Military physical training risk assessment method and system based on deep learning

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a military physical training risk assessment method and system based on deep learning, and the method comprises the steps: S1, collecting fixed influence factors; s2, calculating an altitude attenuation coefficient function; s3, obtaining a terrain attenuation coefficient function; s4, constructing a deep learning model; s5, collecting physical sign data; s6, risk assessment; s7, training feedback; and S8, updating and evaluating the subjective attenuation coefficient function. The system comprises an influence data acquisition module, an altitude module, a terrain module, a model construction module, a physical sign data acquisition module, a risk assessment module, a training feedback module and an updating module. The terrain attenuation coefficient and the altitude attenuation coefficient are calculated by introducing the influence factors of the terrain flatness and the altitude, so that the comprehensiveness and the accuracy of risk assessment are improved, and the safety and the efficiency of training are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a military physical training risk assessment method and system based on deep learning. Background Art

[0002] Military physical training is the basic training for the army to improve the physical fitness and combat effectiveness of soldiers. However, due to the high intensity and high pressure of training and the complex environment and equipment involved, there are certain safety risks in the training process. Therefore, risk assessment is a necessary means to ensure training safety and prevent accidents.

[0003] In the existing military physical training risk assessment methods, although there are some technologies that can combine physical sign data for preliminary risk assessment, these methods often ignore environmental factors, especially the important impact of terrain conditions on training risks. The flatness of different terrains will directly affect the training personnel's movement state, force distribution and movement stability, which will increase the risk of injury or safety accidents.

[0004] Therefore, there is an urgent need for a military physical training risk assessment method and system that can comprehensively consider environmental data (including altitude and terrain, etc.) on the basis of physical sign data and has physical fitness feedback updates for trainees, so as to improve the accuracy of risk assessment, enhance the safety of training, and facilitate instructors to provide personalized training guidance to trainees. Summary of the invention

[0005] To solve the above problems, the present invention provides a military physical training risk assessment method and system based on deep learning. By introducing the influencing factors of terrain flatness and altitude, the terrain attenuation coefficient and the altitude attenuation coefficient are calculated, thereby improving the comprehensiveness and accuracy of risk assessment and enhancing the safety and efficiency of training.

[0006] In order to achieve the above object, the technical solution of the present invention is as follows: A military physical training risk assessment method based on deep learning comprises the following steps:

[0007] S1, collection of fixed influencing factors: collection of the altitude and oxygen supply of the target training location;

[0008] S2, calculating the altitude attenuation coefficient function: calculating the constant in the altitude attenuation coefficient function corresponding to the target training location according to the altitude and oxygen supply of the target training location;

[0009] S3, terrain attenuation coefficient function: use optical instruments to scan the ground three-dimensional image data of the target training site, and then use image algorithms to analyze and calculate the surface flatness results of the target training site, and calculate the terrain attenuation coefficient function;

[0010] S4, construction of deep learning model: using the standard range average value of each vital sign data as input, supplemented by the altitude attenuation coefficient function and the terrain attenuation coefficient function as a coefficient in the evaluation formula, and using low-level injury risk, low-level physical insufficiency risk and low-level safety risk as output results to construct the deep learning model;

[0011] S5, physical sign data collection: collect the real-time temperature of the target training site and the physical sign data of each trainee;

[0012] S6, risk assessment: input the temperature data into the deep learning model, and then input the physical sign data of each trainee into the deep learning model to assess the risk of injury, physical deficiency risk and safety risk respectively;

[0013] S7, Training Feedback: After the trainers have completed the training, they will provide feedback on the risk assessment;

[0014] S8, update the evaluation subjective attenuation coefficient function: calculate the subjective attenuation coefficient function based on the feedback from the trainers and the evaluation given by the deep learning model as parameters, and perform evaluation calculation adjustments on the deep learning model for each trainer.

[0015] Furthermore, in S2, the calculation formula of the altitude attenuation coefficient function is as follows:

[0016]

[0017] Where O2(A) is the oxygen supply at the target training location, O2(R) is the oxygen supply at a reference area, and h A is the altitude of the target training site, h R is the altitude of a reference area, k2 is the altitude attenuation coefficient; k2 is calculated by this formula, that is, the altitude attenuation coefficient function f1(O2(A),h A ) value, which usually evaluates to a constant.

[0018] Furthermore, in S3, the ground flatness is presented in the form of a percentage, indicating the proportion of the area in the target training venue that meets the flatness requirement.

[0019] Furthermore, in S3, the calculation formula of the terrain attenuation coefficient function is as follows:

[0020]

[0021] In the formula, s is the ground flatness, and the calculation result of f2(s) is the terrain attenuation coefficient k1.

[0022] Furthermore, in S4, the vital signs data include heart rate, blood oxygen, body temperature, respiratory rate, blood pressure and limb electromyographic signal frequency, among which the heart rate is 80 times / minute, the blood oxygen is 97.5%, the body temperature is 37°C, the respiratory rate is 16 times / minute, the blood pressure is 120 / 80 mmHg, the arm electromyographic signal frequency is 80 Hz, and the leg electromyographic signal frequency is 120 Hz.

[0023] Furthermore, in S6, the temperature data is input to change the risk assessment standard. When the temperature is greater than 33°C or the temperature is less than -1°C, the risk assessment result is correspondingly improved.

[0024] Furthermore, in S6, the risk of injury, risk of insufficient physical fitness and safety risk in the risk assessment are assessed with a risk probability of 0-20% as low risk, a risk probability of 21-40% as medium risk and a risk probability of 41-60% as high risk. When the risk probability is assessed to be above 61%, the trainee should not conduct training.

[0025] Furthermore, in S7, after the trainee completes the training, feedback is provided on the fatigue after the training, and the feedback includes mild, moderate, severe and extreme. The feedback fatigue data is used for subsequent risk assessment of physical insufficiency for the trainee.

[0026] Furthermore, in S8, the feedback of mild fatigue corresponds to a 4-5% reduction in the probability of the trainee's risk of insufficient physical fitness, the feedback of moderate fatigue corresponds to a 1-2% reduction in the probability of the trainee's risk of insufficient physical fitness, the feedback of severe fatigue corresponds to an increase in the probability of the trainee's risk of insufficient physical fitness by 2-3%, and the feedback of extreme fatigue corresponds to a 6-7% reduction in the probability of the trainee's risk of insufficient physical fitness.

[0027] A military physical training risk assessment method based on deep learning is operated on the basis of the military physical training risk assessment method based on deep learning, including:

[0028] Impact data collection module, used to collect the altitude and oxygen supply of the target training site;

[0029] An altitude module is used to calculate a constant corresponding to the target training location in the altitude attenuation coefficient function according to the altitude of the target training location and the oxygen supply;

[0030] A terrain module is used to use an optical instrument to scan the ground three-dimensional image data of the target training site, and then use an image algorithm to analyze and calculate the surface flatness result of the target training site, and calculate the terrain attenuation coefficient function;

[0031] A model building module is used to construct a deep learning model using the standard range average value of each vital sign data as input, supplemented by the altitude attenuation coefficient function and the terrain attenuation coefficient function as a coefficient in the evaluation formula, and using low-level injury risk, low-level physical insufficiency risk and low-level safety risk as output results;

[0032] The physical sign data collection module is used to collect the real-time temperature of the target training site and the physical sign data of each trainee;

[0033] The risk assessment module is used to input the temperature data into the deep learning model, and then input the physical sign data of each trainee into the deep learning model to evaluate the risk of injury, physical deficiency risk and safety risk respectively;

[0034] The training feedback module is used for trainers to provide feedback on risk assessment after they have completed the training;

[0035] The updating module is used to calculate the subjective attenuation coefficient function based on the feedback from the trainers and the evaluation given by the deep learning model as parameters, and to adjust the evaluation calculation of the deep learning model for each trainer.

[0036] The above scheme can achieve the following beneficial effects:

[0037] 1. In military physical training, the accuracy of risk assessment is crucial. Traditional risk assessment methods are often based on empirical judgment or a single physiological indicator, which makes it difficult to fully and accurately reflect the various risk factors in the training process. However, through the deep learning model combined with physical sign data and environmental data, the environmental data includes altitude and terrain data, which can achieve a comprehensive and integrated assessment of training risks. In traditional assessments, on the basis of considering the individual differences of trainees, the impact of environmental factors on training risks is also considered, which greatly improves the accuracy of risk assessment and helps trainers and supervisors better understand the potential risks in the training process, so as to formulate targeted preventive measures and reduce the probability of accidents.

[0038] 2. Since the deep learning model has the function of continuous learning and updating, in order to improve the accuracy and personalization of risk assessment, through the design of feedback mechanism, the system can make advance estimates of unexpected risk situations. When trainees have unexpected physical deficiency, abnormal heart rate or safety hazards during training, feedback can make the risk assessment more cautious during the next training, improve the assessment of unexpected situations, further enhance the safety of training, reduce accidents during training, and ensure the life safety and physical health of trainees.

[0039] 3. Since the physical fitness of trainees will increase with the increase of training times, the feedback mechanism after training can also capture the signal of the physical fitness improvement of trainees. Based on this feedback information, the difficulty and intensity of subsequent training are automatically adjusted to ensure that the training is always maintained at a level that can promote physical fitness improvement without causing overtraining or injury, which helps trainees gradually break through themselves and achieve the maximum growth of physical fitness.

[0040] 4. Each trainer has different physical condition, training experience and adaptability. The design conducts accurate assessment based on the physical data of each trainer to achieve personalized training, better meet the actual needs of the trainer, and improve training efficiency and effectiveness. In addition, personalized training plans can also help trainers better adapt to the training environment and reduce health risks caused by over- or under-training.

[0041] 5. Different terrain flatness will have a direct impact on the training personnel's movement state, force distribution and movement stability. The introduction of terrain factors, taking terrain flatness as a key indicator of risk assessment, and assessing safety risks based on the flatness of the training site can significantly improve the comprehensiveness of risk assessment, and more accurately reflect the actual risk status during training, reduce the occurrence of accidental injuries during training, and provide more comprehensive safety protection for trainees.

[0042] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a flowchart of an embodiment of a method for risk assessment of military physical training based on deep learning of the present invention;

[0044] Figure 2 This is a schematic diagram of the operation of an embodiment of a military physical training risk assessment system based on deep learning according to the present invention. DETAILED DESCRIPTION

[0045] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only 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.

[0046] The following is further described in detail through specific implementation methods:

[0047] Embodiment 1:

[0048] As attached Figure 1As shown: A military physical training risk assessment method based on deep learning, comprising the following steps:

[0049] S1. Collection of fixed influencing factors: By finding the Surveying and Mapping Bureau of the target training site, the altitude of the target training site can be collected, and the average oxygen supply of the target site can be collected through relevant environmental monitoring agencies. Reasonable values ​​can be taken as reference data. Understanding the altitude and oxygen supply can help trainers and supervisors better plan training plans, especially in high-altitude areas, so that targeted adaptive training can be carried out to improve training efficiency.

[0050] S2, calculate the altitude attenuation coefficient function: through the collected altitude of the target training site and the oxygen supply, through the following formula:

[0051]

[0052] Where O2(A) is the oxygen supply at the target training location, O2(R) is the oxygen supply at a reference area, and h A is the altitude of the target training site, h R is the altitude of a reference area, k2 is the altitude attenuation coefficient; k2 is calculated by this formula, that is, the altitude attenuation coefficient function f1(O2(A),h A ) value, which is usually calculated as a constant. Calculate the constant corresponding to the target training location in the altitude attenuation coefficient function. By calculating the altitude attenuation coefficient and adjusting the training plan accordingly, the health risks of athletes caused by hypoxia due to high altitude can be reduced.

[0053] S3, terrain attenuation coefficient function: Use optical instruments to scan the ground three-dimensional image data of the target training site, and then use image algorithms to analyze and calculate the surface flatness results of the target training site. The ground flatness is presented in the form of a percentage, indicating the proportion of the area in the target training site that meets the flatness requirements, and is expressed by the following formula:

[0054]

[0055] In the formula, s is the ground flatness, and the calculation result of f2(s) is the terrain attenuation coefficient k1. The terrain attenuation coefficient function is calculated. By calculating the ground flatness and the terrain attenuation coefficient k1, the ground conditions of the target training site can be quantitatively evaluated, providing terrain condition factors for subsequent risk assessment and increasing the rationality of risk assessment.

[0056] S4, construction of deep learning model: the standard range average value of each physical sign data is used as input, where each physical sign data includes heart rate, blood oxygen, body temperature, respiratory rate, blood pressure and limb electromyographic signal frequency, where the heart rate is 80 times / minute, blood oxygen is 97.5%, body temperature is 37℃, respiratory rate is 16 times / minute, blood pressure is 120 / 80mmHg, arm electromyographic signal frequency is 80Hz, and leg electromyographic signal frequency is 120Hz. With the help of altitude attenuation coefficient function and terrain attenuation coefficient function as a coefficient in the evaluation formula, the construction of deep learning model is constructed with low-level injury risk, low-level physical insufficiency risk and low-level safety risk as output results. The deep learning model can automatically learn the complex relationship between input data and risk assessment, thereby improving the accuracy of risk assessment. By introducing altitude and terrain attenuation coefficients, the model can more comprehensively consider the impact of environmental factors on risk assessment. Standardizing input feature data helps enhance the generalization ability of the model, enabling the deep learning model to more quickly train risk assessments for healthy, normal humans, making the model applicable to different training locations and more quickly assessing trainees with individual differences, thereby improving the versatility of risk assessments.

[0057] S5, physical sign data collection: collect the real-time temperature of the target training site and the physical sign data of each trainee.

[0058] S6, risk assessment: input the temperature data into the deep learning model to change the risk assessment standard. When the temperature is greater than 33°C or less than -1°C, the risk assessment result is improved accordingly. Then input the physical sign data of each trainee into the deep learning model to evaluate the risk of injury, physical insufficiency and safety risk. The risk of injury, physical insufficiency and safety risk in the risk assessment are evaluated with a risk probability of 0-20% as low risk, 21-40% as medium risk, and 41-60% as high risk. When the risk probability is assessed to be more than 61%, the trainee should not be trained;

[0059] S7, training feedback: After the trainee completes the training, he / she will provide feedback on the fatigue after the training. The feedback includes mild, moderate, severe and extreme. The feedback fatigue data is used for the subsequent physical fitness risk assessment of the trainee.

[0060] S8, update the evaluation of the subjective attenuation coefficient function: based on the feedback from the trainees and the evaluation given by the deep learning model as parameters, the feedback of mild fatigue corresponds to a 4-5% reduction in the probability of the trainee's risk of insufficient physical fitness, the feedback of moderate fatigue corresponds to a 1-2% reduction in the probability of the trainee's risk of insufficient physical fitness, the feedback of severe fatigue corresponds to an increase in the probability of the trainee's risk of insufficient physical fitness by 2-3%, and the feedback of extreme fatigue corresponds to a 6-7% reduction in the probability of the trainee's risk of insufficient physical fitness. The subjective attenuation coefficient function is calculated, and the deep learning model is used for evaluation and calculation adjustments for each trainee. Take any trainee as an example. Before a training session, various instruments are used to detect the heart rate, blood oxygen, body temperature, respiratory rate, blood pressure and limb electromyographic signal frequency data before training, and the data are input into the deep learning model. The deep learning model evaluates the trainee's risk of insufficient physical fitness as low risk based on the physical sign data. After the trainee completes the training, the feedback is slightly tired after the training (not tired during the training). After the deep learning model receives the slight fatigue, the next pre-training risk assessment of the trainee's insufficient capacity risk assessment will be reduced by 4-5%, which means that the member's physical fitness has improved. When the intensity of the next training increases, the trainee's risk of insufficient capacity should be assessed as medium risk (i.e., the training intensity is too high, and the trainee will have a bit of difficulty in completing the training). However, since the deep learning model records the reduction of the trainee's risk of insufficient capacity in the last training (i.e., the trainee's physical fitness has improved), the trainee's risk of insufficient capacity in this training is assessed as low risk (i.e., the training intensity is assessed to be suitable for the trainee). On the contrary, feedback of extreme fatigue indicates that the trainee's physical fitness has decreased or the training intensity needs to be reduced.

[0061] Embodiment 2:

[0062] As attached Figure 2 As shown, a military physical training risk assessment system based on deep learning is designed according to the military physical training risk assessment method based on deep learning described in Example 1, including an impact data acquisition module, an altitude module, a terrain module, a model building module, a vital sign data acquisition module, a risk assessment module, a training feedback module and an update module.

[0063] The impact data acquisition module collects the altitude and oxygen supply of the target training site, and transmits the altitude and oxygen supply to the altitude module for calculation.

[0064] The altitude module calculates the constant corresponding to the target training site in the altitude attenuation coefficient function through the altitude and oxygen supply of the target training site. The calculated constant is used to build the deep learning model in the model building module.

[0065] The terrain module uses optical instruments to scan the three-dimensional ground image data of the target training site, and then uses image algorithms to analyze and calculate the surface flatness results of the target training site, and calculates the terrain attenuation coefficient function. The calculated constants are also used for model construction to continue evaluating safety risks considering terrain factors.

[0066] The model building module takes the standard range average value of each vital sign data as input, supplemented by the altitude attenuation coefficient function and the terrain attenuation coefficient function as a coefficient in the evaluation formula, and constructs a deep learning model with low-level injury risk, low-level physical fitness risk and low-level safety risk as output results.

[0067] The physical sign data collection module collects the real-time temperature of the target training site and the physical sign data of each trainee.

[0068] The risk assessment module inputs the temperature data into the deep learning model, and then inputs the physical sign data of each trainee into the deep learning model to assess the risk of injury, physical fitness risk and safety risk respectively.

[0069] Training Feedback Module After the trainers complete the training, they will provide feedback on the risk assessment.

[0070] The update module calculates the subjective attenuation coefficient function based on the feedback from the trainers and the evaluation given by the deep learning model as parameters, and performs evaluation calculation adjustments on the deep learning model for each trainer.

[0071] Obviously, the above embodiments are merely examples for the purpose of clear explanation, and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the invention.

Claims

1. A military physical training risk assessment method based on deep learning, characterized in that: The following steps are involved: S1, collection of fixed influencing factors: collection of the altitude and oxygen supply of the target training location; S2, calculating the altitude attenuation coefficient function: calculating the constant in the altitude attenuation coefficient function corresponding to the target training location according to the altitude and oxygen supply of the target training location; S3, terrain attenuation coefficient function: use optical instruments to scan the ground three-dimensional image data of the target training site, and then use image algorithms to analyze and calculate the surface flatness results of the target training site, and calculate the terrain attenuation coefficient function; S4, construction of deep learning model: using the standard range average value of each vital sign data as input, supplemented by the altitude attenuation coefficient function and the terrain attenuation coefficient function as a coefficient in the evaluation formula, and using low-level injury risk, low-level physical insufficiency risk and low-level safety risk as output results to construct the deep learning model; S5, physical sign data collection: collect the real-time temperature of the target training site and the physical sign data of each trainee; S6, risk assessment: input the temperature data into the deep learning model, and then input the physical sign data of each trainee into the deep learning model to assess the risk of injury, physical deficiency risk and safety risk respectively; S7, Training Feedback: After the trainers have completed the training, they will provide feedback on the risk assessment; S8, update the evaluation subjective attenuation coefficient function: calculate the subjective attenuation coefficient function based on the feedback from the trainers and the evaluation given by the deep learning model as parameters, and perform evaluation calculation adjustments on the deep learning model for each trainer.

2. The military physical training risk assessment method based on deep learning according to claim 1 is characterized in that: In S2, the calculation formula of the altitude attenuation coefficient function is as follows: Where O2(A) is the oxygen supply at the target training location, O2(R) is the oxygen supply at a reference area, and h A is the altitude of the target training site, h R is the altitude of a reference area, k2 is the altitude attenuation coefficient; k2 is calculated by this formula, that is, the altitude attenuation coefficient function f1(O2(A),h A ) value, which usually evaluates to a constant.

3. The military physical training risk assessment method based on deep learning according to claim 2 is characterized in that: In S3, the ground flatness is presented in the form of a percentage, indicating the proportion of the area in the target training venue that meets the flatness requirements.

4. The military physical training risk assessment method based on deep learning according to claim 3 is characterized in that: In S3, the calculation formula of the terrain attenuation coefficient function is as follows: In the formula, s is the ground flatness, and the calculation result of f2(s) is the terrain attenuation coefficient k1.

5. The military physical training risk assessment method based on deep learning according to claim 4 is characterized in that: In S4, the vital signs data include heart rate, blood oxygen, body temperature, respiratory rate, blood pressure and EMG signal frequency of the limbs, among which the heart rate is 80 times / minute, the blood oxygen is 97.5%, the body temperature is 37°C, the respiratory rate is 16 times / minute, the blood pressure is 120 / 80 mmHg, the arm EMG signal frequency is 80 Hz, and the leg EMG signal frequency is 120 Hz.

6. The military physical training risk assessment method based on deep learning according to claim 5 is characterized in that: In S6, the temperature data is input to change the risk assessment standard. When the temperature is greater than 33°C or less than -1°C, the risk assessment result is improved accordingly.

7. The military physical training risk assessment method based on deep learning according to claim 6 is characterized in that: In S6, the risk of injury, risk of insufficient physical fitness and safety risk in the risk assessment are assessed with a risk probability of 0-20% as low risk, a risk probability of 21-40% as medium risk and a risk probability of 41-60% as high risk. When the risk probability is assessed to be above 61%, the trainee should not conduct training.

8. The military physical training risk assessment method based on deep learning according to claim 7 is characterized in that: In S7, after the trainees complete the training, they will provide feedback on their fatigue after the training, which includes mild, moderate, severe and extreme. The feedback fatigue data is used for subsequent physical fitness risk assessment of the trainees.

9. The military physical training risk assessment method based on deep learning according to claim 8 is characterized in that: In S8, the feedback of mild fatigue corresponds to a 4-5% reduction in the probability of the trainee's risk of insufficient physical fitness, the feedback of moderate fatigue corresponds to a 1-2% reduction in the probability of the trainee's risk of insufficient physical fitness, the feedback of severe fatigue corresponds to an increase in the probability of the trainee's risk of insufficient physical fitness by 2-3%, and the feedback of extreme fatigue corresponds to a 6-7% reduction in the probability of the trainee's risk of insufficient physical fitness.

10. A military physical training risk assessment system based on deep learning, which operates on the basis of a military physical training risk assessment method based on deep learning, characterized in that: include: Impact data collection module, used to collect the altitude and oxygen supply of the target training site; An altitude module is used to calculate a constant corresponding to the target training location in the altitude attenuation coefficient function according to the altitude of the target training location and the oxygen supply; A terrain module is used to use an optical instrument to scan the ground three-dimensional image data of the target training site, and then use an image algorithm to analyze and calculate the surface flatness result of the target training site, and calculate the terrain attenuation coefficient function; A model building module is used to construct a deep learning model using the standard range average value of each vital sign data as input, supplemented by the altitude attenuation coefficient function and the terrain attenuation coefficient function as a coefficient in the evaluation formula, and using low-level injury risk, low-level physical insufficiency risk and low-level safety risk as output results; The physical sign data collection module is used to collect the real-time temperature of the target training site and the physical sign data of each trainee; The risk assessment module is used to input the temperature data into the deep learning model, and then input the physical sign data of each trainee into the deep learning model to evaluate the risk of injury, physical deficiency risk and safety risk respectively; The training feedback module is used for trainers to provide feedback on risk assessment after they have completed the training; The updating module is used to calculate the subjective attenuation coefficient function based on the feedback from the trainers and the evaluation given by the deep learning model as parameters, and to adjust the evaluation calculation of the deep learning model for each trainer.

Images