Police physical training examination management system

Abstract

The invention relates to the technical field of assessment management, and discloses a police physical training assessment management system which comprises a training data acquisition module, a correlation analysis module, a feature extraction module, a model training module and a training scheme generation module. The method comprises the steps of obtaining data relevance, determining strong relevance data according to the data relevance, performing feature extraction on the strong relevance data to obtain data features of the strong relevance data, and determining data tags of the data features according to assessment score data. Performing model training on a pre-constructed portrait model according to the data features, the data labels and a preset multi-task loss function to obtain a trained portrait model, generating a physical ability hologram of the police according to the trained portrait model, and generating a training scheme of the police according to the physical ability hologram, physical training is carried out on the police according to the training scheme, and the physical training assessment management efficiency of the police is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of assessment management, and in particular to a police physical training assessment management system. Background Art

[0002] It is very important to improve the efficiency of police physical training and assessment management. Traditional police physical training management has problems of inefficiency and information lag, scattered data and failure to reflect the actual status of police officers in real time, resulting in poor training results.

[0003] Police officers’ physical fitness data and training results are often stored in multiple systems and manually recorded, lacking centralized management and making it impossible to track officers’ physical fitness in real time. Traditional training plans are often “one size fits all” and fail to be adjusted according to the physical fitness levels and needs of different police officers, resulting in unsatisfactory training results. At the same time, a large amount of manual recording and analysis of data is inefficient, prone to errors, and cannot reflect the actual physical fitness of police officers in a timely manner. Summary of the invention

[0004] The present invention provides a police physical training assessment management system, the main purpose of which is to solve the problem of low efficiency in police physical training assessment management.

[0005] To achieve the above-mentioned purpose, the present invention provides a police physical training assessment management system, which includes a training data collection module, a correlation analysis module, a feature extraction module, a model training module and a training program generation module, wherein:

[0006] The training data collection module is used to collect multi-source training data of police officers, wherein the multi-source training data includes: training project completion time, training intensity index, physical function parameters and assessment performance data;

[0007] The correlation analysis module is used to perform a dual correlation analysis on the multi-source training data to obtain data correlation of the multi-source training data;

[0008] The feature extraction module is used to determine the strongly correlated data in the multi-source training data according to the data correlation, perform feature extraction on the strongly correlated data, and obtain data features of the strongly correlated data;

[0009] The model training module is used to determine the data label of the data feature according to the assessment score data, and to perform model training on the pre-built portrait model according to the data feature, the data label and the preset multi-task loss function to obtain a trained portrait model, wherein the preset multi-task loss function for:

[0010]

[0011] Among them, Y1 is the target value of task one, is the model's predicted value for task one, and Y2 is the target value for task two. is the model's predicted value for task 2, α is the weight coefficient for task 1, and β is the weight coefficient for task 2. is the loss function of task one, is the loss function of task 2;

[0012] The training program generation module is used to generate a physical hologram of the policeman according to the trained portrait model, generate a training program for the policeman according to the physical hologram, and perform physical training on the policeman according to the training program.

[0013] Optionally, the training data collection module collects multi-source training data of police officers, including:

[0014] Collect the time it takes police officers to complete each physical training item one by one;

[0015] Collect training intensity indicators for each physical training;

[0016] Collect the physical function parameters of police officers one by one;

[0017] Collect the assessment performance data of police officers in physical training one by one.

[0018] Optionally, when the correlation analysis module performs a dual correlation analysis on the multi-source training data to obtain data correlation of the multi-source training data, the correlation analysis module includes:

[0019] Calculate item correlations for different training items;

[0020] Calculate the functional relevance of training programs and physical performance parameters;

[0021] A correlation matrix of the multi-source training data is generated according to the item correlation and the function correlation, thereby determining the data relevance of the multi-source training data.

[0022] Optionally, when the correlation analysis module calculates the item correlation of different training items, it includes:

[0023] The project relevance of different training projects is calculated using a preset relevance algorithm, wherein the preset relevance algorithm is:

[0024]

[0025] Among them, r ij is the item correlation between the i-th and j-th training items, N is the total number of data in the training items, k is the data identifier, and Xik is the value of the kth data of the i-th training item, is the data mean of the i-th training item, is the data mean of the i-th training item, X jk is the value of the kth data of the jth training item, i is the identifier of the training item, and j is the identifier of the training item.

[0026] Optionally, when the feature extraction module determines the strongly correlated data in the multi-source training data according to the data correlation, it includes:

[0027] Based on the data correlation, the training items and body function parameters that are strongly correlated with the assessment performance data are determined as strongly correlated data.

[0028] Optionally, when the feature extraction module extracts features from the strongly correlated data to obtain data features of the strongly correlated data, the feature extraction module includes:

[0029] The strongly correlated data are vector-converted, and the vector-converted strongly correlated data are vector-concatenated to obtain data features of the strongly correlated data.

[0030] Optionally, when the model training module performs model training on the pre-built portrait model according to the data features, the data labels and a preset multi-task loss function to obtain a trained portrait model, the model training module includes:

[0031] Establishing a corresponding relationship between the data feature and the data label;

[0032] Generate a pre-built training set of the portrait model according to the corresponding relationship, the data label and the data feature;

[0033] The pre-built portrait model is trained using the training set and a preset multi-task loss function.

[0034] Optionally, when the training program generation module generates the policeman's physical hologram according to the trained portrait model, the training program generation module includes:

[0035] Based on the trained portrait model, data mapping is performed on the pre-collected police data to obtain mapping data of the police data;

[0036] A body-powered hologram of the police officer is generated based on the mapping data.

[0037] Optionally, when the training program generating module generates the training program for the police according to the physical energy hologram, the training program generating module includes:

[0038] collecting real-time performance data of the police officers during training;

[0039] Generating the physical state of the police officer according to the real-time functional data and the preset training intensity;

[0040] The training intensity of the physical fitness hologram is adjusted according to the physical fitness state to obtain the training plan for the police officer.

[0041] Optionally, when the training program generation module adjusts the training intensity of the physical fitness hologram according to the physical fitness state to obtain the training program for the police, it includes:

[0042] The training intensity of the physical fitness hologram is adjusted according to the physical fitness state and a preset adjustment formula to obtain the training plan for the police, wherein the preset adjustment formula is:

[0043]

[0044] Among them, S′ t is the adjusted training intensity, S t is the preset training intensity, t is the adjustment coefficient, HR t is the physical state, HR max It is the maximum physical state.

[0045] The present invention significantly improves the efficiency of assessment management through comprehensive data collection, in-depth correlation analysis, precise feature extraction and intelligent personalized training program formulation. First, the data collection module comprehensively records various data of the police during the training process, including the completion time of the training project, training intensity, physical function parameters and assessment results, etc., providing a sufficient source of information for subsequent analysis and decision-making. Then, the correlation analysis module performs double correlation analysis on multi-source data, deeply explores the relationship between different training projects and the physical performance of the police, finds out the key factors affecting physical fitness improvement, and thus provides a scientific basis for the optimization of the training program. Then, the feature extraction module extracts the features reflecting the police's physical fitness according to the analysis results. The key features of development can help better optimize the training content and methods. Through the model training module, the system combines the assessment performance data and adopts a multi-task loss function to adjust the model prediction effect to ensure that the training plan is more accurate in personalization and effect. According to the physical fitness hologram of each policeman, the system can generate personalized training plans to avoid a one-size-fits-all training method, ensuring that each policeman's training plan can be tailored to their personal physical status and development needs, thereby improving the effectiveness and pertinence of training. Finally, through intelligent decision-making support and real-time feedback, managers can adjust training strategies in a timely manner according to the analysis results of the system, and continuously optimize training plans, thereby improving the effectiveness and management efficiency of police physical training. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A system architecture diagram of a police physical training and assessment management system provided in one embodiment of the present invention.

[0047] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "said" and "the" used in the embodiments of the present invention are also intended to include plural forms, unless the context clearly indicates other meanings, and "multiple" generally includes at least two.

[0050] As used herein, the words “if” or “if” may be interpreted as “at the time of” or “when” or “in response to determining” or “in response to detecting”, depending on the context. Similarly, the phrases “if it is determined” or “if (stated condition or event) is detected” may be interpreted as “when it is determined” or “in response to determining” or “when detecting (stated condition or event)” or “in response to detecting (stated condition or event)”, depending on the context.

[0051] In fact, the server-side device deployed by the police physical training assessment management system may be composed of one or more devices. The above-mentioned police physical training assessment management system can be implemented as: business instance, virtual machine, hardware device. For example, the police physical training assessment management system can be implemented as a business instance deployed on one or more devices in the cloud node. In simple terms, the police physical training assessment management system can be understood as a software deployed on the cloud node, which is used to provide the police physical training assessment management system for each user terminal. Alternatively, the police physical training assessment management system can also be implemented as a virtual machine deployed on one or more devices in the cloud node. The virtual machine is installed with application software for managing each user terminal. Alternatively, the police physical training assessment management system can also be implemented as a server composed of many hardware devices of the same or different types, and one or more hardware devices are set to provide the police physical training assessment management system for each user terminal.

[0052] In terms of implementation, the police physical training assessment management system and the user end are adapted to each other. That is, if the police physical training assessment management system is an application installed on the cloud service platform, the user end is a client that establishes a communication connection with the application; or if the police physical training assessment management system is implemented as a website, the user end is implemented as a web page; or if the police physical training assessment management system is implemented as a cloud service platform, the user end is implemented as a small program in an instant messaging application.

[0053] like Figure 1 1 is a system architecture diagram of a police physical training assessment management system provided by an embodiment of the present invention.

[0054] The police physical training assessment management system 100 of the present invention can be set in a cloud server. In terms of implementation, it can be used as one or more service devices, or as an application installed in the cloud (such as a server of a mobile service operator, a server cluster, etc.), or it can be developed as a website. According to the functions implemented, the police physical training assessment management system 100 can include a training data acquisition module 101, a correlation analysis module 102, a feature extraction module 103, a model training module 104 and a training program generation module 105. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0055] In the embodiment of the present invention, in the police physical training assessment management system, each of the above modules can be implemented independently and called with other modules. The call here can be understood as a module that can connect to multiple modules of another type and provide corresponding services to the multiple modules connected to it. For example, the sharing evaluation module can call the same information collection module to obtain the information collected by the information collection module. Based on the above characteristics, in the police physical training assessment management system provided by the embodiment of the present invention, the scope of application of the police physical training assessment management system architecture can be adjusted by adding modules and directly calling them without modifying the program code, so as to achieve cluster-based horizontal expansion, so as to achieve the purpose of quickly and flexibly expanding the police physical training assessment management system. In actual applications, the above modules can be set in the same device or different devices, or they can be set in virtual devices, such as service instances in cloud servers.

[0056] The following is a description of each component and specific workflow of the police physical training assessment management system in conjunction with a specific embodiment:

[0057] The training data collection module 101 is used to collect multi-source training data of police officers, wherein the multi-source training data includes: training project completion time, training intensity index, physical function parameters and assessment performance data.

[0058] In the embodiment of the present invention, the training data collection module collects multi-source training data of police officers, including:

[0059] Collect the time it takes police officers to complete each physical training item one by one;

[0060] Collect training intensity indicators for each physical training;

[0061] Collect the physical function parameters of police officers one by one;

[0062] Collect the assessment performance data of police officers in physical training one by one.

[0063] In detail, the completion time of each police officer in various physical training exercises is recorded, which is mainly used to evaluate the police's physical performance in terms of speed, endurance, etc.

[0064] Furthermore, devices such as stopwatches, smart watches, treadmills, heart rate monitors, etc. are used to collect data on each police officer while performing specific training (such as 1000-meter running, push-ups, sit-ups, etc.), that is, to record their completion time.

[0065] For example: Police number: P001; Training item: 1000-meter run; Completion time: 300 seconds. Through these data, the speed and endurance of the police can be intuitively evaluated.

[0066] In detail, collecting the training intensity index of each physical training can quantify the intensity level of each training, so that in subsequent analysis it can be determined whether the training has reached a certain intensity or whether it needs to be adjusted.

[0067] Furthermore, the intensity of training can be determined by the police's heart rate, acceleration sensor, motion status and other data during training. For example, the intensity of training can be divided into three levels: "low", "medium" and "high". Low intensity: easy to carry out without obvious fatigue; medium intensity: feel a certain degree of fatigue, but can continue; high intensity: significant fatigue appears, close to the limit.

[0068] For example: Police number: P001; Training item: 1000-meter run; Training intensity: high. Through these intensity indicators, the training intensity of the police can be controlled more accurately to avoid overtraining or insufficient training.

[0069] In detail, the police's body function parameters are collected one by one, and the police's physiological data, such as heart rate, blood pressure, weight, etc., are collected to help analyze their health status and physical potential.

[0070] Furthermore, data collection is performed using smart wearable devices, body fat scales, blood pressure monitors, etc.

[0071] For example: the maximum heart rate is usually obtained through heart rate monitoring during exercise; the minimum heart rate is usually measured at rest; weight and height refer to basic body parameters; blood pressure is measured using an electronic blood pressure monitor.

[0072] For example: Police number: P001; Maximum heart rate: 180 beats / minute; Minimum heart rate: 110 beats / minute; Weight: 75kg; Blood pressure: 120 / 80mmHg. These data will help judge the physical adaptability of the police and their safety during training.

[0073] In detail, the assessment performance data of police officers in physical training are collected one by one, and the comprehensive results of each police physical training are recorded as the basis for their training evaluation.

[0074] Furthermore, the data collection system automatically records the completion status and scores of each physical fitness test. The comprehensive evaluation results of each physical fitness training, such as the completion time of running, the number of push-ups, etc. The scores can be divided into "excellent", "good", "qualified" or "unqualified" grades according to the physical fitness test scoring standards.

[0075] For example: Police number: P001; Training item: 1000-meter run; Assessment result: Pass; Training item: Push-ups; Assessment result: Excellent. These results will be used to assess the physical fitness level of police officers and generate personalized training plans.

[0076] In detail, the multi-source training data includes: training project completion time, training intensity index, physical function parameters and assessment performance data.

[0077] Specifically, the training data collection module collects multiple training data of police officers in a detailed and comprehensive manner, providing the necessary foundation for subsequent data analysis, model training, and generation of personalized training programs. By collecting training time, training intensity, physical function data, and assessment results item by item, a comprehensive assessment of the physical condition of police officers can be conducted, and combined with artificial intelligence or machine learning models, the training content can be further optimized to ensure the scientificity and personalization of the training.

[0078] The correlation analysis module 102 is used to perform a dual correlation analysis on the multi-source training data to obtain data correlation of the multi-source training data.

[0079] In the embodiment of the present invention, when the correlation analysis module performs a dual correlation analysis on the multi-source training data to obtain the data correlation of the multi-source training data, it includes:

[0080] Calculate item correlations for different training items;

[0081] Calculate the functional relevance of training programs and physical performance parameters;

[0082] A correlation matrix of the multi-source training data is generated according to the item correlation and the function correlation, thereby determining the data relevance of the multi-source training data.

[0083] In detail, assume there are the following two types of data:

[0084] The training project data include: Training project 1 (running): the results are: 12 minutes, 11 minutes, 13 minutes respectively; Training project 2 (push-ups): the results are: 30 times, 32 times, 29 times respectively; Training project 3 (sit-ups): the results are: 40 times, 42 times, 38 times respectively.

[0085] Physical function parameter data include: heart rate (average heart rate measured after each training): 150 beats / minute, 145 beats / minute, 148 beats / minute.

[0086] In detail, when calculating the project relevance of different training projects, the relevance analysis module includes:

[0087] The project relevance of different training projects is calculated using a preset relevance algorithm, wherein the preset relevance algorithm is:

[0088]

[0089] Among them, r ij is the item correlation between the i-th and j-th training items, N is the total number of data in the training items, k is the data identifier, and X ik is the value of the kth data of the i-th training item, is the data mean of the i-th training item, is the data mean of the i-th training item, X jk is the value of the kth data of the jth training item, i is the identifier of the training item, and j is the identifier of the training item.

[0090] Specifically, the correlation between the training items is calculated. For example, the correlation between the training item 1 (running) and the training item 2 (push-ups) is calculated.

[0091] First, calculate the mean for each training item:

[0092] Mean values ​​for training event 1 (running):

[0093]

[0094] Mean values ​​for Exercise 2 (Push-ups):

[0095]

[0096] Second, calculate the difference between each data point and the mean

[0097] Differences for Training 1 (Running):

[0098]

[0099] Differences for Exercise 2 (Push-ups):

[0100]

[0101] Then, the product of each pair of difference terms is calculated, where

[0102] The first set of data:

[0103] The second set of data:

[0104] The third set of data:

[0105] Second, the sum of squares of the difference terms is calculated, where

[0106] The sum of squares for training item 1 (running):

[0107]

[0108] The sum of squares of training item 2 (push-ups):

[0109]

[0110] Finally, calculate the correlation coefficient between training item 1 (running) and training item 2 (push-ups):

[0111]

[0112] In summary, the correlation coefficient between training item 1 (running) and training item 2 (push-ups) is approximately -0.983, indicating that there is a very strong negative correlation between the two training items.

[0113] In detail, the body function parameters are heart rate, weight, height and blood pressure.

[0114] In detail, in addition to calculating the correlation between the training items, it is also necessary to analyze the relationship between the training items and the police's physical function parameters (such as heart rate, blood pressure, weight, etc.). This part of the calculation can also use a similar method, that is, by calculating the correlation between the training item data and the physical function parameter data.

[0115] In detail, based on the calculated correlation coefficients, a simple correlation matrix can be generated. Assuming that only four variables, running, push-ups, sit-ups, and heart rate, are considered, the correlation matrix is ​​as follows:

[0116]

[0117]

[0118] Furthermore, based on the correlation matrix generated above, the strength of the correlation between different data can be evaluated. This process usually includes: setting a correlation threshold (such as 0.8 or 0.9). When the correlation coefficient is greater than this value, it is considered that there is a strong correlation between these training items or physical function parameters. Based on the correlation matrix, cluster analysis can be performed to group training items and physical function parameters with high correlation, and further discover potential training patterns or rules.

[0119] The feature extraction module 103 is used to determine the strongly correlated data in the multi-source training data according to the data correlation, perform feature extraction on the strongly correlated data, and obtain data features of the strongly correlated data.

[0120] In the embodiment of the present invention, when the feature extraction module determines the strongly correlated data in the multi-source training data according to the data correlation, it includes:

[0121] Based on the data correlation, the training items and body function parameters that are strongly correlated with the assessment performance data are determined as strongly correlated data.

[0122] In detail, assume that the police's assessment results are as follows:

[0123] police 100m running performance (seconds) Number of push-ups done Endurance running results (minutes / 5 km) A 12.5 50 24 B 11.8 60 22 C 13.2 45 26

[0124] In detail, the training project data is as follows:

[0125]

[0126] In detail, the body function parameter data are as follows:

[0127]

[0128]

[0129] In detail, based on the data correlation, the training items and physical function parameters that are strongly correlated with the assessment performance data are determined as strongly correlated data, as shown in the table:

[0130] Related items Example of results Running training and 100-meter running results Policeman B performed best with a longer and faster run. Strength training and number of push-ups Policeman B trained more and did the most push-ups Physical function and endurance running performance Policeman C has a lower body fat percentage and the best endurance running performance

[0131] In general, through these tables, we can intuitively see how different training programs and physical function data affect the assessment results of police officers, while revealing the correlation between different factors.

[0132] For example: If the analysis results show that the distance and speed of running training have a strong positive correlation with the 100-meter running performance, and the body fat percentage and maximum oxygen uptake have a strong positive correlation with the endurance running performance, then the feature extraction module will use "running training distance", "average running training speed", "body fat percentage" and "maximum oxygen uptake" as strongly correlated data.

[0133] In the embodiment of the present invention, when the feature extraction module extracts features from the strongly correlated data to obtain data features of the strongly correlated data, the feature extraction module includes:

[0134] The strongly correlated data are vector-converted, and the vector-converted strongly correlated data are vector-concatenated to obtain data features of the strongly correlated data.

[0135] In detail, assuming that after data correlation analysis, the strongly correlated data selected include "running training distance", "running speed" and "body fat percentage", which are numerical data respectively. After vectorization and standardization processing: running training distance: [3.5, 4.2] (indicating two different training data), running speed: [5.6, 6.0], body fat percentage: [0.22]. After vector splicing operation, the final data features are: [3.5, 4.2, 5.6, 6.0, 0.22].

[0136] The model training module 104 is used to determine the data labels of the data features according to the assessment performance data, and to perform model training on the pre-constructed portrait model according to the data features, the data labels and a preset multi-task loss function to obtain a trained portrait model.

[0137] In detail, the preset multi-task loss function for:

[0138]

[0139] Among them, Y1 is the target value of task one, is the model's predicted value for task one, and Y2 is the target value for task two. is the model's predicted value for task 2, α is the weight coefficient for task 1, and β is the weight coefficient for task 2. is the loss function of task one, is the loss function of task 2.

[0140] In detail, suppose there are two tasks, Task 1 is to predict the police's physical fitness test scores, and Task 2 is to assess the police's physical health status.

[0141] In detail, the loss function of task one and the loss function of task two can be mean square error or cross entropy, and the weight coefficient is used to balance the influence of different tasks.

[0142] In general, the innovation of this multi-task loss function lies in optimizing multiple related tasks at the same time, ensuring that the model can comprehensively consider training performance and health assessment.

[0143] Specifically, in the physical fitness test score prediction, Y1 is the actual physical fitness test score (for example, a police officer's running time, number of push-ups, etc.), The mean squared error (MSE) of the fitness performance predicted by the model (e.g., the predicted running time or number of push-ups) is used to measure the difference between the predicted value and the true value.

[0144] Specifically, Task 2 is health status prediction, Y2 is the real health indicator (e.g., body fat percentage, blood pressure, etc.), is the health indicator predicted by the model; assuming it is a regression task, the mean square error (MSE) or other health-related loss functions can be used, or an appropriate loss function can be selected according to the characteristics of health assessment.

[0145] In detail, by adjusting the weight coefficient, the contribution of each task to the total loss can be controlled according to the importance of the task. If the physical performance prediction is more important, α can be set to a larger value, and vice versa. The goal of the training process is to minimize the total loss function, thereby optimizing the performance of task 1 and task 2 at the same time.

[0146] In an embodiment of the present invention, when the model training module determines the data label of the data feature according to the assessment result data, it includes: designing a definition method of the data label according to the feature of the assessment result data. For example, if the assessment result is a score for a certain type of task, the result may be divided into several levels (such as excellent, good, pass, fail), and each level is represented by a numerical value (such as: excellent = 1, good = 2, pass = 3, fail = 4).

[0147] Furthermore, the assessment result data and the preset standards are combined to generate a label for each data point. For example, for each police officer's assessment result, a threshold can be set to distinguish different score intervals according to their scores, and the corresponding label can be assigned to the data point.

[0148] For example, suppose the assessment result data is a comprehensive performance of a police officer, including the following features: running training distance: [3.5, 4.2] (indicating two different training data), running speed: [5.6, 6.0], body fat percentage: [0.22]. At the same time, the assessment result is a numerical final score ranging from 0 to 100.

[0149] In detail, based on the scoring range, the scores can be divided into four levels: excellent (90 points and above), good (70-89 points), passing (50-69 points) and failing (below 50 points). Based on the police performance data (such as training scores), the model can generate the following labels: Police A’s score is 85 and the label is “good” (2), Police B’s score is 55 and the label is “passing” (3), and Police C’s score is 95 and the label is “excellent” (1).

[0150] In an embodiment of the present invention, when the model training module performs model training on the pre-built portrait model according to the data features, the data labels and the preset multi-task loss function to obtain the trained portrait model, it includes:

[0151] Establishing a corresponding relationship between the data feature and the data label;

[0152] Generate a pre-built training set of the portrait model according to the corresponding relationship, the data label and the data feature;

[0153] The pre-built portrait model is trained using the training set and a preset multi-task loss function.

[0154] Specifically, for police physical training data, if data features such as "running training distance" and "running speed" are related to assessment results data labels such as "excellent" and "good", it is necessary to accurately define this correspondence rule. By analyzing the training data and results of a large number of police officers, it is determined that when the running training distance reaches a certain value range and the running speed is in a specific interval, the corresponding assessment result label may be "excellent", thereby establishing a corresponding relationship between the two.

[0155] In detail, according to the corresponding relationship established above, the relevant data labels and data features are combined to form a training set for the portrait model. Taking the police physical training as an example, assuming that there are training data of multiple police officers, for each police officer, the data features such as "running training distance", "running speed", "body fat percentage" and the corresponding data labels such as "excellent", "good", and "pass" together constitute a sample in the training set. By integrating many such samples, a training set that can be used to train the portrait model is generated, ensuring that the training set can fully reflect the inherent laws between data features and labels.

[0156] Specifically, the pre-built portrait model is trained using the generated training set. During the training process, the portrait model learns the mapping pattern between data features and data labels in the training set, and continuously adjusts its own parameters to improve the accuracy of predictions for new data. For example, in the training of the police physical fitness portrait model, the model gradually grasps the complex relationship between different training data features and assessment scores by learning a large amount of police training data and score labels, so that it can accurately predict the possible assessment scores of new police officers based on their training data features, achieve effective evaluation and analysis of the physical fitness of police officers, and provide strong support for the subsequent generation of personalized training plans.

[0157] The training program generation module 105 is used to generate a physical hologram of the policeman according to the trained portrait model, generate a training program for the policeman according to the physical hologram, and perform physical training on the policeman according to the training program.

[0158] In the embodiment of the present invention, when the training program generation module generates the policeman's physical hologram according to the trained portrait model, it includes:

[0159] Based on the trained portrait model, data mapping is performed on the pre-collected police data to obtain mapping data of the police data;

[0160] A body-powered hologram of the police officer is generated based on the mapping data.

[0161] In detail, the training program generation module uses the trained portrait model to process various types of police data collected in advance. These data cover various information such as the completion time of the training project and physical function parameters. For example, for the police's running training data, including the running speed, distance and corresponding physical function indicators such as heart rate and body fat rate in different training stages, the portrait model will convert these raw data into mapping data with specific meanings based on the rules it has learned. It is possible to quantitatively map the correlation between running speed and distance and physical performance, so that the originally isolated training data can reflect the physical characteristics of the police in a new dimension, laying the foundation for the subsequent generation of a physical hologram.

[0162] In detail, based on the data obtained after mapping by the portrait model, a physical fitness hologram of the police is further constructed. The physical fitness hologram is a comprehensive visual presentation of the physical fitness of the police. It integrates the comprehensive information of the police in different training projects and physical functions, and displays the physical fitness advantages and shortcomings of the police in the form of intuitive graphics or charts. For example, in the physical fitness hologram, different colors or areas may be used to represent the physical fitness level of the police in terms of strength, endurance, speed, etc. Through the reasonable layout and visualization of the mapping data, the physical fitness of the police can be seen at a glance, which is convenient for the subsequent formulation of targeted training plans based on the hologram, thereby achieving scientific guidance and precise improvement of the physical fitness training of the police.

[0163] In detail, the data mapping process for the police is as follows:

[0164] The portrait model may learn from a large amount of data that the average running training speed in the range of 10-14km / h is related to the speed dimension of physical fitness. After mapping, 12km / h is mapped to a relative score of the speed dimension, assuming it is 7 points (out of 10 points); the number of push-ups training is 30 / group, mapped to the strength dimension with a score of 6 points; the number of sit-ups training is 40 / group, mapped to the endurance dimension with a score of 7 points; the body fat rate is 15%, mapped to the comprehensive dimension of physical function with a score of 8 points.

[0165] In detail, a physical fitness hologram is generated based on these mapping data. In the physical fitness hologram, different colors or area sizes may be used to represent the scores of each dimension. For example, in the physical fitness hologram of policeman 001, the speed dimension is represented by a blue area, which accounts for 70% of the entire circular area (corresponding to 7 points), the strength dimension is represented by a green area, which accounts for 60%, the endurance dimension is represented by a yellow area, which accounts for 70%, and the comprehensive physical function dimension is represented by a red area, which accounts for 80%. In this way, the advantages and disadvantages of each policeman in different physical fitness dimensions are intuitively displayed, thereby providing a basis for the subsequent formulation of training plans.

[0166] In the embodiment of the present invention, when the training program generation module generates the training program for the police according to the physical energy hologram, the training program generation module includes:

[0167] collecting real-time performance data of the police officers during training;

[0168] Generating the physical state of the police officer according to the real-time functional data and the preset training intensity;

[0169] The training intensity of the physical fitness hologram is adjusted according to the physical fitness state to obtain the training plan for the police officer.

[0170] In detail, based on the physical fitness hologram, targeted training plans are formulated to improve the overall physical fitness level of the police. During the training process, the police's physical fitness data (such as heart rate, body temperature, etc.) is collected in real time, the training status is monitored, and the training plan is adjusted. The physical fitness status is evaluated based on the functional data collected in real time. For example, the current physical condition of the police is evaluated through indicators such as heart rate and body fat percentage.

[0171] In detail, the physical fitness hologram is a three-dimensional map constructed based on the police's physical fitness status and real-time data. It presents various dimensions of the police's physical fitness, including heart rate, endurance, muscle strength, etc.

[0172] In detail, when the training program generation module adjusts the training intensity of the physical fitness hologram according to the physical fitness state to obtain the training program for the police, it includes:

[0173] The training intensity of the physical fitness hologram is adjusted according to the physical fitness state and a preset adjustment formula to obtain the training plan for the police, wherein the preset adjustment formula is:

[0174]

[0175] Among them, S′ t is the adjusted training intensity, S t is the preset training intensity, t is the adjustment coefficient, HR t is the physical state, HR max It is the maximum physical state.

[0176] Specifically, the assumed values ​​are: The preset training intensity is S t =80 (e.g. 80% of initial training intensity), adjustment factor t = 0.15, maximum physical state HR max =180, current physical condition HR t =160.

[0177] In detail, substitute the values ​​into the formula for calculation: According to the given value, the adjusted training intensity S ′ t It is 81.33, which means that the current physical condition HR t = 160 times, the preset training intensity is S t =80 increased to 81.33 through the adjustment formula.

[0178] Specifically, the training intensity is adjusted according to the police officer's current physical condition. For example, if the police officer's heart rate is close to the maximum, the system will reduce the training intensity; if the police officer's physical strength recovers quickly, the system may increase the training intensity.

[0179] In detail, by comparing the changes in physical status at different time points, it shows how to adjust the training plan based on real-time data.

[0180] In detail, the initial physical state (at the beginning of training) is:

[0181] Heart rate: 125 bpm (in the moderate aerobic zone, about 65% of maximum heart rate);

[0182] Body temperature: 37.2℃ (normal);

[0183] Blood oxygen concentration: 98% (normal);

[0184] Respiratory rate: 18 times / minute (normal);

[0185] Exercise intensity: Light training (such as jogging).

[0186] The training objectives are:

[0187] Improve aerobic endurance, with a target heart rate of 160bpm;

[0188] To increase muscle strength, aim to squat 80 kg and do 40 push-ups.

[0189] Furthermore, the training adjustment process is divided into aerobic training and strength training, wherein aerobic training is:

[0190] Training content: jogging for 30 minutes, with the target heart rate maintained between 130-140bpm (60%-70% of maximum heart rate); actual data: the heart rate reached 130bpm and the exercise intensity was moderate; adjustment: if the heart rate of policeman A is lower than 130bpm, the system recommends increasing the running speed to ensure the training intensity.

[0191] Furthermore, the strength training is as follows: Training content: performing squats, using a 60 kg weight, and completing 3 sets of 10 times; Actual data: After completing 3 sets of squats, Policeman A's muscle soreness score was 6 (1-10 points, 10 being the strongest).

[0192] Furthermore, if the soreness score exceeds 7, the system recommends reducing the weight or increasing the rest time.

[0193] In detail, during the training process, the aerobic training of Policeman A gradually increased from the initial low intensity to the target heart rate of 160bpm. After each week of training, the system made adjustments based on data such as heart rate, body temperature, and respiratory rate to ensure appropriate intensity and avoid overtraining. The weight of strength training was gradually increased from 60 kg to 80 kg, and the system dynamically adjusted the load according to the degree of muscle soreness. By adjusting the training intensity, Policeman A's physical fitness was gradually improved, and he was able to adapt to higher-intensity training, and his physical recovery was also relatively smooth.

[0194] In an embodiment of the present invention, when the training program generation module executes physical training for the police according to the training program, it includes: ensuring that the training program can promote the continuous improvement of the police's physical fitness; avoiding over-training according to the actual physical condition of the police; and performing personalized optimization according to the police's personal health status, physical fitness level, training experience, etc.

[0195] In detail, combined with the adjusted physical fitness hologram and training intensity, a training program for police officers is generated. The training program usually includes: aerobic training: improve cardiopulmonary function and endurance; strength training: enhance muscle strength; flexibility training: increase joint and muscle flexibility; recovery training: help police officers better recover physical fitness and avoid excessive fatigue. The program is dynamic and will be adjusted in real time as the police training progress and physical fitness status change to ensure the effectiveness and safety of the training.

[0196] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0197] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0198] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A police physical training assessment management system, characterized in that: The system includes a training data acquisition module, a correlation analysis module, a feature extraction module, a model training module and a training scheme generation module, wherein: The training data collection module is used to collect multi-source training data of police officers, wherein the multi-source training data includes: training project completion time, training intensity index, physical function parameters and assessment performance data; The correlation analysis module is used to perform a dual correlation analysis on the multi-source training data to obtain data correlation of the multi-source training data; The feature extraction module is used to determine the strongly correlated data in the multi-source training data according to the data correlation, perform feature extraction on the strongly correlated data, and obtain data features of the strongly correlated data; The model training module is used to determine the data label of the data feature according to the assessment score data, and to perform model training on the pre-built portrait model according to the data feature, the data label and the preset multi-task loss function to obtain a trained portrait model, wherein the preset multi-task loss function for: Among them, Y1 is the target value of task one, is the model's predicted value for task one, and Y2 is the target value for task two. is the model's predicted value for task 2, α is the weight coefficient for task 1, and β is the weight coefficient for task 2. is the loss function of task one, is the loss function of task 2; The training program generation module is used to generate a physical hologram of the policeman according to the trained portrait model, generate a training program for the policeman according to the physical hologram, and perform physical training on the policeman according to the training program.

2. The police physical training assessment management system as claimed in claim 1, characterized in that: The training data collection module collects multi-source training data of police officers, including: Collect the time it takes police officers to complete each physical training item one by one; Collect training intensity indicators for each physical training; Collect the physical function parameters of police officers one by one; Collect the assessment performance data of police officers in physical training one by one.

3. The police physical training assessment management system as claimed in claim 1, characterized in that: When the correlation analysis module performs a dual correlation analysis on the multi-source training data to obtain the data correlation of the multi-source training data, it includes: Calculate item correlations for different training items; Calculate the functional relevance of training programs and physical performance parameters; A correlation matrix of the multi-source training data is generated according to the item correlation and the function correlation, thereby determining the data relevance of the multi-source training data.

4. The police physical training assessment management system as claimed in claim 3, characterized in that: When the correlation analysis module calculates the correlation between different training items, it includes: The project relevance of different training projects is calculated using a preset relevance algorithm, wherein the preset relevance algorithm is: Among them, r ij is the item correlation between the i-th and j-th training items, N is the total number of data in the training items, k is the data identifier, and X ik is the value of the kth data of the i-th training item, is the data mean of the i-th training item, is the data mean of the i-th training item, X jk is the value of the kth data of the jth training item, i is the identifier of the training item, and j is the identifier of the training item.

5. The police physical training assessment management system as claimed in claim 1, characterized in that: When the feature extraction module determines the strongly correlated data in the multi-source training data according to the data correlation, it includes: Based on the data correlation, the training items and body function parameters that are strongly correlated with the assessment performance data are determined as strongly correlated data.

6. The police physical training assessment management system as claimed in claim 1, characterized in that: When the feature extraction module extracts features from the strongly correlated data to obtain data features of the strongly correlated data, the feature extraction module includes: The strongly correlated data are vector-converted, and the vector-converted strongly correlated data are vector-concatenated to obtain data features of the strongly correlated data.

7. The police physical training assessment management system as claimed in claim 1, characterized in that: The model training module performs model training on the pre-built portrait model according to the data features, the data labels and the preset multi-task loss function to obtain the trained portrait model, including: Establishing a corresponding relationship between the data feature and the data label; Generate a pre-built training set of the portrait model according to the corresponding relationship, the data label and the data feature; The pre-built portrait model is trained using the training set and a preset multi-task loss function.

8. The police physical training assessment management system as claimed in claim 1, characterized in that: When the training program generation module generates the policeman's physical hologram according to the trained portrait model, the training program generation module includes: Based on the trained portrait model, data mapping is performed on the pre-collected police data to obtain mapping data of the police data; A body-powered hologram of the police officer is generated based on the mapping data.

9. The police physical training assessment management system as claimed in claim 1, characterized in that: When the training program generation module generates the training program for the police according to the physical energy hologram, the training program generation module includes: collecting real-time performance data of the police officers during training; Generating the physical state of the police officer according to the real-time functional data and the preset training intensity; The training intensity of the physical fitness hologram is adjusted according to the physical fitness state to obtain the training plan for the police officer.

10. The police physical training assessment management system as claimed in claim 9, characterized in that: When the training program generation module adjusts the training intensity of the physical fitness hologram according to the physical fitness state to obtain the training program for the police, the training program generation module includes: The training intensity of the physical fitness hologram is adjusted according to the physical fitness state and a preset adjustment formula to obtain the training plan for the police, wherein the preset adjustment formula is: Among them, S′ t is the adjusted training intensity, S t is the preset training intensity, t is the adjustment coefficient, HR t is the physical state, HR max It is the maximum physical state.

Images