Intelligent wearable pelvic posture real-time correction method and equipment

Abstract

The invention discloses an intelligent wearable pelvic posture real-time correction method and equipment, and relates to the technical field of intelligent wearable equipment, and the method comprises the steps: integrating a pelvic posture monitoring sensor in the intelligent wearable equipment, and collecting pelvic motion data of a user in real time; processing and analyzing the pelvic motion data, and extracting a feature parameter set related to pelvic postures; carrying out comparative analysis on the extracted characteristic parameter set and a pre-stored normal pelvic posture; judging whether the current pelvic posture is in a normal range or not by adopting a machine learning algorithm and a mode recognition technology; if it is detected that the pelvic posture is abnormal, the abnormal type and the abnormal degree of the pelvic posture are judged; according to the judgment result, different posture abnormity reminding signals are sent to the user. According to the invention, fine pelvic motion features can be captured in real time, various abnormal types are accurately identified and graded and judged in combination with an advanced machine learning algorithm, and personalized reminding is realized by dynamically adjusting feedback parameters.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent wearable devices, and more particularly to an intelligent wearable pelvic posture real-time correction method and device. Background Art

[0002] The pelvis occupies a central position in the human body, its structure and function evolving continuously throughout human evolution. In modern society, the pelvis is particularly important for maintaining normal physiological function and postural balance. However, with the dramatic changes in modern lifestyles, people's daily activity patterns have shifted significantly, with unhealthy habits such as prolonged sitting at work and prolonged periods of looking down at electronic devices becoming increasingly common. These behaviors pose unprecedented challenges to pelvic stability, leading to a sharp increase in the incidence of abnormal pelvic posture each year.

[0003] According to statistics from the World Health Organization, approximately 42% of adults worldwide experience varying degrees of anterior or posterior pelvic tilt. Abnormal pelvic posture not only affects aesthetics but can also lead to a range of serious health issues, such as low back pain and scoliosis. These conditions not only cause physical pain but also create significant inconvenience in daily life and work, becoming a global public health concern.

[0004] Traditional pelvic correction methods rely primarily on manual intervention or static correction devices. Manual intervention requires professional medical guidance and operation, which is not only costly but also difficult to achieve real-time, continuous monitoring and correction. Static correction devices can only be used in specific environments and cannot monitor and adjust the user's dynamic pelvic posture in daily life, making them unable to meet the modern demand for convenient and efficient health management.

[0005] In existing technologies, some wearable devices attempt to monitor pelvic posture using a single sensor. However, this approach has significant limitations. The data dimensions that a single sensor can collect are limited, making it impossible to fully and accurately capture the subtle changes in pelvic movements. Furthermore, its ability to identify abnormal pelvic posture is weak, making it difficult to accurately distinguish between different types of abnormalities, such as forward tilt, backward tilt, and lateral tilt, and to accurately determine the degree of abnormality.

[0006] Therefore, how to provide an intelligent wearable pelvic posture real-time correction method and equipment, capture the subtle movement characteristics of the pelvis in real time, combine advanced machine learning algorithms to accurately identify and grade various abnormal types, and realize personalized reminders by dynamically adjusting feedback parameters is a problem that technical personnel in this field urgently need to solve. Summary of the Invention

[0007] In view of this, the present invention provides an intelligent wearable pelvic posture real-time correction method and device, which can capture the subtle movement characteristics of the pelvis in real time and comprehensively, and use machine learning algorithms and pattern recognition technology to accurately identify and grade abnormal types of pelvic tilt, backward tilt, and lateral tilt. By dynamically adjusting the vibration feedback parameters, it sends personalized posture abnormality reminder signals to the user, guiding the user to adjust the posture in a timely and autonomous manner, thereby effectively preventing and correcting pelvic posture abnormalities and improving the user's physical health.

[0008] In order to achieve the above objectives, the present invention adopts the following technical solution: a smart wearable pelvic posture real-time correction method, comprising:

[0009] Integrate pelvic posture monitoring sensors into smart wearable devices to collect users' pelvic movement data in real time;

[0010] Processing and analyzing the pelvic motion data to extract a set of characteristic parameters related to the pelvic posture;

[0011] Compare and analyze the extracted feature parameter set with the pre-stored normal pelvic posture;

[0012] Using machine learning algorithms and pattern recognition technology to determine whether the current pelvic posture is within the normal range;

[0013] If an abnormal pelvic posture is detected, the abnormal type and degree of the pelvic posture are determined; based on the determination result, different abnormal posture reminder signals are sent to the user.

[0014] Preferably, the method further includes: transmitting and storing the real-time pelvic movement data and correction execution status to the user terminal device.

[0015] Preferably, the feature parameter set related to the pelvic posture is extracted, including:

[0016] Based on the feature pyramid network, an improved side fusion network is introduced to establish a pelvic posture feature extraction network;

[0017] A feature parameter set related to the pelvic posture is extracted through the pelvic posture feature extraction network.

[0018] Preferably, based on the feature pyramid network, an improved side fusion network is introduced to establish a pelvic posture feature extraction network, including:

[0019] The feature pyramid network is used to process the pelvic motion data in stages. In each stage, convolution operations are performed in sequence, followed by pooling operations to obtain the feature parameters {C1, C2, C3, C4, C5} of different stages.

[0020] A top-down fusion path is adopted, and the top-level feature parameters {C5} are amplified by upsampling method to make their dimensions consistent with the feature parameters of the previous stage;

[0021] Add the corresponding elements of the amplified feature parameters to the feature parameters of the previous stage to obtain the fused feature parameters {P2, P3, P4, P5};

[0022] Introducing a side fusion network structure, between feature layers at the same stage and the same scale, {N1} is directly set to {C1}, and {N5} is directly set to {C5}, keeping them unchanged;

[0023] On this basis, a jump connection path is added from {C2, C3, C4} to {N2, N3, N4}; the feature parameters of {C2, C3, C4} are side-fused with the feature parameters of {P2, P3, P4} after top-down fusion through the jump connection path to generate a feature parameter set of {N1, N2, N3, N4, N5} containing the feature information of this layer and the fusion of the top-level feature information.

[0024] Preferably, machine learning algorithms and pattern recognition technology are used to determine whether the current pelvic posture is within a normal range, including:

[0025] A feature parameter in the feature parameter set is selected to construct a new optimization problem, and the new optimization problem is used to replace the basic optimization problem in the traditional SVM model;

[0026] The sigmoid kernel function is selected as the kernel function of the traditional SVM model to obtain a preliminary improved SVM model;

[0027] Perform hyperparameter optimization on the preliminary improved SVM model, update its optimal weight vector ω and bias term b, and obtain the trained improved SVM model;

[0028] Based on the mean absolute error and root mean square error, the correlation coefficient R and relative error σ are obtained. R and σ are used to evaluate the performance of the improved SVM model after training. At the same time, the complexity of the improved SVM model after training is controlled by adjusting the regularization parameter C.

[0029] Apply the evaluated improved SVM model to new data in the feature parameter set and calculate ω·x in the evaluated improved SVM model i +b value, the data are assigned to the category with a calculated value greater than 0.05, and the improved SVM model after classification is obtained;

[0030] Repeat the above operation for other feature parameters in the feature parameter set, output and update the weights and bias of the preliminary improved SVM model, complete the learning of the preliminary improved SVM model, and obtain the final improved SVM prediction model.

[0031] Preferably, the hyperparameter optimization is performed on the preliminary improved SVM model, and its optimal weight vector ω and bias term b are updated to obtain the trained improved SVM model, including:

[0032] Randomly select multiple hyperparameters in the preliminary improved SVM model to form the first generation of data sets;

[0033] The fitness function is used to calculate the fitness of each individual in the first generation data set, and to determine whether the individual meets the termination condition at this time. If so, the optimal weight vector ω and bias term b in the preliminary improved SVM model are output and updated;

[0034] If it is not satisfied, the probability of each individual being selected in the next selection is calculated based on the obtained fitness level and the probability formula of individual i being selected;

[0035] According to the calculated probability, M individuals with large probabilities are selected to form the second generation data set; among them, a certain data can be selected repeatedly, and M <N;

[0036] The data in the second-generation dataset are subjected to a "crossover" operation based on the mating probability pc to generate NM data, and then a "mutation" operation is performed on the generated NM data to obtain the mutated NM dataset; the data in the second-generation dataset are then "copied" to generate L data, and the L data are merged with the mutated NM dataset to generate the third-generation dataset;

[0037] Based on the third-generation data set, it returns to determine whether the individual meets the termination condition at this time. If it does, it outputs and updates the optimal weight vector ω and bias term b in the preliminary improved SVM model to achieve hyperparameter optimization of the preliminary improved SVM model and obtain the improved SVM model after training.

[0038] Preferably, the performance method of the final improved SVM prediction model includes: calculating the correlation coefficient R and the relative error σ based on the mean absolute error and the root mean square error, and measuring the performance of the final improved SVM prediction model by the correlation coefficient R and the relative error σ.

[0039] Preferably, an intelligent wearable pelvic posture real-time correction device comprises:

[0040] A smart belt made of elastic material, having a built-in pelvic posture monitoring sensor, a microprocessor, a Bluetooth module, and a battery;

[0041] A mobile terminal is connected to the smart belt via Bluetooth and is used to receive and store data transmitted by the smart belt.

[0042] A vibration motor is installed in the inner ring of the smart belt. When an abnormal pelvic posture is detected, the vibration motor is controlled to generate vibrations of different frequencies and intensities according to the different types of abnormalities, reminding the user to adjust the posture in time.

[0043] Preferably, the pelvic posture monitoring sensor is used to collect pelvic movement, angle and pressure data in real time and transmit the data to a microprocessor for analysis and processing.

[0044] Through the above technical solutions, it can be seen that compared with the existing technology, the present invention discloses a smart wearable pelvic posture real-time correction method and device, which relates to the technical field of smart wearable devices, including: integrating a pelvic posture monitoring sensor in the smart wearable device to collect the user's pelvic movement data in real time; processing and analyzing the pelvic movement data to extract a feature parameter set related to the pelvic posture; comparing and analyzing the extracted feature parameter set with the pre-stored normal pelvic posture; using machine learning algorithms and pattern recognition technology to determine whether the current pelvic posture is within the normal range; if an abnormal pelvic posture is detected, the abnormal type and degree of the pelvic posture are determined; and according to the judgment result, different posture abnormality reminder signals are sent to the user. The present invention can capture the subtle movement characteristics of the pelvis in real time, combine advanced machine learning algorithms to accurately identify and grade various abnormal types, and realize personalized reminders by dynamically adjusting feedback parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0046] Figure 1 This is a flow chart of a smart wearable real-time pelvic posture correction method provided by the present invention.

[0047] Figure 2 This is a schematic structural diagram of an intelligent wearable pelvic posture real-time correction device provided by the present invention. DETAILED DESCRIPTION

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0049] The embodiment of the present invention discloses a smart wearable pelvic posture real-time correction method, such as Figure 1 Shown, including:

[0050] Integrating pelvic posture monitoring sensors into smart wearable devices. These sensors include accelerometers, gyroscopes, and pressure sensors to collect real-time pelvic motion data, including acceleration, angular velocity, and other information. These sensors operate at a high sampling frequency, ensuring they can capture any subtle changes in pelvic movement, providing a rich and accurate data foundation for subsequent analysis and judgment.

[0051] Processing and analyzing the pelvic motion data to extract a set of characteristic parameters related to the pelvic posture;

[0052] Compare and analyze the extracted feature parameter set with the pre-stored normal pelvic posture;

[0053] Using machine learning algorithms and pattern recognition technology to determine whether the current pelvic posture is within the normal range;

[0054] If an abnormal pelvic posture is detected, the abnormal type and degree of the pelvic posture are determined; the abnormal type of the pelvic posture includes forward tilt, backward tilt, lateral tilt, etc., and the abnormal degree of the pelvic posture is determined based on the monitored forward tilt angle, backward tilt angle, and lateral tilt angle;

[0055] According to the judgment results, different posture abnormality reminder signals are sent to the user;

[0056] Based on the judgment results, the corresponding control instructions are sent, and by adjusting the vibration intensity and frequency of the vibration motor, different posture abnormality reminder signals are sent to the user to guide the user to adjust the posture independently.

[0057] Specifically, it also includes: transmitting and storing real-time pelvic movement data and correction execution status to the user's terminal device. Users can intuitively view their pelvic posture change history, correction records, and current posture status through the app.

[0058] Specifically, the pelvic motion data is processed, including filtering, noise reduction and coordinate conversion processing, to remove noise and interference in the data and improve the quality of the data.

[0059] Extract the feature parameter set related to pelvic posture, including:

[0060] Based on the feature pyramid network, an improved side fusion network is introduced to establish a pelvic posture feature extraction network;

[0061] A feature parameter set related to the pelvic posture is extracted through the pelvic posture feature extraction network.

[0062] The pelvic posture feature extraction network extracts characteristic parameters closely related to pelvic posture from the preprocessed data, such as pelvic tilt angle, rotation angle, motion speed, and acceleration rate. These characteristic parameters can accurately reflect pelvic posture information and provide a key basis for subsequent abnormality judgment.

[0063] Specifically, based on the feature pyramid network, an improved side fusion network is introduced to establish a pelvic posture feature extraction network, including:

[0064] The pelvic motion data is processed in stages using a feature pyramid network. At each stage, convolution operations are performed sequentially to extract local features from the data. Pooling operations are then performed to reduce the data dimension and highlight the main features, thereby obtaining the feature parameters {C1, C2, C3, C4, C5} at different stages. The feature parameter {C5} is located at the top layer of the feature pyramid network.

[0065] A top-down fusion path is adopted, and the top-level feature parameters {C5} are amplified by upsampling method to make their dimensions consistent with the feature parameters of the previous stage;

[0066] The amplified feature parameters are added to the feature parameters of the previous stage by corresponding elements to obtain the fused feature parameters {P2, P3, P4, P5}. This process realizes the fusion of feature parameters of different scales and enriches the feature information.

[0067] Introducing a side fusion network structure, between feature layers at the same stage and the same scale, {N1} is directly set to {C1}, and {N5} is directly set to {C5}, keeping them unchanged;

[0068] On this basis, a skip connection path is added from {C2, C3, C4} to {N2, N3, N4}; the feature parameters of {C2, C3, C4} are side-fused with the feature parameters of {P2, P3, P4} after top-down fusion through the skip connection path to generate a feature parameter set of {N1, N2, N3, N4, N5} containing detailed feature information of this layer and fused with the top-level feature information.

[0069] These characteristic parameters will provide key data basis for subsequent accurate judgment of pelvic posture.

[0070] In the convolution operation of the feature pyramid network, a convolution kernel of size 3×3 is used with a step size of 1; the pooling operation adopts a pooling kernel of size 2×2 with a step size of 2; the upsampling multiple in the top-down fusion path is 2.

[0071] Specifically, machine learning algorithms and pattern recognition technology are used to determine whether the current pelvic posture is within the normal range, including:

[0072] A feature parameter in the feature parameter set is selected to construct a new optimization problem, and the new optimization problem is used to replace the basic optimization problem in the traditional SVM model;

[0073] The sigmoid kernel function is selected as the kernel function of the traditional SVM model to obtain a preliminary improved SVM model;

[0074] Perform hyperparameter optimization on the preliminary improved SVM model, update its optimal weight vector ω and bias term b, and obtain the trained improved SVM model;

[0075] Based on the mean absolute error and root mean square error, the correlation coefficient R and relative error σ are obtained. R and σ are used to evaluate the performance of the improved SVM model after training. At the same time, the complexity of the improved SVM model after training is controlled by adjusting the regularization parameter C.

[0076] Apply the evaluated improved SVM model to new data in the feature parameter set and calculate ω·x in the evaluated improved SVM model i +b value, the data are assigned to the category with a calculated value greater than 0.05, and the improved SVM model after classification is obtained;

[0077] Repeat the above operation for other feature parameters in the feature parameter set, output and update the weights and bias of the preliminary improved SVM model, complete the learning of the preliminary improved SVM model, and obtain the final improved SVM prediction model.

[0078] The pelvic posture feature parameter set collected and processed in real time is input into the final improved SVM prediction model, and the model will output a judgment result on whether the current pelvic posture is normal.

[0079] If an abnormal pelvic posture is detected, the type of abnormality (forward tilt, backward tilt, lateral tilt, etc.) and the degree of abnormality are determined based on the monitored forward tilt angle, backward tilt angle, and lateral tilt angle.

[0080] Based on the judgment results, the system sends different posture abnormality warning signals to the user, and guides the user to adjust their posture by adjusting the vibration intensity and frequency of the vibration motor. At the same time, real-time pelvic movement data and correction execution status are transmitted and stored to the user's terminal device. Users can intuitively view pelvic posture change history, correction records, and current posture status through the app.

[0081] Specifically, the hyperparameters of the preliminary improved SVM model are optimized, and its optimal weight vector ω and bias term b are updated to obtain the trained improved SVM model, including:

[0082] Randomly select multiple hyperparameters in the preliminary improved SVM model to form the first generation of data sets;

[0083] The fitness function is used to calculate the fitness of each individual in the first generation data set, and to determine whether the individual meets the termination condition at this time. If so, the optimal weight vector ω and bias term b in the preliminary improved SVM model are output and updated;

[0084] If it is not satisfied, the probability of each individual being selected in the next selection is calculated based on the obtained fitness level and the probability formula of individual i being selected;

[0085] According to the calculated probability, M individuals with large probabilities are selected to form the second generation data set; among them, a certain data can be selected repeatedly, and M <N;

[0086] The data in the second-generation dataset are subjected to a "crossover" operation based on the mating probability pc to generate NM data, and then a "mutation" operation is performed on the generated NM data to obtain the mutated NM dataset; the data in the second-generation dataset are then "copied" to generate L data, and the L data are merged with the mutated NM dataset to generate the third-generation dataset;

[0087] Based on the third-generation data set, it returns to determine whether the individual meets the termination condition at this time. If it does, it outputs and updates the optimal weight vector ω and bias term b in the preliminary improved SVM model to achieve hyperparameter optimization of the preliminary improved SVM model and obtain the improved SVM model after training.

[0088] Specifically, the performance method of the final improved SVM prediction model includes: calculating the correlation coefficient R and the relative error σ based on the mean absolute error and the root mean square error, and measuring the performance of the final improved SVM prediction model by the correlation coefficient R and the relative error σ.

[0089] In a specific embodiment of the present invention, a smart wearable pelvic posture real-time correction device, such as Figure 2 Shown, including:

[0090] A smart belt is made of an adjustable elastic material and has a built-in pelvic posture monitoring sensor, a microprocessor, a Bluetooth module, and a battery; the two sides of the smart belt are connected by an adhesive.

[0091] A mobile terminal is connected to the smart belt via Bluetooth and is used to receive and store data transmitted by the smart belt.

[0092] A small vibration motor is installed in the inner ring of the smart belt. When an abnormal pelvic posture is detected, the vibration motor is controlled to produce vibrations of different frequencies and intensities according to the type of abnormality, prompting the user to adjust their posture in time. The pelvic posture monitoring sensor device is only placed on the lumbar vertebra 5-sacrum to measure the front, back, left, and right pelvic tilt.

[0093] Specifically, the pelvic posture monitoring sensor is used to collect pelvic movement, angle and pressure data in real time, and transmit the data to a microprocessor for analysis and processing.

[0094] Specifically, the data collected by the pelvic posture monitoring sensor is fused using a Kalman filter algorithm. The pelvic posture monitoring sensor includes an accelerometer, a gyroscope, and a pressure sensor. This algorithm performs weighted fusion of the data from its various internal sensors based on the sensor's measurement noise and dynamic characteristics, resulting in more accurate and stable pelvic motion status data. When a user performs complex movements, the accelerometer may be subject to significant external interference, while the gyroscope is relatively stable. The Kalman filter algorithm automatically adjusts the weight based on real-time noise estimation, prioritizing the gyroscope's data to ensure that the final collected data truly reflects the pelvic motion.

[0095] To ensure data synchronization between the three sensors, a trigger mechanism based on clock synchronization is employed. Using the microprocessor's internal clock as a reference, a synchronization signal is sent to the three sensors every 10ms. Upon receiving the synchronization signal, the sensors simultaneously collect and transmit data, ensuring that data from the different sensors is time-aligned and providing accurate time series data for subsequent analysis and processing.

[0096] At the same time, through the data of the acceleration sensor, gyroscope sensor and pressure sensor, when the user wears the belt, due to the physiological structure characteristics of the lumbar vertebra-5-sacrum area, the pressure sensors in different positions will sense different pressure values. Based on the pressure distribution characteristics generated by the belt, the pressure sensor data is analyzed and processed to determine whether the belt is in the correct position.

[0097] Specifically, the microprocessor collects pressure data from the pressure sensor at a certain sampling frequency and stores the data for subsequent analysis.

[0098] Establish a pressure distribution characteristic model and collect sample data: Invite multiple volunteers of different body shapes, genders, and ages to wear the smart belt correctly. Collect the pressure sensor data corresponding to their lumbar vertebra 5-sacrum to form a sample data set.

[0099] The sample data set is processed to extract the characteristics of the pressure distribution, such as the pressure peak position, pressure gradient changes, and the proportional relationship of pressure values in different areas. Statistical analysis methods are used to calculate the mean value, standard deviation and other statistical quantities of each feature to establish a pressure distribution characteristic model under normal conditions.

[0100] The pressure sensor data collected in real time is preprocessed, and the pressure distribution features identical to those in the pressure distribution feature model are extracted from the preprocessed data.

[0101] A reasonable threshold range is set for each pressure distribution feature, and the real-time extracted pressure distribution feature values are compared with the threshold range. If all feature values are within the threshold range, the belt is considered to be in the correct position; otherwise, it is judged to be in an incorrect position and a position adjustment prompt is issued through the vibration motor.

[0102] When the belt is judged to be in an incorrect position, the built-in vibration motor (such as Figure 2 As shown, the two orange vibration motors near the pelvic posture monitoring sensor generate vibrations of specific frequency and intensity, or emit sounds through the mobile terminal APP to remind the user to adjust the belt position.

[0103] While the user is adjusting the belt position, pressure sensor data is continuously collected and analyzed until the belt is determined to be in the correct position.

[0104] Specifically, in addition to allowing users to adjust the belt position, view the history of pelvic posture changes, correction records, current posture status and other information, the mobile terminal APP also supports personalized settings. Users can customize the sensitivity of abnormal reminders based on factors such as their physical condition, exercise habits and work environment; for example, for users who often engage in high-intensity physical labor, the threshold of abnormal reminders can be appropriately raised to avoid frequent reminders due to normal body movements; for users who sit for a long time at work, the threshold can be lowered to detect pelvic posture problems more promptly.

[0105] The app also includes professional exercise rehabilitation courses and advice modules. Based on the type and severity of the user's pelvic posture abnormality, it provides users with personalized exercise rehabilitation plans. For example, for users with anterior pelvic tilt, exercises to strengthen the abdominal muscles and stretch the hip flexors are recommended, such as supine leg raises and hip stretches, accompanied by detailed movement demonstrations and explanation videos. Furthermore, the exercise plan is dynamically adjusted based on the user's correction progress, helping them gradually restore a normal pelvic posture.

[0106] Specifically, the adjustable elastic material used in the smart belt has good breathability, which can effectively reduce the stuffiness when the user wears it; at the same time, it has high wear resistance and can withstand friction and stretching in daily use, ensuring the service life of the product.

[0107] In a specific embodiment of the present invention, the specific mapping relationship between different types of pelvic posture abnormalities and the vibration intensity and frequency of the vibration motor includes:

[0108] 1. Anterior pelvic tilt

[0109] Mild Abnormality: When a mild anterior pelvic tilt abnormality is detected, that is, the tilt angle is between 5° and 10°, the vibration motor vibrates at a frequency of 20Hz and an intensity of 50mN to alert the user. This lower frequency and intensity of vibration can gently recognize minor pelvic posture issues without being overly disturbing.

[0110] Moderate Abnormality: If the forward tilt angle is between 10° and 15°, it is considered moderate abnormality. In this case, the vibration motor frequency is increased to 40Hz and the intensity is increased to 100mN. The increase in frequency and intensity can more clearly draw the user's attention, reminding them to adjust their posture as soon as possible.

[0111] Severe Abnormality: When the posterior tilt angle exceeds 15°, it is considered a severe abnormality. The vibration motor vibrates at a frequency of 60Hz and an intensity of 150mN. This strong vibration allows the user to clearly perceive severe abnormalities in pelvic posture during any activity.

[0112] Posterior pelvic tilt

[0113] Mild Abnormality: A backward tilt angle between 3° and 6° is considered a mild abnormality, and the vibration motor vibrates at a frequency of 18Hz and an intensity of 40mN. The backward tilt angle is smaller than the forward tilt angle, so the initial frequency and intensity are also slightly lower.

[0114] Moderate abnormality: When the tilt angle is between 6° and 9°, it is considered moderate abnormality, the vibration frequency increases to 35Hz, and the intensity becomes 80mN.

[0115] Severe abnormality: A backward tilt angle of more than 9° is considered a severe abnormality, the vibration frequency reaches 55Hz, and the intensity is 120mN.

[0116] Pelvic tilt

[0117] Mild Abnormality: A roll angle between 2° and 4° is considered a mild abnormality, and the vibration motor vibrates at a frequency of 15Hz and an intensity of 30mN. Because roll can be more subtle than forward and backward tilt, a mild abnormality is still indicated.

[0118] Moderate abnormality: When the roll angle is 4°-6°, it is considered moderate abnormality, the vibration frequency is adjusted to 30Hz, and the intensity is 60mN.

[0119] Severe abnormality: A roll angle exceeding 6° is considered a severe abnormality, the vibration frequency increases to 50Hz, and the intensity is 100mN.

[0120] 2. Dynamically adjust the reminder signal according to the change of abnormality

[0121] As the abnormality increases, the vibration intensity and frequency increase linearly. Specifically, for each degree of abnormal angle increase, within the mild to moderate and moderate to severe ranges of the corresponding abnormality type, the frequency increases according to a certain slope, and the intensity also increases according to a fixed increment. For example, in the moderate abnormal range of pelvic anterior tilt (10°-15°), for every 1° increase in the tilt angle, the frequency increases by 4Hz ((40-20) / (15-10)) and the intensity increases by 10mN ((100-50) / (15-10)).

[0122] At the same time, in order to prevent users from adapting to the reminder signal and ignoring the reminder, when the abnormality persists for a certain period of time, while maintaining the basic frequency and intensity corresponding to the current abnormality level, the frequency of 5Hz and the intensity of 10mN are increased every 3 minutes until the user adjusts the posture to reduce or eliminate the abnormality.

[0123] 3. Setting the duration and interval of reminder signals

[0124] The reminder signal lasts for 3 seconds each time, with an interval of 15 seconds. This setting ensures that the user has enough time to perceive the vibration reminder without becoming annoyed by overly frequent reminders. When the abnormality increases, the duration and interval remain unchanged, but the vibration intensity and frequency will increase according to the dynamic adjustment rules described above. If the user still does not adjust their posture to alleviate the abnormality after three consecutive reminders, the interval is shortened to 10 seconds to more strongly urge the user to adjust their pelvic posture. When the abnormality is reduced or eliminated, the duration and interval settings are restored to the original settings.

[0125] The embodiments of the present invention can be applied to places such as gyms and sports training institutions. Coaches can use the system to monitor the pelvic posture of athletes or fitness enthusiasts in real time, promptly discover possible incorrect postures during exercise, and avoid sports injuries caused by long-term incorrect postures. At the same time, coaches can formulate more scientific training plans for users based on monitoring data to improve training effects.

[0126] It can also be used as an auxiliary tool for rehabilitation treatment of patients with pelvic injuries or diseases. Doctors can use the system to understand the patient's pelvic posture recovery in real time and adjust the rehabilitation treatment plan, such as the intensity and frequency of physical therapy, the movements and difficulty of rehabilitation training, etc., to help patients recover faster.

[0127] In nursing homes or home care, the device can be used to monitor the elderly's pelvic posture in real time. When abnormalities occur, caregivers or family members can be promptly alerted to take appropriate measures, such as helping the elderly adjust their posture, arranging further examinations and treatments, etc., to improve the quality of life and safety of the elderly.

[0128] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0129] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A smart wearable pelvic posture real-time correction method, characterized in that: include: Integrate pelvic posture monitoring sensors into smart wearable devices to collect users' pelvic movement data in real time; Processing and analyzing the pelvic motion data to extract a set of characteristic parameters related to the pelvic posture; Compare and analyze the extracted feature parameter set with the pre-stored normal pelvic posture; Using machine learning algorithms and pattern recognition technology to determine whether the current pelvic posture is within the normal range; If an abnormal pelvic posture is detected, the type and degree of abnormality of the pelvic posture are determined; Based on the judgment results, different abnormal posture reminder signals are sent to the user.

2. The intelligent wearable pelvic posture real-time correction method according to claim 1, characterized in that: Also includes: The real-time pelvic movement data and correction execution status are transmitted and stored to the user terminal device.

3. The intelligent wearable pelvic posture real-time correction method according to claim 1, characterized in that: Extract the feature parameter set related to pelvic posture, including: Based on the feature pyramid network, an improved side fusion network is introduced to establish a pelvic posture feature extraction network; A feature parameter set related to the pelvic posture is extracted through the pelvic posture feature extraction network.

4. The intelligent wearable pelvic posture real-time correction method according to claim 3, characterized in that: Based on the feature pyramid network, an improved side fusion network is introduced to establish a pelvic posture feature extraction network, including: The feature pyramid network is used to process the pelvic motion data in stages. In each stage, convolution operations are performed in sequence, followed by pooling operations to obtain the feature parameters {C1, C2, C3, C4, C5} of different stages. A top-down fusion path is adopted, and the top-level feature parameters {C5} are amplified by upsampling method to make their dimensions consistent with the feature parameters of the previous stage; Add the corresponding elements of the amplified feature parameters to the feature parameters of the previous stage to obtain the fused feature parameters {P2, P3, P4, P5}; Introducing a side fusion network structure, between feature layers at the same stage and the same scale, {N1} is directly set to {C1}, and {N5} is directly set to {C5}, keeping them unchanged; On this basis, a jump connection path is added from {C2, C3, C4} to {N2, N3, N4}; the feature parameters of {C2, C3, C4} are side-fused with the feature parameters of {P2, P3, P4} after top-down fusion through the jump connection path to generate a feature parameter set of {N1, N2, N3, N4, N5} containing the feature information of this layer and the fusion of the top-level feature information.

5. The intelligent wearable pelvic posture real-time correction method according to claim 1, characterized in that: Using machine learning algorithms and pattern recognition technology, it determines whether the current pelvic posture is within the normal range, including: A feature parameter in the feature parameter set is selected to construct a new optimization problem, and the new optimization problem is used to replace the basic optimization problem in the traditional SVM model; The sigmoid kernel function is selected as the kernel function of the traditional SVM model to obtain a preliminary improved SVM model; Perform hyperparameter optimization on the preliminary improved SVM model, update its optimal weight vector ω and bias term b, and obtain the trained improved SVM model; Based on the mean absolute error and root mean square error, the correlation coefficient R and relative error σ are obtained. R and σ are used to evaluate the performance of the improved SVM model after training. At the same time, the complexity of the improved SVM model after training is controlled by adjusting the regularization parameter C. Apply the evaluated improved SVM model to new data in the feature parameter set and calculate ω·x in the evaluated improved SVM model i +b value, the data are assigned to the category with a calculated value greater than 0.05, and the improved SVM model after classification is obtained; Repeat the above operation for other feature parameters in the feature parameter set, output and update the weights and bias of the preliminary improved SVM model, complete the learning of the preliminary improved SVM model, and obtain the final improved SVM prediction model.

6. The intelligent wearable pelvic posture real-time correction method according to claim 5, characterized in that: Perform hyperparameter optimization on the preliminary improved SVM model, update its optimal weight vector ω and bias term b, and obtain the trained improved SVM model, including: Randomly select multiple hyperparameters in the preliminary improved SVM model to form the first generation of data sets; The fitness function is used to calculate the fitness of each individual in the first generation data set, and to determine whether the individual meets the termination condition at this time. If so, the optimal weight vector ω and bias term b in the preliminary improved SVM model are output and updated; If it is not satisfied, the probability of each individual being selected in the next selection is calculated based on the obtained fitness level and the probability formula of individual i being selected; According to the calculated probability, M individuals with large probabilities are selected to form the second generation data set; among them, a certain data can be selected repeatedly, and M <N; The data in the second-generation dataset are subjected to a "crossover" operation based on the mating probability pc to generate NM data. The generated NM data are then subjected to a "mutation" operation to obtain the mutated NM dataset. The data in the second-generation dataset are then "copied" to generate L data, and the L data are merged with the mutated NM dataset to generate the third-generation dataset. Based on the third-generation data set, it returns to determine whether the individual meets the termination condition at this time. If it does, it outputs and updates the optimal weight vector ω and bias term b in the preliminary improved SVM model to achieve hyperparameter optimization of the preliminary improved SVM model and obtain the improved SVM model after training.

7. The intelligent wearable pelvic posture real-time correction method according to claim 6, characterized in that: The performance method of the final improved SVM prediction model includes: calculating the correlation coefficient R and the relative error σ based on the mean absolute error and the root mean square error, and measuring the performance of the final improved SVM prediction model by the correlation coefficient R and the relative error σ.

8. An intelligent wearable pelvic posture real-time correction device, applying the intelligent wearable pelvic posture real-time correction method according to any one of claims 1 to 7, characterized in that: include: A smart belt made of elastic material, having a built-in pelvic posture monitoring sensor, a microprocessor, a Bluetooth module, and a battery; a mobile terminal connected to the smart belt via Bluetooth and configured to receive and store data transmitted by the smart belt; A vibration motor is installed in the inner ring of the smart belt. When an abnormal pelvic posture is detected, the vibration motor is controlled to generate vibrations of different frequencies and intensities according to the different types of abnormalities, reminding the user to adjust the posture in time.

9. The intelligent wearable pelvic posture real-time correction device according to claim 8, characterized in that: The pelvic posture monitoring sensor is used to collect pelvic movement, angle and pressure data in real time and transmit the data to the microprocessor for analysis and processing.

Images