Dynamic pressure regulation method and system under step frequency cooperation, and medium

Abstract

The application discloses a dynamic pressure regulation method and system under step frequency cooperation and a medium, relates to the technical field of rehabilitation medical equipment, and comprises the following steps: collecting historical user step frequency pressure data sets of an anti-gravity treadmill, establishing a benchmark library and a step frequency pressure regulation model library through correlation analysis; embedding a high-precision pressure sensor array in the treadmill pedal, collecting foot bottom pressure data of a target user and calculating a step frequency time curve; outputting a benchmark pressure regulation curve through model library matching; obtaining user stride data and a step frequency stride safety threshold, correcting the benchmark pressure regulation curve to obtain a dynamic pressure regulation curve, and then realizing accurate regulation of the treadmill pressure. The application solves the technical problems of poor adaptability and lack of safety constraints of traditional anti-gravity treadmill related regulation methods, achieves accurate regulation in line with individual characteristics and real-time motion states of users, and improves the technical effects of use adaptability and safety of the anti-gravity treadmill.

Description

Technical Field

[0001] This invention relates to the field of rehabilitation medical equipment technology, and in particular to a method, system and medium for dynamic pressure regulation under cadence coordination. Background Technology

[0002] In the biomedical engineering industry, antigravity treadmills are core equipment for rehabilitation training and health management, and their pressure regulation accuracy directly affects rehabilitation outcomes and exercise safety. Current technologies, particularly traditional antigravity treadmills, often employ fixed parameters or single-dimensional adjustment methods. While these methods have some application in basic scenarios, their limitations are becoming increasingly apparent as the industry demands more precise rehabilitation. Individual users in the biomedical engineering field exhibit significant differences in height, health status, and other factors, and their stride frequency and stride length change dynamically during exercise. Traditional methods cannot achieve personalized adaptation and lack stride frequency-stride length coordination constraints, leading to inaccurate pressure regulation, increasing the risk of rehabilitation complications, and failing to meet the industry's requirements for precise rehabilitation and safe training management. Summary of the Invention

[0003] This application provides a dynamic pressure regulation method, system and medium under cadence coordination, which solves the technical problems of poor adaptability and lack of safety constraints of traditional anti-gravity treadmill related adjustment methods.

[0004] The first aspect of this application provides a dynamic pressure regulation method under cadence coordination. The method includes: collecting historical user cadence-pressure datasets from a target anti-gravity treadmill; performing cadence-pressure correlation analysis on the historical user cadence-pressure datasets to establish a cadence-pressure benchmark library; and generating a cadence-pressure regulation model library based on the cadence-pressure benchmark library; embedding a high-precision pressure sensor array into the anti-gravity treadmill pedals; collecting plantar pressure distribution data of the target user through the high-precision pressure sensor array; performing real-time cadence calculation based on the plantar pressure distribution data to obtain a cadence-time curve; performing matching regulation analysis on the target user and the cadence-time curve based on the cadence-pressure regulation model library to output a benchmark pressure regulation curve; acquiring user stride length data and cadence-stride length safety thresholds; using the user stride length data and cadence-stride length safety thresholds to perform coordinated intervention correction on the benchmark pressure regulation curve to determine a dynamic pressure regulation curve; and regulating treadmill pressure based on the dynamic pressure regulation curve.

[0005] A second aspect of this application provides a dynamic pressure regulation system under cadence coordination. The system includes: a cadence-pressure regulation model library construction module, used to collect historical user cadence-pressure datasets from a target anti-gravity treadmill, perform cadence-pressure correlation analysis on the historical user cadence-pressure datasets, establish a cadence-pressure benchmark library, and generate a cadence-pressure regulation model library based on the benchmark library; a cadence-time curve acquisition module, used to embed a high-precision pressure sensor array into the anti-gravity treadmill pedals, collect plantar pressure distribution data of the target user through the high-precision pressure sensor array, perform real-time cadence calculation based on the plantar pressure distribution data, and obtain a cadence-time curve; a benchmark pressure regulation curve acquisition module, used to perform matching and regulation analysis on the target user and the cadence-time curve based on the cadence-pressure regulation model library, and output a benchmark pressure regulation curve; and a dynamic pressure regulation curve acquisition module, used to acquire user stride data and cadence-stride safety thresholds, use the user stride data and cadence-stride safety thresholds to perform collaborative intervention correction on the benchmark pressure regulation curve, determine the dynamic pressure regulation curve, and adjust the treadmill pressure based on the dynamic pressure regulation curve.

[0006] A third aspect of this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the dynamic pressure regulation method under step frequency coordination provided in this application.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application collects historical data from anti-gravity treadmills and establishes a benchmark and model library. It uses a sensor array to collect plantar pressure data of target users and calculates real-time motion data. A benchmark adjustment curve is generated through model matching. Combined with user stride data and safety thresholds, a collaborative correction is made to achieve dynamic adaptation and adjustment of treadmill pressure. This improves the accuracy of adjustment and the safety of use, achieving the technical effect of precise adjustment that conforms to individual user characteristics and real-time motion status, thus improving the usability and safety of anti-gravity treadmills. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0009] Figure 1 This is a flowchart illustrating the dynamic pressure regulation method under step frequency coordination provided in the embodiments of this application.

[0010] Figure 2 This is a schematic diagram of the dynamic pressure regulation system under step frequency coordination provided in the embodiments of this application.

[0011] Figure labeling: Step frequency pressure regulation model library construction module 1, step frequency time curve acquisition module 2, reference pressure regulation curve acquisition module 3, dynamic pressure regulation curve acquisition module 4. Detailed Implementation

[0012] This application provides a dynamic pressure regulation method, system and medium under cadence coordination, which solves the technical problems of poor adaptability and lack of safety constraints of traditional anti-gravity treadmill related adjustment methods.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0015] Example 1, as Figure 1 As shown, a dynamic pressure regulation method under step frequency coordination is described, wherein the method includes: Historical user cadence pressure datasets are collected from the target anti-gravity treadmill. Cadence pressure correlation analysis is performed on the historical user cadence pressure datasets to establish a cadence pressure benchmark library. Based on the cadence pressure benchmark library, a cadence pressure adjustment model library is generated.

[0016] In this embodiment of the application, the anti-gravity treadmill is a special device that reduces the body weight load through anti-gravity technology and is used in the fields of fitness or rehabilitation to carry out running and gait training, thereby reducing joint pressure.

[0017] Specifically, firstly, historical users of the target anti-gravity treadmill are screened to ensure that users cover different height and weight ranges, age distributions, genders, health conditions, and exercise goals, thus guaranteeing the diversity and representativeness of the dataset. Resistive pressure sensors are embedded in the pedal area of ​​the anti-gravity treadmill, and a timing module is integrated into the treadmill control system.

[0018] When a user exercises on the target anti-gravity treadmill, a resistive pressure sensor captures the pressure signal between the foot and the pedal in real time, collecting pressure data every 10 milliseconds and transmitting it synchronously to the treadmill's control terminal. The timing module continuously records the exercise time, and the control terminal stores the pressure data and time data in a one-to-one correspondence through data synchronization software. The cadence is calculated using a peak counting method, which involves analyzing the pressure data curve to identify the pressure peak generated each time the foot lands, and counting the number of pressure peaks per minute to obtain the user's real-time cadence data. Finally, by synchronously storing the pressure data and corresponding cadence data, the user's complete cadence-pressure related raw data is obtained and integrated to form a historical user cadence-pressure dataset.

[0019] Next, a user feature dimension set including height and weight range, age distribution, gender, health status, and exercise goals is obtained. Based on this feature dimension set, the historical user cadence stress dataset is classified, labeled, and summarized. Subsequently, outlier removal and cadence interval division are performed to complete data preprocessing. Finally, based on the preprocessed multi-interval feature dimension parameters—cadence stress dataset—cadence stress correlation analysis is conducted to construct a cadence stress benchmark library. This step will be explained in detail later.

[0020] Next, based on the characteristic information of the multi-interval feature dimension parameters of the optimized cadence pressure dataset in the cadence pressure benchmark library, a corresponding multi-interval feature dimension model architecture set is selected. Then, cadence pressure fitting training is performed on the optimized dataset using this model architecture set to construct a multi-interval feature dimension cadence pressure regulation model set. Finally, this model set is validated, optimized, and integrated with labels to generate a cadence pressure regulation model library. This step will be explained in detail later.

[0021] A high-precision pressure sensor array is embedded in the anti-gravity running platform pedal. The high-precision pressure sensor array collects the plantar pressure distribution data of the target user, and the real-time cadence is calculated based on the plantar pressure distribution data to obtain the cadence-time curve.

[0022] Optionally, a high-precision resistive pressure sensor is selected first. The sensor's pressure measurement range is set to 0-500 kPa, and the measurement accuracy is controlled within ±1 kPa, meeting the precise requirements for plantar pressure detection. A sensor array is laid out on the surface of the anti-gravity treadmill pedal, with sensors evenly distributed in 8 columns horizontally and 12 rows vertically, with a sensor spacing of 3 cm, ensuring comprehensive coverage of the foot contact area for different users. The sensors are fixed to the circuit board inside the pedal by soldering. The circuit board is connected to the treadmill's main control terminal via a wired serial port, ensuring stable data transmission and avoiding delays or interference issues associated with wireless transmission.

[0023] After the anti-gravity treadmill is activated, the main control terminal sends a start command to the pressure sensor array. The sensor array collects plantar pressure data from the target user in real time at a sampling frequency of 100Hz, recording the pressure values ​​of all sensors every 10 milliseconds to form a raw plantar pressure distribution data matrix. During the acquisition process, the main control terminal synchronously records the timestamp corresponding to each data point, with timestamp accuracy down to the millisecond level, ensuring a strict correspondence between pressure data and time information. The raw data is transmitted to the main control terminal via serial port in byte stream format. The terminal uses a circular storage method to temporarily store the pressure data for the most recent 5 minutes to avoid data overflow.

[0024] The collected raw plantar pressure distribution data was then preprocessed. A moving average method was used to remove environmental noise interference. Five consecutive sampling points from each sensor were used as a window, and the arithmetic mean of the pressure values ​​within the window was calculated. This average value was then used to replace the original pressure value at the center of the window, completing the data smoothing process. Subsequently, the total pressure from all sensors at each time point was calculated. If the total pressure was greater than 30% of the sensor's full scale, the plantar surface was considered to be in a ground-based state; if the total pressure was less than 10% of the sensor's full scale, the plantar surface was considered to be in a lift-off state. This threshold judgment was used to identify the plantar contact state.

[0025] Finally, based on the foot contact state recognition results, the real-time cadence is calculated. Timing begins from the first landing state, and the number of landing states occurring per unit time is counted. The cadence calculation is completed every 60 seconds, yielding the cadence value per minute. A cadence-time curve generation rule is constructed with time as the horizontal axis (unit: seconds) and cadence as the vertical axis (unit: steps / minute). The curve data points are updated every second; that is, every second, the current cumulative number of landings is extracted, converted into the corresponding cadence value, and recorded in the curve data. The `plot` function from Python's `matplotlib` library is used to connect all data points in chronological order, generating a continuous cadence-time curve. The curve data is stored in real-time in the local database of the main control terminal for subsequent use.

[0026] By installing a sensor array, acquiring and preprocessing pressure data in real time, identifying contact status, and calculating cadence, a cadence-time curve that accurately reflects the cadence changes during the target user's movement is generated, providing real-time and reliable motion state data support for the subsequent output of the benchmark pressure adjustment curve.

[0027] Based on the step frequency pressure adjustment model library, a matching adjustment analysis is performed on the target user and the step frequency time curve to output a benchmark pressure adjustment curve.

[0028] In one embodiment of this application, the basic information of the target user is first classified according to the aforementioned user feature dimension set to obtain target user feature dimension parameters. Next, based on these target user feature dimension parameters and the cadence interval corresponding to the cadence-time curve, feature matching is performed with a cadence-pressure adjustment model library to determine the corresponding target cadence-pressure adjustment model. Finally, this target cadence-pressure adjustment model is used to perform pressure adjustment analysis on the cadence-time curve, thereby outputting a baseline pressure adjustment curve. This step will be described in detail later.

[0029] The system acquires user stride length data and stride frequency safety thresholds, uses these data to collaboratively intervene and correct the baseline pressure adjustment curve, determines the dynamic pressure adjustment curve, and adjusts the treadmill pressure based on the dynamic pressure adjustment curve.

[0030] Specifically, the stride length calculation formula is first constructed as: Stride length = Pressure center trajectory length × Step frequency correction coefficient, where the step frequency correction coefficient is dynamically adjusted based on the user's height and weight. Next, the user's pressure center trajectory length is obtained using an anti-gravity treadmill. This stride length calculation formula is then applied, combined with the step frequency-time curve, to estimate the user's stride length data. Finally, based on the target user's characteristic parameters, safe thresholds for step frequency and stride length are set, thereby determining the safe thresholds for step frequency and stride length. This step will be explained in detail later.

[0031] Next, it is determined whether the ratio of the cadence-time curve to the user's stride data is within the cadence-stride safety threshold. If it is within the threshold, the baseline pressure adjustment curve is directly determined as the dynamic pressure adjustment curve. If it exceeds the threshold, the motion imbalance deviation between the ratio and the cadence-stride safety threshold is first calculated, and then the correction mechanism is triggered. Based on the motion imbalance deviation, the baseline pressure adjustment curve is collaboratively intervened and corrected, thereby determining the dynamic pressure adjustment curve and performing treadmill pressure adjustment. This step will also be explained in detail later.

[0032] Furthermore, the method provided in this application embodiment includes: Obtain a user feature dimension set, which includes height and weight range, age distribution, gender, health status, and exercise goals; classify and summarize the historical user cadence stress dataset according to the user feature dimension set to obtain a user feature dimension parameter-cadence stress dataset; remove outliers and divide cadence intervals in the user feature dimension parameter-cadence stress dataset to obtain a multi-interval feature dimension parameter-cadence stress dataset; perform cadence stress correlation analysis based on the multi-interval feature dimension parameter-cadence stress dataset to establish a cadence stress benchmark library.

[0033] Specifically, during user registration or first-time use of the anti-gravity treadmill, an authorization prompt interface pops up through the device control system. This clearly informs the user that the collection of information such as height and weight range, age distribution, gender, health status, and exercise goals is solely for optimizing pressure regulation. Users must click the confirmation button to complete the authorization before subsequent data collection can begin. Data from unauthorized users will not be included in the historical dataset. After authorization, relevant information is collected through the device's input interface. The height and weight range are entered directly by the user. Age distribution options include preset age groups such as 18-25, 26-35, 36-45, and 46+. Gender options include male and female. Health status options include simple categories such as healthy, minor joint discomfort, post-operative rehabilitation, and others. Exercise goals include common types such as fat loss, endurance enhancement, muscle training, and rehabilitation training. All collected information is stored and linked to the user's account.

[0034] Next, the historical user cadence and stress dataset is processed. First, a related table is created in the MySQL database, containing user characteristic dimension columns such as height and weight range, age distribution, gender, health status, and exercise goals, as well as data columns such as cadence and stress. The authorized characteristic information of each historical user is entered into the table one-to-one with the corresponding cadence and stress data. Then, the data is hierarchically classified in the order of height and weight range, age distribution, gender, health status, and exercise goals. For example, the height and weight ranges are first divided into interval groups by 10 cm of height and 5 kg of weight, and then further subdivided into preset age groups within each interval group. This process is repeated to complete the multi-dimensional classification. Finally, the cadence and stress data under each category are labeled with the corresponding characteristic dimension parameters, and the data are summarized to form the user characteristic dimension parameter - cadence and stress dataset.

[0035] Next, the 3σ criterion was used to remove outliers. Specifically, first, the mean and standard deviation of all pressure and cadence data in the user feature dimension parameter - cadence-pressure dataset were calculated using Excel software. Then, the pressure and cadence values ​​in each data point were compared one by one, and values ​​exceeding the mean plus or minus three times the standard deviation were identified as outliers and removed from the dataset. After outlier removal, the cadence interval method was used to divide the data into intervals: first, the maximum and minimum cadence values ​​in the remaining data were counted, and the difference between the two was calculated. Then, the data was divided into intervals of 10 steps / minute. For example, if the minimum cadence was 80 steps / minute and the maximum was 140 steps / minute, intervals of 80-90 steps / minute, 91-100 steps / minute, and so on, up to 131-140 steps / minute, were created. Then, the cadence value of each data point was assigned to the corresponding interval, resulting in a multi-interval feature dimension parameter - cadence-pressure dataset.

[0036] Finally, based on the exercise objectives of the anti-gravity treadmill, a corresponding set of exercise objective effect evaluation indicators is set. Using this evaluation indicator set, the multi-interval feature dimension parameter-step frequency pressure dataset is evaluated and filtered to obtain the multi-interval feature dimension parameter-optimized step frequency pressure dataset. Finally, according to the user feature dimension parameters and step frequency intervals, this optimized dataset is divided, associated, and integrated to establish a step frequency pressure benchmark library. This step will be explained in detail later.

[0037] Through the above series of specific and operable steps, the process of establishing the cadence-pressure benchmark library is fully disclosed, from user information authorization collection, data classification and summarization, preprocessing, to effect evaluation screening, cadence-pressure correlation analysis and data integration.

[0038] Furthermore, the method provided in this application embodiment includes: Based on the anti-gravity treadmill exercise goals, a set of exercise goal effect evaluation indicators is set; the multi-interval feature dimension parameter-step frequency pressure dataset is evaluated and filtered according to the exercise goal effect evaluation indicator set to obtain the multi-interval feature dimension parameter-preferred step frequency pressure dataset; the multi-interval feature dimension parameter-preferred step frequency pressure dataset is divided, associated and integrated according to user feature dimension parameters and step frequency intervals to establish a step frequency pressure benchmark library.

[0039] Optionally, a set of corresponding effect evaluation indicators can be set according to different exercise goals of the anti-gravity treadmill. If the exercise goal is fat loss, the evaluation indicators include the rate of achieving the target exercise time and the average pressure adaptation degree; if the goal is to improve endurance, the evaluation indicators include cadence stability and pressure duration adaptation time; if the goal is rehabilitation training, the evaluation indicators include the accuracy of pressure peak control and the amplitude of cadence fluctuation. A 100-point scoring method is used to evaluate the effect of each data point in the multi-interval feature dimension parameter - cadence pressure dataset. Each evaluation indicator has a maximum score of 20 points. Scores are given based on the degree to which the data meets the exercise goal. Data with a total score higher than 80 points is considered to have met the target. All data that meet the target are aggregated to obtain the multi-interval feature dimension parameter - optimized cadence pressure dataset.

[0040] Next, Pearson correlation analysis was used to conduct a cadence-pressure correlation analysis. Cadence data for each multi-interval feature dimension was used as the x-axis, and pressure data as the y-axis. The correlation coefficient between cadence and pressure was calculated using the `corrcoef` function in the scipy library via Python programming, clarifying the correlation between cadence and pressure. Subsequently, the multi-interval feature dimension parameter-optimized cadence-pressure dataset was further divided according to user feature dimension parameters and cadence intervals. The correlation analysis results for the same feature dimension parameter and the same cadence interval were integrated into a single data set. All data sets were then aggregated to form a complete cadence-pressure benchmark library.

[0041] Furthermore, the method provided in this application embodiment includes: Based on the characteristic information of the multi-interval feature dimension parameter-preferred cadence pressure dataset in the cadence pressure benchmark library, a multi-interval feature dimension model architecture set is selected; cadence pressure fitting training is performed on the multi-interval feature dimension parameter-preferred cadence pressure dataset according to the multi-interval feature dimension model architecture set to construct a multi-interval feature dimension cadence pressure adjustment model set; the multi-interval feature dimension cadence pressure adjustment model set is verified, optimized, and integrated to generate a cadence pressure adjustment model library.

[0042] Specifically, firstly, the characteristic information of the optimal cadence-pressure dataset—a multi-interval feature dimension parameter—is extracted from the cadence-pressure benchmark database. The correlation coefficient between cadence and pressure in each data set is calculated to determine the type of association. If the absolute value of the correlation coefficient is greater than 0.8, it is considered a linear association; if it is less than 0.8, it is considered a non-linear association. Simultaneously, the sample size of each data set is calculated: groups with more than 1000 samples are considered large data sets, and groups with less than 1000 samples are considered small data sets. Based on these characteristics, corresponding model architectures are selected: linear regression models are used for linear associations with small data sets, ridge regression models are used for linear associations with large data sets, and support vector machine models are used for non-linear associations. This forms a multi-interval feature dimension model architecture set, and all model architectures are publicly available algorithm models.

[0043] Next, the multi-interval feature dimension parameter-optimized step frequency and stress dataset was grouped according to the corresponding model architecture. Each group was randomly divided into a training set and a validation set in a 7:3 ratio. The partitioning process used the `random` function from Python's numpy library for random sampling. The corresponding model architecture's algorithm interface was then called using Python's scikit-learn library, with step frequency data as the input feature (steps / minute) and stress data as the output label (kPa), to perform fitting training on each training set. During training, the linear regression model uses the least squares method to find the optimal coefficients. Parameters are initialized with random small values ​​in the range [-0.01, 0.01]. The iteration termination condition is set to either reaching 1000 iterations or the mean squared error being less than 0.001; either condition is met. The ridge regression model adds an L2 regularization term to avoid overfitting. The regularization parameter α is initially set to 0.5. Five-fold cross-validation is used, taking values ​​in the range α ∈ [0.1, 1.0] at 0.1 intervals. The α value with the smallest mean squared error in the cross-validation set is selected. The iteration termination condition is the same as for the linear regression model. The support vector machine model uses a radial basis function kernel with a kernel parameter γ set to 0.1. The loss function is ε-support vector regression with an ε value set to 0.01. The model complexity is controlled by adjusting the penalty coefficient C. The iteration termination condition is either an error accuracy reaching 1e-3 or the number of iterations reaching 2000. During training, the model's fitting performance on the validation set is monitored in real time to ensure that the model can capture the correlation between step frequency and stress. After training all groups, a multi-interval feature dimension step frequency and stress regulation model set is constructed.

[0044] Finally, mean squared error (MSE) was used as the model validation metric to calculate the MSE between the predicted and actual stress values ​​for each model on the corresponding validation set. If the MSE exceeded a preset threshold of 0.05, the model was tuned. For the linear regression model, the data preprocessing steps were optimized by adding a secondary outlier filter, i.e., the 3σ criterion was used again to screen the training data. For the ridge regression model, the regularization parameter α was adjusted based on 5-fold cross-validation, testing values ​​from 0.1 to 1.0 at 0.1 intervals to select the parameter with the smallest MSE. For the support vector machine model, a grid search method combined with 5-fold cross-validation was used to traverse combinations of C from 1 to 10 and γ from 0.01 to 0.1 to determine the optimal parameter combination. After all models were validated and tuned to meet the standards, each model was labeled with corresponding user feature dimension parameters and step frequency range information, for example: height 170-180cm + weight 60-70kg + step frequency 90-100 steps / minute + ridge regression model. All labeled models were then integrated and stored in a MySQL database to generate a complete step frequency stress regulation model library.

[0045] Through the steps of feature matching to select models, group fitting training, validation and optimization, and label integration, a cadence pressure adjustment model library adapted to different user characteristics and cadence ranges was generated, providing accurate and reliable model support for the subsequent output of benchmark pressure adjustment curves for target users.

[0046] Furthermore, the method provided in this application embodiment includes: The basic information of the target user is classified according to the user feature dimension set to obtain the target user feature dimension parameters; based on the target user feature dimension parameters and the cadence interval of the cadence-time curve, feature matching is performed with the cadence-pressure adjustment model library to determine the target cadence-pressure adjustment model; the target cadence-pressure adjustment model is used to perform pressure adjustment analysis on the cadence-time curve to output the baseline pressure adjustment curve.

[0047] Specifically, before using the anti-gravity treadmill, target users fill in basic information through the treadmill's accompanying touch input interface. The information entered corresponds exactly to the user characteristic dimension set in the previous steps, including specific values ​​for height and weight, selection of the preset age range, gender selection, confirmation of health status options, and selection of exercise goals. The target user's basic information is categorized according to the classification standards of historical user data. Height is divided into intervals of 10 centimeters, weight into intervals of 5 kilograms, and age uses preset ranges of 18-25 years, 26-35 years, 36-45 years, and 46 years and above. Health status and exercise goals are directly matched with corresponding classification tags, and finally integrated to form the target user characteristic dimension parameters, such as height 170-180cm, weight 60-70kg, 26-35 years, male, healthy, and fat loss.

[0048] Next, cadence data is extracted from the generated cadence-time curve, and the cadence distribution of the curve during the target user's exercise period is statistically analyzed to determine the intervals where the main cadence frequencies are located. The cadence interval division standard remains consistent with that used when the dataset was created, i.e., each interval is 10 steps / minute. For example, if more than 90% of the cadence values ​​in the cadence-time curve are concentrated between 90-100 steps / minute, then the cadence interval corresponding to the cadence-time curve is determined to be 90-100 steps / minute. A MySQL database storing the cadence stress regulation model library is connected via Python programming. Feature matching is performed using a method of exact matching combined with priority ranking. Priority is given to matching models whose feature dimension parameters of the target user are completely consistent with the feature dimension parameters of the model annotation, followed by models with the same cadence interval. If no completely matching model is found, the model with the most overlapping feature dimension parameters and adjacent cadence intervals is selected as a candidate. The candidate model with the highest feature overlap score (1 point for each matching feature, maximum 5 points) is selected as the target cadence stress regulation model.

[0049] Finally, the established target cadence-pressure regulation model is invoked, and the cadence values ​​corresponding to each time point in the cadence-time curve are input into the model sequentially. Based on the cadence-pressure correlation learned during training, the model calculates pressure for each input cadence value and outputs the corresponding predicted pressure value. Each cadence value and predicted pressure value have a one-to-one correspondence, forming a three-dimensional data set of time-cadence-pressure. With time as the horizontal axis (unit: seconds) and predicted pressure values ​​as the vertical axis (unit: kPa), the `plot` function from Python's matplotlib library is used to connect the coordinate points corresponding to all predicted pressure values ​​in chronological order, generating a continuous and smooth curve. This curve is the baseline pressure regulation curve, and the curve data is synchronously stored in the local database of the treadmill's main control terminal for subsequent intervention and correction.

[0050] By classifying the basic information of the target users, matching features to determine the target model, and analyzing the step frequency-time curve of the model, a baseline pressure adjustment curve adapted to the characteristics of the target users and the real-time step frequency was output, providing an accurate basis for determining the subsequent dynamic pressure adjustment curve.

[0051] Furthermore, the method provided in this application embodiment includes: A stride length calculation formula is constructed, specifically: stride length = length of pressure center trajectory × cadence correction coefficient, wherein the cadence correction coefficient is dynamically adjusted according to the user's height and weight; the length of the user's pressure center trajectory is obtained through the anti-gravity treadmill, and the stride length is estimated by using the stride length calculation formula and the cadence-time curve to obtain the user's stride length data; based on the target user's characteristic dimension parameters, cadence and stride length safety ratio thresholds are set to determine the cadence and stride length safety thresholds.

[0052] Specifically, a stride length calculation formula is first constructed, defined as stride length equal to the length of the pressure center trajectory multiplied by a cadence correction coefficient. The cadence correction coefficient is dynamically adjusted based on the user's height and weight. Specifically, it is determined by collecting actual stride length data and corresponding pressure center trajectory lengths from a sufficient number of users of different heights and weights, performing linear fitting analysis using Excel software, and obtaining the correlation formula between the correction coefficient and height and weight: cadence correction coefficient = a × height (cm) + b × weight (kg) + c, where a, b, and c are positive constants determined through linear fitting. This formula is derived through statistical fitting of existing data. Those skilled in the art can independently determine the specific values ​​of a, b, and c by collecting sample data from similar users and using the same linear fitting method.

[0053] Next, using a high-precision pressure sensor array installed on the anti-gravity treadmill pedals, real-time data on the plantar pressure distribution during the target user's movement is collected, recording the pressure value and coordinate position of each sensor every 10 milliseconds. A weighted average method is used to calculate the pressure center coordinates at each timestamp. With the sensor's horizontal coordinate as the x-axis and the vertical coordinate as the y-axis, the pressure center x-coordinate = (sum of each sensor's x-coordinate × corresponding pressure value) ÷ sum of all sensor pressure values. The pressure center y-coordinate is calculated using the same method. Within a complete gait cycle, from the first foot strike to the next, all pressure center coordinates are connected chronologically to form a pressure center movement trajectory. The straight-line distance between adjacent coordinate points on the trajectory is calculated using the `distance` function from Python's scipy library, and all distances are summed to obtain the pressure center trajectory length. The cadence value corresponding to the current gait cycle is extracted from the cadence-time curve, substituted into the cadence correction coefficient formula to calculate the corresponding cadence correction coefficient, and then the pressure center trajectory length is multiplied by the cadence correction coefficient to obtain the stride data for that gait cycle. The stride length for each gait cycle is continuously calculated to form a real-time stride data sequence for the user.

[0054] Then, based on the target user's characteristic dimension parameters, set the safety ratio thresholds for cadence and stride: First, clarify the definition of the cadence and stride safety ratio: Cadence and stride safety ratio = real-time cadence (unit: steps / minute) ÷ real-time stride (unit: meters). This ratio reflects the adaptation and coordination between the user's cadence and stride during exercise. Its safety threshold is the ratio range that ensures the user's exercise is injury-free and the exercise efficiency is optimal.

[0055] The process follows a step-by-step logic of referencing general standards → statistically analyzing historical data → matching feature combinations: First, referencing existing general standards in sports medicine, the basic safe range for the cadence-to-stride ratio of healthy adults is determined to be 0.6-1.4. This range originates from publicly available gait safety research conclusions in the field of sports medicine. Second, at least 5000 sets of safe exercise data from historical users are collected, with screening criteria including "no injury feedback during exercise, no discomfort symptoms after exercise, and exercise goal achievement rate ≥80%". Data from abnormal exercise states such as sudden acceleration or deceleration is excluded, retaining only the cadence, stride length, and corresponding feature dimension parameters during uniform motion. Third, the screened historical data is grouped into five-dimensional feature combinations based on: height per 10 cm + weight per 5 kg + age distribution + health status + exercise goal. The mean and standard deviation of the cadence and stride ratio are calculated for each group. The mean ± 1.5 times the standard deviation is used as the initial threshold range for that group. The initial range is then calibrated against sports medicine standards to ensure it does not exceed the basic range of 0.6-1.4, ultimately forming the specific safety threshold for each feature combination.

[0056] For example, for users aged 18-35 with a healthy health condition and an exercise goal of improving endurance, the safe threshold for cadence and stride length ratio is set at 0.8-1.2; for users aged 36-45 with a mild health condition and an exercise goal of fat loss, the threshold is set at 0.7-1.3; and for users over 46 years old or whose health condition requires rehabilitation and whose exercise goal is rehabilitation training, the threshold is set at 0.6-1.4. All threshold ranges are determined by statistically analyzing historical safe exercise data. Technicians can directly match the corresponding thresholds based on the characteristic dimension parameters of the target user to ultimately determine the safe thresholds for cadence and stride length.

[0057] By constructing a correlation formula for step frequency correction coefficient, calculating the length of the pressure center trajectory and estimating stride length using the weighted average method, and matching safety thresholds with user characteristic parameters, the system achieves accurate acquisition of user stride length data and reasonable setting of step frequency and stride length safety thresholds, providing a key safety judgment basis for subsequent correction of dynamic pressure adjustment curves.

[0058] Furthermore, the method provided in this application embodiment includes: If the ratio of the cadence-time curve to the user's stride data is within the cadence-stride safety threshold, then the baseline pressure adjustment curve is determined as the dynamic pressure adjustment curve; if the ratio of the cadence-time curve to the user's stride data exceeds the cadence-stride safety threshold, the deviation value of the movement imbalance from the cadence-stride safety threshold is calculated; a correction mechanism is triggered to perform collaborative intervention correction on the baseline pressure adjustment curve based on the movement imbalance deviation value, thereby determining the dynamic pressure adjustment curve.

[0059] In one embodiment, firstly, cadence data for the most recent 10 seconds from the cadence-time curve is extracted, and the average cadence for this time period is calculated. This average cadence is calculated by dividing the sum of all cadence data points within the 10 seconds by the number of data points, resulting in the real-time average cadence, with the unit being steps per minute. Simultaneously, user stride data for the same 10-second time period is extracted, and the average stride is calculated similarly. This average stride is calculated by dividing the sum of stride data for all gait cycles within the same time period by the number of gait cycles, resulting in the real-time average stride, with the unit being meters. Timestamp alignment ensures consistency in the temporal dimension between cadence and stride data.

[0060] Then, calculate the real-time ratio of cadence to stride length. The formula is: real-time ratio equals the average real-time cadence divided by the average real-time stride length. Obtain the previously matched cadence and stride length safety threshold range for the target user, and compare the calculated real-time ratio with this threshold range one by one. If the real-time ratio is greater than or equal to the lower threshold and less than or equal to the upper threshold, it is determined that the current step frequency and stride length are in a safe matching state. The main control terminal directly calls the stored baseline pressure adjustment curve data, marks it as a dynamic pressure adjustment curve, and updates it synchronously to the treadmill pressure control module.

[0061] If the real-time ratio exceeds the threshold range, the motion imbalance deviation value is calculated according to different cases. When the real-time ratio is higher than the upper threshold, the motion imbalance deviation value is equal to the real-time ratio minus the upper threshold, and then divided by the upper threshold to obtain the relative deviation ratio. When the real-time ratio is lower than the lower threshold, the motion imbalance deviation value is equal to the lower threshold minus the real-time ratio, and then divided by the lower threshold to obtain the relative deviation ratio. Two decimal places are retained during the calculation to ensure that the accuracy of the deviation value meets the subsequent correction requirements. All calculation results are stored in real time in the local database of the main control terminal, and the deviation type is marked as either higher than the upper limit or lower than the lower limit.

[0062] The above process is repeated every 10 seconds to continuously monitor the matching status of the cadence-to-stride ratio and the safety threshold, ensuring dynamic tracking of the user's movement status.

[0063] Finally, a personalized imbalance deviation-pressure adjustment correction model was established for the target user. The correction mechanism then invoked this personalized model to calculate pressure adjustment based on the calculated motion imbalance deviation value, yielding a dynamic pressure adjustment coefficient. Based on this dynamic pressure adjustment coefficient, a collaborative intervention correction was implemented on the baseline pressure adjustment curve, thereby determining the dynamic pressure adjustment curve and performing treadmill pressure adjustment. This step will be explained in detail later.

[0064] By synchronizing data with timestamps, calculating the mean using the arithmetic mean method, determining the ratio and deviation value using basic operations, and comparing the status with threshold intervals, the system achieves accurate judgment of stride frequency and stride length adaptability and quantifies the degree of movement imbalance. This provides real-time and reliable judgment basis and data support for the subsequent collaborative intervention and correction of the baseline pressure adjustment curve.

[0065] Furthermore, the method provided in this application embodiment includes: The test establishes a personalized imbalance deviation-pressure adjustment correction model for the target user; the correction mechanism is triggered to call the personalized imbalance deviation-pressure adjustment correction model to calculate the pressure adjustment of the motion imbalance deviation value and obtain the dynamic pressure adjustment coefficient; based on the dynamic pressure adjustment coefficient, the benchmark pressure adjustment curve is collaboratively intervened and corrected to determine the dynamic pressure adjustment curve.

[0066] Optionally, when target users use the anti-gravity treadmill for the first time, they first undergo a 5-minute personalized test exercise. The treadmill provides three different frequency ranges (low, medium, and high) and three different amplitude ranges of combined exercise modes according to a preset program, covering common exercise states. During the test, the treadmill continuously collects foot pressure distribution data, real-time cadence and stride length data, and exercise comfort scores provided by the user through the treadmill's touch interface, specifically on a scale of 1 to 5, where 1 is extremely uncomfortable and 5 is extremely comfortable. For each exercise mode, a slight cadence and stride length imbalance is introduced, such as intentionally increasing cadence or stride length. The exercise imbalance deviation value, user comfort score, and corresponding optimal pressure correction amount are recorded for different levels of imbalance. The optimal pressure correction amount is determined by the pressure adjustment value when the user reports a comfort score of 4-5. Finally, at least 30 sets of matching data of exercise imbalance deviation value and optimal pressure correction amount are collected.

[0067] Next, a personalized imbalance deviation-pressure regulation correction model was established using the linear regression algorithm from Python's scikit-learn library. The imbalance deviation was used as the input feature, and the optimal pressure correction was used as the output label. The model was then trained on the matching data collected from the test. During training, the training and validation sets were divided in a 7:3 ratio. The model parameters were solved using the least squares method, and the iteration termination condition was set to reach 500 iterations or a mean squared error (MSE) of less than 0.005. After training, the model accuracy was verified using the validation set data. If the MSE was greater than 0.01, five additional test sets were added for retraining until the model met the accuracy requirements. This model was stored on the treadmill's main control terminal and exclusively matched to the target user.

[0068] Upon detecting a deviation in motion imbalance, a correction mechanism is triggered. This mechanism uses Python programming to call a stored personalized imbalance deviation-pressure regulation correction model. The motion imbalance deviation value is input into this model, and the model calculates and outputs a dynamic pressure regulation coefficient in real time based on the mapping relationship learned during training. If the deviation value is positive, meaning the real-time ratio is higher than the upper threshold, the regulation coefficient is set to a value greater than 1, and the larger the deviation value, the larger the coefficient. For example, a deviation value of 0.1 corresponds to a coefficient of 1.1, a deviation value of 0.2 corresponds to a coefficient of 1.2, and so on. If the deviation value is negative, meaning the real-time ratio is lower than the lower threshold, the regulation coefficient is set to a value less than 1, and the larger the absolute value of the deviation value, the smaller the coefficient. For example, a deviation value of -0.1 corresponds to a coefficient of 0.9, a deviation value of -0.2 corresponds to a coefficient of 0.8, and so on. The coefficient range is controlled between 0.6 and 1.4 to avoid over-adjustment of pressure.

[0069] Finally, the baseline pressure adjustment curve is corrected point by point based on the dynamic pressure adjustment coefficient. The correction formula is: Corrected pressure value = Baseline pressure value × Dynamic pressure adjustment coefficient. After correcting the pressure value at each time point in the baseline pressure adjustment curve in chronological order, a new dynamic pressure adjustment curve is formed. The treadmill main control terminal converts the dynamic pressure adjustment curve into a pressure control signal, which is transmitted to the pressure execution module of the treadmill pedal via wired connection. The execution module adjusts the support pressure of each area of ​​the pedal at a response frequency of 10 milliseconds to ensure that the pressure adjustment is synchronized with the user's exercise status in real time.

[0070] By using personalized test modeling, model call to calculate adjustment coefficients, point-by-point correction curves, and real-time pressure execution, precise pressure adjustment based on the user's movement imbalance state is achieved, ensuring that the treadmill pressure always matches the user's stride frequency and stride coordination state, thereby improving exercise safety and comfort.

[0071] In summary, the dynamic pressure regulation method under step frequency coordination provided in this application has the following technical effects: This application establishes a model library by collecting historical user cadence and pressure data, acquires target user plantar pressure data using a sensor array and calculates the cadence curve, outputs a baseline pressure adjustment curve through model library matching, and determines the dynamic curve to adjust the treadmill pressure by combining stride data and safety threshold correction. This achieves the technical effect of achieving precise adjustment that conforms to the individual characteristics and real-time movement status of the user, and improving the usability and safety of the anti-gravity treadmill.

[0072] Example 2, as Figure 2 As shown, based on the same inventive concept as in Embodiment 1 above, this application provides a dynamic pressure regulation system under step frequency coordination, the system comprising: Step frequency pressure adjustment model library construction module 1 is used to collect historical user step frequency pressure datasets of the target anti-gravity treadmill, perform step frequency pressure correlation analysis on the historical user step frequency pressure datasets, establish a step frequency pressure benchmark library, and generate a step frequency pressure adjustment model library based on the step frequency pressure benchmark library.

[0073] The cadence-time curve acquisition module 2 is used to embed a high-precision pressure sensor array into the anti-gravity treadmill pedal, collect the plantar pressure distribution data of the target user through the high-precision pressure sensor array, and perform real-time cadence calculation based on the plantar pressure distribution data to obtain the cadence-time curve.

[0074] The reference pressure adjustment curve acquisition module 3 performs matching adjustment analysis on the target user and the step frequency time curve based on the step frequency pressure adjustment model library, and outputs the reference pressure adjustment curve.

[0075] The dynamic pressure adjustment curve acquisition module 4 is used to acquire user stride data and stride frequency safety thresholds, use the user stride data and stride frequency safety thresholds to perform collaborative intervention correction on the benchmark pressure adjustment curve, determine the dynamic pressure adjustment curve, and adjust the treadmill pressure based on the dynamic pressure adjustment curve.

[0076] Furthermore, the step frequency pressure regulation model library construction module 1 is used to perform the following steps: Obtain a user feature dimension set, which includes height and weight range, age distribution, gender, health status, and exercise goals; classify and summarize the historical user cadence stress dataset according to the user feature dimension set to obtain a user feature dimension parameter-cadence stress dataset; remove outliers and divide cadence intervals in the user feature dimension parameter-cadence stress dataset to obtain a multi-interval feature dimension parameter-cadence stress dataset; perform cadence stress correlation analysis based on the multi-interval feature dimension parameter-cadence stress dataset to establish a cadence stress benchmark library.

[0077] Furthermore, the step frequency pressure regulation model library construction module 1 is used to perform the following steps: Based on the anti-gravity treadmill exercise goals, a set of exercise goal effect evaluation indicators is set; the multi-interval feature dimension parameter-step frequency pressure dataset is evaluated and filtered according to the exercise goal effect evaluation indicator set to obtain the multi-interval feature dimension parameter-preferred step frequency pressure dataset; the multi-interval feature dimension parameter-preferred step frequency pressure dataset is divided, associated and integrated according to user feature dimension parameters and step frequency intervals to establish a step frequency pressure benchmark library.

[0078] Furthermore, the step frequency pressure regulation model library construction module 1 is used to perform the following steps: Based on the characteristic information of the multi-interval feature dimension parameter-preferred cadence pressure dataset in the cadence pressure benchmark library, a multi-interval feature dimension model architecture set is selected; cadence pressure fitting training is performed on the multi-interval feature dimension parameter-preferred cadence pressure dataset according to the multi-interval feature dimension model architecture set to construct a multi-interval feature dimension cadence pressure adjustment model set; the multi-interval feature dimension cadence pressure adjustment model set is verified, optimized, and integrated to generate a cadence pressure adjustment model library.

[0079] Furthermore, the reference pressure adjustment curve acquisition module 3 is used to perform the following steps: The basic information of the target user is classified according to the user feature dimension set to obtain the target user feature dimension parameters; based on the target user feature dimension parameters and the cadence interval of the cadence-time curve, feature matching is performed with the cadence-pressure adjustment model library to determine the target cadence-pressure adjustment model; the target cadence-pressure adjustment model is used to perform pressure adjustment analysis on the cadence-time curve to output the baseline pressure adjustment curve.

[0080] Furthermore, the dynamic pressure regulation curve acquisition module 4 is used to perform the following steps: A stride length calculation formula is constructed, specifically: stride length = length of pressure center trajectory × cadence correction coefficient, wherein the cadence correction coefficient is dynamically adjusted according to the user's height and weight; the length of the user's pressure center trajectory is obtained through the anti-gravity treadmill, and the stride length is estimated by using the stride length calculation formula and the cadence-time curve to obtain the user's stride length data; based on the target user's characteristic dimension parameters, cadence and stride length safety ratio thresholds are set to determine the cadence and stride length safety thresholds.

[0081] Furthermore, the dynamic pressure regulation curve acquisition module 4 is used to perform the following steps: If the ratio of the cadence-time curve to the user's stride data is within the cadence-stride safety threshold, then the baseline pressure adjustment curve is determined as the dynamic pressure adjustment curve; if the ratio of the cadence-time curve to the user's stride data exceeds the cadence-stride safety threshold, the deviation value of the movement imbalance from the cadence-stride safety threshold is calculated; a correction mechanism is triggered to perform collaborative intervention correction on the baseline pressure adjustment curve based on the movement imbalance deviation value, thereby determining the dynamic pressure adjustment curve.

[0082] Furthermore, the dynamic pressure regulation curve acquisition module 4 is used to perform the following steps: The test establishes a personalized imbalance deviation-pressure adjustment correction model for the target user; the correction mechanism is triggered to call the personalized imbalance deviation-pressure adjustment correction model to calculate the pressure adjustment of the motion imbalance deviation value and obtain the dynamic pressure adjustment coefficient; based on the dynamic pressure adjustment coefficient, the benchmark pressure adjustment curve is collaboratively intervened and corrected to determine the dynamic pressure adjustment curve.

[0083] In embodiment three, this application also provides a computer-readable storage medium that can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the dynamic pressure regulation system under step frequency coordination in the embodiments of this application, thereby realizing the above-mentioned dynamic pressure regulation method under step frequency coordination.

[0084] It should be understood that the embodiments disclosed in this application and the above description enable those skilled in the art to implement this application. However, this application is not limited to the embodiments mentioned above. It should be understood that those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this invention, and should all be included within the protection scope of this application.

Claims

1. A dynamic pressure regulation method under step frequency coordination, characterized in that, The method includes: Collect historical user cadence pressure datasets for the target anti-gravity treadmill, perform cadence pressure correlation analysis on the historical user cadence pressure datasets, establish a cadence pressure benchmark library, and generate a cadence pressure adjustment model library based on the cadence pressure benchmark library. A high-precision pressure sensor array is embedded in the anti-gravity running platform. The high-precision pressure sensor array collects the plantar pressure distribution data of the target user. Based on the plantar pressure distribution data, the cadence is calculated in real time to obtain the cadence-time curve. Based on the step frequency pressure adjustment model library, a matching adjustment analysis is performed on the target user and the step frequency time curve to output a benchmark pressure adjustment curve; The system acquires user stride length data and stride frequency safety thresholds, uses these data to collaboratively intervene and correct the baseline pressure adjustment curve, determines the dynamic pressure adjustment curve, and adjusts the treadmill pressure based on the dynamic pressure adjustment curve.

2. The dynamic pressure regulation method under step frequency coordination as described in claim 1, characterized in that, Establish a step frequency pressure benchmark library, including: Obtain a set of user feature dimensions, which includes height and weight range, age distribution, gender, health status, and exercise goals; The historical user cadence stress dataset is classified, labeled, and summarized according to the user feature dimension set to obtain the user feature dimension parameter - cadence stress dataset. The user feature dimension parameter - step frequency stress dataset is subjected to outlier removal and step frequency interval division to obtain a multi-interval feature dimension parameter - step frequency stress dataset; Based on the multi-interval feature dimension parameter-step frequency pressure dataset, step frequency pressure correlation analysis is performed to establish a step frequency pressure benchmark library.

3. The dynamic pressure regulation method under step frequency coordination as described in claim 2, characterized in that, Based on the multi-interval feature dimension parameter-step frequency stress dataset, step frequency stress correlation analysis is performed to establish a step frequency stress benchmark library, including: Based on the motion objectives of the anti-gravity treadmill, a set of evaluation indicators for the effectiveness of the motion objectives is set. The multi-interval feature dimension parameter - cadence pressure dataset is evaluated and filtered according to the set of exercise target effect evaluation indicators to obtain the multi-interval feature dimension parameter - preferred cadence pressure dataset. The multi-interval feature dimension parameter-preferred cadence pressure dataset is divided, associated, and integrated according to user feature dimension parameters and cadence intervals to establish a cadence pressure benchmark library.

4. The dynamic pressure regulation method under step frequency coordination as described in claim 3, characterized in that, Generate a library of step frequency pressure regulation models, including: Based on the characteristic information of the multi-interval feature dimension parameters-preferred step frequency pressure dataset in the step frequency pressure benchmark library, select a multi-interval feature dimension model architecture set; According to the multi-interval feature dimension model architecture set, the multi-interval feature dimension parameter-optimized cadence pressure dataset is used for cadence pressure fitting training to construct a multi-interval feature dimension cadence pressure regulation model set; The multi-interval feature dimension step frequency pressure adjustment model set is verified, optimized, identified, and integrated to generate a step frequency pressure adjustment model library.

5. The dynamic pressure regulation method under step frequency coordination as described in claim 2, characterized in that, Output reference pressure adjustment curve, including: The basic information of the target user is classified according to the user feature dimension set to obtain the target user feature dimension parameters; Based on the target user feature dimension parameters and the cadence interval of the cadence time curve, feature matching is performed with the cadence pressure regulation model library to determine the target cadence pressure regulation model. The target step frequency pressure regulation model is used to perform pressure regulation analysis on the step frequency time curve, and a baseline pressure regulation curve is output.

6. The dynamic pressure regulation method under step frequency coordination as described in claim 5, characterized in that, Obtain user stride data and cadence-stride safety thresholds, including: A stride length calculation formula is constructed, which is specifically: stride length = length of pressure center trajectory × step frequency correction coefficient, wherein the step frequency correction coefficient is dynamically adjusted according to the user's height and weight; The length of the user's center of pressure trajectory is obtained by the anti-gravity treadmill. The stride length is estimated by the stride length of the user's center of pressure trajectory and the stride frequency-time curve using the stride calculation formula to obtain the user's stride data. Based on the target user characteristic dimension parameters, the step frequency and stride length safety ratio thresholds are set to determine the step frequency and stride length safety thresholds.

7. The dynamic pressure regulation method under step frequency coordination as described in claim 1, characterized in that, Determine the dynamic pressure regulation curve, including: If the ratio of the cadence-time curve to the user stride data is within the cadence-stride safety threshold, then the reference pressure adjustment curve is determined as the dynamic pressure adjustment curve. If the ratio of the cadence-time curve to the user stride data exceeds the cadence-stride safety threshold, calculate the deviation value of the movement imbalance from the cadence-stride safety threshold; The trigger correction mechanism performs coordinated intervention correction on the reference pressure regulation curve based on the motion imbalance deviation value to determine the dynamic pressure regulation curve.

8. The dynamic pressure regulation method under step frequency coordination as described in claim 7, characterized in that, The trigger correction mechanism performs coordinated intervention correction on the reference pressure regulation curve based on the motion imbalance deviation value to determine the dynamic pressure regulation curve, including: The test establishes a personalized imbalance deviation-stress regulation correction model for the target user; The trigger correction mechanism invokes the personalized imbalance deviation-pressure regulation correction model to perform pressure regulation calculation on the motion imbalance deviation value, and obtains the dynamic pressure regulation coefficient. Based on the dynamic pressure regulation coefficient, the reference pressure regulation curve is collaboratively intervened and corrected to determine the dynamic pressure regulation curve.

9. A dynamic pressure regulation system under step frequency coordination, characterized in that, For implementing the dynamic pressure regulation method under step frequency coordination as described in any one of claims 1-8, the system comprises: The cadence pressure adjustment model library construction module is used to collect historical user cadence pressure datasets from the target anti-gravity treadmill, perform cadence pressure correlation analysis on the historical user cadence pressure datasets, establish a cadence pressure benchmark library, and generate a cadence pressure adjustment model library based on the cadence pressure benchmark library. The cadence-time curve acquisition module is used to embed a high-precision pressure sensor array into the anti-gravity treadmill pedal, collect the plantar pressure distribution data of the target user through the high-precision pressure sensor array, and perform real-time cadence calculation based on the plantar pressure distribution data to obtain the cadence-time curve. The baseline pressure adjustment curve acquisition module performs matching adjustment analysis on the target user and the step frequency time curve based on the step frequency pressure adjustment model library, and outputs the baseline pressure adjustment curve. The dynamic pressure adjustment curve acquisition module is used to acquire user stride data and stride frequency safety thresholds, use the user stride data and stride frequency safety thresholds to perform collaborative intervention correction on the benchmark pressure adjustment curve, determine the dynamic pressure adjustment curve, and adjust the treadmill pressure based on the dynamic pressure adjustment curve.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the dynamic pressure regulation method under step frequency coordination as described in any one of claims 1 to 8.

Images