Method for determining maximum oxygen uptake based on physiological data of an athlete during training

Abstract

The present application belongs to the technical field of sports health data analysis, and relates to a maximum oxygen uptake determination analysis method based on athlete training physiological data. The present application collects multi-source physiological time series data, divides training units according to speed sequences, extracts submaximal steady-state segments and quasi-limit anchor point segments; based on historical segment extraction, local curvature parameters, phase synchronization index and response delay are extracted, different load intervals are divided, and a limit state extrapolation model containing linear, exponential and phase synchronization index attenuation boundary constraints is constructed; the limit power is obtained by inputting the current cycle segment characteristics into the model, and the maximum oxygen uptake estimate value is output by the individualized mapping function. The problems of traditional test dependence on special equipment affecting daily training, single experience estimation being greatly affected by fatigue, and non-representative data leading to result deviation are solved, so that continuous monitoring can be realized without additional testing, the evaluation stability and cross-cycle comparability are improved, and the real aerobic capacity can be accurately reflected.

Description

Technical Field

[0001] This invention belongs to the field of sports health data analysis technology, and relates to a method for determining and analyzing the maximum oxygen uptake based on athletes' training physiological data. Background Technology

[0002] Maximum oxygen uptake (VO2 max) is an important indicator for evaluating an athlete's aerobic capacity, endurance level, and training adaptation status. It is commonly used in training plan development, periodic physical fitness assessments, and competitive performance analysis. Current methods for measuring VO2 max typically include laboratory incremental load testing, specific scenario testing, and empirical estimation based on training records.

[0003] Laboratory testing typically relies on specialized equipment, fixed procedures, and specific testing environments, requiring athletes to perform additional loads during the testing process. For athletes in continuous training or competition preparation cycles, there is a coordination requirement between testing schedules and daily training schedules, which in turn affects the frequency and continuity of VO2 max data acquisition.

[0004] Estimation methods based on training records often use single-exercise results, interim statistical results, or general empirical formulas for calculation. Because athletes' physiological responses differ at different training stages, fatigue levels, and recovery states, estimations of maximum oxygen uptake based solely on localized training results may exhibit interim fluctuations, affecting the judgment of the actual trend in aerobic capacity changes.

[0005] Furthermore, daily training logs typically include warm-up, cool-down, interval, short-duration high-intensity exercise, and data acquisition disturbances. If continuous training logs are directly used for VO2 max estimation, non-representative data may be involved in the calculation, leading to biased estimation results and reducing the comparability of assessment results between different training cycles. Summary of the Invention

[0006] In view of this, in order to solve the problems mentioned in the background art, a method for determining maximum oxygen uptake based on athletes' training physiological data is proposed.

[0007] The objective of this invention can be achieved through the following technical solution: a method for determining and analyzing the maximum oxygen uptake based on athletes' training physiological data, including: synchronously collecting time-series data of athletes' heart rate, power, speed, blood oxygen saturation, respiratory rate and heart rate variability indices during daily training.

[0008] Training units are divided according to velocity sequence. Submaximal steady-state segments with power and heart rate variability fluctuations less than steady-state thresholds within the sliding decision window are extracted, as well as quasi-limit anchor point segments with power in the high proportion range, heart rate reaching peak heart rate, and blood oxygen saturation continuously below the oxygen drop threshold.

[0009] Local curvature parameters of the heart rate-power relationship, phase synchronization index of respiratory rate and heart rate variability index, and response delay of heart rate variability index recovery slope and power gradient are extracted from submaximal steady state and quasi-limit anchor point segments of historical cycles. Submaximal, quasi-limit to limit transition and limit intervals are divided, and limit state extrapolation model with linear, exponential function and phase synchronization exponential decay boundary constraints is constructed.

[0010] The phase synchronization index, response delay, and local curvature parameters of the current period's quasi-limit anchor point segment are input into the limit state extrapolation model to obtain the limit power estimate, and the maximum oxygen uptake estimate is output according to the individualized power-oxygen uptake mapping function.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention collects time-series data of heart rate, power, speed, blood oxygen saturation, respiratory rate and heart rate variability in athletes' daily training simultaneously, without requiring athletes to undergo additional load tests, thus eliminating the dependence on special equipment and specific testing scenarios. This method solves the problem of conflict between traditional testing arrangements and daily training, improves the acquisition frequency and update continuity of VO2 max data, enables physical fitness assessment to be seamlessly integrated into the regular training cycle, and ensures the real-time nature of monitoring.

[0012] (2) This invention extracts local curvature parameters, phase synchronization index, and response delay from submaximal steady-state segments and quasi-limit anchor point segments based on historical cycles, constructs a limit state extrapolation model, and calculates the estimated maximum oxygen uptake value according to the individualized power-oxygen uptake mapping function. This method overcomes the shortcomings of a single empirical formula that does not take into account the differences in individual fatigue and recovery states, suppresses the stage fluctuations in the estimation results at different training stages, and thus accurately reflects the athlete's true maximum oxygen uptake change trend.

[0013] (3) This invention divides training units according to velocity sequences, extracts submaximal steady-state segments within the sliding decision window where the fluctuations in power and heart rate variability indicators are less than the steady-state threshold, and quasi-limit anchor point segments that meet the conditions of high proportional power, peak heart rate, and blood oxygen saturation. This method filters out interference from non-representative data such as warm-up, adjustment, and equipment acquisition disturbances, solves the problem of result deviation caused by invalid data participating in the calculation, and improves the comparability of maximum oxygen uptake assessment results between different training cycles. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a flowchart of the method for determining and analyzing the maximum oxygen uptake based on athletes' training physiological data in this invention;

[0016] Figure 2 This is a diagram showing the centrality-power relationship curve and load interval division of the present invention;

[0017] Figure 3 This is a box plot of physiological indicators under different load ranges in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0020] The specific scheme of the method for determining and analyzing the maximum oxygen uptake based on athlete training physiological data provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0021] Please see Figure 1 As shown, the implementation of the present invention includes S1 to S4.

[0022] This invention addresses the problem of the difficulty in directly and continuously assessing maximum oxygen uptake under daily training conditions. It utilizes multi-source time-series data such as heart rate, power, speed, blood oxygen saturation, respiratory rate, and heart rate variability obtained synchronously during training to establish an analytical framework that extends from training state identification to extrapolation of limit capabilities.

[0023] First, this invention divides the training process into relatively independent training units based on speed changes. Then, it filters submaximal steady-state segments by combining power and heart rate variability fluctuation characteristics. Furthermore, it identifies quasi-limit anchor point segments that characterize near-limit physiological responses through high power output, peak heart rate characteristics, and continuous decrease in blood oxygen saturation. Based on this, it extracts the curvature of the heart rate-power relationship, the phase coordination between respiratory rate and heart rate variability, and the lag characteristics of heart rate variability response relative to power changes during the recovery phase from historical training cycles. These are used to characterize the nonlinear physiological response of an individual transitioning from a submaximal to a limit state. Further, a partitioned extrapolation model is constructed, subjecting the submaximal, quasi-limit to limit transition, and limit intervals to different functional forms and phase-synchronous decay boundaries. This allows for the estimation of limit power based on the quasi-limit anchor point characteristics of the current cycle without requiring a standard VO2 max test, and outputs an estimated VO2 max estimate by combining the individualized power-VO2 max mapping relationship.

[0024] S1. Synchronously collect time-series data of athletes' heart rate, power, speed, blood oxygen saturation, respiratory rate and heart rate variability during daily training.

[0025] Since maximum oxygen uptake is related to the athlete's cardiopulmonary response, external power output, blood oxygen utilization, and autonomic nervous system regulation under high-intensity load, estimation using only single heart rate or single power data is easily affected by training fatigue, environmental conditions, equipment errors, and individual recovery status, resulting in a lack of stability in the estimation results. At the same time, subsequent training unit division, steady-state segment identification, quasi-limit anchor point segment identification, and limit state extrapolation models all rely on continuous, synchronous, and comparable multi-source time-series data. Therefore, it is necessary to first obtain heart rate, power, velocity, blood oxygen saturation, respiratory rate, and heart rate variability indicators with a unified time reference.

[0026] Therefore, by configuring a sensor group including a heart rate sensor, a power meter, a speed sensor, a blood oxygen saturation sensor, a respiratory rate sensor, and a heart rate variability detection module, the sensor group is installed on the athlete's body and / or training equipment; wherein, the heart rate variability detection module is a signal processing module that collects electrocardiogram signals and detects the peak value of the R wave to extract the RR interval of adjacent cardiac cycles, and then calculates the root mean square of the difference between adjacent RR intervals as a heart rate variability index.

[0027] A synchronous sampling clock is set to synchronously acquire time-series data from the aforementioned sensors at a preset sampling rate, and a unified timestamp is added to the acquired data from each channel. The preset sampling rate is set according to the effective frequency range of each type of signal; the sampling rate for heart rate, blood oxygen saturation, and respiratory rate is no less than 1Hz, the sampling rate for power and velocity is no less than 5Hz, and the sampling rate for the electrocardiogram signal required for heart rate variability indicators is no less than 250Hz. After synchronous acquisition, the data from each channel is interpolated based on the lowest sampling rate to form time-series data with a unified time reference.

[0028] S2. Divide the training units according to the velocity sequence, and extract the submaximal steady-state segments within the sliding judgment window where the fluctuation of power and heart rate variability indicators is less than the steady-state threshold, as well as the quasi-limit anchor point segments where the power is in the high proportion range, the heart rate reaches the peak heart rate, and the blood oxygen saturation is continuously lower than the oxygen drop threshold.

[0029] Athletes' daily training typically includes multiple phases such as warm-up, intervals, continuous load, sprint, and recovery. The speed changes between different phases can reflect the switching of training load structure. At the same time, the estimation of maximum oxygen uptake needs to utilize both the submaximal physiological response under relatively stable load conditions and the anchor point physiological response under near-limit load conditions. Therefore, this step divides the continuous data into several training units and identifies the submaximal steady-state segment and the near-limit anchor point segment within each training unit.

[0030] In one specific embodiment, firstly, a moving average filter with a window length of 5 seconds is used to smooth the velocity sequence in the time series data. Then, the first-order difference of the smoothed velocity sequence is calculated, and the point where the sign of the first-order difference changes from negative to positive is found as the local minimum point of the velocity.

[0031] If the time interval between a local minimum and the previous local minimum is greater than or equal to a preset minimum interval, and the absolute value of the average velocity difference between adjacent time periods before and after the local minimum is greater than or equal to a preset minimum velocity change amplitude, then the time period between adjacent local minimums that meet the above conditions is divided into training units. In this embodiment, the preset minimum interval is 60 seconds, and the preset minimum velocity change amplitude is 0.5 m / s. Furthermore, for time periods without local minimums, the global minimum of the velocity sequence is used for division.

[0032] For each training unit, the coefficient of variation of the overall power of that training unit and the coefficient of variation of the heart rate variability index are calculated. The coefficient of variation is the ratio of the standard deviation to the mean of the corresponding sequence. The larger of the two values ​​is multiplied by a preset scaling factor to obtain the steady-state threshold. The preset scaling factor is a scaling factor used to adjust the sensitivity of local steady-state determination, and its value is less than or equal to 1. In this embodiment, the value is 0.8.

[0033] Within each training unit, a sliding decision window is set with a window length of 60 seconds and a sliding step size of 1 second. This sliding decision window represents the time observation interval used to evaluate the stability of local physiological and load states. The coefficient of variation of power and the coefficient of variation of heart rate variability index are calculated respectively within the current sliding decision window.

[0034] Adjacent sliding decision windows with both power output coefficient of variation and heart rate variability coefficient of variation less than the steady-state threshold are merged, and the resulting continuous time interval is marked as a submaximal steady-state segment. The submaximal steady-state segment represents a continuous training period within the corresponding training unit where both power output and heart rate variability meet the stability criteria, and is used to characterize the relatively stable physiological response state of the athlete under non-maximum load.

[0035] The percentile distribution of power within the training unit is calculated, and the interval where the power value is greater than or equal to the 95th percentile and not lower than a preset proportion of the athlete's historical threshold power is defined as the high proportion interval; in this embodiment, the preset proportion is 0.9. The method for determining the athlete's historical threshold power is as follows: select high-intensity stable segments with a duration of not less than 10 minutes and a power variation coefficient of less than 0.1 from the historical training cycle, calculate the average power of each high-intensity stable segment, and take the median of each average power as the athlete's historical threshold power.

[0036] The time period in which the power output is in the high proportion range is identified as the candidate time period. Within the candidate time period, the difference between the heart rate values ​​of adjacent timestamps is calculated as the heart rate difference sequence. The turning point where the difference value changes from positive to negative is identified as the heart rate peak turning point. The starting point of the time window is shifted 45 seconds before this turning point, and the ending point of the time window is 10 to 30 seconds after this turning point, thus determining the heart rate peak determination time window.

[0037] Within a time window, if the heart rate reaches the peak heart rate of the training unit and the blood oxygen saturation is below the oxygen drop threshold for a preset duration, such as 10 to 30 seconds, the corresponding time period is marked as a quasi-limit anchor point segment. The quasi-limit anchor point segment represents a continuous training period within the corresponding training unit that simultaneously satisfies the conditions of high proportional power output, peak heart rate of the training unit, and continuous decrease in blood oxygen saturation. This segment is used to subsequently extract response delay features and serve as the quasi-limit state input anchor point for the limit state extrapolation model.

[0038] Wherein, the oxygen drop threshold is at least one of the following: the difference between the athlete's individual resting blood oxygen saturation baseline and the historical blood oxygen fluctuation standard deviation, or the difference between the mean and standard deviation of blood oxygen saturation in the training unit. The athlete's individual resting blood oxygen saturation baseline is the mean blood oxygen saturation continuously collected in a resting state before training, and the historical blood oxygen fluctuation standard deviation is the standard deviation of effective blood oxygen saturation data in historical training cycles.

[0039] S3. Extract local curvature parameters of the heart rate-power relationship, phase synchronization index of respiratory rate and heart rate variability index, and response delay of heart rate variability index recovery slope and power gradient from submaximal steady state and quasi-limit anchor point segments of historical cycles. Divide submaximal, quasi-limit to limit transition and limit intervals, and construct a limit state extrapolation model with linear, exponential function and phase synchronization exponential decay boundary constraints.

[0040] Given that athletes' heart rate, power, respiratory rate, and heart rate variability do not maintain the same relationship across different load zones, especially during the transition from the quasi-limit to the limit, where the heart rate-power relationship, phase synchronization relationship, and recovery response process exhibit different characteristics compared to the submaximal phase, using a uniform model for all load zones may reduce the adaptability of the limit boundary estimation. Therefore, this step constructs a limit state extrapolation model describing different intervals based on local curvature parameters, phase synchronization index, and response delay extracted from submaximal steady-state segments and quasi-limit anchor point segments in historical cycles.

[0041] In one specific embodiment, firstly, from the submaximal steady-state segments obtained from historical cycles, with power as the independent variable and heart rate as the dependent variable, a local quadratic polynomial fitting is performed within a sliding power window to obtain the heart rate-power relationship curve. The second derivative of the heart rate-power relationship curve within each sliding power window is calculated as a local curvature parameter. This local curvature parameter represents the degree of nonlinearity of the change in heart rate with increasing power, that is, the bending direction and degree of the heart rate-power relationship curve.

[0042] From the submaximal steady-state segment, the respiratory rate sequence and the heart rate variability index sequence are subjected to discrete Hilbert transform using the Fast Fourier Transform algorithm to construct analytic signals. The amplitude of the analytic signals is calculated to obtain the instantaneous phase. Then, the cumulative average value of the cosine of the instantaneous phase difference between the respiratory rate and the heart rate variability index within a set time window is calculated as the phase synchronization index. This phase synchronization index represents the coupling strength between the respiratory rhythm and autonomic nervous system regulation, and its value ranges from [-1, 1]. The closer the value is to 1, the higher the degree of synchronization, reflecting the coordinated working state of the athlete's cardiopulmonary system under the current load.

[0043] Quasi-limit anchor point segments and their adjacent recovery segments are obtained from historical cycles; wherein, the adjacent recovery segment is the time period from the end of the quasi-limit anchor point segment to the start of the next submaximal steady-state segment. Within this adjacent recovery segment, the heart rate variability index is fitted using linear least squares with a sliding time window, such as a 10-second window. The slope of the fitted line is used as the recovery slope, and the recovery slope sequence is constructed from the recovery slopes of all sliding time windows. Similarly, the power gradient sequence is obtained by fitting the power using linear least squares.

[0044] Cross-correlation was performed on the recovery slope sequence and the power gradient sequence, and the time shift corresponding to the maximum value of the cross-correlation function was taken as the response delay. This response delay represents the time lag of the physiological recovery process of heart rate variability index after the cessation of high-intensity load relative to the decrease in external power, reflecting the regulatory rate of the autonomic nervous system's recovery from stress.

[0045] To enable the model to describe the submaximal stable response, the quasi-limit to limit transition response, and the limit boundary response respectively, it is necessary to divide different load intervals based on the changes in local curvature parameters and phase synchronization index. Therefore, the method for dividing the submaximal, quasi-limit to limit transition, and limit intervals is as follows: On the heart rate-power relationship curve, when the local curvature parameter crosses zero for the first time from a negative value, it indicates that the heart rate-power relationship has entered a new stage of change from its original curvature trend. Therefore, the power value corresponding to this zero-crossing point is determined as the first connection point.

[0046] As the load approaches its limit, the coupling between respiratory rhythm and autonomic nervous system regulation decouples, causing the phase synchronization index to decrease significantly. Therefore, the decay rate of the phase synchronization index with power is calculated. Specifically, the first-order differential derivative of the phase synchronization index sequence with respect to the power sequence is calculated, and the power value corresponding to the point where the absolute value of the decay rate reaches a local maximum is determined as the second connection point. The local maximum is the value corresponding to the point where the absolute value sequence of the first-order differential derivative is differentiated again and the sign of the first-order differential changes from positive to negative.

[0047] The interval where the power is less than the power at the first connection point is divided into a submaximal interval, representing a stable load interval where the heart rate-power relationship can be described linearly. The interval where the power is between the power at the first connection point and the power at the second connection point is divided into a quasi-limit to limit transition interval, representing a load interval where the heart rate-power relationship transitions from a linear response to a nonlinear response. The interval where the power is greater than the power at the second connection point is divided into a limit interval. The limit interval represents a high load interval where the phase synchronization index enters the continuous decay boundary constraint.

[0048] After completing the interval division, a limit state extrapolation model is constructed to extrapolate the limit state of the current period's quasi-limit anchor point segment. Specifically, the arithmetic mean of the absolute values ​​of all power gradients within the recovered segment is taken as the power gradient sequence average. This average value represents the average rate of decrease of external power over time during the recovery phase, which will affect the response delay. With the power gradient sequence average The product of these two terms serves as the power domain response scale. .

[0049] Within the submaximal region, the limit state extrapolation model employs a linear function: .

[0050] in, and The slope and intercept are obtained by fitting power and heart rate data from historical submaximal steady-state segments. The power at the first connection point.

[0051] Within the transition zone from the quasi-limit to the limit, the power value at the first connection point is taken as... and heart rate value The initial value of the exponential function is the power value at the second connection point. and heart rate value As a reference value for the termination of the exponential function, measured by the power domain response. As a power scale parameter, construct an exponential function: .

[0052] Therefore, the exponential function is Corresponding to ,exist Corresponding to This ensures that the segmented model has continuity at the connection points.

[0053] Within the limiting interval, the phase synchronization index at the second connection point is used. The initial boundary value is taken as . A continuous period in the historical training cycle where the power output reaches more than 95% of the athlete's historical maximum power and lasts for no less than 10 seconds is taken as the historical limit anchor point segment, and the low-level stable value of the phase synchronization index in the historical limit anchor segment is used as the boundary value. As the termination boundary value, construct the phase synchronization exponential decay boundary: .

[0054] in, Indicates power as The phase synchronization index boundary value at that time. The low-level stable value of the phase synchronization index in the historical limit anchor point segment. The median of the low-order data of the phase synchronization index within the historical limit anchor point segment can be taken.

[0055] Therefore, the limit state extrapolation model is: when hour, .

[0056] when hour, .

[0057] when At that time, with As a boundary constraint for the phase synchronization exponential decay.

[0058] when and The difference between them is no greater than the preset nearest neighbor threshold. When the power is at that time, the corresponding power is taken as the candidate boundary of the limiting power; where, This model is used to describe the submaximal range using a linear function, the quasi-limit to limit transition range using an exponential function, and to limit the extrapolation range of the limit range using a phase-synchronous exponential decay boundary, thus providing a model basis for subsequent limit power estimation.

[0059] like Figure 2 As shown, this invention divides the heart rate-power relationship curve into three intervals based on the point where the local curvature parameter crosses zero (first connection point P1) and the point where the phase synchronization exponential decay rate reaches its maximum (second connection point P2): the submaximal interval where power is below the first connection point is described by a linear function, with heart rate increasing linearly with power; the quasi-limit to limit transition interval where power is between the two connection points is described by an exponential function, with heart rate growth tending to level off; and the limit interval where power is above the second connection point is constrained by the phase synchronization exponential decay trend. In the figure, the blue scatter dots represent historical training data points, the red curve represents the partition fitting curve, the green dashed line marks the position of the first connection point P1, and the orange dashed line marks the position of the second connection point P2.

[0060] To further verify the effectiveness of the above partitioning model, such as Figure 3 As shown, the physiological indicators of athletes differed significantly across the three load zones: heart rate increased progressively from submaximal to maximal load, with the median rising from approximately 120 bpm to approximately 185 bpm; blood oxygen saturation remained around 98% in the submaximal load zone, but dropped to approximately 90% in the maximal load zone, reflecting changes in oxygen utilization efficiency under high-intensity load; the phase synchronization index decreased from approximately 0.8 in the submaximal load zone to approximately 0.3 in the maximal load zone, indicating that the coupling between respiratory rhythm and autonomic nervous system regulation decoupled near the maximal load.

[0061] S4. Input the phase synchronization index, response delay and local curvature parameters of the current period's quasi-limit anchor point segment into the limit state extrapolation model to obtain the limit power estimate, and output the maximum oxygen uptake estimate according to the individualized power-oxygen uptake mapping function.

[0062] Considering that the extreme state extrapolation model is used to output the extreme power estimate corresponding to the current training cycle, while the VO2 max estimate needs to be represented by the VO2 max dimension; and that the correspondence between power output and VO2 max is affected by the athlete's individual training level, body mass, and historical calibration data, using a fixed conversion relationship may reduce the degree of matching between the VO2 max estimate and the individual's actual state. Therefore, this step obtains the power sequence from the historical calibration data and the synchronously collected VO2 max sequence, obtains the individualized power-VO2 max mapping function through regression calculation, and converts the extreme power estimate into the VO2 max estimate based on this mapping function.

[0063] In one specific embodiment, firstly, a phase synchronization index is extracted from the quasi-limit anchor segment identified in the current training cycle; response delay is extracted from the current quasi-limit anchor segment and its adjacent recovery segments; and local curvature parameters are extracted from the heart rate-power relationship curve corresponding to the current training cycle. The absolute value of the product of the response delay of the current training cycle and the average value of the current power gradient sequence is used as the current power domain response scale, and the local curvature parameters, phase synchronization index, and current power domain response scale are input into the limit state extrapolation model.

[0064] Heart rate is calculated using a linear function within the submaximal range and an exponential function within the transition range from the quasi-limit to the limit. Within the limit range, a power boundary is determined based on a phase-synchronous exponential decay boundary function. When the difference between the phase-synchronous exponential boundary value and the termination boundary value is not greater than a preset proximity threshold, the corresponding power is determined as the limit power estimate. Subsequently, the limit power estimate is substituted into the individualized power-oxygen uptake mapping function to output the maximum oxygen uptake estimate.

[0065] The method for constructing the individualized power-oxygen uptake mapping function is as follows: obtaining the power sequence and the synchronously collected oxygen uptake sequence from the historical calibration data; the historical calibration data is collected by the athlete during stepwise incremental load testing or high-intensity training testing, by wearing a breathing mask of a portable gas metabolism analyzer to collect the athlete's exhaled gas, using a gas sensor to measure the oxygen and carbon dioxide concentrations in the exhaled gas, and using a bidirectional flow sensor to measure the minute ventilation, and then calculating the oxygen uptake sequence based on the difference in oxygen concentration between inhaled and exhaled gas and the minute ventilation, as well as the external mechanical power sequence collected synchronously by a power meter.

[0066] Perform a univariate linear regression on the power series and oxygen uptake series, and output the power-oxygen uptake regression slope. With intercept .

[0067] The slope of the power-oxygen uptake regression ,intercept and athlete's weight Substituting into a linear equation in one variable, we construct an individualized power-oxygen uptake mapping function, the expression of which is: ,in, This is the estimate of the limiting power. This mapping function is used to accurately convert the mechanical work index into an individual's metabolic oxygen consumption index, thereby outputting the final estimate of the maximum oxygen uptake.

[0068] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0069] Those skilled in the art will recognize that the algorithmic steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0070] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0071] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0072] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for determining and analyzing maximal oxygen uptake based on athletes' training physiological data, characterized in that, include: Simultaneously collect time-series data on athletes' heart rate, power, speed, blood oxygen saturation, respiratory rate, and heart rate variability during daily training. Training units are divided according to velocity sequence. Submaximal steady-state segments with power and heart rate variability fluctuations less than steady-state thresholds within the sliding decision window are extracted, as well as quasi-limit anchor point segments with power in the high proportion range, heart rate reaching peak heart rate and blood oxygen saturation continuously below oxygen drop threshold. Local curvature parameters of the heart rate-power relationship, phase synchronization exponent of respiratory rate and heart rate variability index, and response delay of heart rate variability index recovery slope and power gradient are extracted from submaximal steady state and quasi-limit anchor point segments of historical cycles. Submaximal, quasi-limit to limit transition and limit intervals are divided, and limit state extrapolation model with linear, exponential function and phase synchronization exponential decay boundary constraints is constructed. The phase synchronization index, response delay, and local curvature parameters of the current period's quasi-limit anchor point segment are input into the limit state extrapolation model to obtain the limit power estimate, and the maximum oxygen uptake estimate is output according to the individualized power-oxygen uptake mapping function.

2. The method for determining and analyzing maximal oxygen uptake based on athlete training physiological data as described in claim 1, characterized in that, The method for identifying the submaximal steady-state segment is as follows: After smoothing the velocity sequence in the time series data, change point detection is performed. The time interval between adjacent local minima is divided into training units, with the local minima that satisfy the preset minimum interval and the preset minimum velocity change amplitude as the boundary. For each training unit, calculate the coefficient of variation of the overall power of the training unit and the coefficient of variation of the heart rate variability index, and take the larger of the two values ​​and multiply it by a preset scaling factor as the steady-state threshold. A sliding decision window is set up within each training unit, and the coefficient of variation of power and the coefficient of variation of heart rate variability index are calculated within the sliding decision window respectively. Continuous periods within a sliding decision window where both the coefficient of variation of power and the coefficient of variation of heart rate variability are less than the steady-state threshold are marked as submaximal steady-state segments.

3. The method for determining and analyzing maximal oxygen uptake based on athlete training physiological data as described in claim 1, characterized in that, The method for identifying the quasi-limit anchor point segment is as follows: Calculate the percentile distribution of power within the training unit, and define the interval where the power value is greater than or equal to the preset percentile and not lower than the preset proportion of the athlete's historical threshold power as the high proportion interval. The time periods in which the power output is in the high proportion range are identified as candidate time periods; Within the candidate time period, the difference between the heart rate values ​​of adjacent timestamps is calculated as the heart rate difference sequence. The starting point of the time window is determined by pushing forward a preset time length from the turning point where the difference value changes from positive to negative. The ending point of the time window is determined by extending backward a preset time length from this turning point. Within the time window, if the heart rate reaches the peak heart rate of the training unit and the blood oxygen saturation is below the oxygen drop threshold for a preset duration, the corresponding time period will be marked as a quasi-limit anchor point segment.

4. The method for determining and analyzing maximal oxygen uptake based on athlete training physiological data as described in claim 1, characterized in that, In the identification of the quasi-limit anchor point segment, the oxygen drop threshold is determined by at least one of the following methods: the difference between the individual athlete's resting blood oxygen saturation and the standard deviation of historical blood oxygen fluctuations; or the difference between the mean and standard deviation of blood oxygen saturation in the training unit.

5. The method for determining and analyzing maximal oxygen uptake based on athlete training physiological data as described in claim 1, characterized in that, The extraction methods for the local curvature parameters of the heart rate-power relationship, the phase synchronization index of respiratory rate and heart rate variability index, and the response delay of the heart rate variability index recovery slope and power gradient are as follows: From the submaximal steady-state segments obtained from historical cycles, with power as the independent variable and heart rate as the dependent variable, a local quadratic polynomial fitting is performed within the sliding power window to obtain the heart rate-power relationship curve. The second derivative of the heart rate-power relationship curve within each sliding power window is calculated as the local curvature parameter. From the submaximal steady-state segment, the instantaneous phase of the respiratory rate and heart rate variability index is extracted by Hilbert transform, and the cumulative average of the cosine values ​​of the instantaneous phase difference is calculated as the phase synchronization index. Obtain quasi-limit anchor point segments and their adjacent recovered segments from historical cycles; Within the adjacent recovery segment, the heart rate variability index is linearly fitted using a sliding time window to obtain the recovery slope sequence, and the power is linearly fitted to obtain the power gradient sequence. Cross-correlation is performed on the recovery slope sequence and the power gradient sequence, and the time shift corresponding to the maximum value of the cross-correlation function is used as the response delay.

6. The method for determining and analyzing maximal oxygen uptake based on athlete training physiological data as described in claim 5, characterized in that, The adjacent recovery segment is the time period from the end of the quasi-limit anchor point segment to the start of the next submaximal steady-state segment.

7. The method for determining and analyzing maximal oxygen uptake based on athlete training physiological data as described in claim 1, characterized in that, The method for dividing the submaximum, quasi-limit to limit transition, and limit interval is as follows: On the heart rate-power relationship curve, the power value corresponding to the first crossing of zero by the local curvature parameter from a negative value is determined as the first connection point; the decay rate of the phase synchronization index with power is calculated, and the power value corresponding to the local maximum value of the decay rate is determined as the second connection point. The interval where the power is less than the power of the first connection point is divided into the submaximal interval, the interval where the power is between the power of the first connection point and the power of the second connection point is divided into the quasi-limit to limit transition interval, and the interval where the power is greater than the power of the second connection point is divided into the limit interval.

8. The method for determining and analyzing maximal oxygen uptake based on athlete training physiological data as described in claim 1, characterized in that, The method for constructing the limit state extrapolation model is as follows: The product of the response delay and the average power gradient sequence is used as the power domain response scale; Using the power domain response scale as the power scale parameter of the exponential function in the quasi-limit to limit transition interval, the power value and heart rate value at the first connection point are used as the initial values ​​of the exponential function, and the power value and heart rate value at the second connection point are used as the termination reference values ​​of the exponential function. The power domain response scale is used as the power scale parameter to construct the exponential function in the quasi-limit to limit transition interval. The phase synchronization index value at the second connection point is used as the initial boundary value of the phase synchronization index decay trend within the limit interval, and the low-level stable value of the phase synchronization index in the historical limit anchor point segment is used as the termination boundary value. A limit state extrapolation model is constructed. The limit state extrapolation model is a linear function in the submaximal interval and an exponential function in the quasi-limit to limit transition interval. In the limit interval, the trend of the phase synchronization exponent decaying from the initial boundary value to the final boundary value is used as the boundary constraint.

9. The method for determining and analyzing maximal oxygen uptake based on athlete training physiological data as described in claim 1, characterized in that, The method for obtaining the estimated maximum oxygen uptake is as follows: Extract the phase synchronization index, response delay, and local curvature parameters from the quasi-limit anchor point segments identified in the current training cycle; The product of the response delay of the current training cycle and the average value of the current power gradient sequence is used as the current power domain response scale. Input the local curvature parameters, phase synchronization index, and current power domain response scale into the limit state extrapolation model; Heart rate values ​​are calculated using a linear function within the submaximal range and an exponential function within the transition range from the quasi-limit to the limit. Within the limit range, when the absolute value of the difference between the phase synchronization index decaying from the initial boundary value and the termination boundary value is not greater than a preset proximity threshold, the corresponding power value is determined as the limit power estimate. Substitute the limiting power estimate into the individualized power-oxygen uptake mapping function to output the maximum oxygen uptake estimate.

10. The method for determining and analyzing maximal oxygen uptake based on athlete training physiological data as described in claim 1, characterized in that, The method for constructing the individualized power-oxygen uptake mapping function is as follows: Obtain the power sequence and synchronously acquired oxygen uptake sequence from historical calibration data; Perform univariate linear regression on the power series and oxygen uptake series, and output the slope and intercept. Substituting the slope, intercept, and athlete's weight into a linear equation in one variable, we construct an individualized power-oxygen uptake mapping function.

Images