinese patent application with the publication number CN103611234A proposes a method for monitoring sports physiological data, which includes inputting personal information, including gender, age, height, weight, and running speed, through a control panel arranged on the front side of a treadmill; putting a monitoring bracelet on the user's wrist, the monitoring bracelet includes a heart rate monitoring module, a blood pressure monitoring module, and a temperature monitoring module, monitoring the user's physiological data, and uploading the physiological data through a Bluetooth module; receiving the physiological data sent by the monitoring bracelet through a Bluetooth receiving module, analyzing it through analysis software pre-stored in a controller, outputting the physiological status and exercise suggestions to a display, and forcibly controlling the motor speed to a safe speed when the user's physiological data reaches a warning value; however, this method does not take into account the problem that different users have different physiological performances and that the exercise status is different at different times;
[0005] To this end, the present invention proposes a real-time rest decision-making method in sports wireless teaching based on Actor-Critic network. Summary of the invention
[0006] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes a real-time rest decision-making method in sports wireless teaching based on an Actor-Critic network, which intelligently provides rest suggestions for students and adaptively ensures students' sports safety.
[0007] To achieve the above object, according to Embodiment 1 of the present invention, a real-time rest decision method in sports wireless teaching based on an Actor-Critic network is proposed, comprising the following steps:
[0008] Step 1: Before physical education begins, collect each student’s physical performance data in advance;
[0009] Step 2: During physical education teaching, identify each student and the corresponding exercise type, and collect a real-time physiological data sequence set of each student through a wireless network;
[0010] Step 3: Based on the real-time physiological data sequence set of each student, determine whether there are students who need to make a rest decision; if it is determined that there are students who need to make a rest decision, go to step 4;
[0011] Step 4: Based on the real-time physiological data sequence set and sports performance data of the corresponding student, use the Actor network in the Actor-Critic network model to output the decision result of whether the student needs to rest, and train the Actor-Critic network model;
[0012] Step 5: If the decision result is that the student needs to rest, the student is reminded to stop exercising through the wireless network;
[0013] Step 6: Update sports performance data;
[0014] Wherein, the sports performance data of each student includes a set of sports data of each sports type;
[0015] Label the student's number as i and the type of sport as s;
[0016] For the i-th student, the sports data set corresponding to the s-th sports type includes the sports data of the student each time he / she performed the sports type in the past physical education;
[0017] The exercise data includes a set of physiological data sequences during each corresponding exercise type;
[0018] The physiological data series set includes a heart rate time series, a respiratory rate time series and a body temperature time series;
[0019] The heart rate time series is a sequence consisting of heart rate values per unit time monitored in real time by a wearable device worn by the student;
[0020] The respiratory frequency time series is a sequence consisting of respiratory frequency values per unit time monitored in real time by a wearable device worn by the student;
[0021] The body temperature time series is a sequence consisting of body temperatures per unit time monitored in real time by a wearable device worn by the student;
[0022] The way to identify each student and the corresponding sport type is:
[0023] A positioning device is installed in the wearable device worn by each student, and the positioning device sends the number and real-time location of each student to the data processing background in real time;
[0024] N image capture devices are installed in the physical education teaching place, and each image capture device sends the real-time captured image to the data processing background; wherein N is the number of preset image capture devices;
[0025] The data processing backend uses the target recognition algorithm to identify the students in the real-time images captured by each image capture device, and then uses the action recognition algorithm to identify the movement type of each student;
[0026] The data processing background obtains the geographic location coordinates of each student in the corresponding real-time image according to the location and shooting angle of each image capture device;
[0027] The data processing background matches the real-time position sent by the wearable device with the geographical coordinates of the student obtained after real-time image processing; if the distance between the real-time position sent by the wearable device and the geographical coordinates of the student sent by the image capture device is less than a preset distance threshold, the movement type performed by the student corresponding to the wearable device is set to the corresponding movement type, and the corresponding movement type is the movement type of the student identified by the movement recognition algorithm in the real-time image;
[0028] The real-time physiological data series collection is collected in the following way:
[0029] Set the real-time heart rate time series, real-time respiratory rate time series and real-time body temperature time series, and the real-time heart rate time series, real-time respiratory rate time series and real-time body temperature time series are all empty initially;
[0030] The wearable device worn by each student sends the student's real-time heart rate, real-time breathing rate and real-time body temperature per unit time to the data processing background in real time via wireless mode;
[0031] For each student, when the data processing background identifies that the student is doing any of all types of exercise, the real-time heart rate, real-time respiratory rate and real-time body temperature are added to the real-time heart rate time series, the real-time respiratory rate time series and the real-time body temperature time series respectively;
[0032] The real-time physiological data series set includes a real-time heart rate time series, a real-time respiratory rate time series, and a real-time body temperature time series;
[0033] The method for determining whether there are students who need to make a break decision is:
[0034] Preset heart rate threshold, respiratory rate threshold and body temperature threshold;
[0035] If any student's real-time heart rate is greater than the heart rate threshold, real-time respiratory rate is greater than the respiratory rate threshold, or real-time body temperature is greater than the body temperature threshold, it is determined that the student needs to make a rest decision;
[0036] The Actor network in the Actor-Critic network model is used to output the decision result of whether the student needs a break, and the Actor-Critic network model is trained as follows:
[0037] Set up an Actor-Critic network model for each student and initialize the parameters of the Actor network and the Critic network;
[0038] When it is determined that a rest decision is needed, perform the following steps:
[0039] Step 11: mark the number of the student who needs to make a rest decision as i0, mark the type of exercise performed by the i0th student as s0; take the real-time heart rate, real-time breathing rate and real-time body temperature of the i0th student at the current moment as the current state;
[0040] Step 12: The Actor network outputs the decision result on whether to take a break;
[0041] The real-time heart rate, real-time respiratory rate and real-time body temperature of the i0th student in the next unit time after the rest decision is made are taken as the next state;
[0042] Step 13: Calculate the actual reward value Q; the actual reward value Q is the reward value obtained after making a rest decision for the student each time;
[0043] When the decision is that no rest is needed:
[0044] Calculate the historical physiological similarity X, and set the actual reward value Q to X;
[0045] The calculation method of the historical physiological similarity X is:
[0046] The real-time heart rate of the i0th student in the current state is marked as L, the real-time breathing rate in the current state is marked as H, and the real-time body temperature in the current state is marked as T;
[0047] For the real-time physiological data sequence set of the i0th student, the length of any sequence is used as the first exercise duration of the real-time physiological data sequence set, and the first exercise duration is marked as C;
[0048] From the sports performance data, read the sports data of the s0th sports type of the i0th student, mark the set composed of all physiological data sequence sets in the sports data as K, and mark the number of the physiological data sequence set in the sports data as k, k=1,2,3,…,|K|; for the kth physiological data sequence set, take the length of any sequence in it as the second sports duration of the physiological data sequence set, and mark the second sports duration as Ck;
[0049] From all physiological data sequence sets of the motion data, a physiological data sequence set whose second motion duration is greater than or equal to the first motion duration is screened out; a set consisting of the screened physiological data sequence sets is marked as K0, wherein the number of each physiological data sequence set is marked as k0; in the k0-th physiological data sequence set, the C-th heart rate value of the heart rate time series is marked as Lk0, the C-th respiratory frequency value of the respiratory frequency time series is marked as Hk0, and the C-th body temperature in the body temperature time series is marked as Tk0;
[0050] Then the calculation formula of historical physiological similarity X is:
[0051] ; Wherein, b2, b3 and b4 are respectively preset proportional coefficients greater than 0, and b1 is a preset bias coefficient greater than 0, which is used to ensure that X is greater than 0;
[0052] When the decision is to take a break:
[0053] The real-time heart rate of the i0th student in the next state is marked as L1, the real-time breathing rate in the next state is marked as H1, and the real-time body temperature in the next state is marked as T1;
[0054] The actual reward value Q is calculated as:
[0055] ;
[0056] Among them, b5 is the preset proportional coefficient;
[0057] Step 14: Use the update formula of the Critic network to update the value of the reward function to adjust the estimate of the actual reward value Q of the decision result;
[0058] Step 15: Update the parameters of the Actor network using the Actor network’s update formula to increase the probability of selecting a high-reward decision outcome in a given state.
[0059] The way to update sports performance data is:
[0060] When the data processing background determines that the i-th student stops performing the corresponding exercise type, the collection of the real-time physiological data sequence set is stopped;
[0061] A historical exercise number threshold is preset; for the exercise data of the exercise type corresponding to the i-th student, if the number of physiological data sequence sets is less than the historical exercise number threshold, the real-time physiological data sequence set is added to the exercise data;
[0062] If the number of physiological data sequence sets is greater than or equal to the historical exercise number threshold, the first real-time physiological data sequence set in the exercise data is filtered out, and the real-time physiological data sequence set is added to the exercise data.
[0063] According to the embodiment 2 of the present invention, the real-time rest decision system in sports wireless teaching based on Actor-Critic network includes a performance data collection module, a real-time data collection module and a rest decision module; wherein each module is electrically connected;
[0064] The performance data collection module collects the sports performance data of each student in advance before the physical education teaching begins, and sends the sports performance data to the rest decision module;
[0065] A real-time data collection module, during physical education teaching, identifies each student and the corresponding exercise type, collects each student's real-time physiological data sequence set through a wireless network, and sends each student's exercise type and each student's real-time physiological data sequence set to a rest decision module;
[0066] The rest decision module determines whether there are students who need to make a rest decision based on the real-time physiological data sequence set of each student. If it is determined that there are students who need to make a rest decision, based on the corresponding student's real-time physiological data sequence set and sports performance data, the Actor network in the Actor-Critic network model is used to output the decision result of whether the student needs to rest, and the Actor-Critic network model is trained. If the decision result is that rest is needed, the student is reminded to stop exercising through the wireless network, and finally the sports performance data is updated.
[0067] An electronic device according to Embodiment 3 of the present invention includes: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;
[0068] The processor executes the above-mentioned real-time rest decision-making method in sports wireless teaching based on Actor-Critic network by calling the computer program stored in the memory.
[0069] A computer-readable storage medium according to embodiment 4 of the present invention stores a rewritable computer program;
[0070] When the computer program is executed on a computer device, the computer device is enabled to execute the above-mentioned real-time rest decision-making method in sports wireless teaching based on the Actor-Critic network.
[0071] Compared with the prior art, the present invention has the following beneficial effects:
[0072] The present invention collects the sports performance data of each student in advance before the start of physical education teaching, identifies each student and the corresponding exercise type during physical education teaching, collects the real-time physiological data sequence set of each student through a wireless network, judges whether there is a student who needs to make a rest decision based on the real-time physiological data sequence set of each student, and when it is judged that there is a student who needs to make a rest decision, based on the real-time physiological data sequence set and sports performance data of the corresponding student, uses the Actor network in the Actor-Critic network model to output the decision result of whether the student needs to rest, and trains the Actor-Critic network model. When the decision result is that rest is needed, the student is reminded to stop exercising through the wireless network, and finally the sports performance data is updated; rest suggestions are intelligently provided to students according to the performance of physiological characteristics of different students during physical education teaching, and the sports safety of students is adaptively guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS
[0073] Figure 1 It is a flow chart of a real-time rest decision-making method in sports wireless teaching based on Actor-Critic network in Embodiment 1 of the present invention;
[0074] Figure 2 A module connection relationship diagram of a real-time rest decision-making system in sports wireless teaching based on an Actor-Critic network in Embodiment 2 of the present invention;
[0075] Figure 3 This is a schematic diagram of the structure of an electronic device in Embodiment 3 of the present invention;
[0076] Figure 4 This is a schematic diagram of the computer-readable storage medium structure in Example 4 of the present invention. DETAILED DESCRIPTION
[0077] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than 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.
[0078] Example 1
[0079] like Figure 1 As shown, the real-time rest decision-making method in sports wireless teaching based on Actor-Critic network is used for data processing background, including the following steps:
[0080] Step 1: Before physical education begins, collect each student’s physical performance data in advance;
[0081] Step 2: During physical education teaching, identify each student and the corresponding exercise type, and collect a real-time physiological data sequence set of each student through a wireless network;
[0082] Step 3: Based on the real-time physiological data sequence set of each student, determine whether there is a student who needs to make a rest decision. If it is determined that there is no student who needs to make a rest decision, continue to step 3; if it is determined that there is a student who needs to make a rest decision, go to step 4;
[0083] Step 4: Based on the real-time physiological data sequence set and sports performance data of the corresponding student, use the Actor network in the Actor-Critic network model to output the decision result of whether the student needs to rest, and train the Actor-Critic network model;
[0084] Step 5: If the decision result is that a rest is needed, the student is reminded to stop exercising through the wireless network; if the decision result is that a rest is not needed, then step 3 is continued;
[0085] Step 6: Update sports performance data;
[0086] The sports performance data of each student includes a set of sports data of each sports type; it should be noted that the sports type is a sports item included in physical education, and the specific sports type is determined by the education-related institution, including but not limited to running, long jump, high jump, basketball, etc.;
[0087] Preferably, the student number is marked as i, and the sport type number is marked as s;
[0088] Furthermore, for the i-th student, the sports data set corresponding to the s-th sports type includes the sports data of the student each time he / she performed the sports type in the past physical education;
[0089] The exercise data includes a set of physiological data sequences during each corresponding exercise type;
[0090] Preferably, the physiological data series set includes a heart rate time series, a respiratory rate time series and a body temperature time series;
[0091] The heart rate time series is a sequence consisting of heart rate values per unit time monitored in real time by a wearable device worn by the student;
[0092] The respiratory frequency time series is a sequence consisting of respiratory frequency values per unit time monitored in real time by a wearable device worn by the student;
[0093] The body temperature time series is a sequence consisting of body temperatures per unit time monitored in real time by a wearable device worn by the student;
[0094] It is understandable that wearable devices monitoring human heart rate, respiratory rate and body temperature are conventional technical means in the art. For example, smart bracelets or smart watches can achieve the above monitoring functions, and the present invention will not be repeated here.
[0095] The way to identify each student and the corresponding sport type is:
[0096] A positioning device is installed in the wearable device worn by each student, and the positioning device sends the number and real-time position of each student to the data processing background in real time; preferably, the real-time position can be a geographical location coordinate, and the geographical location coordinate can be a longitude and latitude coordinate, or a spatial coordinate in a set of coordinate systems independently constructed using surveying and mapping related technologies in a physical education teaching place;
[0097] N image capture devices are installed in the physical education teaching place, and each image capture device sends the real-time captured image to the data processing background; wherein N is the number of preset image capture devices, and the specific installation position of each image capture device is determined according to the actual building layout or structural diagram of the physical education teaching place, so as to ensure that the image capture device can cover all possible movement positions of students in the physical education teaching place;
[0098] The data processing background uses a target recognition algorithm to identify students in the real-time images captured by each image capture device, and then uses a motion recognition algorithm to identify the type of movement of each student; it should be noted that the target recognition algorithm to identify the human body belongs to conventional technical means in this field, and the motion recognition algorithm to identify the type of movement of students also belongs to conventional technical means in this field, and the present invention will not be repeated here;
[0099] The data processing background obtains the geographic location coordinates of each student in the corresponding real-time image according to the position and shooting angle of each image capture device; it should be noted that calculating the geographic coordinate position of the target in the video based on the surveillance video belongs to the conventional technical means in this field. For example, the paper "Spatial Positioning Method of Dynamic Targets in Surveillance Videos" ("Bulletin of Surveying and Mapping" 2022 Issue 8, Pages 87-92) aims at the spatial positioning problem of dynamic targets in surveillance videos. Under the premise of considering camera distortion, the mapping relationship between surveillance cameras, target pixel coordinates and geographic scenes is studied, and a target positioning algorithm based on digital surface model (DSM) and plane constraint is proposed. First, complete the calibration of the camera and determine the imaging model of the camera; then, extract the target pixel coordinates after distortion correction, and calculate the three-dimensional geographic coordinates of the target through the target positioning algorithm;
[0100] The data processing background matches the real-time position sent by the wearable device with the geographical location coordinates of the student obtained after real-time image processing; specifically, if the distance between the real-time position sent by the wearable device and the geographical location coordinates of the student sent by the image capture device is less than a preset distance threshold, the movement type performed by the student corresponding to the wearable device is set to the corresponding movement type, and the corresponding movement type is the movement type of the student identified by the movement recognition algorithm in the real-time image;
[0101] The real-time physiological data series collection is collected in the following way:
[0102] Set the real-time heart rate time series, real-time respiratory rate time series and real-time body temperature time series, and the real-time heart rate time series, real-time respiratory rate time series and real-time body temperature time series are all empty initially;
[0103] The wearable device worn by each student sends the student's real-time heart rate, real-time breathing rate and real-time body temperature per unit time to the data processing background in real time via wireless mode;
[0104] For each student, when the data processing background identifies that the student is doing any of all types of exercise, the real-time heart rate, real-time respiratory rate and real-time body temperature are added to the real-time heart rate time series, the real-time respiratory rate time series and the real-time body temperature time series respectively;
[0105] The real-time physiological data series set includes real-time heart rate time series, real-time respiratory rate time series and real-time body temperature time series; thereby obtaining the physiological data series of students from the beginning of exercise, reducing the data monitoring of students when they are not exercising, and improving data processing efficiency;
[0106] The method for determining whether there are students who need to make a break decision is:
[0107] Preset heart rate threshold, respiratory rate threshold and body temperature threshold;
[0108] If any student's real-time heart rate is greater than the heart rate threshold, real-time respiratory rate is greater than the respiratory rate threshold, or real-time body temperature is greater than the body temperature threshold, it is determined that the student needs to make a rest decision;
[0109] If the real-time heart rate of all students is less than or equal to the heart rate threshold, the real-time respiratory rate is less than or equal to the respiratory rate threshold, and the real-time body temperature is less than or equal to the body temperature threshold, it is determined that no student needs to make a rest decision;
[0110] The Actor network in the Actor-Critic network model is used to output the decision result of whether the student needs a break, and the Actor-Critic network model is trained as follows:
[0111] Set up an Actor-Critic network model for each student and initialize the parameters of the Actor network and the Critic network; the parameters include but are not limited to the dimensions of the state input layer of the Actor network, the number and size of hidden layers, the dimension of the action output layer, the dimensions of the state input layer of the Critic network, the number and size of hidden layers, the dimension of the reward value function output layer, the learning rate, the discount factor, and the network optimization algorithm (gradient descent method or Adam optimization algorithm, etc.);
[0112] When it is determined that a rest decision is needed, perform the following steps:
[0113] Step 11: mark the number of the student who needs to make a rest decision as i0, mark the type of exercise performed by the i0th student as s0; take the real-time heart rate, real-time breathing rate and real-time body temperature of the i0th student at the current moment as the current state;
[0114] Step 12: The Actor network outputs the decision result on whether to take a break;
[0115] The real-time heart rate, real-time respiratory rate and real-time body temperature of the i0th student in the next unit time after the rest decision is made are taken as the next state;
[0116] Step 13: Calculate the actual reward value Q; the actual reward value Q is the reward value obtained after making a rest decision for the student each time;
[0117] Specifically, when the decision result is that rest is not needed:
[0118] Calculate the historical physiological similarity X, and set the actual reward value Q to X;
[0119] The calculation method of the historical physiological similarity X is:
[0120] The real-time heart rate of the i0th student in the current state is marked as L, the real-time breathing rate in the current state is marked as H, and the real-time body temperature in the current state is marked as T;
[0121] For the real-time physiological data sequence set of the i0th student, the length of any sequence is used as the first exercise duration of the real-time physiological data sequence set, and the first exercise duration is marked as C;
[0122] From the sports performance data, read the sports data of the s0th sports type of the i0th student, mark the set composed of all physiological data sequence sets in the sports data as K, and mark the number of the physiological data sequence set in the sports data as k, k=1,2,3,…,|K|; for the kth physiological data sequence set, take the length of any sequence in it as the second sports duration of the physiological data sequence set, and mark the second sports duration as Ck;
[0123] From all physiological data sequence sets of the motion data, a physiological data sequence set whose second motion duration is greater than or equal to the first motion duration is screened out; a set consisting of the screened physiological data sequence sets is marked as K0, wherein the number of each physiological data sequence set is marked as k0; in the k0-th physiological data sequence set, the C-th heart rate value of the heart rate time series is marked as Lk0, the C-th respiratory frequency value of the respiratory frequency time series is marked as Hk0, and the C-th body temperature in the body temperature time series is marked as Tk0;
[0124] Then the calculation formula of historical physiological similarity X is:
[0125] ; Wherein, b2, b3 and b4 are respectively preset proportional coefficients greater than 0, and b1 is a preset bias coefficient greater than 0, which is used to ensure that X is greater than 0;
[0126] Understandably, It expresses the proportion of students whose exercise duration in the past has reached the first exercise duration C. The larger it is, the more the i0th student is adapted to the exercise time of the first exercise duration and the more inclined he is to not need to rest; It expresses the similarity between the physiological characteristics of the i0th student during the first exercise duration and the past exercise duration of the first exercise duration. The smaller it is, the greater the similarity is, that is, the i0th student is more adapted to the exercise intensity represented by the current physiological characteristics;
[0127] When the decision is to take a break:
[0128] The real-time heart rate of the i0th student in the next state is marked as L1, the real-time breathing rate in the next state is marked as H1, and the real-time body temperature in the next state is marked as T1;
[0129] The actual reward value Q is calculated as:
[0130] ;
[0131] Among them, b5 is the preset proportional coefficient;
[0132] It is understandable that when the real-time heart rate is larger, the real-time breathing rate is larger, or the real-time temperature is higher, it is more necessary to consider rest; and further, when the real-time heart rate changes more, the real-time breathing rate changes more, or the real-time temperature changes more, it means that the physiological characteristics are unstable, and rest is also recommended;
[0133] Step 14: Use the update formula of the Critic network to update the value of the reward value function to adjust the estimate of the actual reward value Q of the decision result; it should be noted that the update formula can be a common update formula used by those skilled in the art, for example: ,in, is the reward value function estimate of the current state a; is the learning rate, which controls the step size of the update; is a discount factor that measures the importance of future rewards; is the next state;
[0134] Step 15: Update the parameters of the Actor network using the Actor network’s update formula to increase the probability of selecting a high-reward decision outcome in a given state.
[0135] The method of reminding the student to stop exercising through the wireless network may be that the data processing background sends a suggestion to stop exercising to the student's wearable device through the wireless network, and the wearable device broadcasts the suggestion to stop exercising in a voice manner or uses vibration to remind the student;
[0136] The way to update sports performance data is:
[0137] When the data processing background determines that the i-th student stops performing the corresponding exercise type, the collection of the real-time physiological data sequence set is stopped;
[0138] A historical exercise number threshold is preset; for the exercise data of the exercise type corresponding to the i-th student, if the number of physiological data sequence sets is less than the historical exercise number threshold, the real-time physiological data sequence set is added to the exercise data;
[0139] If the number of physiological data sequence sets is greater than or equal to the historical exercise number threshold, the first real-time physiological data sequence set in the exercise data is screened out, and the real-time physiological data sequence set is added to the exercise data; thereby, by controlling the number of physiological data sequence sets, the error caused by the difference between physiological characteristic data from too long ago and current physiological characteristic data due to the improvement of students' physical fitness can be reduced.
[0140] Example 2
[0141] like Figure 2 As shown, a real-time rest decision system in sports wireless teaching based on Actor-Critic network includes a performance data collection module, a real-time data collection module and a rest decision module; wherein each module is electrically connected;
[0142] The performance data collection module is mainly used to collect the sports performance data of each student in advance before the physical education teaching begins, and send the sports performance data to the rest decision module;
[0143] The real-time data collection module is mainly used to identify each student and the corresponding exercise type during physical education, collect the real-time physiological data sequence set of each student through a wireless network, and send the exercise type of each student and the real-time physiological data sequence set of each student to the rest decision module;
[0144] Among them, the rest decision module is mainly used to determine whether there are students who need to make a rest decision based on the real-time physiological data sequence set of each student. If it is determined that there are students who need to make a rest decision, based on the corresponding student's real-time physiological data sequence set and sports performance data, the Actor network in the Actor-Critic network model is used to output the decision result of whether the student needs a rest, and the Actor-Critic network model is trained. If the decision result is that a rest is needed, the student is reminded to stop exercising through the wireless network, and finally the sports performance data is updated.
[0145] Example 3
[0146] Figure 3 Schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Figure 3 As shown, according to another aspect of the present application, an electronic device 100 is also provided. The electronic device 100 may include one or more processors and one or more memories. The memories store computer readable codes, and when the computer readable codes are executed by one or more processors, the real-time rest decision method in sports wireless teaching based on the Actor-Critic network as described above can be executed.
[0147] The method or system according to the embodiment of the present application can also be used by Figure 3 The electronic device shown in the figure is implemented by the electronic device architecture. Figure 3 As shown, the electronic device 100 may include a bus 101, one or more CPUs 102, a read-only memory (ROM) 103, a random access memory (RAM) 104, a communication port 105 connected to a network, an input / output component 106, a hard disk 107, etc. The storage device in the electronic device 100, such as ROM 103 or hard disk 107, may store the real-time rest decision-making method in sports wireless teaching based on Actor-Critic network provided in the present application. A real-time rest decision method in wireless sports teaching based on an Actor-Critic network may, for example, include the following steps: Step 1: Before the start of physical education teaching, collect the sports performance data of each student in advance; Step 2: During physical education teaching, identify each student and the corresponding exercise type, and collect the real-time physiological data sequence set of each student through a wireless network; Step 3: Based on the real-time physiological data sequence set of each student, determine whether there is a student who needs to make a rest decision. If it is determined that there is no student who needs to make a rest decision, continue to execute Step 3; If it is determined that there is a student who needs to make a rest decision, go to Step 4; Step 4: Based on the real-time physiological data sequence set and sports performance data of the corresponding student, use the Actor network in the Actor-Critic network model to output the decision result of whether the student needs to rest, and train the Actor-Critic network model; Step 5: If the decision result is that rest is needed, remind the student to stop exercising through the wireless network; If the decision result is that rest is not needed, continue to execute Step 3; Step 6: Update the sports performance data;
[0148] Furthermore, the electronic device 100 may further include a user interface 108. Figure 3 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 3 One or more components of an electronic device are shown.
[0149] Example 4
[0150] Figure 4 Schematic diagram of a computer-readable storage medium structure provided by an embodiment of the present application. Figure 4As shown, a computer-readable storage medium 200 according to an embodiment of the present application is shown. Computer-readable instructions are stored on the computer-readable storage medium 200. When the computer-readable instructions are executed by the processor, the real-time rest decision-making method in sports wireless teaching based on the Actor-Critic network according to the embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium 200 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.
[0151] In addition, according to the implementation of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, the non-transitory machine-readable storage medium stores machine-readable instructions, the machine-readable instructions can be run by a processor to execute instructions corresponding to the method steps provided by the present application, and when the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.
[0152] The methods, apparatuses, and devices of the present application may be implemented in many ways. For example, the methods, apparatuses, and devices of the present application may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present application are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present application may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the method according to the present application. Therefore, the present application also covers recording media storing programs for executing the method according to the present application.
[0153] In addition, the parts of the above-mentioned technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.
[0154] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
[0155] The above preset parameters or preset thresholds are all set by technicians in this field according to actual conditions or obtained through large-scale data simulation.
[0156] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.
Claims
1. A real-time rest decision-making method in sports wireless teaching based on Actor-Critic network, characterized by: The following steps are involved: Step 1: Before physical education begins, collect each student’s physical performance data in advance; Step 2: During physical education teaching, identify each student and the corresponding exercise type, and collect a real-time physiological data sequence set of each student through a wireless network; Step 3: Based on the real-time physiological data sequence set of each student, determine whether there are students who need to make a rest decision; if it is determined that there are students who need to make a rest decision, go to step 4; Step 4: Based on the real-time physiological data sequence set and sports performance data of the corresponding student, use the Actor network in the Actor-Critic network model to output the decision result of whether the student needs to rest, and train the Actor-Critic network model; Step 5: If the decision result is that the student needs to rest, the student is reminded to stop exercising through the wireless network; Step 6: Update sports performance data; The Actor network in the Actor-Critic network model is used to output the decision result of whether the student needs a break, and the Actor-Critic network model is trained as follows: Set up an Actor-Critic network model for each student and initialize the parameters of the Actor network and the Critic network; When it is determined that a rest decision is needed, perform the following steps: Step 11: mark the number of the student who needs to make a rest decision as i0, mark the type of exercise performed by the i0th student as s0; take the real-time heart rate, real-time breathing rate and real-time body temperature of the i0th student at the current moment as the current state; Step 12: The Actor network outputs the decision result on whether to take a break; The real-time heart rate, real-time respiratory rate and real-time body temperature of the i0th student in the next unit time after the rest decision is made are taken as the next state; Step 13: Calculate the actual reward value Q; the actual reward value Q is the reward value obtained after making a rest decision for the student each time; Step 14: Use the update formula of the Critic network to update the value of the reward function to adjust the estimate of the actual reward value Q of the decision result; Step 15: Update the parameters of the Actor network using the Actor network’s update formula to increase the probability of selecting a high-reward decision outcome in a given state. The reward value Q is calculated as: When the decision is that no rest is needed: Calculate the historical physiological similarity X, and set the actual reward value Q to X; When the decision is to take a break: The real-time heart rate of the i0th student in the current state is marked as L, the real-time breathing rate in the current state is marked as H, and the real-time body temperature in the current state is marked as T; The real-time heart rate of the i0th student in the next state is marked as L1, the real-time breathing rate in the next state is marked as H1, and the real-time body temperature in the next state is marked as T1; The actual reward value Q is calculated as: ; Among them, b2, b3, b4 and b5 are respectively preset proportional coefficients greater than 0.
2. The real-time rest decision-making method in sports wireless teaching based on Actor-Critic network according to claim 1 is characterized in that: The sports performance data of each student includes a set of sports data for each sports type; Label the student's number as i and the type of sport as s; For the i-th student, the sports data set corresponding to the s-th sports type includes the sports data of the student each time he / she performed the sports type in the past physical education; The exercise data includes a set of physiological data sequences during each corresponding exercise type.
3. The real-time rest decision-making method in sports wireless teaching based on Actor-Critic network according to claim 2 is characterized in that: The physiological data series set includes a heart rate time series, a respiratory rate time series and a body temperature time series; The heart rate time series is a sequence consisting of heart rate values per unit time monitored in real time by a wearable device worn by the student; The respiratory frequency time series is a sequence consisting of respiratory frequency values per unit time monitored in real time by a wearable device worn by the student; The body temperature time series is a sequence consisting of body temperature per unit time monitored in real time by a wearable device worn by the student.
4. The real-time rest decision-making method in sports wireless teaching based on Actor-Critic network according to claim 3 is characterized in that: The way to identify each student and the corresponding sport type is: A positioning device is installed in the wearable device worn by each student, and the positioning device sends the number and real-time location of each student to the data processing background in real time; N image capture devices are installed in the physical education teaching place, and each image capture device sends the real-time captured image to the data processing background; wherein N is the number of preset image capture devices; The data processing backend uses the target recognition algorithm to identify the students in the real-time images captured by each image capture device, and then uses the action recognition algorithm to identify the movement type of each student; The data processing background obtains the geographic location coordinates of each student in the corresponding real-time image according to the location and shooting angle of each image capture device; The data processing background matches the real-time position sent by the wearable device with the geographic location coordinates of the student obtained after real-time image processing; if the distance between the real-time position sent by the wearable device and the geographic location coordinates of the student sent by the image capture device is less than a preset distance threshold, the movement type performed by the student corresponding to the wearable device is set to the corresponding movement type, and the corresponding movement type is the movement type of the student identified by the motion recognition algorithm in the real-time image.
5. The real-time rest decision-making method in sports wireless teaching based on Actor-Critic network according to claim 4 is characterized in that: The real-time physiological data series collection is collected in the following way: Set the real-time heart rate time series, real-time respiratory rate time series and real-time body temperature time series, and the real-time heart rate time series, real-time respiratory rate time series and real-time body temperature time series are all empty initially; The wearable device worn by each student sends the student's real-time heart rate, real-time breathing rate and real-time body temperature per unit time to the data processing background in real time via wireless mode; For each student, when the data processing background identifies that the student is doing any of all types of exercise, the real-time heart rate, real-time respiratory rate and real-time body temperature are added to the real-time heart rate time series, the real-time respiratory rate time series and the real-time body temperature time series respectively; The real-time physiological data series set includes a real-time heart rate time series, a real-time respiratory rate time series, and a real-time body temperature time series.
6. The real-time rest decision-making method in sports wireless teaching based on Actor-Critic network according to claim 5 is characterized in that: The method for determining whether there are students who need to make a break decision is: Preset heart rate threshold, respiratory rate threshold and body temperature threshold; If any student's real-time heart rate is greater than the heart rate threshold, real-time breathing rate is greater than the breathing rate threshold, or real-time body temperature is greater than the body temperature threshold, it is judged that the student needs to take a rest decision.
7. The real-time rest decision-making method in sports wireless teaching based on Actor-Critic network according to claim 1 is characterized in that: The calculation method of the historical physiological similarity X is: For the real-time physiological data sequence set of the i0th student, the length of any sequence is used as the first exercise duration of the real-time physiological data sequence set, and the first exercise duration is marked as C; From the sports performance data, read the sports data of the s0th sports type of the i0th student, mark the set composed of all physiological data sequence sets in the sports data as K, and mark the number of the physiological data sequence set in the sports data as k, k=1,2,3,…,|K|; for the kth physiological data sequence set, take the length of any sequence in it as the second sports duration of the physiological data sequence set, and mark the second sports duration as Ck; From all physiological data sequence sets of the motion data, a physiological data sequence set whose second motion duration is greater than or equal to the first motion duration is screened out; a set consisting of the screened physiological data sequence sets is marked as K0, wherein the number of each physiological data sequence set is marked as k0; in the k0-th physiological data sequence set, the C-th heart rate value of the heart rate time series is marked as Lk0, the C-th respiratory frequency value of the respiratory frequency time series is marked as Hk0, and the C-th body temperature in the body temperature time series is marked as Tk0; Then the calculation formula of historical physiological similarity X is: ; b1 is a preset bias coefficient greater than 0.
8. A real-time rest decision system in sports wireless teaching based on Actor-Critic network, which is implemented based on the real-time rest decision method in sports wireless teaching based on Actor-Critic network according to any one of claims 1 to 7, characterized in that: It includes a performance data collection module, a real-time data collection module and a rest decision module; wherein each module is electrically connected; A performance data collection module collects the sports performance data of each student in advance before the physical education teaching begins, and sends the sports performance data to the rest decision module; A real-time data collection module, during physical education teaching, identifies each student and the corresponding exercise type, collects each student's real-time physiological data sequence set through a wireless network, and sends each student's exercise type and each student's real-time physiological data sequence set to a rest decision module; The rest decision module determines whether there are students who need to make a rest decision based on the real-time physiological data sequence set of each student. If it is determined that there are students who need to make a rest decision, based on the corresponding student's real-time physiological data sequence set and sports performance data, the Actor network in the Actor-Critic network model is used to output the decision result of whether the student needs to rest, and the Actor-Critic network model is trained. If the decision result is that rest is needed, the student is reminded to stop exercising through the wireless network, and finally the sports performance data is updated.
9. An electronic device, characterized in that: include: A processor and a memory, wherein: The memory stores a computer program that can be called by the processor; The processor executes the real-time rest decision-making method in sports wireless teaching based on the Actor-Critic network as described in any one of claims 1 to 7 by calling the computer program stored in the memory.
10. A computer-readable storage medium, characterized in that: A rewritable computer program is stored thereon; When the computer program runs on a computer device, the computer device executes the real-time rest decision-making method in sports wireless teaching based on the Actor-Critic network as described in any one of claims 1 to 7.
