Intelligent motion environment prediction system based on Internet of Things

Abstract

The invention relates to the technical field of data processing, in particular to a sports environment intelligent prediction system based on the Internet of Things, and the system comprises a data collection module which is used for collecting the data of each environment parameter at different moments in each time period under each monitoring period in the environment of a sports place; the environment prediction module is used for predicting the environment parameters, including the steps of determining the environment fluctuation degree, the long-term deflection degree, the evaluation coefficient and the effective weight of each time period, and predicting each environment parameter by using a neural network model; and the evaluation and recommendation module is used for evaluating the environment of the sports place and recommending the sports mode to the user, and comprises the following steps: determining the first suitability, the second suitability and the aerobic recommendation coefficient of the next monitoring period, and evaluating and recommending the sports mode of the user in the sports place. According to the method, the accuracy of environment parameter prediction can be improved, so that the condition that the environment condition is suitable for aerobic exercise is evaluated, the exercise effect of rehabilitation training is improved, and exercise injuries are reduced.

Description

Technical Field

[0001] This application relates to the field of data processing technology, specifically to an intelligent prediction system for motion environment based on the Internet of Things. Background Technology

[0002] With the rapid development and deep integration of artificial intelligence, the Internet of Things and big data technologies, the field of medical rehabilitation is also facing a challenging transformation. In the context of clinical rehabilitation and community-based health care, scientific rehabilitation training, maximizing rehabilitation effects and effectively avoiding secondary injuries have become core issues of common concern.

[0003] Traditional sports environment prediction systems typically only provide static prediction data displays, lacking the ability to intelligently integrate and dynamically predict multi-source heterogeneous data. They struggle to capture instantaneous changes in environmental parameters and abrupt changes in local microclimates, leading to inaccurate predictions of the sports environment in sports venues. Furthermore, they cannot provide users with personalized rehabilitation training suggestions and high-value sports decision support based on environmental changes, causing users to exercise in unsuitable environmental conditions, affecting their health and exercise performance. Summary of the Invention

[0004] To address the aforementioned technical issues, an IoT-based intelligent motion environment prediction system is provided to resolve existing problems.

[0005] The solution to the technical problem of this application is to provide an intelligent prediction system for motion environment based on the Internet of Things, the system comprising: The data acquisition module is used to collect data on each environmental parameter at different times within each monitoring period of the sports venue. These environmental parameters include temperature, humidity, wind speed, ultraviolet intensity, and PM2.5. 2.5 concentration; The environmental prediction module is used to predict environmental parameters, including: Anomaly detection was performed on the data of each environmental parameter at all times within each time period to obtain the abnormal times. The average difference between the data of different environmental parameters at the abnormal times and other times was analyzed. Combined with the periodic changes of humidity and wind speed at the abnormal times, the environmental fluctuation of each time period was determined. Through ultraviolet intensity, PM 2.5 The distribution shift and trend of concentration data at all times within each time period are analyzed. The long-term bias of each time period is calculated, and the evaluation coefficient of each time period is obtained by combining the environmental fluctuation. Based on the interval between each time period in each monitoring cycle and the next monitoring cycle, and the evaluation coefficient, the effective weight of each time period is calculated. Based on the effective weight of the time period to which each moment belongs, a neural network model is used to predict each environmental parameter in the next monitoring cycle. The assessment and recommendation module is used to evaluate the environment of sports venues and recommend exercise methods to users, including: For the next monitoring cycle, analyze the deviation and dispersion of the predicted values ​​for temperature and humidity, and determine the first suitability for the next monitoring cycle; By wind speed and PM 2.5 The differences in the trends of predicted values ​​between concentrations, and PM2.5 concentrations. 2.5 The average level of the predicted values ​​for concentration and ultraviolet intensity is used to determine the second fitness level for the next monitoring period. Combined with the first fitness level, the aerobic recommendation coefficient for the next monitoring period is obtained, and the exercise methods of users in sports venues are evaluated and recommended.

[0006] Preferably, determining the environmental fluctuations for each time period includes: Calculate the ratio between the mean of the data at all abnormal times in each time period and the mean of the data at all other times except for abnormal times for each environmental parameter. Use this ratio as the relative change gradient of each environmental parameter in each time period. Calculate the sum of the relative change gradients of all environmental parameters in each time period. Using the autocorrelation function, based on the data of all abnormal moments corresponding to humidity and wind speed in each time period, the autocorrelation coefficients corresponding to multiple preset lag orders are obtained respectively. The mode of the autocorrelation coefficients corresponding to all preset lag orders is used as the first correlation and the second correlation for each time period. The average value of the first correlation and the second correlation is calculated. The environmental volatility is the product of the average value and the summation.

[0007] Preferably, the calculation of long-term bias for each time period includes: The absolute value of the skewness of the ultraviolet intensity at all times within each time period is compared with the PM... 2.5 The sum of the absolute values ​​of the skewness of the concentration is denoted as the first sum. Calculate the Hurst index and PM2.5 for all anomalous times corresponding to ultraviolet intensity within each time period. 2.5 The sum of the Hurst exponents of the data at all anomalous times corresponding to the concentration is denoted as the second sum. The long-term bias is the product of the first sum and the second sum.

[0008] Preferably, the evaluation coefficient is the normalized result of the product of the environmental volatility and the long-term bias.

[0009] Preferably, the calculation of the effective weights for each time period includes: Obtain the first moment of the next monitoring cycle within each monitoring cycle and record it as the initial prediction moment; use the reciprocal of the time interval between the last moment of each time period within each monitoring cycle and the initial prediction moment as the time weight of each time period. The effective weight is the product of the time weight and the evaluation coefficient.

[0010] Preferably, the prediction of each environmental parameter in the next monitoring period includes: based on the data of each environmental parameter at all times in each monitoring period, using the effective weight of the time period to which each time belongs as the attention weight of the neural network model with the attention mechanism, using the neural network model to predict each environmental parameter, and obtaining the predicted value of each environmental parameter at each time in the next monitoring period.

[0011] Preferably, determining the first suitability for the next monitoring cycle includes: The rate of change between the average predicted temperature at each moment in the next monitoring period and the preset suitable temperature is recorded as the temperature change rate; the rate of change between the average predicted humidity at each moment in the next monitoring period and the preset suitable humidity is recorded as the humidity change rate; the sum of the temperature change rate and the humidity change rate is used as the temperature-humidity deviation. Calculate the interquartile range of the predicted values ​​of temperature and humidity at all times in the next monitoring period, and perform positive mapping on the interquartile ranges respectively; the sum of the positive mapping results of temperature and humidity in the next monitoring period is taken as the temperature and humidity dispersion. The first suitability is the result of negative mapping of the product of the temperature and humidity deviation and the temperature and humidity dispersion.

[0012] Preferably, determining the second suitability for the next monitoring cycle includes: Wind speed and PM2.5 in the next monitoring cycle 2.5 The predicted concentration values ​​at each time point were curve-fitted, and the fitted curves corresponding to wind speed and PM2.5 were plotted separately. 2.5 The predicted values ​​corresponding to the moments when the slope of the fitted curve corresponding to the concentration is greater than 0 are used to form the wind speed growth sequence and the concentration growth sequence; the distance between the wind speed growth sequence and the concentration growth sequence is calculated. Calculate PM2.5 for the next monitoring period 2.5 The ratio of the average predicted concentration at each time point to a preset first reference threshold is denoted as the concentration ratio. The ratio of the average predicted value of ultraviolet intensity at each moment in the next monitoring cycle to a preset second reference threshold is calculated and denoted as the intensity ratio; the sum of the concentration ratio and the intensity ratio is taken as the environmental risk level. The second suitability is the ratio of the distance to the environmental risk level.

[0013] Preferably, the aerobic recommendation coefficient is the normalized result of the product of the first fitness level and the second fitness level.

[0014] Preferably, the evaluation and recommendation of users' exercise methods in the sports venue includes: if the aerobic recommendation coefficient is greater than or equal to a preset threshold, recommending users to perform aerobic exercise in the sports venue; otherwise, recommending users to perform short-term anaerobic exercise in the sports venue.

[0015] This application has at least the following beneficial effects: This application analyzes anomalous changes in various environmental parameters, the differences in average levels between anomalous and non-nominal data, and the periodic variations in humidity and wind speed at anomalous times to determine the environmental volatility over different time periods. Its beneficial effect lies in considering the drastic changes in environmental parameters within sports venues over short periods, and the periodic disturbances in humidity and wind speed caused by the sensitivity to climate change abruptly, reflecting the significant environmental changes within that time period and assessing its instability. The application also calculates the long-term bias for each time period, which is beneficial because it considers factors such as ultraviolet intensity and PM2.5. 2.5 The distribution shifts and long-term trends of concentrations reflect the risks of long-term inhalable particulate matter pollution and ultraviolet radiation exposure to the exercise environment during this period, assessing the existing health risks. Determining the evaluation coefficients for each time period provides a comprehensive assessment of the drastic changes in environmental parameters, reflecting the instability of the exercise environment. Calculating the effective weights for each time period, and using a neural network model based on these weights, predicts each environmental parameter for the next monitoring cycle. This assesses the impact of different time periods approaching the predicted target, and dynamically adjusts the effective weights based on the instability of the exercise environment across different time periods, making the model focus more on data from time periods with a greater impact on the prediction results, thus improving the accuracy of environmental parameter predictions. Determining the first suitability level for the next monitoring cycle reflects the deviation and instability of predicted temperature and humidity values, assessing the suitability of the environment for aerobic exercise in the next monitoring cycle, by analyzing the deviations of predicted temperature and humidity from perceived comfort levels and the dispersion of predicted values. Determining the second suitability level for the next monitoring cycle takes into account PM2.5 levels. 2.5 The increasing trend in concentration is influenced by the diffusion intensity caused by wind speed, and PM2.5 concentration. 2.5The average concentration and intensity of ultraviolet radiation reflect the potential health risks of environmental conditions to users in the next monitoring period, further assessing whether environmental conditions are suitable for aerobic exercise; obtaining the aerobic recommendation coefficient for the next monitoring period, evaluating and recommending exercise methods for users in sports venues. Its beneficial effect is that it comprehensively assesses whether environmental conditions are suitable for aerobic exercise, provides users with scientific and personalized rehabilitation exercise suggestions, avoids potential sports environment risks in advance, ensures that users can carry out rehabilitation exercises in a suitable environment, improves the exercise effect of rehabilitation training, and reduces sports injuries. Attached Figure Description

[0016] The following description, in conjunction with the accompanying drawings, provides a more detailed explanation of the IoT-based intelligent motion environment prediction system of this application.

[0017] Figure 1 This is a block diagram of an IoT-based intelligent prediction system for motion environments provided in one embodiment of this application; Figure 2 This is a flowchart illustrating the steps of an environmental prediction module provided in one embodiment of this application. Figure 3 This is a flowchart illustrating the steps of a method for obtaining an aerobic recommendation coefficient according to an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of the IoT-based intelligent motion environment prediction system proposed in this application, in conjunction with the accompanying drawings and implementation examples, is provided. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0020] Please see Figure 1 The diagram illustrates a block diagram of an IoT-based intelligent prediction system for motion environments provided in an embodiment of this application. The system includes: a data acquisition module, an environment prediction module, and an evaluation and recommendation module.

[0021] The data acquisition module is used to collect data on each environmental parameter at different times within each time period during each monitoring cycle in the sports venue environment.

[0022] With the development of rehabilitation medicine, users need to recover their functions through exercise therapy. During rehabilitation training, the effectiveness and safety depend not only on the user's physical condition and training intensity, but also on the exercise environment. Different rehabilitation training methods are suitable for different exercise environments. When the environment is suitable, such as when the temperature and humidity are comfortable and the air quality is good, long-duration aerobic exercise is recommended to improve cardiopulmonary function and muscle endurance. Conversely, when the temperature and humidity are high and the air quality is poor, short-duration aerobic exercise is recommended to reduce health risks and the probability of complications.

[0023] Based on the above analysis, temperature and humidity sensors, wind speed sensors, ultraviolet sensors, and PM2.5 sensors were used. 2.5 The sensors detect the temperature, humidity, wind speed, UV intensity, and PM2.5 at various times within the sports venue. 2.5 concentration; In this embodiment, the data acquisition frequency is 1Hz. As for other implementation methods, the implementer can set the frequency according to the actual situation.

[0024] The S12SD ultraviolet sensor outputs an analog voltage signal, and the voltage at each moment is obtained through AD conversion. Therefore, the formula for calculating the ultraviolet intensity is: in, For the first Light intensity at any given time This is a preset proportional coefficient. For the first Voltage at time, The preset conversion coefficients are, where, The value is 307 Lux / V. The value is 200 Lux, where Lux is a unit of light intensity.

[0025] Among them, the PM using model GP2Y1023AUOF 2.5 The sensor outputs a PWM signal indicating the concentration of inhalable particulate matter. The PM2.5 concentration is measured by analyzing the proportion of low-level signals in the PWM signal. 2.5 Regarding concentration, it should be noted that the PM2.5 model GP2Y1023AUOF... 2.5 Sensors acquire PM 2.5 The concentration process is a well-known technique and will not be elaborated here.

[0026] The collected data underwent missing value imputation and normalization. Multiple moments were grouped into a time period, and multiple time periods were grouped into a monitoring cycle. Data for each environmental parameter at each moment within each time period of each monitoring cycle was obtained. These environmental parameters included temperature, humidity, wind speed, ultraviolet intensity, and PM2.5. 2.5 concentration; In this embodiment, the time period is 5 minutes and the monitoring cycle is 3 hours. As other implementation methods, the implementer can set them according to the actual situation. Secondly, bilinear interpolation is used to fill missing values, and Z-Score standardization is used for normalization. Bilinear interpolation and Z-Score standardization are well-known technologies and will not be described in detail here. As other implementation methods, the implementer can use other methods of existing technology, such as cubic spline interpolation to fill missing values ​​and maximum-minimum normalization to normalize. As other implementation methods, the implementer can set them according to the actual situation.

[0027] Thus, we have obtained the data for each environmental parameter at each time point within each monitoring cycle.

[0028] The environmental prediction module is used to predict environmental parameters.

[0029] Furthermore, the flowchart of the environmental prediction module provided in this application is as follows: Figure 2 As shown.

[0030] Step 1: Perform anomaly detection on the data of each environmental parameter at all times within each time period, obtain the abnormal times, analyze the average difference between the data of different environmental parameters at the abnormal times and other times, and combine the periodic changes of humidity and wind speed at the abnormal times to determine the environmental fluctuation of each time period.

[0031] Furthermore, the vast amount of environmental parameters in sports venues cannot directly provide valuable information to users and will increase the complexity of intelligent prediction systems for sports environments. Secondly, environmental parameters with high gradient dynamic changes can usually reflect the instantaneous environmental evolution trend and local microclimate change characteristics of sports venues. They should have a higher weight in the prediction process to make more accurate predictions and provide data support for users' sports needs.

[0032] Secondly, the more pronounced the instantaneous environmental evolution trend and local microclimate abrupt changes in the sports venue, the more obvious the gradient changes in temperature and humidity will be due to the abrupt changes in microclimate. Furthermore, humidity and wind speed are more sensitive to changes in air pressure, and their abrupt changes are often accompanied by periodic disturbances, resulting in high autocorrelation between humidity and wind speed. Simultaneously, the ultraviolet intensity and PM2.5 in the sports venue... 2.5Concentrations will deviate from a normal distribution, and UV intensity will correlate with PM2.5 concentration. 2.5 The more significant the long-term trend of a concentration mutation, the better.

[0033] Based on the above analysis, the degree of environmental anomaly is calculated by analyzing the abrupt changes of each environmental parameter in different time periods. Specifically: Anomaly detection is performed on the data of each environmental parameter at all times within each time period to obtain the abnormal times corresponding to each environmental parameter in each time period. In this embodiment, the BG (Bernaola Galvan) segmentation algorithm is used for anomaly detection. The BG segmentation algorithm is a well-known technology and will not be described in detail here. As other implementation methods, implementers may use other methods of existing technology, such as Bayesian mutation point detection algorithms, etc. This embodiment does not impose any special restrictions on this.

[0034] Calculate the mean of the data for each environmental parameter at all times except for abnormal times within each time period, and denote it as the first mean. Calculate the mean of the data for each environmental parameter at all abnormal times in each time period, and denote it as the second mean; The ratio of the second mean to the first mean is used as the relative change gradient of each environmental parameter in each time period; Calculate the sum of the relative change gradients of all environmental parameters in each time period; It should be noted that the larger the relative change gradient, the more drastic the environmental parameter in the sports venue has changed in a short period of time; the larger the cumulative sum, the more unstable the microclimate or local environment changes are during this period.

[0035] Using the autocorrelation function, based on the data of all abnormal moments corresponding to humidity in each time period, the autocorrelation coefficients corresponding to multiple preset lag orders are obtained, and the mode of all preset lag orders corresponding to the autocorrelation coefficients is used as the first correlation of each time period. Based on the data of all abnormal moments corresponding to wind speed in each time period, the autocorrelation coefficients corresponding to multiple preset lag orders are obtained, and the mode of all preset lag orders corresponding to the autocorrelation coefficients is used as the second correlation of each time period. In this embodiment, the autocorrelation function is a well-known technique and will not be described in detail here. Secondly, the lag order is taken as an integer between 1 and 30. As for other implementation methods, the implementer can set it according to the actual situation.

[0036] Calculate the average of the first correlation and the second correlation, and multiply the average by the sum to obtain the environmental volatility for each time period; It should be noted that the greater the first correlation and the greater the second correlation, the more likely the humidity and wind speed are to have significant periodic disturbances and changes during the time period, and the more likely they are to be affected by sudden changes in local microclimate. Secondly, the greater the environmental volatility, the more likely the environmental parameters have undergone significant dynamic changes during the time period, and there may be periodic disturbances, reflecting the instability of environmental changes during the time period.

[0037] Step 2, through ultraviolet intensity, PM 2.5 The distribution shift and trend of concentration data at all times within each time period are analyzed. The long-term bias of each time period is calculated, and the evaluation coefficient of each time period is obtained by combining the environmental fluctuation.

[0038] Secondly, analyze the ultraviolet intensity and PM2.5 concentration at different time periods. 2.5 The distribution deviation and trend of concentration are analyzed to calculate the long-term bias, specifically: The absolute value of the skewness of the ultraviolet intensity at all times within each time period is compared with the PM... 2.5 The sum of the absolute values ​​of the skewness of the concentration is denoted as the first sum. Calculate the Hurst index and PM2.5 for all anomalous times corresponding to ultraviolet intensity within each time period. 2.5 The sum of the Hurst exponents of the data at all anomalous times corresponding to the concentration is denoted as the second sum. It should be noted that the calculation of skewness and Hurst exponent are both well-known techniques, and will not be elaborated upon here.

[0039] The product of the first sum and the second sum is taken as the long-term bias for each time period; It should be noted that the larger the first sum, the higher the ultraviolet intensity and PM2.5 concentration. 2.5 The more asymmetrical the concentration distribution, the more skewed the distribution; the larger the second sum, the higher the UV intensity and PM2.5 concentration. 2.5 The concentration mutations have a strong long-term dependence, and the more significant the persistence of the change trend, the higher the obtained long-term bias, indicating that there is a higher health risk in the exercise environment during this period.

[0040] Furthermore, based on the environmental volatility and the long-term bias, an evaluation coefficient is determined, specifically as follows: The normalized result of the product of the environmental volatility and the long-term bias is used as the evaluation coefficient for each time period. In this embodiment, the sigmoid function is used for normalization. The sigmoid function is a well-known technique and will not be described in detail here. As other implementation methods, implementers may use other methods of the prior art, such as the softmax function, tanh function, etc. This embodiment does not impose any special restrictions on this.

[0041] It should be noted that the larger the evaluation coefficient, the more drastic the changes in environmental parameters during that period, the greater the possibility of sudden changes in microclimate, and the more unstable the motion environment.

[0042] Thus, the evaluation coefficients for each time period are obtained.

[0043] Step 3: Based on the interval between each time period in each monitoring cycle and the next monitoring cycle, and the evaluation coefficient, calculate the effective weight of each time period. Based on the effective weight of the time period to which each moment belongs, use a neural network model to predict each environmental parameter in the next monitoring cycle.

[0044] Furthermore, when making predictions using environmental parameters from each monitoring period, data that is closer to the prediction target has higher reference value. Therefore, based on the last moment of different time periods within each monitoring period and the evaluation coefficient, the effective weights for each time period are obtained, specifically as follows: Obtain the first moment of the next monitoring cycle within each monitoring cycle, and record it as the initial prediction moment; The reciprocal of the time interval between the last moment of each time period within each monitoring cycle and the initial prediction moment is used as the time weight of each time period; The product of the time weight and the evaluation coefficient is used as the effective weight for each time period. Collect data on each environmental parameter at each time point in all time periods within the historical period, and use the above method to calculate the effective weights of each time period within the historical period, form a training set, and train the neural network model with the attention mechanism. In this embodiment, an Attention-LSTM (Long Short-term Memory Networks) model is used for training. Attention-LSTM refers to a model that combines an attention mechanism with an LSTM model. The LSTM model uses Adam as the optimization algorithm and MSE as the loss function. The Attention-LSTM model is a well-known technology and will not be described in detail here. As for other implementation methods, implementers can use other methods of existing technology, such as the Attention-CNN model, etc. This embodiment does not impose any special restrictions on this.

[0045] Based on the data of each environmental parameter at all times in each monitoring cycle, the effective weight of the time period to which each time belongs is used as the attention weight of the neural network model with the attention mechanism. The neural network model is used to predict each environmental parameter and obtain the predicted value of each environmental parameter at each time in the next monitoring cycle. It should be noted that a larger time weight indicates that the time period is closer to the prediction target and has a greater impact on the prediction; a larger evaluation coefficient indicates that the environmental changes within the time period are more significant and have a greater impact on the prediction; a larger effective weight indicates that the environmental parameters within the time period are not only close to the prediction target but also that the environmental changes are significant and have a higher impact on the prediction results; the attention mechanism dynamically allocates different attention weights to different parts of the input data, enabling the model to more effectively capture key features and time steps, thereby improving the accuracy of the prediction.

[0046] The assessment and recommendation module is used to evaluate the environment of sports venues and recommend exercise methods to users.

[0047] Step 1: For the next monitoring cycle, analyze the deviation and dispersion of the predicted values ​​of temperature and humidity, and determine the first suitability for the next monitoring cycle.

[0048] Furthermore, relying solely on the prediction results of each environmental parameter cannot intuitively provide users with personalized suggestions to meet their rehabilitation training needs. It cannot be directly translated into decision recommendations before users begin rehabilitation exercise, and it lacks correlation analysis between the user's exercise needs and the changing characteristics of environmental parameters. For example, regarding the type of rehabilitation exercise, the user's rehabilitation exercise needs can be divided into anaerobic and aerobic exercise. In aerobic exercise, the stability of temperature and humidity is crucial for the user's thermoregulation. When temperature and humidity are within a comfortable range and change steadily, it indicates that the environment is suitable for aerobic exercise, which is beneficial for improving aerobic endurance. If temperature and humidity change too much during exercise, it may cause discomfort to the user and affect the exercise effect. Secondly, ultraviolet intensity and PM2.5... 2.5 Concentration directly affects user health; for example, excessively high UV intensity may cause skin damage, and PM2.5 concentration may also be a concern. 2.5 Excessive concentrations may lead to respiratory illnesses; therefore, when UV intensity and PM2.5 concentrations are high... 2.5 A lower concentration indicates a lower health risk from the environment, making it suitable for aerobic exercise.

[0049] First, by analyzing the changes in the predicted values ​​of temperature and humidity at various times during the next monitoring cycle, the first fitness level is calculated, specifically as follows: The average predicted temperature is the average value of the predicted temperature at each time point in the next monitoring cycle. The average predicted humidity is the mean of the predicted humidity values ​​at each time point in the next monitoring cycle. Calculate the rate of change between the average predicted temperature and the preset suitable temperature, and denot it as the temperature change rate. Calculate the rate of change between the average predicted humidity and the preset suitable humidity, and denot it as the humidity change rate; In this embodiment, the suitable temperature range for human movement is 17°C. ~24 The suitable humidity range is 30%~50%, therefore, the preset suitable temperature is 22. The preset suitable humidity value is 40%, but for other implementation methods, the implementer can set it according to the actual situation; secondly, the temperature change rate. The calculation process is as follows: ,in, Indicates the average predicted temperature. If the preset suitable temperature is indicated, then the humidity change rate is adopted in relation to the temperature change rate. The same method is used for calculation. It should be noted that the calculation of the rate of change is a well-known technique and will not be elaborated here.

[0050] The sum of the temperature change rate and the humidity change rate is taken as the temperature-humidity deviation. Calculate the interquartile range of the predicted temperature at all times during the next monitoring period, and perform a positive mapping on the interquartile range; Calculate the interquartile range of the predicted humidity values ​​at all times during the next monitoring cycle, and perform a positive mapping on the interquartile range; The sum of the positive mapping results of temperature and humidity in the next monitoring cycle is taken as the temperature and humidity dispersion. In this embodiment, the calculation of the interquartile range (IQR) is a well-known technique and will not be elaborated further. Secondly, the specific process of the positive mapping is as follows: a positive mapping is performed using a logarithmic function. Let the interquartile range be denoted as IQR. The result is taken as the result of the positive mapping, where, It is a logarithmic function with the natural constant as the base. Through the process of positive mapping, the dimensional problem between temperature and humidity is eliminated.

[0051] The product of the temperature and humidity deviation and the temperature and humidity dispersion is negatively mapped and used as the first suitability for the next monitoring cycle; In this embodiment, the specific process of negative mapping is as follows: the reciprocal of the product of the temperature and humidity deviation and the temperature and humidity dispersion is taken as the result of negative mapping. In other implementations, the implementer can use an exponential function for negative mapping. Let's assume the product of the temperature and humidity deviation and the temperature and humidity dispersion is denoted as... ,Will The result is used as the result of the negative mapping, where, It is an exponential function with the natural constant as the base.

[0052] It should be noted that the greater the temperature and humidity deviation, the higher the degree to which the predicted temperature and humidity in the next monitoring period deviates from the comfortable temperature and humidity for the human body, indicating that it is less suitable for aerobic exercise. The greater the temperature and humidity dispersion, the more dispersed and unstable the predicted temperature and humidity distribution in the next monitoring period, indicating that it is less suitable for aerobic training. The lower the first suitability, the less suitable the environmental conditions are for aerobic training in the next monitoring period. Conversely, the higher the first suitability, the more suitable the environmental conditions are for aerobic training in the next monitoring period.

[0053] Step 2, by measuring wind speed and PM2.5 2.5 The differences in the trends of predicted values ​​between concentrations, and PM2.5 concentrations. 2.5 The average level of the predicted values ​​for concentration and ultraviolet intensity is used to determine the second fitness level for the next monitoring period. Combined with the first fitness level, the aerobic recommendation coefficient for the next monitoring period is obtained, and the exercise methods of users in sports venues are evaluated and recommended.

[0054] Secondly, analyze PM2.5 levels in the next monitoring cycle. 2.5 The diffusion of concentration affected by wind speed changes, and PM2.5 concentration. 2.5 The second fitness is calculated based on the average concentration and UV intensity, specifically as follows: Curve fitting is performed on the predicted wind speed values ​​at each time point in the next monitoring period. The slope of the fitted curve corresponding to the wind speed at each time point is calculated. The predicted values ​​corresponding to all times with a slope greater than 0 are used to form a wind speed growth sequence. For the next monitoring period, PM 2.5 The predicted concentration values ​​at each time point were curve-fitted to calculate PM2.5. 2.5 The slope of the fitted curve corresponding to the concentration at each time point, and the predicted values ​​corresponding to all times when the slope is greater than 0, are used to form the concentration growth sequence; It should be noted that the least squares method is used for curve fitting. The least squares method and the calculation of the slope of the curve are well-known techniques and will not be elaborated here.

[0055] Calculate the distance between the wind speed growth sequence and the concentration growth sequence; In this embodiment, the distance is measured by calculating the DTW distance between the wind speed growth sequence and the concentration growth sequence. The DTW distance is a well-known technique and will not be described in detail here. As other implementation methods, implementers may use other methods of the prior art, such as Euclidean distance, etc. This embodiment does not impose any special restrictions on this.

[0056] Calculate PM2.5 for the next monitoring period 2.5 The ratio of the average predicted concentration at each time point to a preset first reference threshold is denoted as the concentration ratio. The ratio of the average predicted value of ultraviolet intensity at each time point in the next monitoring cycle to the preset second reference threshold is denoted as the intensity ratio. The sum of the concentration ratio and the intensity ratio is taken as the environmental risk level; In this embodiment, due to PM 2.5 When the concentration does not exceed 35 micrograms per cubic meter, the air quality is excellent and the ultraviolet index does not exceed 3, making it safe to carry out outdoor activities. Therefore, the preset first reference threshold value is 35 and the preset second reference threshold value is 3. As other implementation methods, implementers can set them according to the actual situation.

[0057] The ratio of the distance to the environmental risk level is used as the second suitability for the next monitoring cycle; It should be noted that the greater the distance, the closer the predicted wind speed is to the PM2.5 concentration in the next monitoring period. 2.5 The predicted concentration trends vary considerably, reflecting the diverse trends in PM2.5 concentrations. 2.5 The less the diffusion intensity is affected by wind speed, the more suitable the environmental conditions are for aerobic exercise; the lower the environmental risk level, the better the PM2.5 concentration. 2.5 Lower average concentrations and UV intensities indicate lower potential health risks to users and are more suitable for aerobic exercise; a higher second suitability indicates more suitable environmental conditions for aerobic exercise.

[0058] Furthermore, based on the first fitness level and the second fitness level, an aerobic recommendation coefficient is determined, specifically as follows: The normalized result of the product of the first fitness level and the second fitness level is used as the aerobic recommendation coefficient for the next monitoring cycle. In this embodiment, the sigmoid function is used for normalization. The sigmoid function is a well-known technique and will not be described in detail here. As other implementation methods, implementers may use other methods of the prior art, such as the softmax function, tanh function, etc. This embodiment does not impose any special restrictions on this.

[0059] It should be noted that the higher the aerobic recommendation coefficient, the more recommended aerobic rehabilitation training should be for the user. The flowchart of the method for obtaining the aerobic recommendation coefficient provided in this application is as follows: Figure 3 As shown.

[0060] Furthermore, based on the aforementioned aerobic recommendation coefficient, the user's exercise methods are evaluated and recommended, specifically as follows: If the aerobic recommendation coefficient is greater than or equal to the preset threshold, the user is recommended to do aerobic exercise in this sports venue; otherwise, the user is recommended to do short-term anaerobic exercise in this sports venue. In this embodiment, the preset threshold value is 0.6. As for other implementation methods, the implementer can set it according to the actual situation.

[0061] It should be noted that a higher aerobic recommendation coefficient indicates that the temperature and humidity of the exercise venue are suitable and stable in the next monitoring period, and that the ultraviolet intensity and concentration of inhalable particulate matter pose a lower health risk to prolonged aerobic exercise. Conversely, if the temperature and humidity of the exercise venue are unsuitable for aerobic exercise in the next monitoring period, users may experience higher stress on their body temperature regulation mechanisms when engaging in aerobic exercise, and the ultraviolet intensity and concentration of inhalable particulate matter in the exercise venue pose a higher health risk to prolonged aerobic exercise.

[0062] Therefore, the Internet of Things (IoT) connects sensors, devices, and networks to achieve real-time data collection, transmission, and processing. The data collection module transmits the collected data to the IoT platform for storage via the TCP / IP protocol. The environmental prediction module makes predictions by calling the data, and the evaluation and recommendation module evaluates and recommends based on the prediction results. The recommended movement methods are then uploaded to the IoT platform, and users can access the IoT platform through communication devices to obtain the recommendation results.

[0063] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0064] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application, without departing from the content of the technical solution of this application, shall fall within the protection scope of the technical solution of this application.

Claims

1. A motion environment intelligent prediction system based on the Internet of Things, characterized in that, The system includes: The data acquisition module is used to collect data on each environmental parameter at different times within each monitoring period of the sports venue. These environmental parameters include temperature, humidity, wind speed, ultraviolet intensity, and PM2.

5. 2.5 concentration; The environmental prediction module is used to predict environmental parameters, including: Anomaly detection was performed on the data of each environmental parameter at all times within each time period to obtain the abnormal times. The average difference between the data of different environmental parameters at the abnormal times and other times was analyzed. Combined with the periodic changes of humidity and wind speed at the abnormal times, the environmental fluctuation of each time period was determined. Through ultraviolet intensity, PM 2.5 The distribution shift and trend of concentration data at all times within each time period are analyzed. The long-term bias of each time period is calculated, and the evaluation coefficient of each time period is obtained by combining the environmental fluctuation. Based on the interval between each time period in each monitoring cycle and the next monitoring cycle, and the evaluation coefficient, the effective weight of each time period is calculated. Based on the effective weight of the time period to which each moment belongs, a neural network model is used to predict each environmental parameter in the next monitoring cycle. The assessment and recommendation module is used to evaluate the environment of sports venues and recommend exercise methods to users, including: For the next monitoring cycle, analyze the deviation and dispersion of the predicted values ​​for temperature and humidity, and determine the first suitability for the next monitoring cycle; By wind speed and PM 2.5 The differences in the trends of predicted values ​​between concentrations, and PM2.5 concentrations. 2.5 The average level of the predicted values ​​for concentration and ultraviolet intensity is used to determine the second fitness level for the next monitoring period. Combined with the first fitness level, the aerobic recommendation coefficient for the next monitoring period is obtained, and the exercise methods of users in sports venues are evaluated and recommended.

2. The IoT-based intelligent motion environment prediction system as described in claim 1, characterized in that, Determining the environmental fluctuations for each time period includes: Calculate the ratio between the mean of the data at all abnormal times in each time period and the mean of the data at all other times except for abnormal times for each environmental parameter. Use this ratio as the relative change gradient of each environmental parameter in each time period. Calculate the sum of the relative change gradients of all environmental parameters in each time period. Using the autocorrelation function, based on the data of all abnormal moments corresponding to humidity and wind speed in each time period, the autocorrelation coefficients corresponding to multiple preset lag orders are obtained respectively. The mode of the autocorrelation coefficients corresponding to all preset lag orders is used as the first correlation and the second correlation for each time period. The average value of the first correlation and the second correlation is calculated. The environmental volatility is the product of the average value and the summation.

3. The IoT-based intelligent motion environment prediction system as described in claim 1, characterized in that, The calculation of long-term biases for each time period includes: The absolute value of the skewness of the ultraviolet intensity at all times within each time period is compared with the PM... 2.5 The sum of the absolute values ​​of the skewness of the concentration is denoted as the first sum. Calculate the Hurst index and PM2.5 for all anomalous times corresponding to ultraviolet intensity within each time period. 2.5 The sum of the Hurst exponents of the data at all anomalous times corresponding to the concentration is denoted as the second sum. The long-term bias is the product of the first sum and the second sum.

4. The IoT-based intelligent motion environment prediction system as described in claim 1, characterized in that, The evaluation coefficient is the normalized result of the product of the environmental volatility and the long-term bias.

5. The IoT-based intelligent motion environment prediction system as described in claim 1, characterized in that, The calculation of the effective weights for each time period includes: Obtain the first moment of the next monitoring cycle within each monitoring cycle and record it as the initial prediction moment; use the reciprocal of the time interval between the last moment of each time period within each monitoring cycle and the initial prediction moment as the time weight of each time period. The effective weight is the product of the time weight and the evaluation coefficient.

6. The IoT-based intelligent motion environment prediction system as described in claim 1, characterized in that, The prediction of each environmental parameter in the next monitoring period includes: based on the data of each environmental parameter at all times in each monitoring period, using the effective weight of the time period to which each time belongs as the attention weight of the neural network model with the attention mechanism, using the neural network model to predict each environmental parameter, and obtaining the predicted value of each environmental parameter at each time in the next monitoring period.

7. The IoT-based intelligent motion environment prediction system as described in claim 1, characterized in that, The determination of the first suitability for the next monitoring cycle includes: The rate of change between the average predicted temperature at each moment in the next monitoring period and the preset suitable temperature is recorded as the temperature change rate; the rate of change between the average predicted humidity at each moment in the next monitoring period and the preset suitable humidity is recorded as the humidity change rate; the sum of the temperature change rate and the humidity change rate is used as the temperature-humidity deviation. Calculate the interquartile range of the predicted values ​​of temperature and humidity at all times in the next monitoring period, and perform positive mapping on the interquartile ranges respectively; the sum of the positive mapping results of temperature and humidity in the next monitoring period is taken as the temperature and humidity dispersion. The first suitability is the result of negative mapping of the product of the temperature and humidity deviation and the temperature and humidity dispersion.

8. The IoT-based intelligent motion environment prediction system as described in claim 1, characterized in that, The determination of the second suitability for the next monitoring cycle includes: Wind speed and PM2.5 in the next monitoring cycle 2.5 The predicted concentration values ​​at each time point were curve-fitted, and the fitted curves corresponding to wind speed and PM2.5 were plotted separately. 2.5 The predicted values ​​corresponding to the moments when the slope of the fitted curve corresponding to the concentration is greater than 0 are used to form the wind speed growth sequence and the concentration growth sequence; the distance between the wind speed growth sequence and the concentration growth sequence is calculated. Calculate PM2.5 for the next monitoring period 2.5 The ratio of the average predicted concentration at each time point to a preset first reference threshold is denoted as the concentration ratio. The ratio of the average predicted value of ultraviolet intensity at each moment in the next monitoring cycle to a preset second reference threshold is calculated and denoted as the intensity ratio; the sum of the concentration ratio and the intensity ratio is taken as the environmental risk level. The second suitability is the ratio of the distance to the environmental risk level.

9. The IoT-based intelligent motion environment prediction system as described in claim 1, characterized in that, The aerobic recommendation coefficient is the normalized result of the product of the first fitness level and the second fitness level.

10. The IoT-based intelligent motion environment prediction system as described in claim 1, characterized in that, The evaluation and recommendation of users' exercise methods in sports venues includes: if the aerobic recommendation coefficient is greater than or equal to a preset threshold, recommending users to perform aerobic exercise in the sports venue; otherwise, recommending users to perform short-term anaerobic exercise in the sports venue.

Images