Dynamic comprehensive training system for highland operation ability reinforcement

Abstract

The application aims to provide a dynamic comprehensive training system for enhancing high field operation capability; comprising multiple functional modules in a closed loop linkage of "perception-judgment-execution-feedback": a data analysis module for calculating the deviation of each monitored quantity relative to the target or safety threshold and performing state recognition; a prediction decision module for generating training stage switching and parameter adjustment strategies based on the analysis results; a control execution module for automatically controlling continuous variables such as treadmill speed / slope, oxygen concentration / air pressure, etc.; a feedback and alarm module for real-time sampling correction of the execution results and issuing an alarm when an anomaly occurs. The "individualized threshold-real-time deviation-predicted trend-execution amplitude-feedback correction" can be applied to every training moment, thereby realizing the progressive enhancement and steady-state maintenance of intensity under the premise of ensuring safety.

Description

Technical Field

[0001] This invention belongs to the field of high-altitude training, specifically relating to a dynamic comprehensive training system for enhancing high-altitude operational capabilities. Background Technology

[0002] For a long time, training for military and police personnel and those working in high-altitude areas has been conducted primarily in plains or simplified simulated conditions. Training intensity and environmental factors (such as oxygen concentration, air pressure, temperature, humidity, wind speed, and illumination) have largely relied on coaching experience or single-variable thresholds for manual grading and adjustment. Exercise loads (such as treadmill speed, incline, or machine resistance) have also been roughly set using stage charts or heart rate limit methods. This "offline planning—online execution" approach lacks continuous quantitative capture of individual responses to multiple physiological indicators (heart rate HR, blood oxygen saturation SpO2, ventilation, etc.) under hypoxic conditions. Status assessment during training relies heavily on single-point thresholds or subjective observation, resulting in discontinuous increases in training intensity, inaccurate maintenance of steady state, and delayed responses to sudden discomfort. This not only affects training efficiency but also creates safety hazards.

[0003] Although related systems have begun to introduce environmental simulation devices and wearable sensors in recent years, they generally still remain in an open process of "collection - display - manual judgment - manual adjustment", lacking a fully closed-loop control framework of "perception - judgment - execution - feedback"; at the algorithm level, they are mostly guided by fixed thresholds or simple rules, lacking the predictive and adaptive capabilities to couple real-time deviations with historical trends. Specifically: (1) Insufficient fusion of multi-source data: Heart rate, blood oxygen and environmental variables are often evaluated separately, making it difficult to form a joint judgment on the coupling relationship of "load-oxygen consumption-oxygen supply"; (2) Limited control precision: The environment and equipment are often adjusted in a step-by-step or manual manner, without the introduction of continuous control strategies such as PID to suppress overshoot and steady-state error, making it difficult to smoothly approach and stably maintain the training target range (such as the target heart rate band); (3) Lack of intelligent decision-making: There is insufficient comprehensive analysis and fuzzy reasoning / machine learning classification of "deviation size × trend of change", and the switching between the reinforcement stage and the maintenance stage often occurs prematurely or delayed; (4) Delayed safety response: When SpO2 drops rapidly below the lower limit or the heart rate overshoots, there is a lack of millisecond-level emergency load reduction, oxygenation and shutdown linkage based on rules and permission management, and abnormal control still relies on manual handling, which poses a risk window.

[0004] Therefore, a dynamic and comprehensive training technology approach deeply coupled with the training process is urgently needed: based on parallel data acquisition from both physiological and environmental sensors, a system architecture is constructed with data analysis and predictive decision-making at its core, and execution control and feedback monitoring forming a closed loop; on the algorithm side, SVM / regression is used for state recognition and short-term trend prediction, fuzzy rules are used to achieve multi-indicator collaborative parameter tuning, and PID is used to achieve fine control of continuous variables such as treadmill speed / incline and oxygen concentration / air pressure; on the safety side, anomaly thresholds and priority strategies are introduced to achieve automated linkage from early warning to emergency shutdown and rapid oxygenation. The above closed-loop mechanism can connect "individualized threshold—real-time deviation—predictive trend—execution amplitude—feedback correction" to every training moment, thereby achieving gradual intensity enhancement and steady-state maintenance while ensuring safety. This is precisely the technical gap and application pain point that the specific implementation method of this application aims to solve. Summary of the Invention

[0005] The purpose of this invention is to provide a dynamic comprehensive training system for enhancing high-altitude work capabilities. This invention can integrate "individualized threshold - real-time deviation - predicted trend - execution range - feedback correction" into every training moment, thereby achieving gradual enhancement and steady-state maintenance of intensity while ensuring safety.

[0006] To achieve the above objectives, the present invention provides the following technical solution, comprising: a dynamic comprehensive training system for enhancing high-altitude work capabilities, including multiple functional modules linked in a closed loop according to "perception-judgment-execution-feedback": a physiological and environmental data acquisition module for acquiring and preprocessing heart rate, blood oxygen saturation (SpO2), lung ventilation, and oxygen concentration, air pressure, and temperature in the training room; a data analysis module for calculating the deviation of each monitored quantity relative to the target or safety threshold and performing state identification; a prediction and decision module for generating training phase switching and parameter adjustment strategies based on the analysis results; a control execution module for automatically controlling continuous variables such as treadmill speed / incline and oxygen concentration / air pressure; and a feedback and alarm module for real-time sampling and correction of execution results and issuing alarms when abnormalities occur; wherein, the data analysis module uses support vector machine (SVM) to perform pattern recognition and classification of multiple physiological parameters, and the control execution module supports continuous fine adjustment based on PID and / or fuzzy control rules, thereby forming a closed-loop control structure for the training process.

[0007] Furthermore, the control execution module is based on the error between the controlled target and the reference heart rate. The control quantity u(t) for the treadmill speed is calculated, and overshoot and steady-state error are suppressed by a combination of proportional, integral and derivative terms, so as to achieve smooth approximation and stable maintenance of heart rate to the reference value.

[0008] Furthermore, the prediction and decision-making module incorporates fuzzy control rules, including at least a collaborative constraint that prohibits increasing training intensity or slightly decreasing slope / speed when heart rate is high and blood oxygen is low, in order to stably maintain intensity in the critical zone of high-altitude endurance training.

[0009] Furthermore, when the feedback and alarm module detects a combination of abnormalities, such as SpO2 momentarily dropping below the safety lower limit and heart rate exceeding the target range, it triggers an emergency response procedure: the control execution module issues an emergency stop / reduction command with the highest authority, while simultaneously rapidly increasing the oxygen concentration to the plain level and raising the air pressure, during which saturation control is used to respond in milliseconds.

[0010] Furthermore, the system supports individualized models and threshold adaptation: before training, the controller retrieves individual resting heart rate, previous training data, or maximum oxygen uptake information from the database to dynamically adjust the SVM classification threshold and the safety / target interval boundary.

[0011] Furthermore, the feedback and alarm module is equipped with an acoustic alarm device, which can simultaneously issue a high-intensity audible alarm to alert on-site personnel after a dangerous situation is identified.

[0012] Furthermore, the system includes a training data storage module, which archives heart rate curves, blood oxygen changes, environmental parameters, and execution control records after training to support subsequent training prescription optimization and evaluation review.

[0013] Furthermore, dynamic and comprehensive training methods to enhance high-altitude work capabilities include: S1. Training preparation: Set the initial training room to a plain level (approximately 21% oxygen concentration, standard air pressure). Trainees wear heart rate and blood oxygenation sensors and load their individual historical data. The goal is to gradually reduce the oxygen concentration to approximately 15%, increase the running speed to approximately 12km / h, and at the same time constrain SpO2 to not be lower than 90% and heart rate to not exceed the aerobic limit. S2, Training Start and Climb: The control execution module reduces the oxygen concentration at a preset slope and accelerates uniformly to the target speed, while the analysis module continuously calculates the deviation. S3, Phase Maintenance: When approaching the critical zone, the fuzzy rule of "no increase in intensity if heart rate is high and blood oxygen is low" is implemented, and the slope / speed is finely adjusted to maintain stability. S4. Termination and Recovery: Smoothly revert the environment and speed to a flat and resting state in a stepwise manner and continue monitoring until the normal range is restored.

[0014] Furthermore, in the S3 phase, the slight reduction in slope and the slight decrease in speed are determined by the prediction and decision-making module based on short-term trends and deviations, so that the heart rate is maintained at 148~152 bpm, blood oxygen is maintained at 90%~91%, and no alarm is triggered.

[0015] Furthermore, when the abnormal combination of "SpO2≤90% and heart rate exceeds the target" occurs at any time from S1 to S3, the system directly enters the emergency response branch: interrupts the treadmill motor, activates oxygen enrichment and pressurization to restore the maximum throughput quickly, and samples blood oxygen and heart rate at a high frequency during the execution until the danger zone is cleared, and then enters the recovery process.

[0016] This invention constructs a fully closed-loop training system of "perception-judgment-execution-feedback". Through parallel acquisition, fusion analysis and predictive decision-making of multi-source physiological / environmental data, it links the treadmill and the execution end such as oxygen supply / pressure to achieve continuous and precise control. Under the premise of ensuring that safety thresholds (such as SpO2, heart rate limit) are not exceeded, it achieves progressive enhancement and steady-state maintenance of training intensity. Compared with the existing open process of "acquisition-display-manual judgment-manual adjustment", it significantly reduces the overshoot / undershoot and risk window caused by the lag of manual intervention, and improves the repeatability, individualization and traceability of training.

[0017] Multi-source fusion and state recognition: By unifying the modeling of heart rate, blood oxygen, ventilation volume and environmental quantities such as oxygen concentration, air pressure, and temperature, and using SVM / rule base for state recognition, this system overcomes the shortcomings of existing technologies in "fragmented evaluation of indicators". It can provide training judgments that are more in line with the physiological response at high altitude under the coupling relationship of "load-oxygen consumption-oxygen supply", reduce misjudgment and wrong judgment, and improve the effectiveness of training prescriptions and current parameter adjustments.

[0018] Continuous control suppresses overshoot and steady-state error; PID / fuzzy control is used to fine-tune continuous variables such as speed, slope, and oxygen concentration with small steps and high frequency, which can effectively suppress overshoot and steady-state error of heart rate / SpO2 on the target band, improve the common problem of "not being able to go up, not being able to stay stable, and not being able to come down", make the heart rate approach and stay stable in the target range more quickly, and reduce fluctuations and discomfort during training.

[0019] The system maintains a safe steady state through synergistic constraints. It introduces a synergistic constraint that prohibits increasing intensity / reverse fine-tuning when heart rate is high and blood oxygen is low. This solves the sawtooth oscillation problem of "increasing intensity - deoxygenating - decreasing - increasing intensity" caused by the use of a single index threshold in the existing system. It promotes smooth maintenance in the critical range, thereby achieving a longer effective aerobic load and better training adaptation.

[0020] Rapidly linked anomaly handling: When a combination of abnormalities occurs, such as SpO2 falling below the lower limit and heart rate exceeding the target, the system automatically triggers a priority linkage of "emergency stop / deload + oxygen enrichment + pressurization". The response link is short and the authority is the highest, which significantly shortens the time from anomaly identification to effective handling, reduces the risk of hypoxia-related events and exercise-related accidents, and makes up for the shortcomings of manual handling, such as "slow identification and slow action".

[0021] Individualized thresholds and model adaptation: before training, individual resting heart rate, VO2 information and historical curves are loaded, the target band and safety limit are dynamically set, and the classification threshold and control gain are adjusted according to short-term trends during training. This overcomes the crude setting of "one-size-fits-all" and enables people with different altitude adaptation levels to obtain more suitable intensity curves and more controllable physiological responses.

[0022] Acoustic alarm and human-machine co-governance: While the system automatically handles the situation, a high-intensity audible alarm alerts on-site personnel, realizing a two-level safety mechanism of "machine control as the main approach and human backup," reducing the risk of missed or delayed judgments due to distraction or excessive personnel workload, and meeting the high safety redundancy requirements of military, police, and training scenarios.

[0023] Training data is retained and reviewed throughout the entire training process; all pre-training settings, process deviations, execution instructions, and physiological responses are fully recorded, facilitating review analysis, threshold recalibration, and prescription optimization. Each training session can progressively improve the efficiency of individual adaptation paths; at the same time, it provides objective evidence for follow-up evaluation, team stratification, and equipment iteration.

[0024] Progressive reinforcement and fatigue management; the methodology breaks down the process into stages of “start-climb-maintenance-recovery”, and during the maintenance period, the adjustment range and direction are determined based on “deviation × trend”. This not only ensures a steady output of effective training load, but also avoids premature load reduction or late shutdown due to short-term fluctuations, thus improving the contradiction of “coarse intensity steps and rapid fatigue accumulation” in traditional fixed-program training.

[0025] The system is easy to implement and promote; the control objects and interfaces (treadmill, oxygen supply / pressurization, monitoring terminal) are all standardized modules, and the algorithms and rules can be deployed on embedded / host computer platforms; the system has good scalability and portability, is compatible with different brands of equipment, reduces modification and maintenance costs, and is easy to implement quickly in training rooms, mobile cabins or hospital rehabilitation departments.

[0026] Compliance and objective verifiability; key performance indicators can be evaluated using engineering quantification indicators (such as residence time within the target heart rate band, number and duration of SpO2 punctures, overshoot amplitude of control commands, response delay from abnormality to treatment, frequency of manual intervention, etc.), which facilitates objective comparison with existing processes, supports empirical evidence and clinical translation, and meets the requirements of patent implementation feasibility and examination verifiability.

[0027] In summary, this invention, with its closed-loop structure of "multi-source fusion + predictive decision-making + continuous control + abnormal linkage," systematically solves the core problems of data fragmentation, human lag, coarse control, and slow safety response in existing technologies. It achieves significant technical effects and application value in terms of safety, effectiveness, personalization, and traceability. Attached Figure Description

[0028] Figure 1The system's overall structure is illustrated in the schematic diagram, showing the modular architecture and data / control flow of a dynamic integrated training system for enhancing high-altitude operational capabilities. The left side represents the input layer: physiological signal acquisition module 100, environmental perception module 110, cognitive testing module 120, and historical data and parameter database 290. The middle layer is the processing layer: data transmission and aggregation 200, standardization processor 210, and sampling strategy management (adaptive) 280. The right side is the evaluation and fusion layer: multidimensional coordinate mapping 220, regression evaluation 230, and neural network evaluation 240. The far right layer is the result and interaction layer: fusion and grading (Z–I–V) 250 and display / alarm / uplink communication 260 (which may include an uplink submodule 270). Solid arrows in the diagram represent data flow and execution command flow, while dashed arrows represent control and feedback loops from 260 to 230 / 240 / 280, forming a closed loop of "perception—judgment—execution—feedback".

[0029] Figure 2 The control logic flowchart illustrates the timing control flow from "start" to "display / record" during system operation: multi-source data acquisition → standardization / filtering → state recognition (SVM / rule-based) → short-term trend prediction → control decision (PID / fuzzy logic) → execution end (treadmill / oxygen supply / pressure boost, etc.) → display and recording. A "safety check (SpO2 / HR)" branch is included in the flow: if safe, the closed loop continues; if unsafe, it reverts to the control decision node via the "load reduction / emergency stop" path, triggering parameter reduction or emergency shutdown and re-entering the closed loop, reflecting real-time adaptation and safety priority.

[0030] Figure 3 Safety linkage and priority strategy diagram; illustrating the hierarchical safety response mechanism and the trigger-action correspondence. L1 (Warning): When the indicator approaches the threshold (e.g., SpO2≤92% or HR≥HR_ref+5%), the action is an audible and visual warning, prompting attention; L2 (Automatic load reduction): When a minor over-limit occurs (e.g., SpO2≤90% or HR≥HR_ref+10%), the action is to automatically reduce speed / gradient and slightly increase oxygen; L3 (Emergency shutdown + oxygen enrichment and pressurization): When a severe combined anomaly occurs (e.g., SpO2≤88% and HR≥HR_ref+10%), the action is an emergency shutdown and maximum flux oxygenation / pressurization. The diagram uses vertical arrows to represent the "trigger condition → action execution" chain for each level, and horizontal arrows to represent the direct escalation principle when a higher-level condition is met (priority L3>L2>L1).

[0031] Figure 4The training phase sequence diagram shows the division of the entire training process and key physiological thresholds in timeline form: S1 Initiation (establishing baseline and slow start), S2 Climbing (gradually increasing intensity and hypoxia), S3 Maintenance (minor adjustments within the "target heart rate band" to stabilize output), and S4 Recovery (smooth reduction of environment and load). The target heart rate band and the lower limit of SpO2 are marked on the diagram, indicating that the heart rate curve is maintained within the target range during the maintenance period through minor parameter adjustments, while avoiding falling below the safe threshold, reflecting the control characteristics of "gradual enhancement - steady-state maintenance - smooth recovery". Detailed Implementation

[0032] The system structure, control logic, and application scenarios of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. The present invention provides a dynamic comprehensive training system for enhancing high-altitude work capabilities. Through the organic cooperation of various functional modules, a closed-loop control structure of "perception-judgment-execution-feedback" is constructed to achieve intelligent monitoring and adaptive regulation of the training process. The system modules are clearly defined and have distinct responsibilities. The following describes the system module composition and functions, the interactive control relationships between modules, the implementation of key control algorithms, and specific application embodiments.

[0033] The high-altitude operation capability enhancement training system comprises multiple functional modules, which work together to form a complete closed-loop control system. The composition and function of each module are as follows: Physiological data acquisition module: This module collects various physiological parameters of trainees and training environment parameters in real time. For example, it acquires physiological data such as heart rate, blood oxygen saturation (SpO2), and lung ventilation through wearable sensors, and environmental data such as oxygen concentration, air pressure, and temperature in the training room through environmental sensors. The physiological data acquisition module performs preliminary filtering and formatting of the collected raw data to form a standardized data structure, and then sends the data to the analysis module for further processing.

[0034] The data analysis module evaluates and judges the data uploaded by the acquisition module, including current training status analysis and safety status monitoring. This module receives real-time data such as heart rate and SpO2, calculates the deviation from preset safety thresholds or target values, and extracts characteristic indicators such as trainee fatigue level and cardiopulmonary load. Based on machine learning algorithms such as Support Vector Machine (SVM), the data analysis module performs pattern recognition and classification on multiple physiological parameters to determine the trainee's current state (e.g., normal, fatigue warning, or hypoxia risk). The analysis results are output to the prediction and control execution modules as a decision-making basis. If a physiological indicator is detected to be outside the normal range, the analysis module will also generate an alarm signal and transmit it to the feedback module to trigger timely safety protection measures.

[0035] The predictive decision-making module is used to make proactive judgments and plans for the training process based on data and historical trends provided by the analysis module. This module continuously receives the latest analysis results and historical data, and estimates the development trends of key physiological indicators in the near future using predictive models. For example, using SVM regression or neural network models built from high-altitude training big data, it predicts the potential levels of the trainee's heart rate and blood oxygen after maintaining the current training intensity for several minutes. Based on the combined analysis and prediction results, the predictive decision-making module generates corresponding control strategies and adjustment plans (decision outputs), including instructions for adjusting the training environment and training load. When the execution conditions are met, the output of this module triggers the control execution module to implement corresponding actions, realizing the transformation from "judgment" to "execution".

[0036] The control execution module receives control commands from the predictive decision module and drives the training equipment and environmental simulation devices to perform corresponding adjustments. This module includes an environmental control submodule and a training equipment control submodule. The environmental control submodule adjusts the simulated environmental parameters in the training room according to commands, such as regulating oxygen concentration and air pressure (simulating different altitudes), temperature, humidity, wind speed, and light intensity, to dynamically approximate the preset plateau target values. The training equipment control submodule adjusts the working status of the training equipment according to commands, such as treadmill speed and incline, and training load weight. The control execution module integrates a PID controller and fuzzy control logic to finely control key variables in the above execution process (see the control algorithm section below for details), ensuring that the actual execution effect matches the decision requirements. When a control command is issued, the execution module executes the actions of each submodule sequentially and transmits the feedback data generated during the execution process to the feedback module.

[0037] Feedback Monitoring Module: This module aggregates sensor feedback information after control execution to achieve closed-loop regulation. It receives real-time data from physiological and environmental sensors (reflecting the physiological response and environmental changes after the action), compares this data with the expected target of the control command, and generates a feedback error signal. The feedback monitoring module returns this error signal to the data analysis module for evaluation and correction of the control effect. Furthermore, if an anomaly occurs during execution (e.g., the trainee's physiological indicators continue to deteriorate or the equipment fails to reach the target state), the feedback module immediately issues a warning via alarms and can trigger the control execution module to enter a safety mode (e.g., reducing exercise intensity or increasing oxygen supply). Through the feedback monitoring module, data from each module circulates, prompting the system to continuously correct and optimize the control strategy, thus forming a complete closed-loop control chain.

[0038] Each module has its own function and works in concert: the physiological data acquisition module provides sensory information input; the data analysis and prediction decision-making module completes the judgment and decision-making functions; the control execution module implements action control; and the feedback monitoring module performs effect evaluation and feedback loop. The modules are connected via wired or wireless communication networks, and control variables and data are transmitted between modules according to predetermined protocols, achieving high-speed information flow and timely response. For example, the physiological data acquisition module transmits sensor data such as heart rate to the analysis module in real time. After judgment, the analysis module generates control variables (such as heart rate error values) which are then processed by the prediction decision-making module, ultimately forming a control signal that is sent to the execution module. The actions of the execution module cause changes in the environment and physiological state. After the feedback module detects the new data, it sends it back to the analysis module for the next round of processing, thus maintaining a stable and controllable training state. This continuous loop system, consisting of data collection, analysis and prediction, decision-making, and execution feedback, ensures efficient closed-loop regulation and real-time response capabilities during the training process.

[0039] Based on the above module division, the control flow of the system of the present invention is as follows, with each step depending on the others in sequence and forming a dynamic closed loop: Step 1: Sensor Acquisition. At the start of training, the physiological data acquisition module activates, monitoring and collecting key parameters of the trainee and environment in real time. For example, surface sensors continuously measure the trainee's heart rate (HR(t), blood oxygen saturation (SpO2(t)), and lung ventilation; ground pressure sensors measure the force applied to the treadmill pedals; and environmental sensors measure oxygen concentration, air pressure, and temperature. After initial filtering, noise reduction, and format conversion, the collected data is sent to the data analysis module at predetermined time intervals. The acquisition module operates continuously throughout the training process, ensuring a constant flow of sensory information into the system, providing foundational data for subsequent analysis.

[0040] Step 2: Data Analysis. The data analysis module receives the multi-source data from Step 1 and immediately assesses the current training status. First, the module compares each physiological indicator with the normal range or training target, calculating the deviation value. For example, let the heart rate deviation eHR = HR(t) - HRtarget, and the blood oxygen deviation eSpO2 = SpO2(t) - SpO 2,target(Using target heart rate and blood oxygen as references). Through this type of difference calculation, the analysis module obtains the degree of deviation of the trainee's physiological state. Furthermore, the system can pre-set safety thresholds for each indicator (such as blood oxygen below 90%, heart rate above a certain upper limit, etc.), and the analysis module judges whether the indicators exceed the limits in real time. Once eHR or eSpO2 is found to exceed the safe range, the current state is immediately marked as abnormal. Next, the analysis module uses a Support Vector Machine (SVM) model to perform pattern classification on the current multidimensional physiological data: the real-time acquired feature vector X=[HR,SpO2,...] is input into the trained SVM classifier, and the discriminant function f(X) is calculated to identify the trainee's state category. The discriminant function is in the form of the formula: In the formula, X i y represents the support vectors of the SVM model. i The corresponding category labels are defined (e.g., +1 for "normal state", -1 for "state warning"), α is the weight coefficient of the support vector, K() is the kernel function (e.g., RBF kernel), and b is the model bias. By calculating the sign of f(X), the analysis module determines whether the trainee is in a safe physiological state at the current moment. If the judgment result is positive, it means that each indicator is within the normal range; if it is negative, it indicates that there is a risk such as fatigue or hypoxia that requires attention. The analysis module organizes the judged state labels, the current values ​​and deviations of each indicator, and the changing trends over a period of time into an analysis result data package and sends it to the prediction and decision module. At the same time, if an abnormality occurs (e.g., blood oxygen is below 90%), the analysis module triggers an alarm signal to the feedback module to control the alarm to sound, reminding on-site personnel to pay attention to safety.

[0041] Step 3: Predictive Decision-Making. After receiving the analysis results, the predictive decision-making module makes forward-looking control decisions regarding the training process. First, the module uses a built-in predictive model to make short-term predictions of key physiological parameters. For example, it uses support vector regression (SVR) or fuzzy neural network models to estimate HR(t+Δt) and SpO2(t+Δt) after a future time interval Δ using recent heart rate and blood oxygen data sequences. Through prediction, the system can anticipate the changing trends of the trainee's condition (e.g., whether the heart rate will continue to rise excessively, or whether blood oxygen will further decrease). Next, the predictive decision-making module combines the current analysis status and prediction results to formulate corresponding control strategies: when the prediction shows that all indicators remain safe and the trainee's condition is good, the strategy may be to gradually increase the training intensity (e.g., slightly increase the treadmill speed or decrease the ambient oxygen concentration to increase the load); if it is predicted that a certain indicator is about to exceed the safe threshold, the strategy is to reduce the current training difficulty (e.g., decrease the treadmill speed or suspend further reduction of environmental pressure) to prevent risks from occurring. During the decision-making process, the module employs a fuzzy control algorithm to map the combined situation of multiple indicators to specific adjustment ranges. For example, fuzzy rules are set such as "if the heart rate deviation is large and the blood oxygen level is low, then significantly reduce the training intensity," and "if the heart rate deviation is moderate and the blood oxygen level is normal, then slightly increase the training intensity." When multiple rules are satisfied simultaneously, the final adjustment instruction value is obtained through fuzzy inference. After the above judgment, the prediction and decision-making module outputs a specific set of control instructions, including environmental parameter adjustment instructions (such as how much to increase / decrease oxygen concentration, wind speed adjustment, etc.) and training equipment parameter instructions (such as how much to increase / decrease treadmill speed in m / s, how much to change in incline, etc.). These instructions, along with the key variables in the prediction process, are sent to the control execution module for execution.

[0042] Step 4: Control Execution. After receiving the decision command, the control execution module performs control operations on the corresponding equipment according to the predetermined priority and sequence. First, the environmental control submodule adjusts the various subsystems of the plateau environment simulation equipment according to the command. For example, when the predictive decision requires "reducing oxygen concentration to simulate higher altitude," the control air-oxygen mixing system or air pressure control system begins to operate: by extracting air from the training room or reducing the oxygen supply, the air pressure / oxygen content in the training room gradually decreases; to avoid discomfort caused by sudden changes, the system uses PID control to finely adjust the rate of air pressure decrease, ensuring that the actual oxygen concentration decreases smoothly along the reference trajectory. Similarly, the temperature control subsystem, humidity control subsystem, and wind and light control subsystem adjust to the specified value range after the command is issued. At the same time, the training equipment control submodule executes adjustment commands for equipment such as treadmills. For example, when the decision command requires "increasing the treadmill speed by 0.5 m / s," the control module drives the treadmill motor to accelerate and adjusts the incline angle accordingly to increase the treadmill exercise intensity. During execution, the control module internally runs a PID closed-loop control algorithm: using heart rate deviation as feedback, it adjusts the treadmill speed in real time to ensure the trainee's heart rate gradually approaches the target heart rate range. The mathematical expression of the PID control algorithm is as follows: Where e(t) is the current control error (e.g., heart rate error eHR), u(t) is the PID control output (e.g., treadmill speed adjustment), and KP, KI, and KD are the proportional, integral, and derivative coefficients, respectively. By adjusting these three parameters, the system response speed, steady-state error, and overshoot can be changed, thereby precisely controlling the treadmill speed changes. After each sub-action is completed, the control execution module records the new environmental state and equipment operating state parameters and sends the relevant data to the feedback monitoring module.

[0043] Step 5: Feedback Adjustment. While the execution module's adjustments take effect, the feedback monitoring module continuously collects feedback on changes in the trainee's physiological signals and environmental parameters. At this point, the physiological data acquisition function from Step 1 works in conjunction with the feedback module, sending a new round of heart rate, blood oxygen, and other data after the action to the feedback module. The feedback module compares this data with the state before execution to evaluate the effectiveness of the control command. For example, if, 5 minutes after the treadmill speed is increased by 0.5 m / s, the trainee's heart rate rises from 140 bpm to 150 bpm and approaches the target heart rate of 155 bpm, while blood oxygen remains at 95%, it indicates that the speed increase command has achieved the expected effect and is safe; the feedback module generates a positive feedback signal accordingly, indicating that the current strategy is effective. Conversely, if the feedback monitoring detects a heart rate spike exceeding the target upper limit and blood oxygen dropping below 90%, it considers the control action too strong, posing a potential risk. At this point, the feedback module generates negative feedback: on the one hand, it immediately notifies the analysis module to mark the state as dangerous, triggering an alarm and notifying the execution module to make emergency corrections (such as immediately reducing the treadmill speed or increasing oxygen supply); on the other hand, it transmits the feedback error information to the prediction and decision-making module to correct subsequent control strategy parameters (such as reducing the next speed increase) to avoid similar overshooting from happening again. Through this feedback mechanism, the system achieves supervision and adjustment of its own control behavior, enabling training control to form a closed-loop adaptive process. Afterward, the system re-enters a new cycle of step 1, continuously updating the state based on feedback, iterating continuously until training ends.

[0044] Through the cyclical operation of steps 1 to 5 above, the present invention constructs an intelligent decision-making closed loop for high-altitude training: from acquiring sensory information, to state judgment and prediction, and then to control execution and feedback correction, each link is dynamically connected to form a closed-loop control system. This allows the system to autonomously adjust the training plan based on the trainee's real-time state, ensuring safe and efficient training while enhancing high-altitude operational capabilities.

[0045] Control Algorithm Design and Implementation To ensure the accuracy and intelligence of the control process, this system comprehensively utilizes various technologies, including traditional PID control, fuzzy control, and machine learning algorithms (support vector machines). The application principles and parameter settings of each algorithm in this system are as follows: PID control algorithm PID (Proportional-Integral-Derivative) control is a classic closed-loop control algorithm, primarily used in this system for fine-tuning continuous variables, such as the automatic control of treadmill speed and oxygen concentration in a training room. The PID controller continuously calculates the control output u(t) based on the real-time feedback error e(t), and its algorithm formula is as follows: As shown. The proportional term coefficient KP determines the responsiveness to instantaneous errors, the integral term coefficient KI is used to eliminate steady-state errors, and the differential term coefficient KD is used to suppress overshoot caused by rapid error changes. In implementation, initial parameters KP, KI, and KD are first selected based on experience and experimental data, and then continuously adjusted through feedback during training: for example, when the system response is found to be slow or has a large steady-state error, KP or KI can be appropriately increased; when overshoot and instability occur, KD can be increased to enhance damping. Taking the control of treadmill speed as an example, let the target heart rate be HRref, and the current measured heart rate be HR(t), then the error e(t) = HRref - HR(t). The PID controller calculates the speed adjustment u(t) in real time and applies it to the treadmill, making HR(t) slowly approach HRref, and finally stabilizing the error near zero. Through the PID algorithm, precise control and smooth adjustment of training equipment and environmental parameters can be achieved, ensuring training effectiveness while avoiding the impact of drastic changes on the human body.

[0046] Fuzzy control algorithm Given the complexity of physiological responses and individual differences during high-altitude training, this system introduces a fuzzy control algorithm for comprehensive decision-making control of multiple variables. Fuzzy control does not require a precise mathematical model; instead, it utilizes expert experience rules and fuzzy logic reasoning to achieve intelligent adjustment of the system. In this invention, fuzzy control is mainly used by the decision-making module to comprehensively adjust training intensity and environmental parameters. Its design steps are as follows: First, determine the input and output linguistic variables for fuzzy control. Inputs may include "heart rate deviation," "blood oxygen deviation," "fatigue level," etc., and outputs may be "training load adjustment range" or "oxygen supply adjustment level," etc. Then, define a fuzzy set and membership function for each linguistic variable. For example, heart rate deviation is divided into three levels: low, medium, and high, and its fuzziness level is represented by a triangular membership function; blood oxygen deviation is divided into normal, slightly low, and low. Next, formulate a fuzzy control rule table, such as: Rule 1: If the "Heart Rate Deviation" is "High" and the "Blood Oxygen Deviation" is "Low", then the "Training Load Adjustment Range" should be "Significantly Reduced". Rule 2: If the "Heart Rate Deviation" is "Low" and the "Blood Oxygen Deviation" is "Normal", then the "Training Load Adjustment Range" should be "Appropriately Increased". Rule 3: If the "Heart Rate Deviation" is "Medium" and the "Blood Oxygen Deviation" is "Low", then the "Training Load Adjustment Range" should remain unchanged. The aforementioned fuzzy rule base can be set by professional coaches and medical experts, covering various typical situations. During runtime, the system fuzzifies the currently detected heart rate deviation, blood oxygen deviation, etc., into corresponding membership values, which are then substituted into the rule base for reasoning. Each rule that meets the conditions will produce a fuzzy conclusion for the output (e.g., simultaneously obtaining suggestions of "significantly reduce" and "appropriately reduce"). Finally, the fuzzy output is transformed into precise control command values ​​through defuzzification (e.g., weighted average method). For example, if the rule reasoning yields a membership degree μ1=0.7 corresponding to "reduce speed by 10%" and μ2=0.3 corresponding to "reduce speed by 5%", then the defuzzification result can be calculated as follows: This indicates that the final treadmill speed was adjusted to be reduced by approximately 8.5%. Through fuzzy control, the system can achieve smooth and reasonable control results when dealing with uncertain control problems involving multiple inputs and multiple outputs, enabling intelligent and delicate adjustment of training intensity.

[0047] Support Vector Machine (SVM), as a supervised machine learning algorithm, plays a crucial role in the data analysis and prediction modules of this system. The data analysis module utilizes the SVM classification model to perform real-time identification of the trainee's state (see...). The prediction module can use SVM to perform regression prediction of future trends. The construction process of the SVM classification model is as follows: First, a large amount of historical physiological data during high-altitude training is collected, and the categories are labeled according to the trainee's state at that time (e.g., normal, mild discomfort, severe discomfort). Then, the data is preprocessed and features are extracted, and key features (heart rate, blood oxygen, altitude, training time, etc.) are used as the input vector of the SVM. By selecting an appropriate kernel function K(xi,xj) (such as Gaussian radial basis kernel) and regularization parameter C, the SVM model is trained to maximize the classification margin and minimize the classification error. The training results in a set of support vectors, weights α, and biases b, forming a decision function such as... As shown, when new data arrives, the analysis module quickly calculates f(X) to determine its category, achieving intelligent judgment of the current state. Experiments show that SVM can effectively integrate information from multiple physiological indicators, accurately classify altitude sickness risks, and provide a more reliable early warning basis than a single threshold judgment.

[0048] In the prediction module, a Support Vector Regression (SVR) model can be used to predict future indicators. Its basic idea is to fit the data within a certain error range to ensure the model has good generalization ability. The optimization objective of SVR is to find the function... To ensure that the bias of all training data does not exceed the tolerance range ϵ and the model is as smooth as possible, i.e., to minimize w2, the constraint is as follows: By introducing slack variables and solving the dual problem, support vectors and their corresponding weights can be obtained, thus constructing a predictive model. In practical applications, inputting recent heart rate, SpO2, and other time series data into the SVR model allows for the prediction of index values ​​shortly in the future, guiding the lead time for control decisions. Due to the excellent nonlinear mapping ability and generalization performance of SVM / SVR, it effectively improves the pattern recognition and trend prediction capabilities for complex physiological states in this system, providing intelligent decision support for the training process.

[0049] In summary, PID control ensures precise adjustment, fuzzy control provides flexibility for intelligent decision-making, and the SVM algorithm provides the theoretical basis for state evaluation and prediction. The combination of these three technologies enables the dynamic training system of this invention to respond quickly and robustly to changes in the trainee's physiological state, dynamically optimize training intensity, and maximize the training effect for high-altitude work capabilities.

[0050] Example 1: Closed-loop control of endurance training in high-altitude environments This embodiment focuses on high-altitude endurance training for ordinary personnel, illustrating the workflow and closed-loop control effect of the system of the present invention. The training scenario simulates an oxygen-deficient environment at an altitude of 4000 meters, with trainees performing aerobic endurance training on a treadmill. The various modules of the system operate collaboratively according to the following process: Pre-training preparation: Set the initial parameters of the training room environment to near-plain levels (e.g., atmospheric pressure at 0m altitude and 21% oxygen concentration). Trainees wear heart rate and blood oxygen sensors and stand on the treadmill. The controller retrieves the trainee's basic information and past training data from the database module. The SVM analysis model adaptively adjusts the classification decision threshold accordingly (e.g., adjusting the normal / abnormal judgment boundary based on individual resting heart rate, VO2 max, etc.). The training goal is defined as gradually reducing the oxygen concentration to 15% (equivalent to an altitude of 4000m) and increasing the running speed to 12km / h, while maintaining the trainee's heart rate within their aerobic heart rate limit and blood oxygen saturation above 90%. Once everything is ready, start the training program.

[0051] Training begins (Phase 0): The physiological data acquisition module reads the initial resting heart rate (e.g., 80 bpm), resting blood oxygen saturation (98%), etc., and the analysis module determines that the trainee's condition is normal. The prediction module provides an initial strategy: environmental parameters begin to adjust towards plateau values ​​(e.g., reducing oxygen concentration from 21% to 18% within 5 minutes), and the treadmill starts at a slow speed of 6 km / h. The control execution module accordingly adjusts the oxygen supply valve and the air pressure control pump to smoothly reduce the oxygen concentration using PID control, decreasing it by approximately 0.6% per minute; simultaneously, it drives the treadmill to accelerate uniformly to the target speed. The trainee begins jogging slowly in a slightly hypoxic environment, and the system enters closed-loop monitoring.

[0052] Mid-training (Phase 1): As the treadmill speed and hypoxia level increase, the trainee's heart rate rises to 120 bpm, and blood oxygen drops to 94%. The physiological data acquisition module uploads data in real time, and the analysis module calculates the heart rate deviation (still a positive deviation of 30 bpm relative to the target aerobic heart rate of 150 bpm) and blood oxygen deviation (still a margin of 4 percentage points relative to the 90% lower limit). SVM analysis determines the condition is normal and risk-free. Based on this, the prediction module formulates an enhancement strategy: the training intensity can be moderately increased. The specific instructions are: continue to reduce the oxygen concentration from 18% to 16% (simulating an altitude increase from approximately 2000m to 3000m), while increasing the treadmill speed from 6km / h to 9km / h over the next 5 minutes. Fuzzy decision-making, considering that the heart rate is still below the target and blood oxygen is sufficient, outputs "moderate enhancement" training. The control execution module executes the instructions one by one: lowering the oxygen concentration and increasing the treadmill speed, both using a PID controller to prevent overshoot. The feedback module monitors that the heart rate gradually rises to around 140 bpm, and the blood oxygen drops to 92%. The result was as expected, and the feedback indicated that the control was effective, with a positive signal transmitted back, requiring no correction. The system then proceeded to the next cycle.

[0053] Late Training Phase (Phase 2): When the ambient oxygen concentration drops to 15% (target altitude achieved) and the treadmill speed increases to 12km / h (planned intensity achieved), the trainee's heart rate rises to nearly 150bpm, and blood oxygen saturation is approximately 90%, close to the safe lower limit. At this point, the analysis module detects that the blood oxygen deviation is approaching 0 with slight fluctuations, and the SVM model may interpret this as a warning sign. Based on this, the prediction module adjusts its strategy to maintain intensity: no further increase in difficulty, entering a stable training phase. On one hand, it instructs the environmental simulation to maintain a constant 15% oxygen concentration; on the other hand, it uses the fuzzy control rule "high heart rate and low blood oxygen, no increase in intensity" to constrain the treadmill speed from increasing further and slightly lowers the incline by 1° to prevent the heart rate from continuing to surge. The execution module accordingly slows down the treadmill's acceleration, or even slightly reduces the speed or incline to stabilize the heart rate at around 150bpm. The feedback module closely monitors the process. During this phase, the heart rate fluctuates slightly between 148 and 152 bpm, blood oxygen saturation remains between 90% and 91%, and the system alarm does not trigger, indicating that the trainee is in a critical stable state. The training has reached its maximum intensity but is still within a safe range. After maintaining this intensity for a period of time, the training reaches its predetermined duration.

[0054] Training End and Recovery: Upon reaching the training time limit, the system issues an end command. The control module gradually adjusts the oxygen concentration back to the baseline level (21%) and slowly reduces the treadmill speed to 0, ensuring a smooth transition to a resting state for the trainee and avoiding dizziness or blood pressure fluctuations caused by sudden stops. The physiological data acquisition module continues monitoring for a period of time after training stops until heart rate and blood oxygen levels return to normal resting ranges. All training data collected throughout the process (including heart rate curves, blood oxygen changes, speed and oxygen concentration adjustment records, etc.) is aggregated and stored in the database module by the central processing module for future analysis, evaluation, and optimization of personalized training programs. This completes the closed-loop process of endurance training in this embodiment.

[0055] As can be seen from Example 1, the system of the present invention can dynamically adjust the environment and exercise load according to the trainee's real-time condition, realizing personalized control of high-altitude hypoxia endurance training. When the trainee is in good condition, the system gradually increases the intensity, and when approaching the limit, the system automatically slows down and adjusts to ensure safety. The entire training is carried out smoothly under closed-loop control, which greatly improves training efficiency and safety, and lays a solid foundation for improving high-altitude work capabilities.

[0056] Example 2: Safety Response Control under Abnormal Conditions This embodiment simulates a physiological abnormality during training to illustrate the system's safety feedback mechanism and emergency control strategy. Suppose a trainee suddenly experiences severe signs of hypoxia (e.g., a sharp drop in blood oxygen) during high-intensity high-altitude training; how will the system detect and respond? Scenario: The trainee is undergoing high-intensity interval training in an environment simulating an altitude of 4500 meters (oxygen concentration of approximately 14%), with a treadmill speed of 12 km / h. The trainee is nearing their fatigue limit, with a heart rate of approximately 155 bpm. Suddenly, the trainee may experience insufficient ventilation due to overexertion, causing their blood oxygen saturation to drop further from 88% to 85% within a short period, accompanied by mild dizziness.

[0057] Anomaly Detection: The physiological data acquisition module reads the SpO2 value every second, immediately detecting a drop in blood oxygen saturation below 90%. The analysis module compares the current blood oxygen saturation of 85% with the safe lower limit of 90%, finding a negative deviation of eSpO2 = -5%, far exceeding the permissible range, and the heart rate deviation is also positive (exceeding the target heart rate by approximately 10 bpm). The SVM classification model determines the status as "-1" (dangerous) based on real-time data, while the analysis module determines the risk of acute altitude sickness according to preset rules. Within less than one second, the analysis module triggers an anomaly alarm: on the one hand, it triggers a piercing alarm sound via the controller; on the other hand, it immediately transmits the anomaly status marker and related data to the predictive decision-making module, requesting emergency response.

[0058] Emergency Decision-Making: Upon receiving an abnormal signal, the predictive decision-making module generates an emergency shutdown strategy directly based on the safety-first principle, without requiring complex predictions. Fuzzy control rules are also defined in this case: for example, "If blood oxygen is very low and heart rate is too high, then the training intensity is set to the lowest." Therefore, the decision-making module's instructions include: immediately stopping the treadmill or reducing the speed to an extremely low level, and rapidly increasing the oxygen concentration and air pressure in the training room to a stable level. Simultaneously, the module also instructs the activation of the pre-set emergency oxygen generation system (such as an oxygen-enriched breathing mask or oxygen cylinder) in the training room for the trainees to recover with oxygen. This emergency plan is formulated in milliseconds and rapidly implemented through the control execution module.

[0059] Execution Control: Upon receiving the shutdown and oxygenation commands, the control execution module immediately interrupts the current training process with the highest privileges and performs the following actions: It sends an emergency stop signal to power off the treadmill motor, reducing the speed to 0; it controls the oxygen supply valve to fully open, stops the air pump from drawing air, and reverses the flow to replenish oxygen to the training room and increase air pressure at the fastest possible rate. Due to the critical situation, the control algorithm adopts saturation control instead of a smooth, gradual change; that is, it does not use the slow adjustment of PID, but directly sets the execution quantity to the maximum (e.g., increasing the oxygen concentration from 14% to 21% within 5 seconds). During execution, the feedback module samples and monitors the recovery of blood oxygen and heart rate at a high frequency.

[0060] Feedback Recovery: Under emergency control, the trainee quickly received a sufficient oxygen supply. Within tens of seconds, blood oxygen saturation rose from 85% to over 95%, and heart rate gradually decreased. The feedback monitoring module sent this recovery data to the analysis module, which confirmed that blood oxygen saturation had returned to normal (deviation reduced to +5%) and heart rate had decreased to a safe range. The system then determined that the danger had passed, and the alarm automatically stopped sounding. After ensuring the trainee was unharmed, the system recorded the entire process of this abnormal event and entered the termination procedure. The trainee, accompanied by medical personnel, stopped training and entered a rest and observation period. The central processing module marked and stored the data of this abnormal case for future reference in improving training programs and risk models.

[0061] As demonstrated in Example 2, this system possesses rapid anomaly detection and automatic safety response capabilities. When a trainee exhibits dangerous signs, the system can instantly capture and interrupt the original control plan within a closed loop, switching to a safety mode for emergency intervention, significantly reducing the risks of high-altitude training. The entire emergency response process does not rely on manual intervention; the system automatically completes the judgment and execution, reflecting the intelligence and safety reliability of this invention in high-altitude operational capability training.

[0062] In summary, the dynamic integrated training system of this invention achieves intelligent monitoring and adaptive adjustment of the entire high-altitude training process through the interaction and closed-loop control of the aforementioned modules. In each specific embodiment, the system demonstrates excellent autonomous decision-making and control capabilities, whether optimizing intensity during normal training phases or handling abnormal situations safely. Compared with existing traditional manual adjustment methods, this system can more accurately maintain the trainee's physiological indicators within the ideal range, thereby enhancing the training effect of high-altitude operational capabilities. Simultaneously, multi-level feedback control ensures safe and controllable training, greatly reducing the risks associated with sudden environmental changes or human discomfort during high-altitude training. It should be noted that the above embodiments are intended to illustrate the principles and advantages of this invention. In practical applications, the parameters of this invention can be adjusted or functions expanded according to different training subjects and individual circumstances without departing from the spirit and scope of this invention. The specific implementation of each module can be hardware circuitry, software programs, or a combination of both. Any equivalent transformations or substitutions made using the ideas of this invention fall within the scope of protection claimed by this invention.

Claims

1. A dynamic comprehensive training system for enhancing high-altitude work capabilities, characterized in that, It includes multiple functional modules linked in a closed loop of "perception-judgment-execution-feedback": The physiological and environmental data acquisition module is used to acquire and preprocess heart rate, blood oxygen saturation (SpO2), lung ventilation, as well as oxygen concentration, air pressure, and temperature in the training room. The data analysis module is used to calculate the deviation of each monitored quantity relative to the target or safety threshold and to identify the status. The predictive decision-making module is used to generate training phase switching and parameter adjustment strategies based on the analysis results. The control execution module is used to automatically control continuous variables such as treadmill speed / incline and oxygen concentration / air pressure. The feedback and alarm module is used to sample and correct the execution results in real time and issue an alarm when there is an anomaly. The data analysis module uses support vector machine (SVM) to perform pattern recognition and classification on multiple physiological parameters. The control execution module supports continuous fine adjustment based on PID and / or fuzzy control rules, thus forming a closed-loop control structure for the training process.

2. The dynamic comprehensive training system for enhancing high-altitude work capabilities according to claim 1, characterized in that, The control execution module is based on the error between the controlled target and the reference heart rate. The control quantity u(t) for the treadmill speed is calculated, and overshoot and steady-state error are suppressed by a combination of proportional, integral and derivative terms, so as to achieve smooth approximation and stable maintenance of heart rate to the reference value.

3. The dynamic comprehensive training system for enhancing high-altitude work capabilities according to claim 1, characterized in that, The prediction and decision-making module incorporates fuzzy control rules, including a collaborative constraint that prohibits increasing training intensity or slightly decreasing slope / speed when heart rate is high and blood oxygen is low. This constraint is used to maintain intensity stably in the critical zone of high-altitude endurance training.

4. The dynamic comprehensive training system for enhancing high-altitude work capabilities according to claim 1, characterized in that, When the feedback and alarm module detects a combination of abnormalities, such as SpO2 momentarily dropping below the safety lower limit and heart rate exceeding the target range, it triggers an emergency response procedure: the control execution module issues an emergency stop / reduction command with the highest authority, while rapidly increasing the oxygen concentration to the plain level and raising the air pressure, during which saturation control is used to respond in milliseconds.

5. The dynamic comprehensive training system for enhancing high-altitude work capabilities as described in claim 2, characterized in that, The control execution module also supports individualized models and threshold adaptation; before training, the control execution module retrieves individual resting heart rate, previous training data or maximum oxygen uptake information from the database and dynamically adjusts the SVM classification threshold and the safety / target interval boundary.

6. The dynamic comprehensive training system for enhancing high-altitude work capabilities as described in claim 4, characterized in that, The feedback and alarm module is equipped with an acoustic alarm device, which can simultaneously issue a high-intensity audible alarm to alert on-site personnel after a dangerous situation is identified.

7. The dynamic comprehensive training system for enhancing high-altitude work capabilities as described in claim 1, characterized in that, It also includes a training data storage module, which archives heart rate curves, blood oxygen changes, environmental parameters, and execution control records after training to support subsequent training prescription optimization and evaluation review.

8. A dynamic comprehensive training method for enhancing high-altitude work capability using the system described in any one of claims 1-7, characterized in that, include: S1. Training preparation: Set the initial training room to a plain level (approximately 21% oxygen concentration, standard air pressure). Trainees wear heart rate and blood oxygenation sensors and load their individual historical data. The goal is to gradually reduce the oxygen concentration to approximately 15%, increase the running speed to approximately 12km / h, and at the same time constrain SpO2 to not be lower than 90% and heart rate to not exceed the aerobic limit. S2, Training Start and Climb: The control execution module reduces the oxygen concentration at a preset slope and accelerates uniformly to the target speed, while the analysis module continuously calculates the deviation. S3, Phase Maintenance: When approaching the critical zone, the fuzzy rule of "no increase in intensity if heart rate is high and blood oxygen is low" is implemented, and the slope / speed is finely adjusted to maintain stability; S4. Termination and Recovery: Smoothly revert the environment and speed to a flat and resting state in a stepwise manner and continue monitoring until the normal range is restored.

9. The dynamic comprehensive training method for enhancing high-altitude work capabilities as described in claim 8, characterized in that, In the S3 phase, the slight reduction in slope and speed are determined by the prediction and decision-making module based on short-term trends and deviations, so that the heart rate is maintained at 148~152 bpm, blood oxygen is maintained at 90%~91%, and no alarm is triggered.

10. The dynamic comprehensive training method for enhancing high-altitude work capabilities as described in claim 9, characterized in that, When the "SpO2≤90% and heart rate exceeds target" combination abnormality occurs at any time from S1 to S3, the system directly enters the emergency response branch: interrupts the treadmill motor, activates oxygen enrichment and pressurization to restore the maximum throughput quickly, and samples blood oxygen and heart rate at a high frequency during the execution until the danger zone is out, and then enters the recovery process.

Images