A multi-parameter coordinated sports medical nursing scheme personalized generation system

Abstract

The application discloses a kind of multi-parameter coordinated sports medical care scheme personalized generation system, it is related to sports medical care technical field, the multi-source sports medical parameter of target individual is collected, parameter sample sequence is constructed, and influence factor parameter table is generated based on nursing model, utilize sports load prediction model to carry out multi-scene deduction to parameter sample, obtain nursing risk response sequence, and generate corresponding nursing intervention duration sequence, by the joint coding input nursing scheme reconstruction process of influence factor parameter table, nursing risk response and intervention duration, filter and determine each parameter sample corresponding optimal nursing measure, finally form sports medical care scheme;Through the application, nursing scheme can respond to the change of target individual's sports load and physiological state difference in real time, reach the effect of dynamic regulation and control nursing strategy;While it can realize nursing risk early assessment and graded intervention, improve the accuracy and safety of sports injury prevention and rehabilitation nursing.

Description

Technical Field

[0001] This invention relates to the field of sports medicine and nursing technology, specifically to a personalized sports medicine and nursing plan generation system with multi-parameter synergy. Background Technology

[0002] With the popularization of competitive sports and mass sports, sports injuries are showing a trend of high incidence, diversification, and insidious occurrence. Sports medicine nursing plans typically require a comprehensive assessment combining multiple dimensions of information, including kinematic parameters, biomechanical parameters, physiological monitoring parameters, and individual medical history. However, in the intersection of sports training, rehabilitation training, and medical care, significant temporal variability, spatial abrupt changes, and individual adaptability exist among different parameters. For example, after strenuous exercise, microstructural damage occurs in joint soft tissues, and the inflammatory response window typically lasts only a few hours, while the metabolic recovery process from muscle fatigue can span 24-48 hours. Furthermore, differences in neuromuscular control abilities among individuals during the same exercise can lead to irregular fluctuations in the exercise load curve. These variability prevents traditional nursing plans from providing a dynamic response tailored to the individual, resulting in delayed and insufficiently precise nursing interventions.

[0003] On the other hand, current commonly used sports medicine nursing plans are mostly based on empirical models or single-parameter risk assessments, such as judging injury risk solely based on heart rate zones or electromyography (EMG) signals. However, in real-world scenarios, injuries in athletes or fitness enthusiasts are often triggered by a combination of factors, including fluctuations in ambient temperature and humidity, changes in training movement structure, resistance training load drift, and increases in local tissue temperature. Once nonlinear coupling occurs between these parameters, such as a rapid increase in deep muscle temperature without a timely response from superficial muscle EMG signals, irreversible small tendon fasciculations or attachment point strains can easily occur, severely impacting the athlete's training progress and rehabilitation outcomes.

[0004] To address the risk of sports injuries caused by the aforementioned lack of coordination among multiple parameters, some systems attempt to incorporate historical training data as compensation. However, these systems often rely on fixed templates and cannot dynamically adjust nursing plans based on real-time monitoring results. Furthermore, they cannot develop personalized intervention models in the early stages of parameter abnormalities, leading to delayed intervention and limited nursing effectiveness. Especially in professional sports rehabilitation settings, sports medicine nursing plans need to accurately respond to the combined changes in multiple physiological, biomechanical, and training parameters within a short period; otherwise, serious consequences such as increased tissue damage, joint instability, and decreased motor function may occur. Summary of the Invention

[0005] The purpose of this invention is to provide a personalized sports medicine nursing plan generation system with multi-parameter synergy to address the shortcomings of the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a personalized generation system for multi-parameter collaborative sports medicine nursing plans, comprising:

[0007] Parameter acquisition module: Acquires historical sports medicine monitoring data of the target individual during the preset exercise period to determine the initial range of multiple parameters related to the nursing plan, including the range of exercise biomechanical parameters, physiological load parameters, and tissue state parameters, and determines the maximum value Lmax and minimum value Lmin of the corresponding range. Parameter sample construction module: The intervals corresponding to Lmin and Lmax are uniformly divided to obtain the parameter sample sequence set P=(P1,P2,…,Pi,…,Pn), where Pi is the i-th parameter sample obtained by uniform division within the interval, and n is the number of parameter samples within the interval. Nursing Influence Factor Calculation Module: Based on the parameter sample sequence P, obtain several preset nursing influence factors of the sports medicine nursing model under each parameter sample Pi, and generate the corresponding influence factor parameter table M=(M1,M2,…,Mi,…,Mn), where Mi is the influence factor parameter table of the nursing model under sample Pi, including the expected muscle group load change rate, joint stability change factor and tissue temperature drift coefficient. Nursing risk prediction module: Based on the parameter sample sequence P and the influencing factor parameter table M, the exercise load prediction model is called to perform training scenario simulation to obtain the prediction result sequence set Q=(Q1,Q2,…,Qp,…,Qq) of the target individual's exercise nursing risk response, where Qp is the nursing risk response generated by the prediction model under the p-th parameter combination, and q is the number of prediction scenarios set by the model. Nursing intervention duration generation module: Obtain the intervention duration Tp corresponding to each risk response Qp in the Q sequence, and generate an intervention duration sequence table T=(t1,t2,…,tp,…,tq), where tp is the suggested nursing intervention duration for the corresponding Qp; Nursing plan generation module: Based on M, Q and T, the sports medicine nursing model is re-invoked to reconstruct the plan, and the individual's multi-parameter status is dynamically updated during the reconstruction process to generate a personalized sports medicine nursing plan for the target individual.

[0008] Preferably, the intervals corresponding to Lmin and Lmax are uniformly divided to obtain the parameter sample sequence set P, including: After obtaining the minimum value Lmin and the maximum value Lmax, the number of segments is selected based on the sensitivity of the target individual's motion parameters to change, so that the entire interval is divided into several segments with equal spacing characteristics, and each segment corresponds to a candidate parameter value range. After forming the segmented structure, representative parameter points are selected from each segment according to the principle of equal amplitude variation; The continuity and stability of the selected representative parameter points are verified. By comparing whether the change range of adjacent representative points remains smooth, abnormal representative points that do not conform to the normal change law of sports medicine parameters are eliminated. The verified representative parameter points are arranged in the order of parameter intervals to form a parameter sample sequence set P.

[0009] Preferably, generating the corresponding impact factor parameter table M further includes: After obtaining the parameter sample sequence P, the parameter input interface of the sports medicine nursing model is called to input each parameter sample Pi into the nursing model in sequence, and Pi is normalized according to the target individual's exercise ability level before input. After the sample input is completed, based on the structural characteristics of the nursing model, the muscle group load analysis process, joint stability analysis process and tissue temperature response analysis process are performed on each parameter sample Pi, and the expected muscle group load change rate, joint stability change factor and tissue temperature drift coefficient are output respectively. The consistency of the three types of influencing factors output by the nursing model is checked. By comparing their internal trends with the positional relationship of Pi in the interval, output results that do not conform to the laws of exercise physiology are eliminated, and a new set of checked influencing factors is generated. The verified set of impact factors is arranged into an impact factor parameter table Mi according to the one-to-one correspondence of parameter samples Pi, and then the impact factor parameter table M is formed in sequence.

[0010] Preferably, the method for obtaining the expected rate of change of muscle group load further includes: After obtaining the original mechanical parameters of the muscle groups output by the nursing model, based on the muscle group composition characteristics of the target individual, the mechanical feature data of the muscle groups, including the degree of muscle fiber activation, the distribution of muscle force direction and the temporal curve of muscle tension, are extracted, and the feature data are time-normalized. After completing the time adjustment, a muscle group load change analysis model is constructed based on muscle group biomechanical characteristic data. By comparing the synchronicity between the local slope changes of the muscle group tension time-series curve and the degree of muscle fiber activation, the muscle group load increase and decrease segments are identified. After identifying each segment, amplitude assessment is performed on the ascending and descending segments respectively. By calculating the average increase or decrease in tension within the segment, the corresponding muscle group load change amplitude value is generated. The expected rate of change of muscle load is formed by combining the amplitude of muscle load change with the stability parameter of the force distribution of the muscle group.

[0011] Preferably, the method for obtaining the joint stability change factor further includes: After obtaining the joint-related motion parameters output by the nursing model, joint dynamic feature data, including joint angle change trajectory, joint rotation speed curve and joint force distribution map, are extracted, and noise suppression and time alignment processing are performed on the feature data. After completing the data preprocessing, a joint stability analysis model is constructed based on the joint dynamics characteristic data. By identifying discontinuities in the trajectory of joint angle changes and abnormal peaks in the rotational velocity curve, key locations that may lead to a decrease in stability are located. After locating the key positions, stress concentration is assessed on the joint stress distribution diagram of the corresponding time period. Based on the deviation between the joint stress peak and uniformity, the joint stability change value is generated. The joint stability change amplitude value is combined with the smoothness parameter of the joint angle change trajectory to form the joint stability change factor.

[0012] Preferably, the method for obtaining the tissue temperature drift coefficient further includes: After obtaining the raw monitoring data related to tissue temperature output by the nursing model, the superficial tissue temperature curve and the deep tissue temperature curve are extracted respectively, and the two types of temperature curves are stabilized based on a sliding window. After stabilization, a temperature change gradient analysis model was constructed to compare the time difference sequences of the temperature curves of superficial and deep tissues, and to identify the synchronicity and lag characteristics of temperature changes in the two types of tissues. Based on the identified synchronicity and lag characteristics, the fluctuation amplitude of the temperature difference sequence between deep and shallow tissues is evaluated, and the tissue temperature drift amplitude value is generated according to the deviation between the maximum fluctuation amplitude and the average fluctuation amplitude of the temperature difference. The tissue temperature drift amplitude value is combined with the smoothness parameter of the synchronous change output by the temperature change gradient analysis model to form the tissue temperature drift coefficient.

[0013] Preferably, the exercise load prediction model is invoked to perform training scenario simulation, resulting in a prediction result sequence set Q of the target individual's exercise care risk response, including: After obtaining the parameter sample sequence P and the influence factor parameter table M, input sample pairs for training the exercise load prediction model are constructed based on the one-to-one correspondence between the two. Each parameter sample Pi is combined with its corresponding influence factor parameter table Mi to form a training input unit. After constructing the training input unit, based on the motion load prediction model, feature enhancement processing is performed before scenario simulation. An extended feature set is generated by calculating the position coefficient of Pi in the interval and the gradient of the changes of each influencing factor in Mi. After feature enhancement is completed, the training input unit and the extended feature set are input into the exercise load prediction model. By executing multiple rounds of training scenario simulation, the changes in nursing risk response of the target individual under different assumed exercise load levels are simulated, and multiple stages of risk response output results are obtained. Each simulation output nursing risk response is numbered and organized according to its corresponding input sample Pi to form a prediction result sequence set Q.

[0014] Preferably, the treatment plan is reconstructed based on the sports medicine nursing model according to M, Q, and T, including: After obtaining the influencing factor parameter table M, the nursing risk response sequence Q, and the intervention duration sequence T, the nursing plan reconstruction input unit is constructed based on the correspondence between the three on the parameter sample Pi. By jointly encoding the influencing factor set, the risk change curve, and the intervention duration structure, an input feature sequence for reconstruction is formed. After constructing the input feature sequence, the scheme generation structure of the sports medicine nursing model is called to perform intervention method screening calculation for each input feature sequence. By comparing the adaptability of different nursing methods under the current risk level, a set of candidate nursing measures is generated. After obtaining the set of candidate nursing measures, the candidate nursing measures are matched and ranked according to the intervention duration sequence T. The duration matching scoring algorithm is used to calculate the suitability score of each nursing measure, thereby determining the optimal nursing measure corresponding to each parameter sample Pi. The optimal nursing measures corresponding to each parameter sample Pi are combined in interval order to form the final personalized sports medicine nursing plan.

[0015] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention constructs a multi-parameter collaborative method for generating sports medicine nursing plans, establishing a continuous and quantifiable mapping relationship between parameter sample sequences, influencing factor parameter tables, nursing risk responses, and intervention duration sequences. This enables the sports medicine nursing model to dynamically extrapolate nursing care based on the target individual's real-time physiological characteristics, exercise load trends, and tissue response changes. By introducing muscle group biomechanics analysis models, joint dynamic stability analysis models, and tissue temperature gradient analysis models, the construction of nursing plans no longer relies on fixed templates but is generated based on real parameter changes. This significantly improves the adaptability of nursing plans to the individual's actual exercise state, enabling individualized and progressive nursing decisions that traditional nursing plans cannot cover.

[0016] 2. This invention achieves early identification and quantitative assessment of potential sports injury risks through the collaborative calculation of nursing risk response prediction and intervention duration generation, avoiding secondary injuries caused by delayed or inappropriate intervention intensity. By jointly inputting the set of influencing factors, risk response curves, and intervention duration structure into the nursing plan reconstruction process, this invention can generate optimal nursing measures for each parameter sample, enabling the nursing plan to possess continuity, logical consistency, and adaptive duration. Compared to traditional experience-based nursing methods, this invention can output more stable, scientific, and feasible sports medicine nursing plans under complex coupling of multidimensional parameters, significantly improving the accuracy and safety of nursing interventions. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0018] Figure 1 This is a flowchart of a multi-parameter collaborative personalized sports medicine nursing plan generation system module of the present invention.

[0019] Figure 2 This is a flowchart illustrating the method for generating the influence factor parameter table M of the present invention.

[0020] Figure 3 This is a flowchart of the method for reconstructing a treatment plan using a sports medicine nursing model, as described in this invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0022] For examples, please refer to Figure 1 As shown in this embodiment, a personalized sports medicine nursing plan generation system with multi-parameter collaborative methods includes: Parameter acquisition module: Acquires historical sports medicine monitoring data of the target individual during a preset exercise period to determine the initial range of multiple parameters related to the nursing plan, including the range of exercise biomechanical parameters, physiological load parameters, and tissue state parameters, and determines the maximum value Lmax and minimum value Lmin of the corresponding range.

[0023] In this embodiment, during the generation of a sports medicine care plan for the target individual, noise suppression processing is first required on the raw sports medicine monitoring data collected during the preset exercise period. The raw monitoring data includes exercise biomechanical parameter data (e.g., real-time sequences of joint posture angles, limb acceleration sequences, and joint torque sequences), physiological load parameter data (e.g., heart rate waveform sequences, electromyographic signal sequences, and respiratory rate sequences), and tissue state parameter data (e.g., superficial tissue temperature sequences and deep tissue temperature sequences).

[0024] To ensure the accuracy of various parameters in subsequent interval division, a bandpass filtering algorithm is used for noise suppression for each type of parameter. The bandpass filtering algorithm constructs a bandpass filter function composed of a lower cutoff frequency and an upper cutoff frequency to simultaneously suppress high-frequency noise and low-frequency drift components of the original signal.

[0025] The bandpass filtering algorithm is constructed as follows: Let the original sequence of acquired biomechanical parameters be S(t), and construct a bandpass filter function H(f), where f is the frequency variable. The lower cutoff frequency f1 and the upper cutoff frequency f2 are selected according to the actual frequency band of the biomechanical parameters. For example, f1 can be set to 0.5 Hz, and f2 can be set to 20 Hz. The filtered biomechanical parameter sequence is: Sfiltered(t) = inverse Fourier transform (Fourier transform (S(t)) × H(f)); where "Fourier transform" is the calculation method for converting the time-domain signal to the frequency-domain signal, and "inverse Fourier transform" is the calculation method for restoring the frequency-domain signal back to the time-domain signal.

[0026] The above methods can be used to obtain preprocessed sequences of exercise biomechanical parameters, physiological load parameters, and tissue state parameters, providing a parameter basis after noise suppression for subsequent steps.

[0027] Because different sensors have different sampling frequencies, various monitoring parameters have time offsets within the same time period. Therefore, it is necessary to perform time axis alignment on the preprocessed parameter sequence. The time axis alignment is constructed using a dynamic time warping algorithm.

[0028] The dynamic time warping algorithm is constructed as follows: Let the preprocessing sequence of exercise biomechanical parameters be X(i), the preprocessing sequence of physiological load parameters be Y(j), and the preprocessing sequence of tissue state parameters be Z(k), where i, j, and k represent the sampling point indices. By constructing the cost matrix D(i,j), the absolute difference |X(i) is used as the basis for the algorithm. Using Y(j)| as the cost function and taking the path with the minimum cumulative cost as the optimal path, the best time mapping relationship between sequence X(i) and Y(j) is obtained.

[0029] Similarly, physiological load parameters and tissue state parameters are normalized so that the three types of parameters are mapped to the same reference time scale Tref, forming aligned parameter sequences Xaligned(t), Yaligned(t), and Zaligned(t). This time axis alignment process ensures that all types of parameters can participate in interval analysis within the same time window during subsequent interval boundary calculations, providing a consistent data foundation for the interval scanning algorithm.

[0030] After obtaining the aligned parameter sequence, it is necessary to scan the instantaneous fluctuations within the exercise biomechanical parameter range, physiological load parameter range, and tissue state parameter range to identify abnormal drift points within the range.

[0031] The interval scanning algorithm is constructed as follows: On the aligned time axis Tref, intervals are traversed using a sliding window Wsize of fixed length. Let the data subsequence within the window be Xwindow, and calculate the instantaneous fluctuation value V of the parameters within this window as follows: V = maximum value (Xwindow). Minimum value (Xwindow): By comparing the instantaneous fluctuation value V with the preset fluctuation threshold Vthreshold, when V > Vthreshold, the time point corresponding to the center of this window is recorded as the abnormal drift point. The fluctuation threshold Vthreshold is set by statistically analyzing the normal fluctuation range of this type of parameter in the past three motion cycles of the target individual, and taking twice the average fluctuation value as the threshold.

[0032] The above scans yield a set of abnormal drift points in each parameter range, providing a basis for subsequent boundary correction.

[0033] Once abnormal drift points of exercise biomechanical parameters, physiological load parameters, and tissue state parameters are obtained, the initial interval boundaries need to be corrected based on these drift points to obtain an interval that better reflects the actual changes in exercise state.

[0034] The boundary correction method is as follows: The lower boundary of the original interval is Lminraw, and the upper boundary is Lmaxraw. When the parameter value corresponding to an abnormal drift point exceeds the original interval boundary, the parameter value of that drift point is used as the new interval boundary. For example, when the drift point parameter value is less than Lminraw, it is replaced with the new minimum value Lmin. When the drift point parameter value is greater than Lmaxraw, it is replaced with the new maximum value Lmax.

[0035] By examining each drift point individually, updated intervals for exercise biomechanical parameters, physiological load parameters, and tissue state parameters can be obtained. Furthermore, the final maximum value Lmax and minimum value Lmin for each interval can be calculated.

[0036] Parameter sample construction module: The intervals corresponding to Lmin and Lmax are uniformly divided to obtain the parameter sample sequence set P=(P1,P2,…,Pi,…,Pn), where Pi is the i-th parameter sample obtained by uniform division within the interval, and n is the number of parameter samples within the interval.

[0037] After determining the minimum value Lmin and the maximum value Lmax, the number of segments needs to be selected based on the sensitivity of the target individual's motion parameters to change. Sensitivity describes the rate of change of the target individual's parameters over multiple motion cycles, and it is calculated as follows: Sensitivity = (maximum rate of change + average rate of change) / 2; where the maximum rate of change is the maximum increase in parameter value per unit time in the historical parameter sequence, and the average rate of change is the average of all rates of change.

[0038] Based on sensitivity to change (Ssensitivity), sensitivity is divided into three levels: high sensitivity, medium sensitivity, and low sensitivity. When Ssensitivity is greater than the first threshold (Shigh), it is classified as high sensitivity; when Ssensitivity is between the first and second thresholds, it is classified as medium sensitivity; and when Ssensitivity is less than the second threshold, it is classified as low sensitivity.

[0039] The number of segments N is selected based on the sensitivity level, with a larger N for high sensitivity, a medium-sized N for medium sensitivity, and a smaller N for low sensitivity. The segmentation step size Δ is determined by the following formula: Δ = (Lmax...) Lmin)÷N; The segmentation step size Δ is used as the segment spacing to divide the parameter interval into N continuous segments with equal spacing. Each segment is defined as: Each Segmenti represents the range of candidate parameter values.

[0040] After forming the segmented structure, a representative parameter point Ti needs to be selected from each Segmenti. The principle of constant amplitude variation is used to ensure that the representative parameter point can reflect the local variation trend of the segment.

[0041] When constructing the principle of constant amplitude variation, the first step is to analyze the dataset Hi corresponding to the historical parameter sequences within Segmenti. Based on the trend category of Hi (upward trend, downward trend, or steady-state trend), the location of representative parameter points is determined. When Hi shows an upward trend: Ti = Segmenti upper boundary value; When Hi shows a downward trend: Ti = Segmenti lower boundary value; When Hi shows a steady-state trend: Ti = midpoint value of Segmenti; The trend category is determined using the following expression: Trendi = Average Slope = (Hi last term) Hi (first term) ÷ Number of sampling points within Segmenti; Trendi>0 → Upward trend; Trendi < 0 → Downward trend; Trendi≈0 → Steady-state trend.

[0042] The representative parameter point set {T1,T2,…,TN} corresponding to each segment is obtained through the above method.

[0043] After obtaining the representative parameter point Ti, a continuity check is performed. The continuity check determines whether any adjacent points Ti and T(i+1) maintain a smooth transition. The expression for the magnitude of the change is used: Di is then compared with the continuity threshold Dthreshold. Dthreshold is calculated from the average fluctuation range of the historical parameters of the target individual: Dthreshold = average historical fluctuation range × safety factor; if Di > Dthreshold, then it is considered that there is a sudden change that does not conform to the normal movement law, and T(i+1) is marked as an abnormal representative point.

[0044] Stability checks are used to determine whether Ti falls within the acceptable range Arange of the target individual's historical parameters. Arange is constructed by statistically analyzing the minimum and maximum values ​​of parameters over multiple past exercise cycles: Arange = [Amin, Amax]. If Ti does not belong to Arange, it is also marked as an anomalous representative point. All Ti points marked as anomalous representative points are removed to ensure the stationarity and physiological rationality of the final sample set.

[0045] After removing outlier representative points, the remaining representative parameter points are reordered in ascending order from Lmin to Lmax to ensure a strictly increasing sequence. The final parameter sample sequence set P is represented as follows: Where M is the number of representative parameter points remaining after continuity and stability checks.

[0046] The final parameter sample sequence set P has the characteristics of equal spacing, smooth change trend, and consistency with the physiological changes of the target individual, and can be directly used as input for the subsequent step of constructing the influencing factor parameter table.

[0047] Nursing Influence Factor Calculation Module: Based on the parameter sample sequence P, it obtains several preset nursing influence factors of the sports medicine nursing model under each parameter sample Pi, and generates the corresponding influence factor parameter table M=(M1,M2,…,Mi,…,Mn), where Mi is the influence factor parameter table of the nursing model under sample Pi, including the expected muscle group load change rate, joint stability change factor and tissue temperature drift coefficient.

[0048] After obtaining the parameter sample sequence P, each parameter sample Pi in the sequence needs to be fed into the sports medicine nursing model as an input signal. To ensure that different parameter samples participate in the model operation on the same computational scale, the parameter sample Pi needs to be normalized.

[0049] Normalization is performed using an interval scaling algorithm, and the normalized parameter samples are denoted as Pinorm. The normalization calculation expression is as follows: Pinorm = (Pi Pmin) ÷ (Pmax) Pmin), where Pmin is the minimum value in the P sequence, used to determine the lower bound of the normalization interval; Pmax is the maximum value in the P sequence, used to determine the upper bound of the normalization interval. The interval scaling algorithm ensures that all Pinorms are within the range [0,1].

[0050] Pinorm is input into the sports medicine nursing model in sequence. The model performs muscle group load analysis, joint stability analysis and tissue temperature response analysis according to the predetermined calculation structure, in preparation for the subsequent generation of influencing factors.

[0051] After receiving the parameter sample Pinorm, the model outputs the corresponding raw mechanical parameters related to the muscle group, including the muscle fiber activation sequence, the muscle group force direction distribution sequence, and the muscle group tension time series curve. The above three types of data are key variables in muscle dynamics.

[0052] To ensure that different types of data can be compared on the same time axis, time normalization is required. Time normalization uses a linear interpolation algorithm to map data with different sampling frequencies to a unified time series Tref, thereby forming muscle group biomechanical feature sequences F(t), D(t), and T(t), where F(t) is the muscle fiber activation function, D(t) is the force direction distribution function, and T(t) is the tension time series curve function.

[0053] A muscle group load change analysis model is used to identify the increasing and decreasing segments of load during muscle tension changes. The model employs a local slope analysis algorithm, with the slope calculation expression being: Slope(t) = dT(t) ÷ dt; where T(t) is the function of muscle tension change over time; dT(t) is the tension derivative; and dt is the time derivative unit. When Slope(t) is greater than zero, it indicates an increase in tension, which can be classified as an increasing segment; when Slope(t) is less than zero, it indicates a decrease in tension, which can be classified as a decreasing segment. To avoid short-term noise interference, a minimum duration threshold Tmin needs to be set. When the duration of a segment is less than Tmin, it is considered an abnormal segment and is removed.

[0054] Calculate the average tension change amplitude for each of the identified ascending and descending segments. The expression for the segment amplitude is: Asegment = (Tend Tstart ÷ Duration segment; where Tstart is the initial tension value of the segment, Tend is the final tension value of the segment, and Duration segment is the duration of the segment. Simultaneously, the stability parameter of the muscle group force direction distribution D(t) is included in the calculation. The stability parameter Sdirection is generated based on the standard deviation of the force direction: Sdirection = Std(D(t)); where Std is the standard deviation function used to describe the magnitude of the force direction change. Combining Asegment and Sdirection yields the expected muscle group load change rate Rmuscle: Rmuscle = Asegment × (1 ÷ (1 + Sdirection)).

[0055] The joint-related motion parameters output by the nursing model include the joint angle change trajectory θ(t), joint rotation velocity ω(t), and joint force distribution diagram Fjoint(t). Noise suppression processing is required to improve the accuracy of subsequent analyses.

[0056] Noise suppression employs a bandpass filtering algorithm, which constructs a bandpass function H(f) to ensure that the angle trajectory and velocity curve retain only the effective components within the actual motion frequency band. The filtered sequences are denoted as θfiltered(t) and ωfiltered(t).

[0057] The time alignment process uses a dynamic time warping algorithm to map Fjoint(t) and the two types of sequences mentioned above to the Tref time axis.

[0058] The joint stability analysis model is constructed using a combination of a continuity detection algorithm and an anomaly peak detection algorithm. Continuity detection is based on the change of the second derivative of the angle change function θfiltered(t). If the second derivative shows an abrupt change, it is marked as a discontinuity segment.

[0059] Abnormal peak detection is based on the maximum value of rotational velocity ωfiltered(t). When ωfiltered(t) exceeds the normal upper limit threshold Wthreshold, it is marked as an abnormal peak.

[0060] Multiple markers form a set K of critical locations that may lead to a decrease in stability.

[0061] Extract the joint stress distribution map Fjoint(t) near the critical location K, and calculate the stress concentration Cstress of this section: Cstress = Fpeak Fmean; where Fpeak is the maximum stress value in this section, and Fmean is the average stress value. The greater the deviation, the more significant the decrease in joint stability. Combining the stress concentration Cstress with the angular trajectory smoothness Sangle (determined by the standard deviation of the first derivative of θfiltered(t), we obtain the joint stability variation factor Rjoint: Rjoint = Cstress × (1 + Sangle).

[0062] The model outputs the superficial tissue temperature Ts(t) and the deep tissue temperature Td(t). A sliding window W is used for stabilization to smooth out short-period temperature noise.

[0063] A temperature change gradient analysis model is constructed, and the temperature difference sequence ΔT(t) is calculated: ΔT(t) = Td(t) Ts(t); and calculate the temperature gradient GT: GT = dΔT(t) ÷ dt. When GT is in the same direction as the shallow or deep temperature gradient, it is determined to be a synchronous change; when the direction is opposite or the delay exceeds the time threshold Tdelay, it is determined to be a lagging change.

[0064] Obtain the maximum fluctuation amplitude ATmax and the average fluctuation amplitude ATavg of ΔT(t), and calculate the fluctuation deviation DT: DT = ATmax ATavg; Combine the fluctuation deviation DT with the temperature change gradient smoothness ST (determined by the standard deviation of GT) to obtain the tissue temperature drift coefficient Rtemp: Rtemp=DT×(1+ST).

[0065] The three types of influencing factors Rmuscle, Rjoint, and Rtemp are combined into an influencing factor set Fi. A consistency check is performed on Fi. The check model uses a trend comparison algorithm: Deviation = |SlopeF| SlopeP|; SlopeF is the slope of Fi along the sample sequence, and SlopeP is the slope of the Pinorm sequence. If the deviation is greater than the deviation threshold, it is judged as an abnormal influence factor and removed.

[0066] The verified effective impact factors are arranged in the order of the parameter samples Pi to form the impact factor parameter table Mi; all Mi are composed of: M=(M1,M2,…,Mi,…,Mn); which serves as the input for subsequent risk prediction steps.

[0067] Nursing risk prediction module: Based on the parameter sample sequence P and the influencing factor parameter table M, the exercise load prediction model is called to perform training scenario simulation, and the prediction result sequence set Q=(Q1,Q2,…,Qp,…,Qq) of the target individual's exercise nursing risk response is obtained, where Qp is the nursing risk response generated by the prediction model under the p-th parameter combination, and q is the number of prediction scenarios set by the model.

[0068] After obtaining the parameter sample sequence P and the influencing factor parameter table M, it is necessary to construct input sample pairs for training the exercise load prediction model based on their one-to-one correspondence. For the parameter sample Pi, its corresponding influencing factor parameter table Mi consists of the expected muscle group load change rate, joint stability change factor, and tissue temperature drift coefficient.

[0069] To enable the prediction model to simultaneously receive multiple types of parameter features, Pi and Mi need to be combined to construct the training input unit Ui. The combination method employs feature vector concatenation, arranging each influencing factor value in Pi and Mi in a fixed order to form the training input vector Ui={Pi,Mi1,Mi2,Mi3}. Here, Mi1 is the expected rate of change of muscle load, Mi2 is the joint stability change factor, and Mi3 is the tissue temperature drift coefficient.

[0070] The training input unit Ui is used as the direct input to the exercise load prediction model to ensure that the model can simultaneously consider the interval characteristics of the parameter samples and the physical meaning of the influencing factors during the training process, thereby improving the physiological adaptability of the prediction.

[0071] After constructing the training input units, feature enhancement processing is needed to improve the ability of the motion load prediction model to identify different motion states. The main purpose of feature enhancement processing is to add higher-order features that better reflect the changing patterns of motion load based on the original input units.

[0072] Feature enhancement employs two algorithms: position coefficient calculation and gradient calculation. The position coefficient Cposition is obtained through the following expression: Cposition = (Pi... Pmin) ÷ (Pmax) Pmin); where Pmin is the minimum value in the parameter sample sequence, Pmax is the maximum value, and Cposition reflects the relative position of Pi in the entire interval.

[0073] The gradient Gfactor is used to describe the changing trend of various influencing factors within Mi as the sample sequence changes. It is calculated as follows: Gfactor = (Mi...) Miprevious) ÷ ΔP; where Miprevious is the influence factor parameter value corresponding to the previous sample, and ΔP is the parameter difference between adjacent samples. The gradient of change can reveal the sensitivity of the influence factor parameter table to changes over intervals.

[0074] The final extended feature set Ei is formed, consisting of the position coefficient Cposition and the gradients of three influencing factors, i.e., Ei = {Cposition, Gfactor1, Gfactor2, Gfactor3}. After feature enhancement, the training input unit Ui and the extended feature set Ei are input into the exercise load prediction model. This model employs a multi-scenario simulation algorithm, aiming to extrapolate the nursing risk response of the target individual under various exercise conditions by adjusting the assumed levels of exercise load parameters.

[0075] The exercise load prediction model employs a multi-stage recursive calculation structure. Let the calculated nursing risk response be Qp, where p is the number of the current simulation scenario. By setting the number of prediction scenarios q, the model will output the risk response under q different assumptions.

[0076] The model calculation process uses a risk response function: Qp=f(Ui,Ei,Lp); where Ui is the training input unit; Ei is the extended feature set; and Lp is the motion load level variable for the p-th scenario. Lp is constructed by setting different intensities, durations, and frequencies to enable the model to simulate different motion scenarios. The function f is constructed based on the mechanical, dynamic, and thermophysiological rules of the nursing model.

[0077] After the model completes multiple rounds of training scenario simulations, multiple nursing risk response values ​​corresponding to different exercise load levels can be obtained. The risk response result Qp from each simulation is numbered and organized according to its corresponding training input unit Ui to form a prediction result sequence set Q. The prediction result sequence set Q is represented as: Q=(Q1,Q2,…,Qp,…,Qq); where Qp is the nursing risk response under the p-th prediction scenario, and q is the set total number of simulation scenarios.

[0078] The predicted result sequence set Q will serve as the input basis for the subsequent nursing intervention duration generation step, used to determine the most suitable sports medicine nursing intervention for the target individual, so that nursing decisions can accurately reflect the risk trends of the predictive model.

[0079] Nursing intervention duration generation module: Obtain the intervention duration Tp corresponding to each risk response Qp in the Q sequence, and generate an intervention duration sequence table T=(t1,t2,…,tp,…,tq), where tp is the recommended nursing intervention duration corresponding to Qp.

[0080] After obtaining the nursing risk response sequence Q, it is necessary to generate the corresponding nursing intervention duration Tp for each nursing risk response Qp. For this purpose, a nursing intervention duration calculation model needs to be constructed. This model is used to describe the quantitative relationship between the degree of nursing risk and the required intervention time.

[0081] The nursing intervention duration calculation model is constructed using a risk grading weighted algorithm. First, the value range of Qp is graded and divided into a low-risk area, a high-risk area, and an extremely high-risk area. The risk area division threshold is determined using a statistical distribution method: Qlow is the twenty-fifth percentile of the Q sequence, Qhigh is the seventy-fifth percentile, and Qextreme is the mean of the maximum value interval. The basic calculation expression for the nursing intervention duration Tp is: Tp = Wp × Btime; where, Wp is the risk weighting coefficient and Btime is the basic intervention duration. The basic intervention duration Btime is preset according to the movement ability level of the target individual. The risk weighting coefficient Wp is determined according to the risk interval to which Qp belongs: When Qp ≤ Qlow, the value of Wp is 1; When Qlow < Qp ≤ Qhigh, the value of Wp is between 1 and 2 and is determined by linear interpolation; When Qp > Qhigh, the value of Wp is an enhancement coefficient greater than 2.

[0082] Since the nursing risk response Qp has a continuous characteristic, the nursing intervention duration Tp also needs to be dynamically calibrated according to the change trend of Qp. Dynamic calibration uses a trend derivative analysis algorithm.

[0083] The trend derivative SlopeQp of Qp is defined as: SlopeQp = (Qp - Qpprevious) ÷ Δp; where, Qpprevious is the nursing risk response corresponding to the previous parameter sample; Δp is the serial number difference between adjacent nursing risk responses, usually taking 1. SlopeQp is used to describe the risk change speed.

[0084] When SlopeQp is greater than the trend threshold Slopethreshold, it indicates that the current risk is in a rapid upward stage. At this time, the nursing intervention duration Tp needs to increase the dynamic compensation duration DeltaT: Tpadjusted = Tp + DeltaT; the dynamic compensation duration DeltaT is generated based on the difference between SlopeQp and Slopethreshold, and the compensation ratio uses a linear enhancement method to enable timely nursing for rapidly rising risks.

[0085] To avoid abrupt changes in nursing intervention duration between consecutive samples, it is necessary to smooth the execution time of all Tp to form an intervention duration sequence T that is more consistent with the execution patterns of medical nursing.

[0086] The time smoothing algorithm is constructed using a weighted moving average method. Let the final smoothed duration be tp, which is expressed as: tp = (α × Tp) + (β × Tpprevious) + (γ × Tpnext); where Tpprevious is the intervention duration at the previous time step, and Tpnext is the intervention duration at the next time step; α, β, and γ are smoothing weight parameters, satisfying α + β + γ = 1. The smoothing weight parameters are determined empirically, with α preferentially used to maintain current risk information, and β and γ used to stabilize sequence changes.

[0087] After performing dynamic calibration and time smoothing on all Tp, all tp are arranged sequentially according to the order of parameter samples Pi to form the intervention duration sequence T.

[0088] The intervention duration sequence T is represented as: T=(t1,t2,…,tp,…,tq); where tp is the final nursing intervention duration corresponding to the p-th nursing risk response Qp; and q is the total number of simulation scenarios set by the exercise load prediction model.

[0089] The final generated intervention duration sequence T serves as a key input for the generation of subsequent nursing plans, enabling the nursing plans to present a refined and gradual intervention rhythm based on the trend of risk changes.

[0090] Nursing plan generation module: Based on M, Q and T, the sports medicine nursing model is re-invoked to reconstruct the plan, and the individual's multi-parameter status is dynamically updated during the reconstruction process to generate a personalized sports medicine nursing plan for the target individual.

[0091] After obtaining the influencing factor parameter table M, the nursing risk response sequence Q, and the intervention duration sequence T, it is necessary to construct the nursing plan reconstruction input unit based on the one-to-one correspondence between the three on the parameter sample Pi. To this end, the influencing factor set Fi={Rmuscle,Rjoint,Rtemp}, the nursing risk response Qp, and the intervention duration tp corresponding to the same parameter sample Pi are jointly encoded to form the input feature sequence Xi for nursing plan reconstruction.

[0092] The joint encoding employs a vector concatenation method, constructing the input vector Xi={Rmuscle,Rjoint,Rtemp,Qp,tp} by arranging Fi, Qp, and tp in a fixed order; where: Rmuscle is the expected rate of change of muscle load; Rjoint is the joint stability change factor; Rtemp is the tissue temperature drift coefficient; Qp is the risk response value; and tp is the duration of nursing intervention. The above input feature sequence Xi comprehensively describes the physiological state, risk state, and intensity of nursing needs of the target individual under parameter sample Pi.

[0093] After constructing the input feature sequence, the protocol generation structure of the sports medicine nursing model is invoked to perform intervention method selection calculations for each Xi. The protocol generation structure calculates the protocol using a nursing fitness function, which evaluates the reliability and effectiveness of different nursing interventions under the current feature input. The nursing fitness function Fcare is constructed as follows: Fcare = w1 × Rmuscle + w2 × Rjoint + w3 × Rtemp + w4 × Qp; where w1, w2, w3, and w4 are nursing fitness weight parameters used to control the relative importance of different influencing factors; all weights satisfy w1 + w2 + w3 + w4 = 1; the weight parameters are set based on the target individual's sports type and injury history.

[0094] The set of candidate nursing interventions is denoted as C = {C1, C2, ..., Ck}. The model calculates the corresponding nursing fitness value Fci for each nursing intervention Ci.

[0095] If Fci exceeds the nursing fitness threshold Fthreshold, the nursing intervention is determined to be suitable for the current feature sequence Xi and included in the candidate nursing intervention set.

[0096] The nursing fitness threshold Fthreshold, based on statistics from a large number of sports medicine nursing cases, represents the minimum fitness level at which nursing interventions can produce significant intervention effects.

[0097] After obtaining the candidate nursing intervention set C, the candidate nursing interventions need to be matched and ranked according to the intervention duration tp. The matching calculation uses a duration matching scoring algorithm, the purpose of which is to ensure that the selected nursing intervention can produce the best intervention effect within the time length corresponding to tp. The duration matching score Smatch is calculated as follows: Where: DurationCi is the standard execution time of nursing intervention Ci; tp is the nursing intervention time corresponding to the current parameter sample Pi; the absolute difference in the denominator represents the degree of deviation between the candidate nursing intervention and the target intervention time.

[0098] For each nursing intervention Ci, the corresponding nursing fitness value Fci is combined with the duration matching score Smatch to generate a comprehensive score Sfinal: Sfinal = α × Fci + β × Smatch; where α and β are the comprehensive score weights, satisfying α + β = 1, used to balance nursing effectiveness and duration fit. The nursing intervention with the highest Sfinal value is determined as the optimal nursing intervention for the parameter sample Pi.

[0099] After determining the optimal nursing intervention for each parameter sample Pi, these optimal nursing interventions are combined according to the interval order of the parameter sample sequence P to form the final personalized sports medicine nursing plan. The personalized nursing plan Splan is represented as: Splan={Care1,Care2,…,Carei,…,Caren}; where: Carei is the optimal nursing intervention corresponding to the parameter sample Pi; n is the number of parameter sample sequences P.

[0100] The resulting personalized sports medicine care plan can be used to guide target individuals to implement precise and highly appropriate medical care interventions at different stages of exercise, thereby reducing the risk of sports injuries and improving rehabilitation efficiency.

[0101] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes 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.

Claims

1. A personalized sports medicine nursing plan generation system with multi-parameter collaborative mechanism, characterized in that: include: Parameter acquisition module: Acquires historical sports medicine monitoring data of the target individual during the preset exercise period to determine the initial range of multiple parameters related to the nursing plan, including the range of exercise biomechanical parameters, physiological load parameters, and tissue state parameters, and determines the maximum value Lmax and minimum value Lmin of the corresponding range. Parameter sample construction module: The intervals corresponding to Lmin and Lmax are uniformly divided to obtain the parameter sample sequence set P=(P1,P2,…,Pi,…,Pn), where Pi is the i-th parameter sample obtained by uniform division within the interval, and n is the number of parameter samples within the interval. Nursing Influence Factor Calculation Module: Based on the parameter sample sequence P, obtain several preset nursing influence factors of the sports medicine nursing model under each parameter sample Pi, and generate the corresponding influence factor parameter table M=(M1,M2,…,Mi,…,Mn), where Mi is the influence factor parameter table of the nursing model under sample Pi, including the expected muscle group load change rate, joint stability change factor and tissue temperature drift coefficient. Nursing risk prediction module: Based on the parameter sample sequence P and the influencing factor parameter table M, the exercise load prediction model is called to perform training scenario simulation to obtain the prediction result sequence set Q=(Q1,Q2,…,Qp,…,Qq) of the target individual's exercise nursing risk response, where Qp is the nursing risk response generated by the prediction model under the p-th parameter combination, and q is the number of prediction scenarios set by the model. Nursing intervention duration generation module: Obtain the intervention duration Tp corresponding to each risk response Qp in the Q sequence, and generate an intervention duration sequence table T=(t1,t2,…,tp,…,tq), where tp is the suggested nursing intervention duration for the corresponding Qp; Nursing plan generation module: Based on M, Q and T, the sports medicine nursing model is re-invoked to reconstruct the plan, and the individual's multi-parameter status is dynamically updated during the reconstruction process to generate a personalized sports medicine nursing plan for the target individual.

2. The personalized sports medicine nursing plan generation system with multi-parameter collaborative method according to claim 1, characterized in that: in, The intervals corresponding to Lmin and Lmax are uniformly divided to obtain the parameter sample sequence set P, which includes: After obtaining the minimum value Lmin and the maximum value Lmax, the number of segments is selected based on the sensitivity of the target individual's motion parameters to change, so that the entire interval is divided into several segments with equal spacing characteristics, and each segment corresponds to a candidate parameter value range. After forming the segmented structure, representative parameter points are selected from each segment according to the principle of equal amplitude variation; The continuity and stability of the selected representative parameter points are verified. By comparing whether the change range of adjacent representative points remains smooth, abnormal representative points that do not conform to the normal change law of sports medicine parameters are eliminated. The verified representative parameter points are arranged in the order of parameter intervals to form a parameter sample sequence set P.

3. The personalized sports medicine nursing plan generation system with multi-parameter collaborative mechanism according to claim 1, characterized in that: The corresponding impact factor parameter table M is further generated including: After obtaining the parameter sample sequence P, the parameter input interface of the sports medicine nursing model is called to input each parameter sample Pi into the nursing model in sequence, and Pi is normalized according to the target individual's exercise ability level before input. After the sample input is completed, based on the structural characteristics of the nursing model, the muscle group load analysis process, joint stability analysis process and tissue temperature response analysis process are performed on each parameter sample Pi, and the expected muscle group load change rate, joint stability change factor and tissue temperature drift coefficient are output respectively. The consistency of the three types of influencing factors output by the nursing model is checked. By comparing their internal trends with the positional relationship of Pi in the interval, output results that do not conform to the laws of exercise physiology are eliminated, and a new set of checked influencing factors is generated. The verified set of impact factors is arranged into an impact factor parameter table Mi according to the one-to-one correspondence of parameter samples Pi, and then the impact factor parameter table M is formed in sequence.

4. The personalized sports medicine nursing plan generation system with multi-parameter collaborative method according to claim 3, characterized in that: in, The method for obtaining the expected rate of change in muscle group load further includes: After obtaining the original mechanical parameters of the muscle groups output by the nursing model, based on the muscle group composition characteristics of the target individual, the mechanical feature data of the muscle groups, including the degree of muscle fiber activation, the distribution of muscle force direction and the temporal curve of muscle tension, are extracted, and the feature data are time-normalized. After completing the time adjustment, a muscle group load change analysis model is constructed based on muscle group biomechanical characteristic data. By comparing the synchronicity between the local slope changes of the muscle group tension time-series curve and the degree of muscle fiber activation, the muscle group load increase and decrease segments are identified. After identifying each segment, amplitude assessment is performed on the ascending and descending segments respectively. By calculating the average increase or decrease in tension within the segment, the corresponding muscle group load change amplitude value is generated. The expected rate of change of muscle load is formed by combining the amplitude of muscle load change with the stability parameter of the force distribution of the muscle group.

5. The personalized sports medicine nursing plan generation system according to claim 3, characterized in that: The method for obtaining the joint stability change factor further includes: After obtaining the joint-related motion parameters output by the nursing model, joint dynamic feature data, including joint angle change trajectory, joint rotation speed curve and joint force distribution map, are extracted, and noise suppression and time alignment processing are performed on the feature data. After completing the data preprocessing, a joint stability analysis model is constructed based on the joint dynamics characteristic data. By identifying discontinuities in the trajectory of joint angle changes and abnormal peaks in the rotational velocity curve, key locations that may lead to a decrease in stability are located. After locating the key positions, stress concentration is assessed on the joint stress distribution diagram of the corresponding time period. Based on the deviation between the joint stress peak and uniformity, the joint stability change value is generated. The joint stability change amplitude value is combined with the smoothness parameter of the joint angle change trajectory to form the joint stability change factor.

6. The personalized sports medicine nursing plan generation system according to claim 3, characterized in that: The method for obtaining the tissue temperature drift coefficient further includes: After obtaining the raw monitoring data related to tissue temperature output by the nursing model, the superficial tissue temperature curve and the deep tissue temperature curve are extracted respectively, and the two types of temperature curves are stabilized based on a sliding window. After stabilization, a temperature change gradient analysis model was constructed to compare the time difference sequences of the temperature curves of superficial and deep tissues, and to identify the synchronicity and lag characteristics of temperature changes in the two types of tissues. Based on the identified synchronicity and lag characteristics, the fluctuation amplitude of the temperature difference sequence between deep and shallow tissues is evaluated, and the tissue temperature drift amplitude value is generated according to the deviation between the maximum fluctuation amplitude and the average fluctuation amplitude of the temperature difference. The tissue temperature drift amplitude value is combined with the smoothness parameter of the synchronous change output by the temperature change gradient analysis model to form the tissue temperature drift coefficient.

7. The personalized sports medicine nursing plan generation system with multi-parameter synergy according to claim 1, characterized in that: The exercise load prediction model is invoked to perform training scenario simulations, resulting in a prediction result sequence set Q of the target individual's exercise care risk response, including: After obtaining the parameter sample sequence P and the influence factor parameter table M, input sample pairs for training the exercise load prediction model are constructed based on the one-to-one correspondence between the two. Each parameter sample Pi is combined with its corresponding influence factor parameter table Mi to form a training input unit. After constructing the training input unit, based on the motion load prediction model, feature enhancement processing is performed before scenario simulation. An extended feature set is generated by calculating the position coefficient of Pi in the interval and the gradient of the changes of each influencing factor in Mi. After feature enhancement is completed, the training input unit and the extended feature set are input into the exercise load prediction model. By executing multiple rounds of training scenario simulation, the changes in nursing risk response of the target individual under different assumed exercise load levels are simulated, and multiple stages of risk response output results are obtained. Each simulation output nursing risk response is numbered and organized according to its corresponding input sample Pi to form a prediction result sequence set Q.

8. The personalized sports medicine nursing plan generation system with multi-parameter collaborative method according to claim 1, characterized in that: Based on M, Q, and T, the sports medicine nursing model was re-invoked to reconstruct the treatment plan, including: After obtaining the influencing factor parameter table M, the nursing risk response sequence Q, and the intervention duration sequence T, the nursing plan reconstruction input unit is constructed based on the correspondence between the three on the parameter sample Pi. By jointly encoding the influencing factor set, the risk change curve, and the intervention duration structure, an input feature sequence for reconstruction is formed. After constructing the input feature sequence, the scheme generation structure of the sports medicine nursing model is called to perform intervention method screening calculation for each input feature sequence. By comparing the adaptability of different nursing methods under the current risk level, a set of candidate nursing measures is generated. After obtaining the set of candidate nursing measures, the candidate nursing measures are matched and ranked according to the intervention duration sequence T. The duration matching scoring algorithm is used to calculate the suitability score of each nursing measure, thereby determining the optimal nursing measure corresponding to each parameter sample Pi. The optimal nursing measures corresponding to each parameter sample Pi are combined in interval order to form the final personalized sports medicine nursing plan.

Images