Pressure monitoring system and method based on knee joint rehabilitation

Abstract

The invention relates to the technical field of rehabilitation training, in particular to a pressure monitoring system and method based on knee joint rehabilitation. Firstly, quadriceps femoris physiological data of a healthy person during knee joint contraction are collected and uploaded to a database, and a pressure threshold value is calculated based on the data; secondly, obtaining new pressure and myoelectricity feedback data by gradually increasing the systolic pressure of quadriceps femoris muscles of the patient, and storing the new pressure and myoelectricity feedback data and basic data of the patient into a database; then, analyzing the data by using a graph factorization machine algorithm, so as to generate a personalized recommended systolic pressure; and finally, when the patient implements recommended pressure training, the system records data such as systolic pressure, time and frequency of the patient in real time, an adaptive adjustment model is established based on the data, and the recommended pressure is continuously optimized to adapt to the rehabilitation progress of the patient.

Description

Technical Field

[0001] The present invention relates to the technical field of rehabilitation training, and in particular to a pressure monitoring system and method based on knee joint rehabilitation. Background Art

[0002] Patients with orthopedic lower limb diseases often have reduced activity due to inflammation pain, fracture immobilization, surgical trauma, etc., which can cause joint-derived muscle inhibition, resulting in lower limb muscle atrophy, decreased muscle strength, etc., which in turn leads to reduced body activity, abnormal gait, decreased balance, etc., which have serious adverse effects on the rehabilitation of lower limbs, especially joint function. Early rehabilitation exercises can ensure the effect of surgery, ensure effective joint mobility, maintain joint stability, and gradually restore joint functional rehabilitation. Among them, the quadriceps femoris is the only knee extensor muscle of the knee joint and is considered to be the focus of rehabilitation nursing for orthopedic lower limb diseases. However, there are currently no specific requirements for the intensity, time and frequency of muscle contraction during the training of the quadriceps femoris of the lower limb knee joint in clinical nursing, lack of professional evaluation basis, and strong arbitrariness. There is no real-time dynamic and specific monitoring tool for the muscle contraction intensity and effect evaluation of the quadriceps femoris rehabilitation exercise of the knee joint.

[0003] To this end, a pressure monitoring system and method based on knee joint rehabilitation are proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a pressure monitoring system and method based on knee joint rehabilitation to achieve a scientific and reliable knee joint rehabilitation program. In order to solve the problems existing in the prior art, the present invention proposes a pressure calculation model, which obtains the contraction pressure of the quadriceps femoris in the knee joint movement in real time by stretching the fabric and the displacement and angle changes of the IMU sensor group; secondly, the present invention proposes a graph factor decomposition machine algorithm, which intelligently recommends appropriate contraction pressure by combining the patient's own basic data, basic data in the patient database, contraction pressure data and electromyographic feedback data; finally, the present invention establishes a pressure adaptive adjustment model based on real-time data, dynamically adjusts the contraction pressure, and ensures pressure optimization and patient comfort during the rehabilitation process.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A pressure monitoring system based on knee joint rehabilitation, comprising:

[0007] A first pressure acquisition module, used for acquiring physiological data generated by the quadriceps femoris when the knee joint of a healthy person contracts through a pressure acquisition device, and uploading the physiological data to a healthy person database; the physiological data includes a first contraction pressure and a first electromyographic feedback;

[0008] A pressure analysis module, configured to obtain a first contraction pressure threshold and a second contraction pressure threshold according to the physiological data;

[0009] A second pressure acquisition module is used to gradually increase the contraction pressure at the quadriceps femoris of the patient according to the first contraction pressure threshold and the second contraction pressure threshold, obtain a second contraction pressure data group and a second myoelectric feedback data group, and upload the second contraction pressure data group, the second myoelectric feedback data group and the patient's basic data to a patient database;

[0010] A pressure recommendation module, configured to obtain a recommended contraction pressure using a graph factorization machine algorithm according to the second contraction pressure data group, the second electromyography feedback data group and the patient basic data in the patient database;

[0011] An effective counting module, used to record the real-time contraction pressure, contraction time, contraction times and effective contraction times of the quadriceps femoris during the implementation phase of the recommended contraction pressure, and to display the real-time contraction pressure, the recommended contraction pressure and the effective contraction times;

[0012] The adaptive adjustment module is used to establish a pressure adaptive adjustment model according to the real-time contraction pressure, the contraction time, the number of contractions and the effective number of contractions, and adaptively update the recommended contraction pressure.

[0013] Furthermore, the pressure acquisition device includes a stretch fabric, a surface electromyography sensor and an IMU sensor group, the IMU sensor group includes a main sensor unit and multiple slave sensor units, and the slave sensor units are evenly and equidistantly distributed with the main sensor unit as the center.

[0014] Furthermore, the physiological data generated by the quadriceps femoris when the knee joint of a healthy person contracts obtained by the pressure acquisition device include:

[0015] Wrapping the pressure collection device on the quadriceps femoris of a healthy person;

[0016] collecting a first state of the slave sensor unit when the knee joint of a healthy person is in an extended state, and a second state of the slave sensor unit when the knee joint of a healthy person is in a contracted state;

[0017] Using a pressure calculation model to obtain the first contraction pressure;

[0018] Collecting the first electromyographic feedback of the surface electromyographic sensor of a healthy person in a state of knee joint contraction;

[0019] The first state includes a first distance from the slave sensor unit to the master sensor unit, and a first horizontal angle and a first vertical angle of the slave sensor unit relative to the master sensor unit; the second state includes a second distance from the slave sensor unit to the master sensor unit, and a second horizontal angle and a second vertical angle of the slave sensor unit relative to the master sensor unit; the calculation formula of the pressure calculation model is:

[0020]

[0021] Wherein, T represents the first contraction pressure, n represents the number of the slave sensing units, k represents the tensile elastic coefficient of the stretched fabric, i represents the index number of the slave sensing unit, Δd i represents the distance change between the first distance and the second distance from the sensing unit i, θ i represents the horizontal change amount of the first horizontal angle and the second horizontal angle of the slave sensing unit i, Indicates the vertical changes of the first vertical angle and the second vertical angle of the slave sensing unit i.

[0022] Furthermore, the first pressure threshold and the second pressure threshold are obtained through normal distribution, and the first contraction pressure threshold P lower =μ-2σ, the second contraction pressure threshold P upper =μ+2σ, where μ represents the mean contraction pressure and σ represents the standard deviation of pressure.

[0023] Further, obtaining the recommended contraction pressure using a graph factorization machine algorithm includes:

[0024] Constructing a graph structure according to the second contraction pressure data group, the second myoelectric feedback data group and the patient basic data in the patient database;

[0025] Using a graph neural network to embed the nodes in the graph structure, and converting the second contraction pressure data group, the second electromyography feedback data group and the patient basic data into a low-dimensional embedding vector;

[0026] Based on the embedded representation, the factorization machine algorithm is used to learn the feature interaction information between features;

[0027] The low-dimensional embedding vector processed by the graph neural network is integrated with the feature interaction information of the factorization machine to predict the electromyographic feedback data of the patient under different contraction pressures, thereby obtaining the predicted electromyographic feedback data;

[0028] Minimizing the difference between the predicted electromyographic feedback data and the second electromyographic feedback data group by optimizing the loss function;

[0029] The patient's basic data is input, and a graph factorization machine algorithm is used to obtain the recommended contraction pressure generated by the quadriceps femoris when the patient's knee joint contracts.

[0030] Furthermore, the effective counting module includes a real-time contraction pressure display device, a recommended contraction pressure display device, an effective contraction number display device and an effective number prompting device; the real-time contraction pressure display device is used to display the real-time contraction pressure; the recommended contraction pressure display device is used to display the recommended contraction pressure; the effective contraction number display device is used to display the effective contraction number, and the effective contraction number is the number of times the real-time contraction pressure is greater than the recommended contraction pressure and the contraction time is greater than the contraction time threshold; the effective number prompting device is used to prompt whether the real-time contraction pressure is greater than the recommended contraction pressure.

[0031] Furthermore, the pressure adaptive adjustment model calculation formula is:

[0032]

[0033] Among them, P recommended represents the updated recommended systolic pressure, P current represents the real-time contraction pressure, α0 represents the adjustment coefficient, ln(·) represents the logarithmic function, e represents the natural exponent, T represents the contraction time, N represents the number of contractions, N effective Indicates the effective contraction number.

[0034] Furthermore, the monitoring system further comprises:

[0035] A manual adjustment module is used to manually adjust the recommended contraction pressure when the recommended contraction pressure is not within the user comfort range, and the manual adjustment module includes an adjustment device.

[0036] A pressure monitoring method based on knee joint rehabilitation, comprising:

[0037] Acquire physiological data generated by the quadriceps femoris when the knee joint of a healthy person contracts, and upload the physiological data to a database of healthy persons; the physiological data includes a first contraction pressure and a first electromyographic feedback;

[0038] Acquire a first contraction pressure threshold and a second contraction pressure threshold according to the first contraction pressure and the first electromyographic feedback;

[0039] According to the first contraction pressure threshold and the second contraction pressure threshold, the contraction pressure at the quadriceps femoris of the patient is increased gradually, a second contraction pressure data group and a second myoelectric feedback data group are obtained, and the second contraction pressure data group, the second myoelectric feedback data group and the patient's basic data are uploaded to the patient database;

[0040] Obtaining a recommended contraction pressure using a graph factorization machine algorithm according to the second contraction pressure data group, the second electromyography feedback data group and the patient's basic data in the patient database;

[0041] The real-time contraction pressure, contraction time, number of contractions and effective number of contractions of the quadriceps femoris during the implementation of the recommended contraction pressure were recorded, and a pressure adaptive adjustment model was established based on the real-time contraction pressure, contraction time, number of contractions and effective number of contractions to adaptively update the recommended pressing pressure.

[0042] Furthermore, the method further comprises:

[0043] Used to manually adjust the recommended contraction pressure when the recommended contraction pressure is not within the user comfort range.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] 1. The pressure calculation model proposed in the present invention can obtain the contraction pressure generated by the quadriceps femoris during the contraction of the knee joint in real time, analyze the displacement and angle changes of the sensor units distributed on the muscle surface, accurately calculate the pressure distribution and contraction intensity of the quadriceps femoris in different contraction states, and combine the feedback information of the electromyographic signal to provide data support for subsequent contraction pressure recommendations.

[0046] 2. The graph factorization machine algorithm proposed in the present invention obtains the basic data of the patient and combines a large amount of basic patient data, contraction pressure data and electromyographic feedback data in the patient database to intelligently calculate the most suitable recommended contraction pressure for the patient. The algorithm uses graph neural networks to capture low-dimensional feature relationships between different features, and obtains higher-order feature interaction information through factorization mechanisms, ensuring that the recommended contraction pressure is more personalized and accurate, thereby adapting to the specific needs of patients.

[0047] 3. The pressure adaptive adjustment model proposed in the present invention can dynamically and adaptively adjust the recommended contraction pressure according to the real-time contraction pressure, contraction time, contraction number and effective contraction number of the patient's knee joint rehabilitation process. Through real-time monitoring and feedback, the model ensures that the applied contraction pressure can be adjusted individually according to the specific situation of the patient to optimize the control of contraction pressure during rehabilitation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A structural diagram of a pressure monitoring system based on knee joint rehabilitation provided by the present invention;

[0049] Figure 2 A diagram of some modules of a pressure monitoring system based on knee joint rehabilitation provided by the present invention;

[0050] Figure 3A flow chart of a pressure monitoring method based on knee joint rehabilitation provided by the present invention;

[0051] Figure numerals: 21-stretching fabric, 22-slave sensing unit, 23-master sensing unit, 24-surface electromyography sensor, 61-real-time contraction pressure display device, 62-recommended contraction pressure display device, 63-effective contraction number display device, 64-effective number prompting device, 81-adjustment device. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0053] A pressure monitoring system based on knee joint rehabilitation, such as Figure 1 As shown, including:

[0054] A first pressure acquisition module, used for acquiring physiological data generated by the quadriceps femoris when the knee joint of a healthy person contracts through a pressure acquisition device, and uploading the physiological data to a healthy person database; the physiological data includes a first contraction pressure and a first electromyographic feedback;

[0055] A pressure analysis module, configured to obtain a first contraction pressure threshold and a second contraction pressure threshold according to the physiological data;

[0056] A second pressure acquisition module is used to gradually increase the contraction pressure at the quadriceps femoris of the patient according to the first contraction pressure threshold and the second contraction pressure threshold, obtain a second contraction pressure data group and a second myoelectric feedback data group, and upload the second contraction pressure data group, the second myoelectric feedback data group and the patient's basic data to a patient database;

[0057] A pressure recommendation module, configured to obtain a recommended contraction pressure using a graph factorization machine algorithm according to the second contraction pressure data group, the second electromyography feedback data group and the patient basic data in the patient database;

[0058] An effective counting module, used to record the real-time contraction pressure, contraction time, contraction times and effective contraction times of the quadriceps femoris during the implementation phase of the recommended contraction pressure, and to display the real-time contraction pressure, the recommended contraction pressure and the effective contraction times;

[0059] The adaptive adjustment module is used to establish a pressure adaptive adjustment model according to the real-time contraction pressure, the contraction time, the number of contractions and the effective number of contractions, and adaptively update the recommended contraction pressure.

[0060] See also Figure 1 The first pressure acquisition module in:

[0061] Furthermore, if Figure 2 As shown, the pressure acquisition device includes a stretch fabric, a surface electromyography sensor and an IMU sensor group. The IMU sensor group includes a main sensor unit and a slave sensor unit. The slave sensor units are evenly and equidistantly distributed with the main sensor unit as the center.

[0062] Specifically, the IMU sensor group includes a master sensor unit and a slave sensor unit. The number of the master sensor unit is 1, and the number of the slave sensor units can be 4, 5, 6, ... 12. The slave sensor units are evenly and equidistantly distributed with the main sensor unit as the center. The surface electromyography sensor is located on the side of the main sensor unit, and there is no overlapping part between the surface electromyography sensor and the main sensor unit or the slave sensor unit. In a feasible embodiment, Figure 2 As shown, the pressure acquisition device includes a stretch fabric, a surface electromyography sensor and an IMU sensor group, wherein the IMU sensor group includes a main sensor unit and 8 slave sensor units. The slave sensor units are evenly and equidistantly distributed with the main sensor unit as the center, and the angle between every two slave sensor units and the main sensor unit is 45°.

[0063] The pressure collection device has a uniform distribution of master and slave sensor units, especially a structure with the master sensor unit as the center and arranged at equal angles, so that the device can more comprehensively and accurately monitor the stretching and displacement changes in all directions, thereby improving the measurement accuracy of tension and angle changes.

[0064] Furthermore, the physiological data generated by the quadriceps femoris when the knee joint of a healthy person contracts obtained by the pressure acquisition device include:

[0065] Wrapping the pressure collection device on the quadriceps femoris of a healthy person;

[0066] collecting a first state of the slave sensor unit when the knee joint of a healthy person is in an extended state, and a second state of the slave sensor unit when the knee joint of a healthy person is in a contracted state;

[0067] Using a pressure calculation model to obtain the first contraction pressure;

[0068] Collecting the first electromyographic feedback of the surface electromyographic sensor of a healthy person in a state of knee joint contraction;

[0069] The first state includes a first distance from the slave sensor unit to the master sensor unit, and a first horizontal angle and a first vertical angle of the slave sensor unit relative to the master sensor unit; the second state includes a second distance from the slave sensor unit to the master sensor unit, and a second horizontal angle and a second vertical angle of the slave sensor unit relative to the master sensor unit; the calculation formula of the pressure calculation model is:

[0070]

[0071] Wherein, T represents the first contraction pressure, n represents the number of the slave sensing units, k represents the tensile elastic coefficient of the stretched fabric, i represents the index number of the slave sensing unit, Δd i represents the distance change between the first distance and the second distance from the sensing unit i, θ i represents the horizontal change amount of the first horizontal angle and the second horizontal angle of the slave sensing unit i, Indicates the vertical changes of the first vertical angle and the second vertical angle of the slave sensing unit i.

[0072] By collecting the position changes of the sensor unit under the state of knee extension and contraction, the tension changes of the quadriceps femoris in different states can be accurately captured. The first contraction pressure obtained by the pressure calculation model analysis and the electromyographic signal fed back by the corresponding electromyographic sensor can fully reflect the relationship between muscle contraction intensity and surface electrical activity.

[0073] See also Figure 1 Pressure analysis module in:

[0074] Furthermore, the first pressure threshold and the second pressure threshold are obtained through normal distribution, and the first contraction pressure threshold P lower =μ-2σ, the second contraction pressure threshold P upper =μ+2σ, where μ represents the mean contraction pressure and σ represents the standard deviation of pressure.

[0075] Specifically, the EMG feedback data can reflect the patient's acceptance of the quadriceps contraction pressure during knee contraction. In a static state, the EMG feedback is between 0-10μV, in a slight muscle contraction, the EMG feedback data is in the range of 10-50μV, in a moderate contraction, the EMG feedback data is between 50-200μV, and in a strong muscle contraction, the EMG feedback data is between 200-500μV or even higher. Extract the quadriceps contraction pressure data P1, P2, ..., P from the healthy database n (first pressure data set) and the corresponding electromyographic feedback data S1, S2, ..., S n(First electromyographic feedback data set), calculate the average contraction pressure μ and standard deviation σ, where Considering that the quadriceps contraction pressure that patients can accept is different from that of healthy people, based on the normal distribution, the first contraction pressure threshold (lower pressure limit) and the second contraction pressure threshold (upper pressure limit) are obtained with a confidence level of 95%. At this time, the first contraction pressure threshold P lower =μ-2σ, the second contraction pressure threshold P upper =μ+2σ.

[0076] By using normal distribution to obtain the first contraction pressure threshold and the second contraction pressure threshold, the rationality and scientificity of the quadriceps contraction pressure range can be ensured. Calculating the threshold based on the average contraction pressure and the pressure standard deviation can effectively cover the comfortable contraction pressure range of most patients and avoid potential discomfort or damage to patients caused by too low or too high contraction pressure.

[0077] See also Figure 1 The second pressure acquisition module and pressure recommendation module in:

[0078] Further, obtaining the recommended contraction pressure using a graph factorization machine algorithm includes:

[0079] Constructing a graph structure according to the second contraction pressure data group, the second myoelectric feedback data group and the patient basic data in the patient database;

[0080] Using a graph neural network to embed the nodes in the graph structure, and converting the second contraction pressure data group, the second electromyography feedback data group and the patient basic data into a low-dimensional embedding vector;

[0081] Based on the embedded representation, the factorization machine algorithm is used to learn the feature interaction information between different features;

[0082] The low-dimensional embedding vector processed by the graph neural network is integrated with the feature interaction information of the factorization machine to predict the electromyographic feedback data of the patient under different contraction pressures, thereby obtaining the predicted electromyographic feedback data;

[0083] Minimizing the difference between the predicted electromyographic feedback data and the second electromyographic feedback data group by optimizing the loss function;

[0084] The patient's basic data is input, and the recommended contraction pressure of the patient's knee joint lesion area is obtained using a graph factorization machine algorithm.

[0085] Specifically, age, gender, weight, medical history, lesion area, lesion index, second contraction pressure data set, and second electromyography feedback data set are used as graph structures, and the relationship between nodes is used as the edge of the graph to construct a graph G = (V, E); a graph neural network is used to process the node information in the graph, and the embedding representation of the nodes in each layer of GNN is the information aggregation of a node to its neighboring nodes. The embedding representation in represents the embedding representation of node v at the kth layer, N(v) represents the set of neighbor nodes of node v, Aggregate represents the weighted sum operation, and Embedding represents h v (k) After GNN processing, the contextual information of the node is extracted and converted into a low-dimensional embedding vector; these embedding vectors are passed as input to the factorization machine layer, which is used to capture the second-order interaction of features. in <v i ,v j > is the inner product of two latent vectors, representing the interaction effect between feature i and feature j; x i x j is the product of the eigenvalues, indicating their interaction; the node representation generated by GNN is fused with the second-order interaction result of the factorization machine to obtain the prediction result Where w0 is the bias term, w i represents the linear weight of feature i, GNN(h i ) After the GNN layer is calculated, the graph structure contribution of node i; by minimizing the loss function This makes the model more accurate in prediction; when the loss function stops iterating, the quadriceps contraction pressure at the highest comfort level recommended by the graph factorization machine algorithm is obtained.

[0086] By constructing the patient's contraction pressure data, electromyographic feedback data, and basic data into a graph structure, and using a graph neural network for low-dimensional embedding representation, combined with a factor decomposition mechanism to capture the interaction between different features, it is possible to more accurately predict the patient's electromyographic feedback at different contraction pressures. Adaptive updates are performed based on big data feedback and optimized loss functions to ensure that the recommended contraction pressure is dynamically adjusted and personalized, ultimately providing each patient with a reasonable recommended quadriceps contraction pressure.

[0087] See also Figure 1 The valid counting module in:

[0088] Furthermore, if Figure 2As shown, the effective counting module includes a real-time contraction pressure display device, a recommended contraction pressure display device, an effective contraction number display device and an effective number prompting device; the real-time contraction pressure display device is used to display the real-time contraction pressure; the recommended contraction pressure display device is used to display the recommended contraction pressure; the effective contraction number display device is used to display the effective contraction number, and the effective contraction number is the number of times the real-time contraction pressure is greater than the recommended contraction pressure and the contraction time is greater than the contraction time threshold; the effective number prompting device is used to prompt whether the real-time contraction pressure is greater than the recommended contraction pressure.

[0089] In a feasible implementation scheme, the effective counting module obtains the first contraction pressure of 50N collected by the first pressure collection module in real time, that is, the real-time contraction pressure is 50N, then the real-time contraction pressure display device will display 50N on the screen; when the pressure recommendation module obtains the recommended contraction pressure, the recommended contraction pressure display device will display 60N on the screen; when the recommended contraction pressure is 60N, the contraction time threshold is 5 seconds, the real-time contraction pressure of the patient's quadriceps is 65N, and the contraction time is 10 seconds, then the effective contraction number displayed by the effective contraction number display device is increased by one; when the real-time contraction pressure of the patient's quadriceps is 65N, the real-time contraction pressure is greater than the recommended contraction pressure of 60N, then the red prompt light of the effective number prompt device lights up, and when the real-time contraction pressure of the patient's quadriceps is 55N, the real-time contraction pressure is less than the recommended contraction pressure of 60N, then the green prompt light of the effective number prompt device lights up.

[0090] The effectiveness of the real-time contraction pressure is monitored and fed back in real time through the effective counting module. The effective contraction number display device provides the number of times the real-time contraction pressure is greater than the recommended contraction pressure, helping users to intuitively understand the contraction effect of the quadriceps femoris; and the effective contraction number prompt device instantly reminds users whether the real-time contraction pressure has reached the recommended contraction pressure, ensuring that it remains within the optimal effect range during the pressing process.

[0091] See also Figure 1 Adaptive adjustment module in:

[0092] Furthermore, the pressure adaptive adjustment model calculation formula is:

[0093]

[0094] Among them, P recommended represents the updated recommended systolic pressure, P current represents the real-time contraction pressure, α0 represents the adjustment coefficient, ln(·) represents the logarithmic function, e represents the natural exponent, T represents the contraction time, N represents the number of contractions, N effective Indicates the effective contraction number.

[0095] By establishing a pressure adaptive adjustment model based on the real-time contraction pressure, contraction time, contraction number and effective contraction number during the knee joint rehabilitation stage, the recommended contraction pressure can be dynamically updated to ensure that the contraction pressure adjustment process changes accurately with the rehabilitation progress. The model continuously optimizes the contraction pressure setting based on the patient's real-time feedback and rehabilitation progress, thereby providing each patient with a more personalized and scientific contraction pressure implementation plan.

[0096] Furthermore, if Figure 1 As shown, the monitoring system also includes:

[0097] The manual adjustment module is used to manually adjust the recommended contraction pressure when the recommended contraction pressure is not within the user comfort range.

[0098] In a feasible implementation scheme, the user comfort is fed back via electromyographic feedback data, and the electromyographic feedback data corresponding to the user comfort range is [10μV, 500μV]. The adjustment device can achieve upward and downward adjustment, and the gradient of each manual adjustment is 1N.

[0099] The manual adjustment module allows users to manually adjust the recommended pressure based on personal feelings or doctor's advice through the up and down functions.

[0100] Embodiment 2

[0101] A pressure monitoring method based on knee joint rehabilitation, such as Figure 3 As shown, including:

[0102] S910: Acquire physiological data generated by the quadriceps femoris when the knee joint of a healthy person contracts, and upload the physiological data to a database of healthy persons; the physiological data includes a first contraction pressure and a first electromyographic feedback;

[0103] S920: Acquire a first contraction pressure threshold and a second contraction pressure threshold according to the first contraction pressure and the first electromyographic feedback;

[0104] S930: Gradually increase the contraction pressure at the quadriceps femoris of the patient according to the first contraction pressure threshold and the second contraction pressure threshold, obtain a second contraction pressure data group and a second myoelectric feedback data group, and upload the second contraction pressure data group, the second myoelectric feedback data group and the patient's basic data to the patient database;

[0105] S940: Obtaining a recommended contraction pressure using a graph factorization machine algorithm according to the second contraction pressure data group, the second electromyography feedback data group, and the patient's basic data in the patient database;

[0106] S950: Record the real-time contraction pressure, contraction time, contraction times and effective contraction times of the quadriceps femoris during the implementation of the recommended contraction pressure, and establish a pressure adaptive adjustment model based on the real-time contraction pressure, contraction time, contraction times and effective contraction times to adaptively update the recommended pressing pressure.

[0107] Furthermore, the pressure acquisition device includes a stretch fabric, a surface electromyography sensor and an IMU sensor group, the IMU sensor group includes a main sensor unit and multiple slave sensor units, and the slave sensor units are evenly and equidistantly distributed with the main sensor unit as the center.

[0108] Furthermore, the physiological data generated by the quadriceps femoris when the knee joint of a healthy person contracts obtained by the pressure acquisition device include:

[0109] Wrapping the pressure collection device on the quadriceps femoris of a healthy person;

[0110] collecting a first state of the slave sensor unit when the knee joint of a healthy person is in an extended state, and a second state of the slave sensor unit when the knee joint of a healthy person is in a contracted state;

[0111] Using a pressure calculation model to obtain the first contraction pressure;

[0112] Collecting the first electromyographic feedback of the surface electromyographic sensor of a healthy person in a state of knee joint contraction;

[0113] The first state includes a first distance from the slave sensor unit to the master sensor unit, and a first horizontal angle and a first vertical angle of the slave sensor unit relative to the master sensor unit; the second state includes a second distance from the slave sensor unit to the master sensor unit, and a second horizontal angle and a second vertical angle of the slave sensor unit relative to the master sensor unit; the calculation formula of the pressure calculation model is:

[0114]

[0115] Wherein, T represents the first contraction pressure, n represents the number of the slave sensing units, k represents the tensile elastic coefficient of the stretched fabric, i represents the index number of the slave sensing unit, Δd i represents the distance change between the first distance and the second distance from the sensing unit i, θ i represents the horizontal change amount of the first horizontal angle and the second horizontal angle of the slave sensing unit i, Indicates the vertical changes of the first vertical angle and the second vertical angle of the slave sensing unit i.

[0116] Furthermore, the first pressure threshold and the second pressure threshold are obtained through normal distribution, and the first contraction pressure threshold Plower =μ-2σ, the second contraction pressure threshold P upper =μ+2σ, where μ represents the mean contraction pressure and σ represents the standard deviation of pressure.

[0117] Further, obtaining the recommended contraction pressure using a graph factorization machine algorithm includes:

[0118] Constructing a graph structure according to the second contraction pressure data group, the second myoelectric feedback data group and the patient basic data in the patient database;

[0119] Using a graph neural network to embed the nodes in the graph structure, and converting the second contraction pressure data group, the second electromyography feedback data group and the patient basic data into a low-dimensional embedding vector;

[0120] Based on the embedded representation, the factorization machine algorithm is used to learn the feature interaction information between features;

[0121] The low-dimensional embedding vector processed by the graph neural network is integrated with the feature interaction information of the factorization machine to predict the electromyographic feedback data of the patient under different contraction pressures, thereby obtaining the predicted electromyographic feedback data;

[0122] Minimizing the difference between the predicted electromyographic feedback data and the second electromyographic feedback data group by optimizing the loss function;

[0123] The patient's basic data is input, and a graph factorization machine algorithm is used to obtain the recommended contraction pressure generated by the quadriceps femoris when the patient's knee joint contracts.

[0124] Furthermore, the effective counting module includes a real-time contraction pressure display device, a recommended contraction pressure display device, an effective contraction number display device and an effective number prompting device; the real-time contraction pressure display device is used to display the real-time contraction pressure; the recommended contraction pressure display device is used to display the recommended contraction pressure; the effective contraction number display device is used to display the effective contraction number, and the effective contraction number is the number of times the real-time contraction pressure is greater than the recommended contraction pressure and the contraction time is greater than the contraction time threshold; the effective number prompting device is used to prompt whether the real-time contraction pressure is greater than the recommended contraction pressure.

[0125] Furthermore, the pressure adaptive adjustment model calculation formula is:

[0126]

[0127] Among them, P recommended represents the updated recommended systolic pressure, P currentrepresents the real-time contraction pressure, α0 represents the adjustment coefficient, ln(·) represents the logarithmic function, e represents the natural exponent, T represents the contraction time, N represents the number of contractions, N effective Indicates the effective contraction number.

[0128] Furthermore, if Figure 3 As shown, the monitoring method also includes:

[0129] S960: When the recommended contraction pressure is not within the user's comfort range, manually adjust the recommended contraction pressure.

[0130] Embodiment 3

[0131] A patient has a knee injury due to excessive exercise. In order to recover as soon as possible, effective quadriceps contraction pressure is needed to control the contraction of the knee joint, thereby assisting the joint to recover quickly. Therefore, a pressure monitoring system based on knee joint rehabilitation proposed in the present invention is adopted.

[0132] Table 1: First state and second state of the slave sensor unit

[0133]

[0134] The first contraction pressure generated by the quadriceps femoris when the knee joint of a healthy person contracts is obtained by a pressure acquisition device, wherein the pressure acquisition device includes a stretch fabric, a surface electromyography sensor and an IMU sensor group; the elastic coefficient of the stretch fabric is 120N / cm, and the IMU sensor group includes 1 main sensor unit and 8 slave sensor units, and the slave sensor units are evenly and equidistantly distributed with the main sensor unit as the center and a radius of 5cm; the first state of the slave sensor unit in the state of straightening the knee joint of a healthy person and the second state of the slave sensor unit in the state of contraction of the knee joint of a healthy person are collected, wherein the data of the first state and the second state of the slave sensor unit are shown in Table 1; according to the tension calculation formula Calculation shows that the first contraction pressure at this moment is 55N.

[0135] The first contraction pressure data and the first electromyographic feedback data generated by the quadriceps femoris when the knee joint of the healthy person contracts are collected, and the first contraction pressure data and the first electromyographic feedback data are uploaded to the healthy person database, wherein some data of the healthy person database are shown in Table 2. According to the first contraction pressure data and the first electromyographic feedback data in the healthy person database, the first contraction pressure threshold and the second contraction pressure threshold are obtained; the first contraction pressure threshold and the second contraction pressure threshold are obtained through normal distribution, and the first contraction pressure threshold P lower =μ-2σ, the second contraction pressure threshold P upper =μ+2σ, the first contraction pressure threshold is 30N and the second contraction pressure threshold is 90N.

[0136] Table 2. Some first contraction pressure data and first electromyographic feedback data in the healthy subjects database

[0137] Healthy person number First contraction pressure (N) First EMG feedback(μV) 1 20 10 1 25 30 ... ... ... 1 60 150 1 65 160 ... ... ... 2 70 135 ... ... ...

[0138] According to the first contraction pressure threshold and the second contraction pressure threshold, the quadriceps contraction pressure of the patient's knee contraction is gradually increased with a gradient of 5N, and the second contraction pressure data group and the second electromyographic feedback data group are obtained. The second contraction pressure data group, the second electromyographic feedback data group and the patient's basic data are uploaded to the patient database. Part of the patient database data is shown in Table 3. According to the second contraction pressure data group, the second electromyographic feedback data group and the patient's basic data in the patient database, the recommended pressing pressure is obtained using the graph factor decomposition machine recommendation algorithm; the current patient's basic data is 28 years old, male, 75KG, and no previous medical history. Based on the patient database and the graph factor decomposition machine recommendation algorithm, the recommended quadriceps contraction pressure is 60N.

[0139] During the implementation phase of the recommended contraction pressure, the pressure monitoring system based on knee joint rehabilitation proposed by the present invention provides a real-time contraction pressure P current The contraction time T is 36 minutes, the number of contractions N is 150 times and the effective number of contractions N is 60N. effective The pressure is 120 times; according to the pressing pressure, pressing time, pressing times and effective times in the knee joint rehabilitation stage, a pressure adaptive adjustment model is established to adaptively update the recommended pressing pressure. The calculation formula of the pressure adaptive adjustment model is:

[0140]

[0141] Adjustment coefficient α0 = 0.8, substitute the recorded data to get:

[0142]

[0143] The patient's electromyographic feedback on the recommended contraction pressure is 100, and the patient's comfort range is [130,200]. Therefore, the recommended contraction pressure can be manually adjusted to obtain a better experience.

[0144] Table 3. Partial data of patient database

[0145]

[0146] The recommended contraction pressures and the corresponding electromyographic feedback under the contraction pressures for a group of patients are shown in Table 4. It can be seen from the table that the recommended contraction pressures are all within the comfort range of the patients, indicating that the system and method proposed in the present invention can accurately recommend quadriceps contraction pressures that meet the needs of patients based on contraction pressure monitoring data and patient basic data.

[0147] Table 4. Recommended contraction pressure for patients and electromyographic feedback at corresponding contraction pressure

[0148]

[0149] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A pressure monitoring system based on knee joint rehabilitation, characterized in that: include: A first pressure acquisition module, used for acquiring physiological data generated by the quadriceps femoris of a healthy person when the knee joint contracts through a pressure acquisition device, and uploading the physiological data to a healthy person database; The physiological data includes a first contraction pressure and a first myoelectric feedback; A pressure analysis module, configured to obtain a first contraction pressure threshold and a second contraction pressure threshold according to the physiological data; A second pressure acquisition module is used to gradually increase the contraction pressure at the quadriceps femoris of the patient according to the first contraction pressure threshold and the second contraction pressure threshold, obtain a second contraction pressure data group and a second myoelectric feedback data group, and upload the second contraction pressure data group, the second myoelectric feedback data group and the patient's basic data to a patient database; A pressure recommendation module, configured to obtain a recommended contraction pressure using a graph factorization machine algorithm according to the second contraction pressure data group, the second electromyography feedback data group and the patient basic data in the patient database; An effective counting module, used to record the real-time contraction pressure, contraction time, contraction times and effective contraction times of the quadriceps femoris during the implementation phase of the recommended contraction pressure, and to display the real-time contraction pressure, the recommended contraction pressure and the effective contraction times; The adaptive adjustment module is used to establish a pressure adaptive adjustment model according to the real-time contraction pressure, the contraction time, the number of contractions and the effective number of contractions, and adaptively update the recommended contraction pressure.

2. A pressure monitoring system based on knee joint rehabilitation according to claim 1, characterized in that: The pressure acquisition device includes a stretch fabric, a surface electromyography sensor and an IMU sensor group. The IMU sensor group includes a main sensor unit and multiple slave sensor units. The slave sensor units are evenly and equidistantly distributed with the main sensor unit as the center.

3. A pressure monitoring system based on knee joint rehabilitation according to claim 1, characterized in that: The physiological data generated by the quadriceps femoris when the knee joint of a healthy person contracts through the pressure acquisition device include: Wrapping the pressure collection device on the quadriceps femoris of a healthy person; collecting a first state of the slave sensor unit when the knee joint of a healthy person is in an extended state, and a second state of the slave sensor unit when the knee joint of a healthy person is in a contracted state; Using a pressure calculation model to obtain the first contraction pressure; Collecting the first electromyographic feedback of the surface electromyographic sensor of a healthy person in a state of knee joint contraction; The first state includes a first distance from the slave sensor unit to the master sensor unit, and a first horizontal angle and a first vertical angle of the slave sensor unit relative to the master sensor unit; the second state includes a second distance from the slave sensor unit to the master sensor unit, and a second horizontal angle and a second vertical angle of the slave sensor unit relative to the master sensor unit; the calculation formula of the pressure calculation model is: Wherein, T represents the first contraction pressure, n represents the number of the slave sensing units, k represents the tensile elastic coefficient of the stretched fabric, i represents the index number of the slave sensing unit, Δd i represents the distance change between the first distance and the second distance from the sensing unit i, θ i represents the horizontal change of the first horizontal angle and the second horizontal angle of the slave sensing unit i, φ i Indicates the vertical changes of the first vertical angle and the second vertical angle of the slave sensing unit i.

4. A pressure monitoring system based on knee joint rehabilitation according to claim 1, characterized in that: The first pressure threshold and the second pressure threshold are obtained through normal distribution. The first contraction pressure threshold P lower =μ-2σ, the second contraction pressure threshold P upper =μ+2σ, where μ represents the mean contraction pressure and σ represents the standard deviation of pressure.

5. A pressure monitoring system based on knee joint rehabilitation according to claim 1, characterized in that: Obtaining the recommended contraction pressure using the graph factorization machine algorithm includes: Constructing a graph structure according to the second contraction pressure data group, the second myoelectric feedback data group and the patient basic data in the patient database; Using a graph neural network to embed the nodes in the graph structure, and converting the second contraction pressure data group, the second electromyography feedback data group and the patient basic data into a low-dimensional embedding vector; Based on the embedded representation, the factorization machine algorithm is used to learn the feature interaction information between features; The low-dimensional embedding vector processed by the graph neural network is integrated with the feature interaction information of the factorization machine to predict the electromyographic feedback data of the patient under different contraction pressures, thereby obtaining the predicted electromyographic feedback data; Minimizing the difference between the predicted electromyographic feedback data and the second electromyographic feedback data group by optimizing the loss function; The patient's basic data is input, and a graph factorization machine algorithm is used to obtain the recommended contraction pressure generated by the quadriceps femoris when the patient's knee joint contracts.

6. A pressure monitoring system based on knee joint rehabilitation according to claim 1, characterized in that: The effective counting module includes a real-time contraction pressure display device, a recommended contraction pressure display device, an effective contraction number display device and an effective number prompt device; the real-time contraction pressure display device is used to display the real-time contraction pressure; the recommended contraction pressure display device is used to display the recommended contraction pressure; The effective contraction number display device is used to display the effective contraction number, and the effective contraction number is the number of times the real-time contraction pressure is greater than the recommended contraction pressure and the contraction time is greater than the contraction time threshold; The effective number prompting device is used to prompt whether the real-time contraction pressure is greater than the recommended contraction pressure.

7. A pressure monitoring system based on knee joint rehabilitation according to claim 1, characterized in that: The pressure adaptive adjustment model calculation formula is: Among them, P recommended represents the updated recommended systolic pressure, P current represents the real-time contraction pressure, α0 represents the adjustment coefficient, ln(·) represents the logarithmic function, e represents the natural exponent, T represents the contraction time, N represents the number of contractions, N effective Indicates the effective contraction number.

8. The pressure monitoring system based on knee joint rehabilitation according to claim 1, characterized in that: The monitoring system further comprises: A manual adjustment module is used to manually adjust the recommended contraction pressure when the recommended contraction pressure is not within the user comfort range, and the manual adjustment module includes an adjustment device.

9. A pressure monitoring method based on knee joint rehabilitation, characterized in that: include: Acquire physiological data generated by the quadriceps muscle when the knee joint of a healthy person contracts, and upload the physiological data to a healthy person database; Physiological data include first contraction pressure and first electromyographic feedback; Acquire a first contraction pressure threshold and a second contraction pressure threshold according to the first contraction pressure and the first electromyographic feedback; According to the first contraction pressure threshold and the second contraction pressure threshold, the contraction pressure at the quadriceps femoris of the patient is increased gradually, a second contraction pressure data group and a second myoelectric feedback data group are obtained, and the second contraction pressure data group, the second myoelectric feedback data group and the patient's basic data are uploaded to the patient database; Obtaining a recommended contraction pressure using a graph factorization machine algorithm according to the second contraction pressure data group, the second electromyography feedback data group and the patient's basic data in the patient database; The real-time contraction pressure, contraction time, number of contractions and effective number of contractions of the quadriceps femoris during the implementation of the recommended contraction pressure were recorded, and a pressure adaptive adjustment model was established based on the real-time contraction pressure, contraction time, number of contractions and effective number of contractions to adaptively update the recommended pressing pressure.

10. A pressure monitoring method based on knee joint rehabilitation according to claim 9, characterized in that: The method further comprises: When the recommended contraction pressure is not within the user's comfort range, manually adjust the recommended contraction pressure.

Images