Search

CN-122022595-A - Physical ability assessment index dynamic generation method for military training outline

CN122022595ACN 122022595 ACN122022595 ACN 122022595ACN-122022595-A

Abstract

The invention discloses a dynamic generation method of physical ability assessment indexes oriented to military training outline, and aims to solve the problems that the existing physical ability assessment indexes are static and uniform and cannot adapt to individual differences and real-time states. The method comprises the steps of digitally analyzing military training outline to extract reference assessment indexes, collecting and constructing individual feature vectors and real-time state data streams of soldiers, calculating dynamic adjustment factors based on adjustment factor calculation models, fusing outline reference, individual features and real-time state data, adjusting the reference indexes by the factors and applying safety boundary constraint to generate personalized dynamic assessment index sets, and carrying out index fine adjustment and recording feedback according to the real-time data in the assessment process. The invention realizes the conversion of the assessment index from static unification to dynamic individuation, improves the scientificity, fairness and safety of the assessment, and forms a continuously optimized closed loop.

Inventors

  • LI MING
  • LIN KANG
  • Huang Shuanglian
  • FANG SHENGXIN
  • HU LIN

Assignees

  • 亚哲科技股份有限公司

Dates

Publication Date
20260512
Application Date
20260309

Claims (10)

  1. 1. A physical ability assessment index dynamic generation method facing military training outline is characterized by comprising the following steps: Step 1, digitally analyzing a military training outline, extracting standard assessment indexes of the objective of each physical assessment class, and forming an outline standard index library; Step 2, collecting individual static characteristic data and dynamic history training data of soldiers to be checked, and constructing individual characteristic vectors; Step 3, synchronously collecting the environment data of the examination site and the real-time physiological monitoring data of soldiers in the examination preparation and examination execution stage to form a real-time state data stream; Step 4, based on the outline reference index library, the individual feature vectors and the real-time state data stream, carrying out fusion analysis through an adjustment factor calculation model, and calculating to obtain a group of dynamic adjustment factors; Step 5, dynamically adjusting the corresponding reference assessment indexes in the outline reference index library by utilizing the dynamic adjustment factors to generate initial dynamic assessment indexes, applying safety and compliance boundary constraints to the initial dynamic assessment indexes, and outputting a final personalized dynamic assessment index set; And 6, continuously performing real-time fine adjustment on the personalized dynamic assessment index set according to the real-time state data stream in the assessment execution process, recording assessment whole process data and adjustment logs, and updating the assessment whole process data and the adjustment logs into the individual feature vectors of the soldiers to be examined.
  2. 2. The method for dynamically generating physical ability assessment indexes for military training outline according to claim 1, wherein in step 2, the dynamic historical training data comprises the historical optimal performance, historical average performance, performance change trend data, historical training damage record and detailed results of recent physical ability test of each physical ability course of soldiers in a past set period; the individual static characteristic data comprise ages, sexes, army professions and job orders of soldiers; The process of constructing the individual feature vector comprises the steps of respectively carrying out numerical coding and normalization processing on the dynamic historical training data and the individual static feature data, and combining all the processed data according to a preset dimension sequence to form the individual feature vector.
  3. 3. The dynamic physical ability assessment index generation method for the military training outline according to claim 1, wherein in step 3, the assessment site environment data are collected through a sensor network deployed at an assessment site, and the sensor network comprises temperature, humidity, altitude and wind speed; the soldier real-time physiological monitoring data are collected through wearable equipment worn by the soldier, and the wearable equipment comprises heart rate, heart rate variability, blood oxygen saturation, core body temperature and real-time exercise intensity; The process for forming the real-time state data stream comprises the steps of sampling the collected examination site environment data and the real-time physiological monitoring data of the soldier at a set frequency, filtering the sampled data, removing abnormal values, preprocessing, and organizing the sampled data into a continuous real-time state data stream according to a time stamp sequence.
  4. 4. The method for dynamically generating physical ability assessment indexes for military training outline according to claim 1, wherein in step 4, the calculation process of the adjustment factor calculation model comprises the following sub-steps: Step 4.1, calculating an adaptive offset according to the difference between the historical achievement data of a specific lesson in the individual feature vector and a corresponding lesson reference value in the outline reference index library to obtain a basic adaptive offset; Step 4.2, environmental compensation calculation, namely inquiring a preset environmental influence relation table according to temperature, humidity and altitude data in the real-time state data stream, and calculating to obtain an expected influence coefficient of the environment on physical performance as an environmental compensation factor; Step 4.3, health load assessment calculation, namely assessing the current physiological load state and health risk level of the soldier according to the heart rate variability, blood oxygen saturation and time sequence change trend of core body temperature data in the real-time state data stream, and generating a health load coefficient; and 4.4, inputting the basic adaptive offset, the environment compensation factor and the health load factor into a comprehensive decision function in the adjustment factor calculation model, and outputting a group of comprehensive dynamic adjustment factors.
  5. 5. The method for dynamically generating physical fitness assessment indicators for military training outline according to claim 1, wherein in step 5, the applying safety and compliance boundary constraints to the initial dynamic assessment indicators comprises: A first layer of boundary constraint, namely outline hard boundary constraint, wherein the initial dynamic assessment index is compared with the minimum standard and the maximum standard of the class specified in the military training outline, and if the initial dynamic assessment index exceeds the range, the initial dynamic assessment index is corrected to the nearest boundary value; and (3) a second layer of boundary constraint, namely health safety boundary constraint, setting a safety intensity upper limit which is inversely related to the health load coefficient according to the health load coefficient calculated in the step (4), and adjusting the assessment intensity to be within the safety intensity upper limit if the assessment intensity represented by the initial dynamic assessment index exceeds the safety intensity upper limit.
  6. 6. The method for dynamically generating physical ability assessment indexes for military training outline according to claim 1, wherein in step 6, the specific process of performing real-time fine adjustment on the personalized dynamic assessment index set is that a lightweight real-time fine adjustment model is established, the input of the real-time fine adjustment model is the continuously input real-time state data stream, and a plurality of fine adjustment triggering conditions are preset in the real-time fine adjustment model; when one or more data in the real-time state data stream meets any one of the fine tuning trigger conditions, the real-time fine tuning model immediately generates a fine tuning instruction; the fine tuning instruction is used for performing real-time adjustment on the course index of the examination being executed, and the real-time adjustment comprises suggesting to pause the examination, adjusting the strength requirement of the subsequent examination paragraphs or modifying the target completion time of the remaining examination distance.
  7. 7. The method for dynamically generating physical ability assessment indexes for military training outline according to claim 4, wherein in step 4.2, the construction method of the preset environmental influence relation table is that collecting statistical variation data of physical ability assessment results of a large number of soldiers relative to results under standard environmental conditions under different temperature, humidity and altitude combination conditions, and establishing a nonlinear mapping relation table between environmental parameters and performance variation rates through a data fitting method, wherein the environmental influence relation table is initialized before system deployment and supports updating according to newly acquired data in later period.
  8. 8. The dynamic physical ability assessment index generation method for the military training outline according to claim 1, wherein the individual feature vectors are constructed in the step 2, and the adjustment factor calculation model in the step 4 is subjected to fusion analysis, and a uniform time window alignment mechanism is adopted; The unified time window alignment mechanism is that when analysis is carried out, historical data in the individual feature vectors and real-time data in the real-time state data stream are all unified and associated to the same analysis time datum point, and for time series data, features in a specific time window with the analysis time datum point as the center are extracted and calculated.
  9. 9. The dynamic generation method of physical ability assessment indexes for military training outline according to claim 1, wherein the method is characterized in that when applied to different types of physical ability assessment classes, the adjustment factor calculation model in step 4 and the boundary constraint rule in step 5 are configured differently according to class types; for endurance class, the adjustment factor calculation model is more focused on the weights of the health load factor and the environmental compensation factor; for a force type or explosive force type class, the adjustment factor calculation model is more focused on the weight of the basic adaptive offset; The threshold value of the safety and compliance boundary constraint is also set differently according to the common damage risk corresponding to the class type.
  10. 10. The method for dynamically generating physical ability assessment indexes for military training outline according to claim 1, wherein in step 6, after updating the assessment whole process data and adjustment log into the individual feature vector of the soldier to be examined, a model optimization process is triggered; And the model optimization process uses the updated individual feature vector containing the current assessment data, the corresponding real-time state data stream and the personalized dynamic assessment index set of the final application as a new training sample to carry out incremental training on the adjustment factor calculation model in the step 4 so as to optimize the parameters of the adjustment factor calculation model, so that the subsequent index dynamic generation is more accurate.

Description

Physical ability assessment index dynamic generation method for military training outline Technical Field The invention relates to the technical field of military training and health management informatization, in particular to a physical ability assessment index dynamic generation method oriented to a military training outline. Background Military physical training is an important foundation for generating combat force of armies, and scientific and reasonable physical ability assessment plays a key role in evaluating training effects and improving training quality. Currently, the army comprehensively implements a new edition of military sports training outline (hereinafter referred to as outline), and the outline constructs a training assessment system of 'basic physical ability + general combat physical ability + army weapon physical ability of military special combat physical ability', and clearly defines assessment standards and scoring methods of various physical ability classes. However, in the practical training and checking organization implementation process, the existing checking mode still has a plurality of technical problems to be solved urgently. Firstly, the existing assessment indexes generally adopt a static and unified setting mode. The assessment indexes are usually set directly according to fixed standards specified in outline, such as 3000 m running completion time, pull-up times and the like, and the standards can not reflect dynamic factors such as actual physical performance conditions, training history, real-time physiological states, environmental conditions and the like of individual soldiers although basic factors such as age, sex and the like are considered. For example, the physical performance of the same soldier in a plateau environment is significantly different from that of the same soldier in a plain environment, but the existing assessment standards often ignore the environmental adaptability difference, so that the assessment result cannot truly reflect the actual physical performance level of the soldier in a specific combat environment. Second, the prior art lacks dynamic tracking and adaptation of individual physical performance development trajectories. Although some physical training management systems can collect training data and generate a training program, different assessment modes can be set according to preset assessment conditions and actual conditions of trained units. However, the setting is still static configuration based on preset conditions, and real-time dynamic adjustment of the assessment index in the assessment process cannot be realized. Although the system refers to 'real-time monitoring of the physical sign state', a direct and automatic association mechanism is not established between the monitored real-time physiological data and the dynamic generation of the assessment index. Furthermore, the existing assessment mode is disjointed with the individual health risk early warning. Military physical training has the characteristics of high strength and high load, and improper training or examination easily causes training damage and even health risks. Most of the current assessment systems focus on the acquisition and judgment of the final results, and lack a mechanism for dynamically adjusting the assessment intensity based on real-time physiological data (such as heart rate, blood oxygen saturation, body temperature and the like) in the assessment process. This can present a safety hazard to the assessment process, especially for soldiers in physical recovery, with injury training or in special physiological states. In addition, the existing assessment index generation method lacks fusion analysis and intelligent decision of multi-source data. The assessment requirements of military training outline, soldier historical training data, real-time physiological monitoring data, environmental sensor data, equipment load data and the like should constitute comprehensive decision basis for dynamic generation of assessment indexes, but the prior art often carries out isolated processing on the data or only carries out simple linear association, and a deep and nonlinear dynamic association model cannot be established, so that the generated assessment indexes are insufficient in scientificity and individuation degree. Finally, from the technical implementation level, the existing scheme mostly adopts a 'modularized' system architecture, which comprises an identity acquisition module, a state monitoring module, a motion recording module, a processing module and the like. Although the function division of the architecture is clear, the data flow among the modules is solidified, the processing logic is stiff, and the requirements of dynamic and self-adaptive generation of the assessment index according to complex and changeable real-time conditions are difficult to support. The cooperation among the modules often needs manual intervention or preset rules, an