2.1 Physical health management in a cloud management model

Cloud-based management platform for student physical health management relies on the mobile internet as a practical service platform, the main role is to provide a convenient sports life for many young people, with a focus on practicality, service, and interactivity [13]. Preventive healthcare, emergency preparation, and centralization of health information are all made possible by a cloud-based platform for managing physical health among students. It involves parents in their child’s health and simplifies reporting and administrative duties. The platform seeks to improve the well-being of students. Youth physical health promotion cloud management mainly includes the management body and the management module. First, the cloud management platform establishes a “four-in-one” management model for students, physical education teachers (class teachers), schools, and authorities to promote the management and sharing of student physical health information (Figure 1). A better, more knowledgeable attitude to physical well-being is promoted when students receive feedback on their physical health. This enhances awareness, motivation, goal setting, education, self-reflection, and self-efficacy, as well as preventing health complications. Based on the evaluation system of the National Standard for Student Physical Fitness, the “verification (monitoring), evaluation, early warning, feedback (information, online platform) and intervention” physical fitness promotion system is established from the perspective of five interlinked and coordinated mechanisms: verification (monitoring), evaluation, early warning, feedback, intervention, and verification (monitoring). The core functions of this dynamic system are in three areas: student physical health data management, student physical health assessment (early warning), student physical health assessment, and student physical health decision-making (profile).

Figure 1

Cloud-based management platform for physical health management system.

In the “Cloud Management Platform + Student Health” model, the data and information processing functions of the cloud platform are fully utilized. Cloud solutions increase the effectiveness of physical health testing by centralizing health data, providing scalability, security, and user-friendly interfaces, and enabling real-time monitoring, automated analysis, remote tracking, and efficient reporting. Scalability, affordability, accessibility, big data analytics, artificial intelligence, disaster recovery, real-time processing, global reach, application programming interfacs, quick application development, data insights, sustainability support, and innovation are all provided by cloud-based data processing. On the one hand, it enhances the timeliness, efficiency, and convenience of overall physical health testing and student health services. On the other hand, it provides a new way of working for managing student health information, making the management of student health more comprehensive and providing alerts and targeted measures for weaknesses in student health.

2.2 Pathway of youth physical health promotion under the cloud management model

The early warning of student physical health is mainly based on the assessment of the National Student Physical Fitness Standard levels, based on student physical fitness test results, a systematic evaluation of adolescent physical fitness test results, and a method to judge the total score (level) of student physical fitness and the unevenness of individual physical fitness and make corresponding early warning warnings [14]. In early warning systems for physical fitness, students are motivated by prompt feedback, concerns identified, goals set, tailored recommendations, health awareness promoted, parents and educators involved, progress tracked, accountability instilled, rewards offered, and positive role models displayed. Informed lifestyle choices are supported, and educators, parents, and legislators may use the data obtained from the assessment of NSPFS levels to inspire kids, improve academic achievement, avoid illnesses, and promote physical health.

The construction of an early warning mechanism for students’ physical fitness consists of two main components (Figure 2). First, based on the requirement of “good or above to participate in the awarding of prizes and merits” in the Standards, an early warning is given to the total score (grade) of students’ physical fitness, and the scores in the “pass” and “fail” stages are given an early warning. The second is the unbalanced warning of individual student indicators, where the warning is about the “weaknesses” of individual student indicators. The second is an unbalanced warning for individual indicators, which is a “weakness,” i.e., a “pass” or a “failure” for individual physical form, physical function, and physical fitness indicators, representing “level 1 warning” and “level 2 warning,” respectively. To identify at-risk kids and offer the appropriate help, early warning systems for students rely on several factors, such as performance benchmarks, historical comparisons, and standardized exams, “Level 1 warning” and “Level 2 warning,” respectively. As the physical condition of students varies from region to region, so do the values set in the Standards when setting early warning parameters for student fitness (Table 1) [15]. When students’ physical fitness test scores fall below the defined values in the Standards, feedback is given to students, physical education teachers, class teachers, and schools to raise students’ awareness of their physical fitness status and to urge them to exercise in a targeted manner, to improve students’ physical fitness.

Figure 2

Youth physical health warning model. (Note: " A+ " , " B+ " , " C+ " represents “normal,” “low weight and overweight,” and “obesity” in the BMI index; " A+ " , " B+ " , " C+ " , " D+ " represents the “excellent,” “good,” “pass,” and “fail” levels in the standard.)

Fundamentally, a single level of warning cannot serve the true purpose of a student’s physical health warning, and only students can become more conscious of their physical health when they are genuinely aware of the status of their bodies at the moment and any potential hazards [16]. This is why it is important to develop a feedback form for physical health warnings. Physical health warning feedback forms enable user engagement, quality enhancement, data collecting, personalized messaging, risk detection, insights into user behavior, empowerment, policy formulation, accountability, and public health results. The focus of the feedback form is to explain the content of the student’s warning level (Table 2) and to provide feedback to students and physical education teachers in the form of a “profile” [17]. On the one hand, the feedback form allows students to know exactly where they are in terms of their fitness indicators and to use the early warnings as a basis for targeted practice to strengthen their “weak” indicators. The feedback on students’ physical health information strengthens the students’ perception of their physical condition, thus creating a sense of urgency at the psychological level and prompting students to re-examine the results of their physical health assessment and engage in physical activity accordingly. On the other hand, physical education teachers can provide targeted counseling and contact with students based on their early warning tips, which not only improves the efficiency of public physical education classes but also ensures the improvement of students’ overall physical fitness level.

By following up on the results of youth physical fitness tests in recent years, we found that youth physical fitness tests have not been improved under the guidance of the national physical health policy; the main reason for this is that although youths can know the items in which they are deficient in physical fitness tests during their four years in university, they should make corresponding exercises to improve these items but are very vague, and some students have insufficient knowledge of physical education and related sports. Several reasons, including insufficient curriculum, socioeconomic inequalities, cultural influences, gender inequities, physical and health problems, limited exposure, peer pressure, digital diversions, and language obstacles, contribute to kids’ lack of understanding about physical education and sports. Some students do not have sufficient knowledge of sports and related exercises, which leads to the inability of young people to choose the right sports to improve their physical fitness [18]. Therefore, the physical health management department of universities should formulate exercise prescriptions according to the physical health test results of youths to guide youths to carry out physical exercises in a planned and targeted manner. Through interviews with experts from the Soochow University Physical Fitness Test Centre and surveys of trainers working in the fitness industry, exercise prescriptions for youth physical fitness programs were developed (Table 3) to provide a theoretical reference for youth to exercise in a more targeted manner [19].

3.1 Analysis of the running model

To estimate the displacement of the mass, the gait cycle of the running process should first be analyzed. Running gait cycle analysis has several advantages, such as biomechanical understanding, injury avoidance, improved performance, rehabilitation assistance, tailored treatments, diagnosis, research, patient education, and data-driven tracking. It is understood that the running process consists of two main parts: one is the support period, and the other is the leap period [21]. When it comes to running biomechanics, researching human gait may help with clinical diagnosis, reduce the risk of injury, improve performance, tailor training, promote biomechanical research, offer data-driven insights, and improve coaching and instruction. Figure 4 shows the sequence of movements of the left and right legs of the human body over some time.

Figure 4

Movement sequence of left and right legs.

According to the different states of the left and right legs, the support period can be divided into two processes: (i) left side support, right side swing; (ii) right side support, left side swing. During running, the left and right legs swing alternately. Running with alternate leg swings enhances forward motion, stability, balance, energy conservation, coordination, endurance, and injury avoidance. It improves running as a technical skill, performance, and mind-body connection. To obtain the displacement of the human body during running, it is necessary to detect the gait pattern and then adaptively apply different displacement estimation methods. Precise displacement estimation in running improves biomechanical insights, injury avoidance, pace, training efficiency, and performance monitoring. It also supports sports science research, motivation, and decision-making.

3.2 Algorithm for centroid displacement during support period

During the support phase, displacement is calculated from the extension and flexion angles of the lower body joints. Figure 5 shows the course of the body movement in the XOY plane view.

Figure 5

The process of human motion on the XOY plane view.

When in a running phase, we use stratification information to estimate displacement. For an accurate estimate of displacement, one must take into account a variety of elements, including geography, weather, impediments, human factors, measuring techniques, time constraints, instruments, outside effects, and data sources. As mentioned above, the root refers to the vector sum of the lengths of the foot, thigh, and knee. As shown in equations (1) and (2):

S foot x S thigh x S root x S shin x represent the vectors on the X-axis for the root, thigh, calf, and foot, respectively, S root y S thigh y , S shin y , S foot y represent the vectors up the Y-axis for the root, thigh, calf, and foot, respectively.

S i x , S i y is the change on the t + Δ t axis from cycle t to X , Y . L k is denoted as the length of joint K, θ i , t + Δ t is the angle at time t + Δ t node, and θ i , t is denoted as the angle at time t node. At time t + Δ t , the expression for the root is shown in equation (4):

S root, t + Δ t x , S root, t + Δ t y are represented as the root vector on the X and Y axes at the time node of t + Δ t and θ root is the angle of the root vector.

3.3 Algorithm for estimating the displacement of the mass in the fly-over period

During the leap period, human motion is only influenced by gravity. Based on Newtonian kinematics, at time node t + Δ t , the root node S root, t + Δ t k has the following equation:

S root, t + Δ t k is denoted as the root vector of the K axis at time node t + Δ t , υ k is the velocity of the root vector that can be solved by integration with a gyroscope, and a t − 1 k is denoted as the integration over the K axis of one of the three axes.

4.1 Exercise individual health evaluation model

The exercise energy expenditure data collected in the real-time exercise data monitoring system is collected at a frequency of once per second. Because data per second exhibits minimal variation, and data may be intermittently missing at specific seconds due to network conditions. For analysis, the collected exercise energy expenditure data is converted into exercise energy expenditure (in kcal/30 s) every 30 s, which is called the exercise energy expenditure rate and is denoted as r n ( n = 1 , 2 , 3 , … ) . In the exercise individual health evaluation model, network problems might result in biased data, lower accuracy, and incomplete evaluations. For a model to be effective, reliable data transmission and gathering methods must be included.

In the process of exercise, as the exercise time advances, the physical exertion of the exerciser first gradually increases and then reaches a relatively stable state, as the human physical strength is limited and finally decreases, the corresponding energy consumption rate will show a gradual increase and then a steady and finally decreasing. On the premise that the exercise state is stable, according to the stability of the Markov chain, we can predict the exercise limit energy consumption rate of 0. If we find that the exercise rate reaches this limit of 0, we can remind the exerciser to reduce the exercise intensity or stop exercising, to achieve the purpose of health evaluation of the exercise individual.

To predict the limiting energy consumption rate of a moving individual based on a Markov model, the current real-time monitored energy consumption of the movement can be calculated to obtain a sequence of energy consumption rates r 1 , r 2 , … r n . The steps of the Markov process for analyzing this sequence are as follows:

1. Divide the state space, the state intervals of the energy consumption rate sequence are divided as follows:

   (6) [ r max , r min + Δ r ] , … , [ r min + ( m − 1 ) ⋅ Δ r , r max ] .

2. Calculate the transfer probability matrix. Analysis of the energy consumption rate state transfer for each successive 30 s (e.g., state transfer from the 30th to the 60th) and calculation of the transfer probability matrix for the energy consumption rate P r ;

3. Find the steady state vector. Using the state of the energy consumption rate at the 30th s as the initial state. In a Markov chain model, the steady-state probability vector – which represents equilibrium distribution and long-term health outcomes – is essential for exercise and health assessment. It helps with model calibration, health forecasting, and policy assessment. Note that this initial state vector is S n , then the following equation holds:

5.1 Health evaluation

The health evaluation model for youth sports individuals uses a health evaluation model based on the exercise energy expenditure rate, which aims to calculate the exercise limit energy expenditure rate Q of the sports individual and to evaluate the health of the individual [23,24]. Numerous techniques, including VO2 max, heart rate monitoring, functional capacity evaluation, wearable fitness trackers, and self-reports, are used in the assessment of health based on exercise limit energy expenditure rate. Calculating the basal metabolic rate, physical activity level, total daily energy expenditure, and resting energy expenditure yields the exercise limit energy expenditure rate. With the use of indirect calorimetry for accurate readings, it is dependent on personal aspects as well as expert advice.

The following is based on the energy expenditure data of an exerciser in a continuous real-time exercise to evaluate this exerciser’s health [25]. The exercise data real-time monitoring system may collect the exercise energy expenditure data of all exercise persons in real-time. Enhancing fitness journeys and user pleasure, a workout data real-time monitoring system offers real-time feedback, performance tracking, safety monitoring, customization, compatibility, user engagement, analytical insights, goal attainment, and support. The energy expenditure rate is set to equal the workout energy expenditure every 30 s to simplify the subsequent analysis and calculation. The energy expenditure rate of a certain exerciser over 10 min is shown in Table 4.

According to the formula, the above energy consumption rate is divided into four state spaces, i.e., m = 4 , then Δ r = ( 7.208 − 2.688 ) / 4 = 1.130 , at which point the state space (noted as states A , B , C , D in that order) is divided as follows:

Based on the above division of the state space, the energy consumption rate transfer statistics for two adjacent 30 s can be derived, as shown in Table 5.

According to the data in Table 5, the transition probability matrix can be obtained as follows:

Next, using the state at 600 s (minute 10) as the initial state, i.e., S n = ( 1 , 0 , 0 , 0 ) , the ensuing rate of energy consumption can be predicted according to the equation S n + 1 = S n ⋅ P .

The probability vector in the steady state, i.e., the stable vector S of this Markov chain, is required to satisfy S = S ⋅ P , i.e., ( E − P T ) ⋅ S T = 0 . It can be seen that to obtain the stable vector S , it is sufficient to obtain the transpose matrix P T of the transfer probability matrix P with eigenvalues of 1 as eigenvector X n = ( x 1 , x 2 , x 3 , x 4 ) , and to obtain the eigenvector X n , which is listed in the characteristic equation as

System analysis relies heavily on the steady-state probability vector for a Markov chain, which is derived by multiplying the transition matrix by the starting probability vector [26]. This vector reveals the long-term behavior of the Markov chain. Since the underlying solution system for this characteristic equation is

The equation has no unique solution, which can be obtained by adding the constraint ∑ i = 1 4 x i = 1 , which is the stable vector S , i.e., the stable vector:

From the resulting stability vector S , the limiting rate of energy expenditure Q can be found for the exerciser

As illustrated in Figure 6, we may plot based on this exercise limit energy consumption rate and the exerciser’s energy consumption rate for each period to determine the constant energy of the exerciser in the current exercise condition. The consumption rate is 5.602 kcal/30 s.

Figure 6

Comparison of energy expenditure rate and ultimate energy expenditure rate of a certain athlete during continuous 10 min of exercise.

Figure 6 shows that the exerciser’s energy expenditure rate starts to drop below the maximum rate of 5.602 kcal/30 s at about the 450-s mark and keeps dropping throughout the exercise, indicating a loss of physical strength.

5.2 Comparative validation

To evaluate the accuracy of the findings of the exercise individual health evaluation model, 20 exercisers were selected to participate in a 2 km run, and the exercise data was monitored using a real-time exercise data monitoring system [27]. The actual exercise time T was obtained when the exercisers felt physically exhausted during the exercise process

The threshold of error E Th is set; if the error is within the value of E Th , the evaluation result is considered correct; otherwise, the evaluation result is incorrect. The accuracy of this health evaluation model can be judged by combining the overall correctness of these 20 exercisers.

The error produced by the exercise individual health evaluation model for each exercise individual may be determined using the test result data in Table 6; the error is then plotted, and the results are displayed in Figure 7. The exercise-stopping time predicted using the exercise individual health evaluation model in this article was within 20% of the actual exercise-stopping time, as shown by the graph, for all 20 study participants. Even after accounting for the influence of arbitrary human factors, the model’s prediction results’ inaccuracy is still within a tolerable range. A key component in evaluating the correctness of a model is the error threshold (E _Th), which establishes the permissible degree of inaccuracy between forecasts and actual observations while also influencing complexity, sensitivity, and precision. If the error threshold E Th is set at 10%, then the accuracy of this test is 85%, which is quite satisfactory. It can be concluded that the health evaluation model proposed in this article can predict the exercise limit of the exercise individual more accurately and give reasonable exercise advice.

Figure 7

Error chart of the recommended time and actual time of the health assessment model for 20 athletes.

5.3 Health evaluation model

Based on the analysis in the preceding section, a group health evaluation model based on the degree of progression of exercise energy expenditure transfer is used for the exercise health evaluation of exercise groups. A group’s health is evaluated using a model that takes into account social determinants, lifestyle decisions, environmental impacts, and individual health measures. It pinpoints risk factors, health trends, and development opportunities. This model allows for the degree of improvement of exercise effect between exercise groups in the same period, concluding the exercise health evaluation of exercise groups.

Based on data obtained during two consecutive weeks of real-time exercise in both classes of A , B , Tables 7 and 8 show the average weekly exercise energy expenditure for each exercising individual.

Tables 7 and 8 show the average weekly exercise energy expenditure data (in kcal) for all exercising individuals within classes A , B during two consecutive weeks of real-time exercise, respectively.

Divide all the average energy consumption of Class A for a fortnight into four state spaces, i.e., m = 4 , then Δ e = ( 550.58 − 348.32 ) / 4 = 50.565 , at which point the state spaces (noted as states 1, 2, 3, 4) are divided as follows:

The average energy consumption for all 2 weeks in Class B is divided into four state spaces, i.e., m = 4 , then Δ e = ( 348.12 − 236.28 ) / 4 = 27.96 , at which point the state spaces (noted as states 1, 2, 3, 4) are divided as follows:

Based on the above division of the state space, statistics on the transfer of average exercise energy expenditure for all individuals in Class A over the adjacent two weeks can be derived, as shown in Table 9.

Comparable transfer statistics may be produced for the average exercise energy consumption of every person in Class B over the preceding 2 weeks, as shown in Table 10.

Based on the transfer statistics in Tables 9 and 10, the transfer probability matrices for the average exercise energy expenditure of all individuals in Classes A , B for the two adjacent weeks can be calculated separately as follows:

Based on the formula for transfer progression, the transfer progression for classes A , B can be calculated separately:

Based on the results of the above calculations, it can be seen that Class A’s transfer progression was −2.73, indicating that Class A exercise regressed in week 2 compared to week 1, while Class B’s transfer progression was 0.51, indicating that Class B exercise improved in week 2 compared to week 1. It can be judged that during the two consecutive weeks, Class B had a better exercise effect than Class A .