Shooting training method and system

Abstract

The invention discloses a shooting training method and system and belongs to the technical field of intelligent training, and the system is composed of a multi-mode sensing unit, a data processing and analyzing unit and a real-time feedback and interaction unit. The sensing unit synchronously collects electromyographic signals, motion postures, holding pressure, early information of non-contact ballistic trajectories and environmental parameters through a sensor array worn on the body of a shooter and a gun. And the analysis unit associates and processes the information through time synchronization and data fusion, establishes a real-time mapping model of a shooter action mode and an impact point prediction result, and generates a specific correction instruction and an adaptive training parameter according to the real-time mapping model. The feedback unit superposes aiming deviation guidance in the field of view of a shooter through an augmented reality display, and provides real-time prompts through a tactile feedback device integrated on a firearm and a wearable device. According to the method, a closed-loop training process from action execution, trajectory prediction to immediate correction is formed, and real-time quantitative diagnosis and intervention of key technical links in the shooting process are achieved.

Description

Technical Field

[0001] This invention relates to the field of intelligent training technology, and in particular to a shooting training method and system. Background Technology

[0002] Machine gun firing training, especially suppressive and dispersion firing training, aims to enable shooters to effectively control the point of impact within a specific area (i.e., the "shooting gate") during continuous fire. Currently, training methods and assessment systems in this area are significantly inadequate.

[0003] Current techniques primarily rely on pre-set fixed-area targets or simple circular targets, with shooters evaluating dispersion effects afterward by observing bullet hole distribution. This method suffers from several fundamental flaws: First, training feedback is severely delayed. Shooters cannot obtain guidance on real-time bullet impact distribution trends during firing; results can only be viewed after the entire firing sequence, making it difficult to establish an immediate causal link between firing actions and dispersion patterns. Second, existing methods heavily depend on instructors' visual observation and experience-based judgment, lacking objective, quantitative data support. They cannot accurately assess key indicators such as dispersion uniformity, density, and center bias, nor can they correlate dispersion effects with shooters' specific operations (such as rifle stability, breathing rhythm, and burst firing rhythm control). Third, the training model is rigid; the firing gate (i.e., the target area) is usually fixed and cannot be dynamically adjusted according to the shooter's skill level and training progress, nor can it simulate the complex scenarios of target area movement or size changes in actual combat. Summary of the Invention

[0004] The main objective of this invention is to provide a shooting training method and system that can effectively solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A shooting training system, comprising: The multimodal sensing unit is used to collect biomechanical data, weapon attitude data, ballistic characteristic data, and environmental parameter data of the shooter in real time during the firing process. The data processing and analysis unit is communicatively connected to the multimodal perception unit. It is used to receive and fuse the multimodal data, analyze and evaluate the shooter's shooting actions, ballistic performance and overall performance based on the artificial intelligence model, and generate personalized real-time guidance instructions and adaptive training schemes. The real-time feedback and interaction unit is communicatively connected to the data processing and analysis unit, and is used to provide the shooter with multi-sensory feedback based on the real-time guidance instructions, and to present the adaptive training scheme.

[0006] Preferably, the multimodal sensing unit includes: Wearable biomechanical sensing modules include electromyography sensors, inertial measurement units, and pressure distribution sensors placed on key parts of the shooter's body and firearms, used to collect muscle activation signals, limb and weapon movement postures, and grip force data. The non-contact ballistic analysis module includes a laser array and a micro-pressure sensor network located near the firing axis. It is used to detect the spatial position and air disturbance of the projectile as it passes through the virtual monitoring surface before hitting the physical target, so as to perform ballistic prediction and early deviation analysis. The environmental sensing module is used to collect data on temperature, humidity, wind speed, wind direction, light intensity, and sound at the training site.

[0007] The non-contact ballistic analysis module is crucial. By setting up a virtual monitoring surface on the trajectory, it enables early flight state capture and deviation prediction of the projectile, thereby advancing the feedback timing from "after impact" to "during flight," providing a data foundation for the core training method.

[0008] Preferably, the data processing and analysis unit includes: The data fusion engine is used to synchronize and register asynchronous data from multiple sources from different sensors in time and space with microsecond-level accuracy, forming a unified spatiotemporal correlated data stream. The skills assessment model, based on transfer learning and the spatiotemporal correlation data stream, constructs a personalized ability benchmark for shooters and dynamically evaluates them from multiple dimensions such as stability, consistency and environmental adaptability. The adaptive planning engine, based on the output of the skill assessment model and preset training objectives, uses reinforcement learning algorithms to dynamically adjust the difficulty of the training task, the complexity of the scenario, and the intensity of feedback, thereby generating a personalized training path.

[0009] Preferably, the real-time feedback and interaction unit includes: Augmented reality display devices are used to overlay and display real-time aiming deviation indicators, action correction guidelines, virtual aiming aids, and key performance indicators in the shooter's field of vision; Haptic feedback devices, integrated into firearm kits and / or shooter wearable devices, including micro-vibration motors and variable-resistance trigger mechanisms, are used to provide tactile cues and simulations related to trigger control, breathing rhythm, and posture adjustment; An immersive environment simulation system is used to generate and project virtual training scenes that include dynamic targets, variable lighting, weather effects, and sound field simulation.

[0010] A shooting training method based on the aforementioned system includes the following steps: S1: Through the multimodal sensing unit, multi-dimensional data of the shooter during the entire process of shooting preparation, firing and subsequent actions are collected synchronously and in real time; S2: In the data processing and analysis unit, the multi-dimensional data is fused and analyzed to identify technical deviations in the shooting action in real time, and the bullet impact point trend is predicted before the bullet hits the physical target. S3: Based on the analysis results of step S2, the real-time feedback and interaction unit provides the shooter with immediate corrective feedback in the form of visual, auditory and / or tactile sensations immediately after shooting. S4: Accumulate data from multiple shots, evaluate the shooter's overall skill level and progress trajectory through the data processing and analysis unit, and automatically generate an adaptive training plan for subsequent stages.

[0011] Preferably, step S2, "predicting the trajectory of the impact point before the projectile hits the target," specifically includes: Using the non-contact ballistic analysis module, the early flight state of the projectile is captured after it leaves the muzzle and before it reaches the target; Combining the weapon attitude data and environmental parameter data, the ballistic trajectory is calculated in real time using an aerodynamic model; During the projectile's flight, its deviation relative to the expected aiming point is continuously predicted, and this predicted deviation is used to generate the real-time feedback in step S3.

[0012] Preferably, it also includes step S5: constructing a digital twin model of the shooter; The digital twin model is built based on historical training data and can simulate the expected performance of shooters in different virtual training scenarios; Before being applied to actual training, the adaptive training plan is first simulated and evaluated on the digital twin model, and the parameters are optimized accordingly.

[0013] Preferably, the logic for generating the adaptive training plan includes: Monitor shooters' real-time performance success rate, physiological fatigue indicators, and psychological load levels during training; When the success rate remains above the first threshold, the difficulty of the training task will be automatically increased or interfering factors will be introduced. When physiological fatigue indicators or psychological load levels exceed the second threshold, the training intensity is automatically reduced or the training content is switched to restorative training. This keeps the training process dynamically within a preset skill challenge range.

[0014] Compared with the prior art, the present invention has the following beneficial effects: This invention constructs a real-time causal mapping and intervention mechanism for "action-trajectory" through synchronous data acquisition by a multimodal sensing unit, real-time correlation analysis by a data processing unit, and instant interaction by a feedback unit. This solution solves the technical problems of delayed data feedback, difficulty in action attribution, and rigid training patterns in traditional shooting training. It enables objective quantitative assessment of the shooter's technical state, early prediction of bullet impact deviation, and dynamic adjustment of training task parameters based on individual performance, thereby improving the scientific rigor, relevance, and efficiency of the training process. Attached Figure Description

[0015] Figure 1 This is a diagram showing the core components of the shooting training system of the present invention. Figure 2 This is a flowchart of the shooting training method of the present invention. Detailed Implementation

[0016] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0017] like Figures 1-2 The diagram shows the core architecture of the system and the shooting training method. The following is a detailed implementation example.

[0018] I. System Composition and Hardware Configuration Details (a) Multimodal sensing unit Wearable biomechanical sensing module: Equipping shooters with customized sensing vests, gloves, and firearm sensing kits. High-precision electromyography (EMG) sensors are embedded in key muscle groups such as the deltoids, trapezius, and rectus abdominis in the shoulder, capturing muscle activation timing and intensity signals in the sensing vest. Pressure distribution sensors are placed at grip points such as the palm, index finger, and middle finger of the sensing gloves to collect the distribution and dynamic changes of firearm grip pressure in real time. Inertial measurement units are installed in the stock, grip, and barrel of the firearm to accurately record the weapon's posture changes, trajectory, and vibration data throughout the firing process.

[0019] Non-contact ballistic analysis module: Laser arrays are symmetrically arranged on both sides of the firing axis at the shooting training range, forming two layers of virtual monitoring surfaces. Micro-pressure sensors are uniformly distributed in the area between the two virtual monitoring surfaces, forming a micro-pressure sensor network. When the projectile passes through the two laser arrays sequentially, the time difference of laser obstruction combined with the array spacing allows for the calculation of the projectile's flight velocity. Simultaneously, the spatial position of the projectile on the monitoring surface is determined by the coordinates of the laser obstruction. The micro-pressure sensor network captures the air disturbance waves generated during projectile flight, assisting in the correction of ballistic parameters.

[0020] Environmental sensing module: Integrated environmental sensors are installed at the four corners of the training ground and above the shooting positions to simultaneously collect data on temperature, humidity, wind speed, wind direction, light intensity, and environmental noise.

[0021] (II) Data Processing and Analysis Unit It adopts a high-performance edge computing server, equipped with a multi-core processor and a dedicated graphics processing unit, to ensure real-time data processing and parallel computing capabilities.

[0022] Data fusion engine: Built-in high-precision time synchronization algorithm and spatial registration model, it can receive multi-source asynchronous data from wearable biomechanical sensing module, non-contact ballistic analysis module and environmental perception module, quickly complete time synchronization and spatial registration, form a unified spatiotemporal correlated data stream, and provide basic data support for subsequent analysis and evaluation.

[0023] Skill assessment model: Based on transfer learning algorithm, it is first pre-trained using massive shooting training sample data, and then combined with the current shooter's initial training data to quickly build a personalized ability benchmark. The shooter is dynamically assessed from multiple dimensions, including the stability of shooting actions (such as the amplitude of gun shaking and the maintenance of posture at the moment of firing), consistency of actions (muscle activation patterns, grip strength changes, and repeatability of weapon posture adjustments in multiple shots), and environmental adaptability (the degree of fluctuation in shooting performance under different temperature, humidity, wind speed, and other environments).

[0024] Adaptive planning engine: Integrating reinforcement learning algorithms, it dynamically adjusts the difficulty, scenario complexity, and feedback intensity of training tasks based on the output of the skill assessment model and preset training objectives. For example, based on the shooter's hit rate against a fixed target, it adjusts the target's movement speed and frequency of appearance, or adds environmental interference factors, to generate a personalized training path that fits the shooter's current skill level.

[0025] (III) Real-time Feedback and Interaction Unit Augmented Reality Display Device: The device uses a lightweight augmented reality helmet, which allows the shooter to overlay various key information into their real field of vision, including real-time aiming deviation indication, action correction guidance, virtual aiming guide lines, and key performance indicators such as shooting accuracy and bullet impact point deviation distance, without affecting the shooter's perception of the real environment.

[0026] Tactile feedback device: Micro-vibration motors are integrated into the gun grip, stock, and the shooting gloves and vest, while a variable-resistance trigger mechanism is installed in the trigger area. Based on real-time feedback commands, tactile cues are transmitted through different vibration frequencies and intensities of the micro-vibration motors, and the variable-resistance trigger mechanism simulates the trigger force of different firearms, assisting the shooter in adjusting the trigger control rhythm, breathing rhythm, and body posture.

[0027] Immersive Environment Simulation System: Composed of high-definition projection equipment, surround sound speakers, and environmental simulation devices. The high-definition projection equipment can project dynamic targets, variable lighting scenes, and simulated weather effects onto the walls and ground of the training area; the surround sound speakers can reproduce sound fields in different environments, such as the sound of wind in the wild and the noise of the urban environment; the environmental simulation devices can adjust the local temperature and humidity in the training area according to training needs, creating a highly realistic virtual training scene for shooters.

[0028] II. Detailed Implementation Procedures of Shooting Training Methods (a) Pre-training preparation stage The shooter puts on a sensor vest and sensor gloves, checks the installation status of the firearm sensor kit, and ensures that all sensors are properly connected and working stably.

[0029] Wearing an augmented reality headset, test the equipment display to ensure the overlaid information is clearly visible; check the vibration response of the haptic feedback device and the working status of the variable resistance trigger mechanism to ensure the feedback function is normal.

[0030] Shooters enter their basic personal information into the system, including height, weight, years of shooting experience, and preferred firearms. Based on this information, the system initially matches the training difficulty range. Simultaneously, shooters set basic training objectives, such as improving accuracy on stationary targets and enhancing their ability to acquire moving targets.

[0031] The system activates the environmental perception module and begins to continuously collect environmental parameter data of the training ground, providing data support for subsequent ballistic analysis and training scenario adjustments.

[0032] (II) Data Collection Phase The shooter enters the training area and begins shooting training according to the initially set scenario. The system uses a multimodal perception unit to synchronously collect multi-dimensional data in real time throughout the entire process of shooting preparation, firing, and subsequent actions: Wearable biomechanical sensing modules capture activation signals of key muscle groups in the shooter in real time, record limb movement posture, dynamic changes in gun grip force, and weapon posture data during firing. Data acquisition is uninterrupted throughout the entire process from preparing to raise the gun, aiming, firing to resetting after firing.

[0033] After the shooter fires, the projectile leaves the muzzle. The non-contact ballistic analysis module uses a laser array to detect the spatial position of the projectile as it passes through two virtual monitoring surfaces. It then calculates the projectile's velocity by combining the time difference with the velocity. At the same time, a micro-pressure sensor network captures the air disturbance data generated by the projectile's flight.

[0034] The environmental perception module continuously collects environmental parameters such as temperature, humidity, wind speed, and wind direction during the training process, and stores them synchronously with shooting action data and ballistic data.

[0035] (III) Data Processing and Analysis Stage The various types of data collected by the multimodal sensing unit are transmitted to the data processing and analysis unit in real time. The data fusion engine first performs time synchronization and spatial registration on the multi-source data, and integrates muscle activation signals, limb and weapon posture data, grip force data, ballistic characteristic data and environmental parameter data to form a unified spatiotemporal correlated data stream.

[0036] The skills assessment model performs in-depth analysis of spatiotemporally correlated data streams to identify technical deviations in shooting actions in real time. For example, by analyzing muscle activation signals and limb posture data, it can determine whether there is excessive tension in the shoulder muscles causing shaky grip when the shooter raises the gun; by comparing the grip force change curves of multiple shots, it can identify whether there are unnecessary fluctuations in grip force at the moment of firing; and by combining weapon posture data, it can determine whether there are problems such as line of sight deviation or gun tilt during the aiming process.

[0037] Before the projectile hits the physical target, the system uses early flight state data captured by the non-contact ballistic analysis module, combined with weapon attitude data and environmental parameter data, to calculate the trajectory in real time using an aerodynamic model. During the projectile's flight, it continuously predicts its deviation from the expected aiming point and transmits this predicted deviation to the real-time feedback and interaction unit in real time, providing a basis for immediate feedback.

[0038] (iv) Real-time feedback stage Based on the results of the data processing and analysis phase, the real-time feedback and interaction unit provides the shooter with instant corrective feedback in a multi-sensory form immediately after firing: Visual feedback: The augmented reality helmet overlays an aiming deviation indicator in the shooter's field of vision, using lines of different colors to mark the direction and distance of the deviation between the actual aiming point and the ideal aiming point; at the same time, it displays action correction guidance, such as text prompts such as "relax shoulder muscles", "grip the gun evenly", and "adjust breathing rhythm", and overlays virtual aiming auxiliary lines to guide the shooter to adjust the aiming posture for the next shot.

[0039] Tactile Feedback: Based on the identified technical deviations, the tactile feedback device activates the corresponding feedback mechanism. If the shooter's grip strength fluctuates too much at the moment of firing, the corresponding part of the sensing glove will vibrate slightly to prompt the shooter to maintain a stable grip. If the breathing rhythm is not properly coordinated with the firing timing, the sensing vest will transmit the correct breathing rhythm prompts through regular vibrations. If the trigger control force is inappropriate, the variable resistance trigger mechanism will adjust the resistance to guide the shooter to use a better trigger control method.

[0040] Auditory feedback: Voice prompts are issued through the built-in speakers of the augmented reality headset, such as "The bullet impact point is 5 cm to the left, it is recommended to adjust the horizontal angle of the gun" and "The stability of holding the gun is insufficient, please pay attention to the core muscle exertion".

[0041] (v) Adaptive training plan generation phase After a shooter completes multiple shooting training sessions, the data processing and analysis unit accumulates all shooting data. The skill assessment model comprehensively analyzes the shooter's shooting accuracy, bullet impact point deviation distribution, frequency of movement technique deviations, and improvement status to evaluate the shooter's overall skill level. Simultaneously, the system records the shooter's performance changes during training, creating a progress trajectory and visually presenting the shooter's improvements in dimensions such as stability, consistency, and environmental adaptability.

[0042] The adaptive planning engine generates an adaptive training plan for subsequent stages based on the shooter's comprehensive skill level assessment, progress trajectory, and initial training objectives. The generation logic is as follows: The system monitors the shooter's performance success rate in training in real time, such as the hit rate of fixed target shooting and the success rate of dynamic target acquisition; it judges physiological fatigue indicators by analyzing data such as the continuous intensity of muscle activation signals and changes in the amplitude of movement; and it assesses the level of psychological load by combining shooting interval time, number of times of hesitation in movement, etc.

[0043] When the shooter's success rate continues to exceed the first threshold, the system automatically increases the difficulty of the training task, such as speeding up the movement speed of dynamic targets, reducing the size of targets, increasing the number of targets appearing at the same time, or introducing stronger environmental interference factors, such as increasing simulated wind speed or changing light intensity.

[0044] When physiological fatigue indicators or psychological load levels exceed the second threshold, the system automatically reduces the training intensity, such as switching to static fixed target training, extending the shooting interval, or switching to restorative training content, such as basic gun-holding stability exercises and breathing rhythm regulation training, to ensure that the training process is dynamically maintained within the preset skill challenge range, thus ensuring training effectiveness while avoiding skill decline or injury risks caused by excessive fatigue.

[0045] (vi) Digital Twin Model Construction and Training Plan Optimization Phase After multiple training sessions, the system constructs a digital twin model of the shooter based on accumulated historical training data, including shooting action data, ballistic data, technical deviation records, and progress trajectories in different scenarios. This model can accurately simulate the shooter's shooting habits, action characteristics, skill shortcomings, and expected performance in different virtual training scenarios.

[0046] The adaptive training plan generated by the adaptive planning engine is first input into the shooter's digital twin model for simulation before being applied to actual training. The system simulates the execution process of the training plan, observes the skill improvement effect of the digital twin model under the training plan, and identifies potential problems, such as whether it leads to excessive fatigue of specific muscles or whether it can effectively target skill weaknesses.

[0047] Based on the simulation results, the system optimizes and adjusts the parameters of the adaptive training plan. For example, if the simulation reveals that the difficulty of a certain training task increases too quickly and the success rate of the digital twin model drops significantly, the system appropriately reduces the difficulty increase for that stage. If it is found that a certain type of restorative training is ineffective in alleviating physiological fatigue, it is replaced with more suitable training content to ensure that the plan ultimately applied to actual training has higher relevance and effectiveness.

[0048] (vii) Cyclic training and continuous optimization The shooter follows the optimized adaptive training plan for subsequent training. The system repeatedly executes processes such as data collection, processing and analysis, real-time feedback, and data accumulation and evaluation, continuously updating the shooter's historical training data and dynamically optimizing the digital twin model. Simultaneously, based on the shooter's real-time training performance, the difficulty, scenarios, and training content of the adaptive training plan are constantly adjusted, forming a closed-loop training model of "training-evaluation-feedback-optimization-retraining" to help the shooter continuously improve their shooting skills.

[0049] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A shooting training system, characterized in that: include: The multimodal sensing unit is used to collect biomechanical data, weapon attitude data, ballistic characteristic data, and environmental parameter data of the shooter in real time during the firing process. The data processing and analysis unit is communicatively connected to the multimodal perception unit. It is used to receive and fuse the multimodal data, analyze and evaluate the shooter's shooting actions, ballistic performance and overall performance based on the artificial intelligence model, and generate personalized real-time guidance instructions and adaptive training schemes. The real-time feedback and interaction unit is communicatively connected to the data processing and analysis unit, and is used to provide the shooter with multi-sensory feedback based on the real-time guidance instructions, and to present the adaptive training scheme.

2. The shooting training system according to claim 1, characterized in that: The multimodal sensing unit includes: Wearable biomechanical sensing modules include electromyography sensors, inertial measurement units, and pressure distribution sensors placed on key parts of the shooter's body and firearms, used to collect muscle activation signals, limb and weapon movement postures, and grip force data. The non-contact ballistic analysis module includes a laser array and a micro-pressure sensor network located near the firing axis. It is used to detect the spatial position and air disturbance of the projectile as it passes through the virtual monitoring surface before hitting the physical target, so as to perform ballistic prediction and early deviation analysis. The environmental sensing module is used to collect data on temperature, humidity, wind speed, wind direction, light intensity, and sound at the training site.

3. The shooting training system according to claim 2, characterized in that: The data processing and analysis unit includes: The data fusion engine is used to synchronize and register asynchronous data from multiple sources from different sensors in time and space with microsecond-level accuracy, forming a unified spatiotemporal correlated data stream. The skills assessment model, based on transfer learning and the spatiotemporal correlation data stream, constructs a personalized ability benchmark for shooters and dynamically evaluates them from multiple dimensions such as stability, consistency and environmental adaptability. The adaptive planning engine, based on the output of the skill assessment model and preset training objectives, uses reinforcement learning algorithms to dynamically adjust the difficulty of the training task, the complexity of the scenario, and the intensity of feedback, thereby generating a personalized training path.

4. The shooting training system according to claim 1, characterized in that: The real-time feedback and interaction unit includes: Augmented reality display devices are used to overlay and display real-time aiming deviation indicators, action correction guidelines, virtual aiming aids, and key performance indicators in the shooter's field of vision; Haptic feedback devices, integrated into firearm kits and / or shooter wearable devices, including micro-vibration motors and variable-resistance trigger mechanisms, are used to provide tactile cues and simulations related to trigger control, breathing rhythm, and posture adjustment; An immersive environment simulation system is used to generate and project virtual training scenes that include dynamic targets, variable lighting, weather effects, and sound field simulation.

5. A shooting training method based on the system described in any one of claims 1-4, characterized in that: Includes the following steps: S1: Through the multimodal sensing unit, multi-dimensional data of the shooter during the entire process of shooting preparation, firing and subsequent actions are collected synchronously and in real time; S2: In the data processing and analysis unit, the multi-dimensional data is fused and analyzed to identify technical deviations in the shooting action in real time, and the bullet impact point trend is predicted before the bullet hits the physical target. S3: Based on the analysis results of step S2, the real-time feedback and interaction unit provides the shooter with immediate corrective feedback in the form of visual, auditory and / or tactile sensations immediately after shooting. S4: Accumulate data from multiple shots, evaluate the shooter's overall skill level and progress trajectory through the data processing and analysis unit, and automatically generate an adaptive training plan for subsequent stages.

6. The method according to claim 5, characterized in that: Step S2, "predicting the trajectory of the impact point before the projectile hits the target," specifically includes: Using the non-contact ballistic analysis module, the early flight state of the projectile is captured after it leaves the muzzle and before it reaches the target; Combining the weapon attitude data and environmental parameter data, the ballistic trajectory is calculated in real time using an aerodynamic model; During the projectile's flight, its deviation relative to the expected aiming point is continuously predicted, and this predicted deviation is used to generate the real-time feedback in step S3.

7. The method according to claim 5, characterized in that: It also includes step S5: constructing a digital twin model of the shooter; The digital twin model is built based on historical training data and can simulate the expected performance of shooters in different virtual training scenarios; Before being applied to actual training, the adaptive training plan is first simulated and evaluated on the digital twin model, and the parameters are optimized accordingly.

8. The method according to claim 5, characterized in that: The logic for generating the adaptive training plan includes: Monitor shooters' real-time performance success rate, physiological fatigue indicators, and psychological load levels during training; When the success rate remains above the first threshold, the difficulty of the training task will be automatically increased or interfering factors will be introduced. When physiological fatigue indicators or psychological load levels exceed the second threshold, the training intensity is automatically reduced or the training content is switched to restorative training. This keeps the training process dynamically within a preset skill challenge range.

Images