Electronic school badge and working method thereof

Abstract

The invention provides an electronic school badge and a working method thereof, and relates to the technical field of school badges. And using a first pose data prediction model to predict first pose data of each target sub-time period in the whole day based on the first pose data change features of each target sub-time period. And according to the first pose data, determining a first exercise amount of the student in each target sub-time period in the whole day. Determining a second pose data change feature of the target sub-time period of the week N in the previous continuous M weeks; and using a second pose data prediction model to predict second pose data of each target sub-time period in the whole day based on the second pose data change features of each target sub-time period. And according to the second pose data, determining a second exercise amount of the student in each target sub-time period in the whole day. And according to the first exercise amount and the second exercise amount, determining the estimated power consumption required by the sampling time period of the whole day. And when the residual electric quantity of the power supply module is lower than the estimated power consumption, the vibration prompt module is controlled to vibrate.

Description

Technical Field

[0001] The present application relates to the technical field of school badges, and in particular to an electronic school badge and a working method thereof. Background Art

[0002] The school emblem is the abbreviation of the school badge, which is one of the symbols of a school. Its main purpose is to distinguish people, keep memories and introduce the nature and disciplines of the school through patterns and texts. At the same time, when wearing the school emblem, it also invisibly increases the discipline constraints of the wearer, regulates students' behavior and increases the school's visibility.

[0003] With the development of technology, electronic school badges are now available, including power supply modules. These can collect students' location information in real time and transmit it to a cloud server via wireless communication modules, allowing parents and teachers to access their daily trajectories. Currently, the need for electronic school badges to frequently report their location to the cloud server results in high power consumption, which can easily lead to the electronic school badge shutting down due to depletion of the power supply module, resulting in a loss of location information. Summary of the Invention

[0004] The present application provides an electronic school badge and a working method thereof, which are used to solve the problem in the prior art that the electronic school badge needs to report location information to the cloud server in real time and frequently, resulting in high power consumption of the electronic school badge, and easily causing the electronic school badge to shut down due to exhaustion of power in the power supply module, resulting in the loss of positioning information collection.

[0005] In the first aspect, the present application provides a working method of an electronic school badge, which is applied to the electronic school badge. The electronic school badge includes a lower cover, an upper cover, a controller, a wireless communication module, a posture acquisition module, a vibration prompt module, and a power supply module. The lower cover and the upper cover are relatively arranged to form a shell. The controller, the wireless communication module, the posture acquisition module, and the power supply module are located in the shell. The shell is also provided with a charging interface. The controller is electrically connected to the wireless communication module, the posture acquisition module, the vibration prompt module, and the power supply module respectively. The charging interface is electrically connected to the power supply module. The posture acquisition module is used to collect student posture data during the sampling time period of 7:00-18:59; the wireless communication module is used to report the student posture data to the cloud server within the sampling time period of 7:00-18:59 according to the reporting time interval associated with the sub-time period at the current time. The method provided by the present application includes:

[0006] At the first specified time today, the controller extracts the pose data for each sub-time period within the sampling period for today and M consecutive days before today, where the first specified time is within the calculation period of 19:00-23:00, the sampling period is 7:00-18:59, and M is an integer greater than 3;

[0007] For each target sub-time period in each sub-time period, the controller determines the first pose data change feature of the consecutive M+1 target sub-time periods based on the pose data of the consecutive M+1 target sub-time periods. The pose data of the consecutive M+1 target sub-time periods are: the pose data of the target sub-time periods today and the consecutive M days before today, where the pose data of each target sub-time period is the average pose data of the pose information sampled at each sampling time in the target sub-time period;

[0008] The controller uses a pre-trained first position data prediction model to predict the first position data of the students in each target sub-time period tomorrow based on the first position data change characteristics of each target sub-time period, where tomorrow is week N, and N is any value from 1 to 5. The position data prediction model is obtained by inputting a plurality of first training samples into a first to-be-trained network, each first training sample including the first position data change characteristics of the position data of the target sub-time period today in history and the position data of the target sub-time period for M consecutive days before today in history, as well as the actual position data of the students in each target sub-time period tomorrow in history;

[0009] The controller determines the first movement amount of the student in each target sub-time period tomorrow based on the predicted first position data of the student in each target sub-time period tomorrow, wherein each target sub-time period is within the sampling time period 7:00-18:59;

[0010] At the first designated moment today, the controller extracts the posture data of each sub-time period in the sampling time period of week N in the previous M consecutive weeks;

[0011] For each target sub-time period in each sub-time period, the controller determines, based on the pose data of the target sub-time period in week N in the previous M consecutive weeks, a second pose data change feature of the target sub-time period in week N of the previous M consecutive weeks, wherein the pose data of each target sub-time period is the average pose data of the pose information sampled at each sampling time in the target sub-time period;

[0012] The controller uses a pre-trained second pose data prediction model to predict the second pose data of the student in each target sub-time period tomorrow based on the second pose data change characteristics of each target sub-time period, wherein the pose data prediction model is obtained by inputting a second to-be-trained network based on a plurality of second training samples, each second training sample including pose data of the target sub-time period in week N of M consecutive weeks in history and actual pose data of the student in each target sub-time period in the corresponding history tomorrow;

[0013] The controller determines the second exercise amount of the student in each target sub-time period tomorrow based on the predicted second posture data of the student in each target sub-time period tomorrow;

[0014] The controller determines the actual predicted exercise amount of the student in each target sub-time period tomorrow based on the first exercise amount and the second exercise amount of the student in each target sub-time period tomorrow;

[0015] The controller determines, based on the actual predicted amount of exercise of the student in each target sub-time period tomorrow, a reporting time interval associated with each target sub-time period tomorrow, wherein the length of the reporting time interval for each target sub-time period tomorrow is negatively correlated with the size of the actual predicted amount of exercise;

[0016] The controller determines the estimated power consumption required for tomorrow's sampling period, 7:00-18:59, based on the reporting intervals associated with each target sub-period within tomorrow's period.

[0017] When the controller detects that the remaining power of the power module is lower than the estimated power consumption required for the sampling period of 7:00-18:59 tomorrow, the controller controls the vibration prompt module to vibrate to prompt charging.

[0018] In one possible implementation, after controlling the vibration prompt module to vibrate to prompt charging, the method provided in this application further includes:

[0019] At the second designated time tomorrow, the controller obtains the actual power consumption of the sampling period of tomorrow from 7:00 to 18:59, where the second designated time is within the calculation period from 19:00 to 23:00.

[0020] The controller determines the difference between the actual power consumption and the estimated power consumption;

[0021] When the difference is greater than the set difference threshold, the difference is input into the network parameter update model so that the network parameter update model optimizes the network parameters of the first pose data prediction model and the network parameters of the second pose data prediction model according to the difference.

[0022] In one possible implementation, for each target sub-time period in each sub-time period, the controller determines, based on the pose data of the consecutive M+1 target sub-time periods, a first pose data change feature of the consecutive M+1 target sub-time periods, including:

[0023] For each target sub-time period in each sub-time period, the controller extracts the temporal dependency of the pose data of the consecutive M+1 target sub-time periods based on the deep confidence DBN network; based on the temporal dependency of the pose data of the consecutive M+1 target sub-time periods, the first pose data change feature of the consecutive M+1 target sub-time periods is determined.

[0024] In one possible implementation, for each target sub-time period in each sub-time period, the controller determines, based on the posture data of the target sub-time period of week N in the previous M consecutive weeks, a second posture data change feature of the target sub-time period of week N in the previous M consecutive weeks, including:

[0025] For each target sub-time period in each sub-time period, the controller extracts the temporal dependency of the posture data of the target sub-time period of week N in the previous M consecutive weeks based on the deep confidence DBN network; based on the temporal dependency of the posture data of the target sub-time period of week N in the previous M consecutive weeks, the controller determines the second posture data change feature of the target sub-time period of week N in the previous M consecutive weeks.

[0026] In one possible implementation, the controller determines the first amount of exercise of the student in each target sub-time period tomorrow based on the predicted first position data of the student in each target sub-time period tomorrow, including:

[0027] The controller inputs the predicted first position data of the student in each target sub-time period tomorrow into a pre-trained exercise amount determination model to determine the first exercise amount of the student in the corresponding target sub-time period, wherein the exercise amount determination model is trained by inputting a plurality of third training samples into the first neural network, each third training sample including the position data of the student in a historical time period and its corresponding historical actual exercise amount;

[0028] The controller determines the second amount of exercise of the student in each target sub-time period tomorrow based on the predicted second posture data of the student in each target sub-time period tomorrow, including:

[0029] The controller inputs the predicted second posture data of the student in each target sub-time period tomorrow into the pre-trained exercise amount determination model to determine the second exercise amount of the student corresponding to the target sub-time period.

[0030] In one possible implementation, the controller determines the actual predicted exercise amount of the student in each target sub-time period tomorrow based on the first exercise amount and the second exercise amount of the student in each target sub-time period tomorrow, including:

[0031] According to the formula G=k1G1+k2G2, the actual predicted exercise volume of the student in each target sub-time period tomorrow is determined, where G is the actual predicted exercise volume, G1 is the first exercise volume, G2 is the second exercise volume, k1 is the first weighting coefficient, k2 is the second weighting coefficient, and 0 <k1<1,0<k2<1,k1+k2=1。

[0032] In a possible implementation, the first weighting coefficient k1 is 0.5, and the second weighting coefficient k2 is 0.5.

[0033] In one possible implementation, the controller determines the estimated power consumption required for tomorrow's sampling period from 7:00 to 18:59 based on the reporting time intervals associated with each target sub-period within tomorrow, including:

[0034] Inputting the predicted reporting time interval associated with each target sub-time period tomorrow into a pre-trained power consumption determination model to determine the sub-power consumption corresponding to the target sub-time period, wherein the power consumption determination model is trained by inputting a plurality of fourth training samples into the second neural network, each fourth training sample including a reporting time interval for a historical time period and its corresponding actual sub-power consumption;

[0035] The power consumption of each target sub-time period is summed to obtain the estimated power consumption required for tomorrow's sampling period of 7:00-18:59.

[0036] In the second aspect, the present application also provides an electronic school badge, including a lower cover, an upper cover, a controller, a wireless communication module, a posture acquisition module, a vibration prompt module, and a power module. The lower cover and the upper cover are arranged relative to each other to form an outer shell. The controller, wireless communication module, posture acquisition module, and power module are located inside the outer shell. The outer shell is also provided with a charging interface. The controller is electrically connected to the wireless communication module, the posture acquisition module, the vibration prompt module, and the power module respectively, and the charging interface is electrically connected to the power module. The posture acquisition module is used to collect students' posture data during the sampling time period of 7:00-18:59; the wireless communication module is used to report students' posture data to the cloud server within the sampling time period of 7:00-18:59 according to the reporting time interval associated with the sub-time period at the current moment. The controller is used to execute the method provided in the first aspect of this application.

[0037] In a possible embodiment, a plurality of threaded grooves are provided on the edge of the upper cover, and a plurality of threaded holes are provided on the edge of the lower cover. Each threaded groove is threadedly connected to a threaded hole via a bolt.

[0038] The present application provides an electronic school badge and a working method thereof, which can extract the posture data of each sub-time period in the sampling time period of today and M consecutive days before today at the first specified moment today, wherein the first specified moment is within the operation time period 19:00-23:00, the sampling time period is 7:00-18:59, and M is an integer greater than 3.

[0039] For each target sub-period within each sub-period, the first position data change characteristics for the M+1 consecutive target sub-periods are determined based on the posture data for the M+1 consecutive target sub-periods. The posture data for the M+1 consecutive target sub-periods is defined as the posture data for today and the M consecutive days before today. A pre-trained first position data prediction model is used to predict the first position data for each target sub-period tomorrow based on the first position data change characteristics for each target sub-period. Tomorrow is week N, and N is a value between 1 and 5. It is understandable that since students' daily lives are usually regular, for example, the target sub-time period 7:00-7:30 is for getting up and packing, the target sub-time period 7:30-8:00 is for going to school, and the target sub-time period 8:00-8:40 is for class; the target sub-time period 8:40-8:50 is for after-get out of class activities; the target sub-time period 8:50-9:30 is for class; the target sub-time period 9:30-9:40 is for after-get out of class activities, and so on; therefore, the first position data change characteristics of the previous M+1 consecutive target sub-time periods (such as the target sub-time period 8:40-8:50 of the previous M+1 days is the after-get out of class activity time) characterize the students' posture data change trends in the most recent M+1 target sub-time periods. Therefore, using the first position data change characteristics based on each target sub-time period (such as the target sub-time period 8:40-8:50), the accuracy of predicting the students' first position data for each target sub-time period (such as the target sub-time period 8:40-8:50) tomorrow is high.

[0040] According to the predicted first position data of students in each target sub-time period tomorrow, the first exercise amount of students in each target sub-time period tomorrow is determined, wherein each target sub-time period is within the sampling time period 7:00-18:59; since the accuracy of the predicted first position data is high, the accuracy of the obtained first exercise amount is also high.

[0041] For each target sub-time period in each sub-time period, the second posture data change characteristics of the target sub-time period in week N in the previous M consecutive weeks are determined according to the posture data of the target sub-time period in week N in the previous M consecutive weeks; and the pre-trained second posture data prediction model is used to predict the second posture data of the students in each target sub-time period tomorrow based on the second posture data change characteristics of each target sub-time period. It is understandable that there is a certain regularity in students' weekly study life. For example, if the Wednesday mornings from 8:00-8:40 in the previous two weeks were Chinese classes, then the Wednesday mornings from 8:00-8:40 in the previous week and the Wednesday mornings of this week were also Chinese classes; if the Wednesday mornings from 8:50-9:30 in the previous two weeks were labor classes, then the Wednesday mornings from 8:50-9:30 in the previous week and the Wednesday mornings of this week were also labor classes; if the Wednesday mornings from 9:40-10:20 in the previous two weeks were physical education classes, then the Wednesday mornings from 9:40-10:20 in the previous week and the Wednesday mornings of this week were also physical education classes, and there is a correlation in the amount of exercise students get in Chinese classes in each week, the amount of exercise in labor classes in each week, and the amount of exercise in physical education classes in each week. The change characteristics of the second posture data of the target sub-time period of week N in the previous M consecutive weeks (such as the labor class from 8:50 to 9:30) represent the change trend of the students' posture data of the target sub-time period of week N in the previous M consecutive weeks. The pre-trained second posture data prediction model is used to predict the second posture data of the students in each target sub-time period tomorrow (i.e., week N) based on the change characteristics of the second posture data of each target sub-time period, with high accuracy.

[0042] According to the predicted second posture data of students in each target sub-time period tomorrow, the second exercise amount of students in each target sub-time period tomorrow is determined; since the accuracy of the predicted second posture data is high, the accuracy of the obtained second exercise amount is also high.

[0043] The actual predicted exercise amount of the student in each target sub-time period tomorrow is determined based on the first exercise amount and the second exercise amount of the student in each target sub-time period tomorrow. Since the accuracy of the first exercise amount and the second exercise amount is high, the accuracy of the actual predicted exercise amount of the student can be further improved.

[0044] Based on the actual predicted amount of exercise of the students in each target sub-time period tomorrow, the reporting time interval associated with each target sub-time period tomorrow is determined respectively, wherein the size of the reporting time interval of each target sub-time period tomorrow is negatively correlated with the size of the actual predicted amount of exercise. It should be noted that at any target sub-time period tomorrow, the electronic school badge can report the posture data to the cloud server according to the associated reporting time interval associated with the target sub-time period. It can be understood that when the actual predicted amount of exercise is greater, the corresponding reporting time interval is smaller, which can ensure the richness and reliability of the normal collection of more valid posture data. Conversely, when the actual predicted amount of exercise is smaller, the corresponding reporting time interval is larger, which can save power consumption without affecting the valid posture data. In this way, it can avoid, to a certain extent, the loss of posture data collection caused by the power supply module of the electronic school badge running out of power and shutting down.

[0045] In addition, due to the high accuracy of the actual predicted amount of exercise, the accuracy of the corresponding determined reporting time interval is also high. According to the reporting time intervals associated with each target sub-time period within tomorrow, the estimated power consumption required for tomorrow's sampling time period of 7:00-18:59 is determined. Due to the high accuracy of the reporting time interval, the accuracy of the corresponding estimated power consumption is also high. When it is detected that the remaining power of the power module is lower than the estimated power consumption required for tomorrow's sampling time period of 7:00-18:59, the vibration prompt module is controlled to vibrate to prompt charging. In this way, students can be accurately reminded to charge the electronic school badge during the operation time period of 19:00-23:00, so as to further avoid the loss of posture data collection due to the power supply module of the electronic school badge running out of power and shutting down. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be 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 application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0047] Figure 1 A schematic diagram of the appearance of the electronic school badge provided in an embodiment of the present application;

[0048] Figure 2 Functional module block diagram of the electronic school badge provided in the embodiment of the present application;

[0049] Figure 3 Flowchart of the working method of the electronic school badge provided in an embodiment of the present application. DETAILED DESCRIPTION

[0050] To make the purpose, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments made by ordinary technicians in this field based on the inspiration of these embodiments fall within the scope of protection of this application.

[0051] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the numbers used in this way are interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0052] The embodiment of the present application provides a method for operating an electronic school badge, which is applied to the electronic school badge. Figure 1 As shown, the electronic school badge includes a lower cover 102, an upper cover 101, a controller, a wireless communication module, a posture acquisition module, a vibration prompt module, and a power module. The lower cover 102 and the upper cover 101 are relatively arranged to form a shell. The upper cover 101 is provided with a QR code area 105 and a student information area 104. The QR code area 105 can be scanned by a scanning terminal so that the scanning terminal displays the scanned student information. Security personnel can compare the scanned student information with the student information displayed in the student information area 104 to see if they are consistent. If they are consistent, students wearing the electronic school badge are allowed to enter the school. The controller, wireless communication module, posture acquisition module, and power module are located in the shell, and the shell is also provided with a charging interface 103. The controller is a processing chip with very strong computing power. As Figure 2 As shown, the controller is electrically connected to the wireless communication module, the posture acquisition module, the vibration prompt module, and the power module respectively, and the charging interface 103 is electrically connected to the power module. The posture acquisition module is used to collect the posture data of students in the sampling time period of 7:00-18:59; the wireless communication module is used to report the posture data of students to the cloud server in the sampling time period of 7:00-18:59 according to the reporting time interval associated with the sub-time period of the current moment. Figure 3 As shown, the method provided in the embodiment of the present application includes:

[0053] S301: At the first designated moment of today, the controller extracts the posture data of each sub-time period in the sampling time period of today and the M consecutive days before today.

[0054] Among them, the first designated time is within the operation time period of 19:00-23:00 (usually students have returned home during this period and there is no need to report posture data, and it is also convenient for charging). The first designated time can be 19:20, 20:00, or 20:30, etc., and is not limited here.

[0055] The sampling period is 7:00-18:59 (usually during this period, students are on their way to school or at school and need to report their posture data). M is an integer greater than 3. For example, M can be equal to 4, 5, or 6, etc., and is not limited here.

[0056] S302: For each target sub-time period in each sub-time period, the controller determines the first position data change feature of the consecutive M+1 target sub-time periods based on the position data of the consecutive M+1 target sub-time periods. The position data of the consecutive M+1 target sub-time periods are: the position data of the target sub-time periods today and M consecutive days before today.

[0057] Specifically, S302 can be implemented as follows: for each target sub-time period in each sub-time period, the controller extracts the temporal dependency of the posture data of the consecutive M+1 target sub-time periods based on the deep confidence DBN network; based on the temporal dependency of the posture data of the consecutive M+1 target sub-time periods, determines the first position posture data change feature of the consecutive M+1 target sub-time periods.

[0058] Understandably, since students' daily lives are usually regular, for example, for today and every day of the M days before today, the target sub-time period 7:00-7:30 is for getting up and packing, the target sub-time period 7:30-8:00 is for going to school, the target sub-time period 8:00-8:40 is for class; the target sub-time period 8:40-8:50 is for after-get out of class activities; the target sub-time period 8:50-9:30 is for class; the target sub-time period 9:30-10:20 is for physical exercises; the target sub-time period 10:20-11:00 is for class; the target sub-time period 11:00-11:10 is for after-get out of class activities; the target sub-time period 11:10-11:50 is for class, the target sub-time period 11:50-12:30 is for meal time, the target sub-time period 12:30-13:40 is for lunch break, and so on.

[0059] The pose data for each target sub-period is the average pose data of the pose information sampled at each sampling time within the target sub-period. For example, the pose data for the target sub-period 9:30-10:20 is the average pose data of the pose information sampled at each sampling time within the target sub-period 9:30-10:20.

[0060] S303: The controller uses the pre-trained first position data prediction model to predict the first position data of the students in each target sub-time period tomorrow based on the first position data change characteristics of each target sub-time period.

[0061] Wherein, tomorrow is week N, where N is any value from 1 to 5. For example, week N can be Monday, Wednesday, Thursday, etc., without limitation. The posture data prediction model is obtained by inputting multiple first training samples into a first network to be trained. Each first training sample includes the posture data of the target sub-time period today in history, the first position data change characteristics of the posture data of the target sub-time period for M consecutive days before today in history, and the actual posture data of the student for each target sub-time period tomorrow in history. The first network to be trained can be, but is not limited to, a gated recurrent neural network (GRU) network.

[0062] For example, the first position data change characteristics of the previous M+1 consecutive target sub-time periods (such as the target sub-time period 8:40-8:50 of the previous M+1 days is the after-get out of class activity time) characterize the students' posture data change trends in the most recent M+1 target sub-time periods. Therefore, by using the first position data change characteristics based on each target sub-time period, the accuracy of predicting the students' first position data in each target sub-time period tomorrow (such as the target sub-time period 8:40-8:50) is high.

[0063] S304: The controller determines the first exercise amount of the student in each target sub-time period tomorrow according to the predicted first position data of the student in each target sub-time period tomorrow.

[0064] Among them, each target sub-time period is within the sampling time period of 7:00-18:59.

[0065] Specifically, S304 may be implemented as follows: the controller inputs the predicted first position data of the student for each target sub-time period tomorrow into a pre-trained exercise volume determination model to determine the first exercise volume of the student corresponding to the target sub-time period. The exercise volume determination model is trained by inputting a plurality of third training samples into the first neural network, where each third training sample includes the position data of the student for a historical time period and its corresponding historical actual exercise volume.

[0066] S305: At the first designated moment today, the controller extracts the posture data of each sub-time period in the sampling time period of week N in the previous M consecutive weeks.

[0067] For example, week N may be Monday, Wednesday, or Thursday. The first designated time may be 19:20, 20:00, or 20:30, etc., which are not limited here.

[0068] S306: For each target sub-time period in each sub-time period, the controller determines the second posture data change feature of the target sub-time period of week N in the previous M consecutive weeks based on the posture data of the target sub-time period of week N in the previous M consecutive weeks.

[0069] The posture data of each target sub-time period is the average posture data of the posture information sampled at each sampling moment within the target sub-time period. Specifically, S306 can be implemented as follows: for each target sub-time period in each sub-time period, the controller extracts the temporal dependency of the posture data of the target sub-time period of week N in the previous M consecutive weeks based on the deep confidence DBN network; and determines the second posture data change feature of the target sub-time period of week N in the previous M consecutive weeks based on the temporal dependency of the posture data of the target sub-time period of week N in the previous M consecutive weeks.

[0070] It is understandable that there is a certain regularity in students' weekly study life. For example, if the Wednesday mornings from 8:00-8:40 in the previous two weeks were Chinese classes, then the Wednesday mornings from 8:00-8:40 in the previous week and the Wednesday mornings of this week were also Chinese classes; if the Wednesday mornings from 8:50-9:30 in the previous two weeks were labor classes, then the Wednesday mornings from 8:50-9:30 in the previous week and the Wednesday mornings of this week were also labor classes; if the Wednesday mornings from 9:40-10:20 in the previous two weeks were physical education classes, then the Wednesday mornings from 9:40-10:20 in the previous week and the Wednesday mornings of this week were also physical education classes, and there is a correlation in the amount of exercise students get in Chinese classes in each week, the amount of exercise in labor classes in each week, and the amount of exercise in physical education classes in each week.

[0071] S307: The controller uses the pre-trained second posture data prediction model to predict the second posture data of the students in each target sub-time period tomorrow based on the second posture data change characteristics of each target sub-time period.

[0072] The posture data prediction model is obtained by inputting multiple second training samples into the second to-be-trained network. Each second training sample includes posture data for a target sub-time period in week N of the previous M consecutive weeks, as well as actual posture data of students for each target sub-time period in the corresponding historical period. The second to-be-trained network may be, but is not limited to, a gated recurrent neural network (GRU) network.

[0073] It can be understood that the change characteristics of the second posture data of the target sub-time period of week N in the previous M consecutive weeks (such as the labor class from 8:50 to 9:30) represent the change trend of the posture data of the target sub-time period of week N in the previous M consecutive weeks of the students. The pre-trained second posture data prediction model is used to predict the second posture data of the students in each target sub-time period tomorrow based on the change characteristics of the second posture data of each target sub-time period, with high accuracy.

[0074] S308: The controller determines the second amount of exercise of the student in each target sub-time period tomorrow based on the predicted second posture data of the student in each target sub-time period tomorrow.

[0075] Specifically, the controller inputs the predicted second posture data of the student in each target sub-time period tomorrow into the pre-trained exercise amount determination model to determine the second exercise amount of the student corresponding to the target sub-time period.

[0076] S309: The controller determines the actual predicted exercise amount of the student in each target sub-time period tomorrow according to the first exercise amount and the second exercise amount of the student in each target sub-time period tomorrow.

[0077] According to the formula G=k1G1+k2G2, the actual predicted exercise volume of the students in each target sub-time period tomorrow, where G is the actual predicted exercise volume, G1 is the first exercise volume, G2 is the second exercise volume, k1 is the first weighting coefficient, k2 is the second weighting coefficient, and 0 <k1<1,0<k2<1,k1+k2=1。

[0078] In some embodiments, the first weighting coefficient k1 is 0.5, and the second weighting coefficient k2 is 0.5. In other embodiments, the first weighting coefficient k1 is 0.4, and the second weighting coefficient k2 is 0.6.

[0079] S310: The controller determines the reporting time interval associated with each target sub-time period tomorrow according to the actual predicted exercise amount of the student in each target sub-time period tomorrow.

[0080] The reporting interval for each target sub-time period tomorrow is negatively correlated with the actual predicted amount of exercise. It should be noted that at any target sub-time period tomorrow, the electronic school badge can report posture data to the cloud server based on the reporting interval associated with that target sub-time period.

[0081] It can be understood that when the actual predicted amount of motion is greater, the corresponding reporting time interval is smaller, which can ensure the richness and reliability of normal collection of more effective posture data. Conversely, when the actual predicted amount of motion is smaller, the corresponding reporting time interval is larger, which can save power consumption without affecting the effective posture data.

[0082] S311: The controller determines the estimated power consumption required for the sampling period 7:00-18:59 tomorrow based on the reporting time intervals associated with each target sub-period within tomorrow.

[0083] Specifically, S311 can be implemented by inputting the predicted reporting interval associated with each target sub-time period tomorrow into a pre-trained power consumption determination model to determine the sub-power consumption for the corresponding target sub-time period. The power consumption determination model is trained by inputting multiple fourth training samples into the second neural network, where each fourth training sample includes the reporting interval for a historical time period and its corresponding actual sub-power consumption. The sub-power consumption for each target sub-time period is summed to obtain the estimated power consumption for tomorrow's sampling period, 7:00 AM - 6:59 PM.

[0084] S312: When the controller detects that the remaining power of the power module is lower than the estimated power consumption required for the sampling period of 7:00-18:59 tomorrow, the controller controls the vibration prompt module to vibrate to prompt charging.

[0085] To sum up, the embodiment of the present application provides a working method for an electronic school badge, which can extract the posture data of each sub-time period in the sampling time period of today and M consecutive days before today at the first specified moment today, wherein the first specified moment is within the operation time period 19:00-23:00, the sampling time period is 7:00-18:59, and M is an integer greater than 3.

[0086] For each target sub-period within each sub-period, the first position data change characteristics for the M+1 consecutive target sub-periods are determined based on the posture data for the M+1 consecutive target sub-periods. The posture data for the M+1 consecutive target sub-periods is defined as the posture data for today and the M consecutive days before today. A pre-trained first position data prediction model is used to predict the first position data for each target sub-period tomorrow based on the first position data change characteristics for each target sub-period. Tomorrow is week N, and N is a value between 1 and 5. It can be understood that since students' daily lives are usually regular, for example, the target sub-time period 7:00-7:30 is for getting up and packing, the target sub-time period 7:30-8:00 is for going to school, and the target sub-time period 8:00-8:40 is for class; the target sub-time period 8:40-8:50 is for after-get out of class activities; the target sub-time period 8:50-9:30 is for class; the target sub-time period 9:30-9:40 is for after-get out of class activities, and so on; therefore, the first position data change characteristics of the previous M+1 consecutive target sub-time periods (such as the target sub-time period 8:40-8:50 of the previous M+1 days is the after-get out of class activity time) characterize the students' posture data change trends in the most recent M+1 target sub-time periods. Therefore, using the first position data change characteristics based on each target sub-time period (such as the target sub-time period 8:40-8:50), the accuracy of predicting the students' first position data for each target sub-time period (such as the target sub-time period 8:40-8:50) in tomorrow (i.e., week N) is high.

[0087] According to the predicted first position data of students in each target sub-time period tomorrow, the first exercise amount of students in each target sub-time period tomorrow is determined, wherein each target sub-time period is within the sampling time period 7:00-18:59; since the accuracy of the predicted first position data is high, the accuracy of the obtained first exercise amount is also high.

[0088] For each target sub-time period in each sub-time period, the second posture data change characteristics of the target sub-time period in week N in the previous M consecutive weeks are determined according to the posture data of the target sub-time period in week N in the previous M consecutive weeks; and the pre-trained second posture data prediction model is used to predict the second posture data of the students in each target sub-time period tomorrow based on the second posture data change characteristics of each target sub-time period. It is understandable that there is a certain regularity in students' weekly study life. For example, if the Wednesday mornings from 8:00-8:40 in the previous two weeks were Chinese classes, then the Wednesday mornings from 8:00-8:40 in the previous week and the Wednesday mornings of this week were also Chinese classes; if the Wednesday mornings from 8:50-9:30 in the previous two weeks were labor classes, then the Wednesday mornings from 8:50-9:30 in the previous week and the Wednesday mornings of this week were also labor classes; if the Wednesday mornings from 9:40-10:20 in the previous two weeks were physical education classes, then the Wednesday mornings from 9:40-10:20 in the previous week and the Wednesday mornings of this week were also physical education classes, and there is a correlation in the amount of exercise students get in Chinese classes in each week, the amount of exercise in labor classes in each week, and the amount of exercise in physical education classes in each week. The change characteristics of the second posture data of the target sub-time period of week N in the previous M consecutive weeks (such as the labor class from 8:50 to 9:30) represent the change trend of the students' posture data of the target sub-time period of week N in the previous M consecutive weeks. The pre-trained second posture data prediction model is used to predict the second posture data of the students in each target sub-time period tomorrow (i.e., week N) based on the change characteristics of the second posture data of each target sub-time period, with high accuracy.

[0089] According to the predicted second posture data of students in each target sub-time period tomorrow, the second exercise amount of students in each target sub-time period tomorrow is determined; since the accuracy of the predicted second posture data is high, the accuracy of the obtained second exercise amount is also high.

[0090] The actual predicted exercise amount of the student in each target sub-time period tomorrow is determined based on the first exercise amount and the second exercise amount of the student in each target sub-time period tomorrow. Since the accuracy of the first exercise amount and the second exercise amount is high, the accuracy of the actual predicted exercise amount of the student can be further improved.

[0091] Based on the student's actual predicted physical activity for each target sub-time period tomorrow, the reporting interval associated with each target sub-time period is determined. The reporting interval for each target sub-time period tomorrow is negatively correlated with the actual predicted physical activity. It should be noted that at any target sub-time period tomorrow, the electronic school badge can report posture data to the cloud server based on the reporting interval associated with that target sub-time period. It is understood that when the actual predicted physical activity is greater, the corresponding reporting interval is shorter, ensuring the richness and reliability of the collected, valid posture data. Conversely, when the actual predicted physical activity is smaller, the corresponding reporting interval is longer, saving power without compromising valid posture data. This can, to a certain extent, prevent the electronic school badge's power supply module from shutting down due to depletion, resulting in the loss of posture data collection. Furthermore, due to the high accuracy of the actual predicted physical activity, the corresponding reporting interval is also highly accurate. Based on the reporting intervals associated with each target sub-time period tomorrow, the estimated power consumption required for tomorrow's sampling period, 7:00 AM - 6:59 PM, is determined. The high accuracy of the reporting intervals leads to a high accuracy in the corresponding power consumption estimates. If the remaining power in the power module is detected to be lower than the estimated power consumption for tomorrow's sampling period, 7:00 AM - 6:59 PM, the vibration reminder module will vibrate to prompt charging. This ensures that students are accurately reminded to charge their electronic school badges during the computing period, 7:00 PM - 11:00 PM, further preventing the electronic school badge from shutting down due to depleted power and resulting in missed pose data collection.

[0092] In addition, in a possible implementation manner, after S312, the method provided in the embodiment of the present application further includes:

[0093] Step 1: The controller obtains the actual power consumption of the sampling period 7:00-18:59 tomorrow at the second designated time tomorrow, wherein the second designated time is within the calculation period 19:00-23:00.

[0094] Step 2: The controller determines the difference between the actual power consumption and the estimated power consumption.

[0095] Step 3: When the difference is greater than the set difference threshold, the difference is input into the network parameter update model so that the network parameter update model optimizes the network parameters of the first pose data prediction model and the network parameters of the second pose data prediction model according to the difference.

[0096] It can be understood that based on the above steps 1 to 3, the updated first posture data prediction model can have high accuracy in the first posture data of students in each target sub-time period in the subsequent prediction tomorrow, and the second posture data prediction model can also have high accuracy in the second posture data of students in each target sub-time period in the subsequent prediction tomorrow.

[0097] In addition, an embodiment of the present application also provides an electronic school badge, including a lower cover 102, an upper cover 101, a controller, a wireless communication module, a posture acquisition module, a vibration prompt module, and a power module. The lower cover 102 and the upper cover 101 are arranged relative to each other to form an outer shell. The controller, wireless communication module, posture acquisition module, and power module are located inside the outer shell. The outer shell is also provided with a charging interface 103. The controller is electrically connected to the wireless communication module, the posture acquisition module, the vibration prompt module, and the power module respectively. The charging interface 103 is electrically connected to the power module. The posture acquisition module is used to collect students' posture data during the sampling time period of 7:00-18:59; the wireless communication module is used to report students' posture data to the cloud server within the sampling time period of 7:00-18:59 according to the reporting time interval associated with the sub-time period at the current moment. The controller is used to execute the method provided in the above embodiment of the present application.

[0098] In a possible embodiment, a plurality of threaded grooves are provided on the edge of the upper cover 101, and a plurality of threaded holes are provided on the edge of the lower cover 102. Each threaded groove is threadedly connected to a threaded hole by a bolt, and has strong tightness.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for working an electronic school badge, characterized in that: Applied to an electronic school badge, the electronic school badge includes a lower cover, an upper cover, a controller, a wireless communication module, a posture acquisition module, a vibration prompt module, and a power module. The lower cover and the upper cover are relatively arranged to form a shell. The controller, the wireless communication module, the posture acquisition module, and the power module are located in the shell. The shell is also provided with a charging interface. The controller is electrically connected to the wireless communication module, the posture acquisition module, the vibration prompt module, and the power module respectively. The charging interface is electrically connected to the power module. The posture acquisition module is used to collect students' posture data during the sampling time period of 7:00-18:59; the wireless communication module is used to report students' posture data to the cloud server within the sampling time period of 7:00-18:59 according to the reporting time interval associated with the sub-time period at the current time. The method includes: The controller extracts, at a first designated time today, posture data for each sub-time period in a sampling time period for today and M consecutive days before today, wherein the first designated time is within a calculation time period of 19:00-23:00, the sampling time period is 7:00-18:59, and M is an integer greater than 3; For each target sub-time period in each sub-time period, the controller determines a first pose data change feature of the continuous M+1 target sub-time periods based on the pose data of the continuous M+1 target sub-time periods, where the pose data of the continuous M+1 target sub-time periods are: the pose data of the target sub-time periods of today and the M consecutive days before today, wherein the pose data of each target sub-time period is the average pose data of the pose information sampled at each sampling time in the target sub-time period; The controller uses a pre-trained first position data prediction model to predict the first position data of the students in each target sub-time period tomorrow based on the first position data change characteristics of each target sub-time period, wherein tomorrow is week N, and N is any value from 1 to 5. The posture data prediction model is obtained by inputting a plurality of first training samples into a first to-be-trained network, each of the first training samples including the first position data change characteristics of the posture data of the target sub-time period today in history and the posture data of the target sub-time periods for M consecutive days before today in history, as well as the actual posture data of the students in each target sub-time period tomorrow in history; The controller determines the first exercise amount of the student in each target sub-time period tomorrow according to the predicted first position data of the student in each target sub-time period tomorrow, wherein each target sub-time period is within the sampling time period of 7:00-18:59; The controller extracts, at a first designated moment of the day, posture data of each sub-time period in a sampling time period of week N of the previous M consecutive weeks; For each of the target sub-time periods in each sub-time period, the controller determines, based on the posture data of the target sub-time period in week N in the previous M consecutive weeks, a second posture data change feature of the target sub-time period in week N of the previous M consecutive weeks, wherein the posture data of each target sub-time period is the average posture data of the posture information sampled at each sampling time in the target sub-time period; The controller uses a pre-trained second posture data prediction model to predict the second posture data of the student in each target sub-time period tomorrow based on the second posture data change characteristics of each target sub-time period, wherein the posture data prediction model is obtained by inputting a second to-be-trained network based on a plurality of second training samples, each of the second training samples including the posture data of the target sub-time period in week N of the previous M consecutive weeks in history, and the actual posture data of the student in each target sub-time period in the corresponding historical tomorrow; The controller determines the second amount of exercise of the student in each target sub-time period tomorrow according to the predicted second posture data of the student in each target sub-time period tomorrow; The controller determines the actual predicted exercise amount of the student in each target sub-time period tomorrow according to the first exercise amount and the second exercise amount of the student in each target sub-time period tomorrow; The controller determines, based on the actual predicted amount of exercise of the student in each target sub-time period tomorrow, a reporting time interval associated with each target sub-time period tomorrow, wherein the length of the reporting time interval of each target sub-time period tomorrow is negatively correlated with the size of the actual predicted amount of exercise; The controller determines the estimated power consumption required for the sampling period 7:00-18:59 tomorrow based on the reporting time intervals associated with each of the target sub-periods tomorrow; When the controller detects that the remaining power of the power module is lower than the estimated power consumption required for the sampling period of 7:00-18:59 tomorrow, the controller controls the vibration prompt module to vibrate to prompt charging.

2. The method according to claim 1, characterized in that After controlling the vibration prompt module to vibrate to prompt charging, the method further includes: The controller obtains the actual power consumption of the sampling period of 7:00-18:59 tomorrow at a second designated time tomorrow, wherein the second designated time is within the calculation period of 19:00-23:00; The controller determines a difference between the actual power consumption and the estimated power consumption; When the difference is greater than the set difference threshold, the difference is input into the network parameter update model so that the network parameter update model optimizes the network parameters of the first pose data prediction model and the network parameters of the second pose data prediction model according to the difference.

3. The method according to claim 1, characterized in that For each target sub-time period in each sub-time period, the controller determines, based on the posture data of the consecutive M+1 target sub-time periods, a first position data change feature of the consecutive M+1 target sub-time periods, including: For each target sub-time period in each sub-time period, the controller extracts the temporal dependency of the posture data of the consecutive M+1 target sub-time periods according to the deep confidence DBN network; and determines the first position data change feature of the consecutive M+1 target sub-time periods according to the temporal dependency of the posture data of the consecutive M+1 target sub-time periods.

4. The method according to claim 1, wherein For each of the target sub-time periods in each sub-time period, the controller determines, based on the posture data of the target sub-time period of week N in the previous M consecutive weeks, a second posture data change feature of the target sub-time period of week N in the previous M consecutive weeks, including: For each of the target sub-time periods in each sub-time period, the controller extracts the temporal dependency of the posture data of the target sub-time period of week N in the previous M consecutive weeks based on the deep confidence DBN network; based on the temporal dependency of the posture data of the target sub-time period of week N in the previous M consecutive weeks, determines the second posture data change feature of the target sub-time period of week N in the previous M consecutive weeks.

5. The method according to claim 1, wherein The controller determines the first amount of exercise of the student in each target sub-time period tomorrow according to the predicted first position data of the student in each target sub-time period tomorrow, including: The controller inputs the predicted first position data of the student in each target sub-time period tomorrow into a pre-trained exercise amount determination model to determine the first exercise amount of the student corresponding to the target sub-time period, wherein the exercise amount determination model is trained by inputting a plurality of third training samples into the first neural network, each of the third training samples including the student's position data in a historical time period and its corresponding historical actual exercise amount; The controller determines the second amount of exercise of the student in each target sub-time period tomorrow according to the predicted second posture data of the student in each target sub-time period tomorrow, including: The controller inputs the predicted second posture data of the student in each target sub-time period tomorrow into the pre-trained exercise amount determination model to determine the second exercise amount of the student corresponding to the target sub-time period.

6. The method according to claim 1, wherein The controller determines the actual predicted exercise amount of the student in each target sub-time period tomorrow according to the first exercise amount and the second exercise amount of the student in each target sub-time period tomorrow, including: According to the formula G=k1G1+k2G2, the actual predicted amount of exercise of the student in each target sub-time period tomorrow is determined, where G is the actual predicted amount of exercise, G1 is the first amount of exercise, G2 is the second amount of exercise, k1 is the first weighting coefficient, k2 is the second weighting coefficient, and 0 <k1<1,0<k2<1,k1+k2=1。 7. The method according to claim 6, characterized in that The first weighting coefficient k1 is 0.5, and the second weighting coefficient k2 is 0.

5.

8. The method according to claim 1, characterized in that The controller determines the estimated power consumption required for the sampling period 7:00-18:59 tomorrow based on the reporting time intervals associated with each of the target sub-periods tomorrow, including: Inputting the predicted reporting time interval associated with each target sub-time period tomorrow into a pre-trained power consumption determination model to determine the sub-power consumption corresponding to the target sub-time period, wherein the power consumption determination model is trained by inputting a plurality of fourth training samples into the second neural network, each of the fourth training samples including the reporting time interval of a historical time period and its corresponding actual sub-power consumption; The power consumption of each target sub-time period is summed to obtain the estimated power consumption required for tomorrow's sampling period of 7:00-18:

59.

9. An electronic school badge, characterized in that: The electronic school badge includes a lower cover, an upper cover, a controller, a wireless communication module, a posture acquisition module, a vibration prompt module, and a power module. The lower cover and the upper cover are arranged relative to each other to form an outer shell. The controller, the wireless communication module, the posture acquisition module, and the power module are located in the outer shell. The outer shell is also provided with a charging interface. The controller is electrically connected to the wireless communication module, the posture acquisition module, the vibration prompt module, and the power module respectively. The charging interface is electrically connected to the power module. The posture acquisition module is used to collect students' posture data during the sampling time period of 7:00-18:59; the wireless communication module is used to report students' posture data to the cloud server within the sampling time period of 7:00-18:59 according to the reporting time interval associated with the sub-time period at the current moment. The controller is used to execute any method described in claims 1-8.

10. The electronic school badge according to claim 9, characterized in that: The edge of the upper cover is provided with a plurality of threaded grooves, and the edge of the lower cover is provided with a plurality of threaded holes. Each of the threaded grooves is threadedly connected to a threaded hole via a bolt.

Images