Ankle-foot orthosis interactive design method and device based on digital twinning

Abstract

The present application belongs to the technical field of ankle rehabilitation, and relates to an interactive design method and device for an ankle-foot orthosis based on digital twinning, which comprises the following steps: obtaining structural data of an ankle-foot and an ankle-foot orthosis and establishing a three-dimensional model; obtaining real-time plantar pressure distribution data and mapping the data onto the three-dimensional model to form a digital twinning model containing dynamic behavior information; extracting plantar pressure features in the digital twinning model and using the plantar pressure features to construct a gait evaluation model and a gait anomaly detection model, the gait evaluation model being used to reflect the stability and pressure distribution of gait, and the gait anomaly detection model being used to detect abnormal gait in real time; when the gait evaluation model is lower than a preset threshold or the gait anomaly detection model detects abnormal gait, feedback information is generated and sent to the ankle-foot orthosis for parameter adjustment of the ankle-foot orthosis. The present application can make the ankle-foot orthosis better adapt to individual differences of users and provide personalized rehabilitation support for users.

Description

Technical Field

[0001] This invention belongs to the field of ankle joint rehabilitation technology, specifically relating to an interactive design method and device for ankle-foot orthotics based on digital twins. Background Technology

[0002] Ankle injuries are a common health problem associated with weakness, muscle dysfunction, spasticity, or limited range of motion. Rehabilitation is necessary to improve a user's quality of life and motor function, and ankle-foot orthoses are common medical devices used to correct ankle and foot injuries, provide support, improve gait, and promote rehabilitation. Based on different design principles, ankle-foot orthoses on the market are mainly divided into three categories: passive, quasi-passive, and active. Passive ankle-foot orthoses do not require any external power source to generate assistance, typically relying on passive elements such as springs, dampers, or elastic bands to provide auxiliary function. Their main advantages are light weight and low cost. However, in most cases, the design parameters of the passive elements can only be optimized under certain special circumstances. Quasi-passive ankle-foot orthoses use pressure sensors and low-energy small motors to adjust the stiffness or damping of the passive elements to improve the functional efficiency of the ankle-foot orthose, which enhances the adaptability of passive ankle-foot orthoses to changes in gait or environment. Due to the limited adaptability and controllability of passive components, they cannot achieve certain motion control and assistance conditions. Therefore, control strategies such as rigid transmission devices, cable drives, and pneumatic muscles are used for the active control of ankle-foot orthoses. The main advantage of active ankle-foot orthoses is that they can provide a large amount of assistance with fully controllable amplitude and time, allowing them to adapt to different speeds or ground conditions. However, their efficiency depends on the performance of the controller and actuator.

[0003] Although the design of existing ankle-foot orthoses has made considerable progress, since they are designed to help users rehabilitate, they often prioritize assisted walking ability. Their structure and function are mainly focused on supporting and restricting ankle joint movement to assist or correct abnormalities in the foot and ankle area. Most passive and semi-passive ankle-foot orthoses can only support dorsiflexion and plantarflexion, which may restrict the user's natural gait and make walking less natural and smooth. At the same time, in order to reduce costs, one type of ankle-foot orthosis is provided to a large number of users, with only simple adjustments to the height and tightness of the orthosis. This results in poor plantar fit of the orthosis, making it difficult to fully adapt to the individual needs of each user and leading to a poor user experience. Summary of the Invention

[0004] In view of this, the present invention provides an interactive design method and device for ankle and foot orthotics based on digital twins, which can improve the plantar fit of ankle and foot orthotics, make ankle and foot orthotics better adapt to individual differences and personalized needs of users, and provide good comfort for users even after long-term use, with good corrective effect and user experience.

[0005] The technical solution of this invention is:

[0006] An interactive design method for ankle-foot orthoses based on digital twins includes the following steps:

[0007] Obtain structural data of the ankle and foot and ankle-foot orthoses;

[0008] A three-dimensional model was built based on the structural data of the ankle and foot and ankle-foot orthotics.

[0009] Real-time pressure distribution data of the soles of the feet is acquired and mapped onto a 3D model to form a digital twin model containing dynamic behavioral information;

[0010] Plantar pressure features are extracted from the digital twin model, and these features are used to construct a gait evaluation model y and a gait anomaly detection model p. c The gait evaluation model y is used to reflect the stability and stress distribution of gait, and the gait anomaly detection model p c Used for real-time detection of abnormal gait;

[0011] When the gait assessment model y is below a preset threshold or the gait abnormality detection model detects a gait abnormality, feedback information is generated and sent to the ankle-foot orthosis for parameter adjustment.

[0012] Preferably, a three-dimensional model is established based on the structural data of the ankle and foot and the ankle-foot orthosis, including the following steps:

[0013] Acquire point cloud data of the ankle and foot from a 3D scanner and point cloud data of the ankle-foot orthosis;

[0014] A three-dimensional model is obtained by performing three-dimensional reconstruction based on the point cloud data of the ankle and foot and the point cloud data of the ankle and foot orthosis.

[0015] Preferably, acquiring real-time plantar pressure distribution data and mapping it onto a 3D model to form a digital twin model containing dynamic behavioral information includes the following steps:

[0016] Preprocess real-time plantar pressure distribution data;

[0017] The preprocessed real-time plantar pressure distribution data is mapped onto a three-dimensional model to form a digital twin model.

[0018] Preferably, the preprocessing includes the following steps:

[0019] Kalman filtering was used to denoise the real-time plantar pressure distribution data.

[0020] The real-time plantar pressure distribution data after noise reduction is standardized.

[0021] Preferably, the plantar pressure feature includes a pressure center point (x c (t),y c (t) and maximum pressure value P max (t), pressure distribution area A(t), gait cycle pressure change rate ΔP(t), and pressure distribution variance

[0022] Preferably, a gait assessment model y and a gait anomaly detection model p are constructed using plantar pressure characteristics. c This includes the following steps:

[0023] The plantar pressure features are used to construct a vector X.

[0024] Preprocess vector X to obtain plantar pressure feature vector Z;

[0025] A neural network is constructed, comprising an input layer, a first hidden layer, a second hidden layer, and an output layer. The input layer is used to input the plantar pressure feature vector Z. The first hidden layer is a fully connected layer with 128 neurons, and the activation function is the Modified Linear Unit (ReLU), calculated as: h1 = ReLU(W1Z + b1), where W1 is a 128x6 weight matrix, b1 is a 128x1 bias vector, and h1 is the output of the first hidden layer of the neural network model. The second hidden layer is a fully connected layer with 64 neurons, and the activation function is the Modified Linear Unit (ReLU), calculated as: h2 = ReLU(W2h1) +b2), W2 is a 64x128 weight matrix, b2 is a 64x1 bias vector, h1 is the output of the first hidden layer of the neural network model, and h2 is the output of the second hidden layer of the neural network model; when the output layer is used for gait evaluation, the number of neurons in the output layer is 1, the activation function is linear, and the gait evaluation model is y = W0h2 + b0, where y is the gait evaluation model, representing the evaluation score, W0 is the weight matrix, h2 is the output of the second hidden layer of the neural network model, and b0 is the bias matrix; when the output layer is used for gait anomaly detection, the number of neurons in the output layer is 2, the activation function is Softmax, and the gait anomaly detection model is p c =Softmax(W c h2+b c ), p c Let W be the gait anomaly detection model, representing the gait detection result. c Here, h1 is the weight matrix, h2 is the output of the second hidden layer of the neural network model, and b... c Here, is the bias vector, and Softmax is the activation function of the output layer of the neural network model.

[0026] Preferably, the feedback information is determined based on the following formula:

[0027]

[0028] Where score(t) is the anomaly level at time t, α and β are adjustment coefficients, and P base It is the maximum pressure reference value, A base It is the reference value for the pressure distribution area, P max (t) is the maximum pressure value, and A(t) is the pressure distribution area.

[0029] An electronic device includes a storage device, a processor, and a computer program stored on the storage device and capable of running on the processor, wherein the processor executes the computer program to implement the digital twin-based interactive design method for ankle-foot orthotics as described above.

[0030] A computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the digital twin-based interactive design method for ankle-foot orthotics as described above.

[0031] Compared with existing technologies, this invention provides an interactive design method and device for ankle-foot orthoses based on digital twins. By integrating pressure sensor technology, digital twin technology, and human-computer interaction design methods, it achieves real-time acquisition, analysis, and feedback of user foot data. This improves the plantar fit of the ankle-foot orthosis, allowing it to better adapt to individual user differences and provide personalized rehabilitation support. It also offers excellent comfort even with prolonged use, resulting in good corrective effects and a superior user experience, thus enhancing the functionality and user experience of the ankle-foot orthosis. Specifically, this invention's solution uses precise modeling to accurately match individual user differences, optimizing the design of the ankle-foot orthosis. It fully utilizes the real-time monitoring and feedback functions of digital twin technology to meet personalized rehabilitation needs, improve user experience, and demonstrates strong practicality, making it worthy of widespread adoption. Attached Figure Description

[0032] Figure 1 This is the main flowchart of the present invention.

[0033] Figure 2 The process flow of this invention Figure 1 .

[0034] Figure 3 The process flow of this invention Figure 2 .

[0035] Figure 4 The process flow of this invention Figure 3 .

[0036] Figure 5 This is a layout diagram of the pressure sensor of the present invention. Detailed Implementation

[0037] Currently, ankle-foot orthoses on the market are mainly divided into three categories: passive, quasi-passive, and active. Passive ankle-foot orthoses refer to those that provide assistance without any external power source, typically relying on passive components such as springs, dampers, or elastic bands to provide auxiliary functions. Their main advantages are light weight and low cost. However, in most cases, the design parameters of passive components can only be optimized under specific conditions. Quasi-passive ankle-foot orthoses use pressure sensors and low-energy small motors to adjust the stiffness or damping of passive components to improve the functional efficiency of the orthoses, enhancing their adaptability to gait or environmental changes. Due to the limited adaptability and controllability of passive components, they cannot achieve certain motion control and assistance conditions. Therefore, control strategies such as rigid transmission devices, cable drives, and pneumatic muscles are used for active control of the ankle-foot orthoses. The main advantage of active ankle-foot orthoses is their ability to provide a large amount of assistance with fully controllable amplitude and time, allowing them to adapt to different speeds or ground conditions; however, their efficiency depends on the performance of the controller and actuator.

[0038] Although the design of existing ankle-foot orthoses has made considerable progress, since they are designed to help users rehabilitate, they often prioritize assisted walking ability. Their structure and function are mainly focused on supporting and restricting ankle joint movement to assist or correct abnormalities in the foot and ankle area. Most passive and semi-passive ankle-foot orthoses can only support dorsiflexion and plantarflexion, which may restrict the user's natural gait and make walking less natural and smooth. At the same time, in order to reduce costs, one type of ankle-foot orthosis is provided to a large number of users, with only simple adjustments to the height and tightness of the orthosis. This results in poor fit of the ankle-foot orthosis, making it difficult to fully adapt to the individual needs of each user. Users experience discomfort or poor corrective effect during prolonged use, resulting in a poor user experience.

[0039] To address the technical problems of current ankle-foot orthoses that restrict users' natural gait, making walking difficult and unnatural, poor fit and comfort, difficulty in fully adapting to each user's individual needs, discomfort or poor corrective effect during prolonged use, and poor user experience, this application presents a technical solution.

[0040] To enable those skilled in the art to better understand and implement the technical solutions of the present invention, the present invention will be further described below in conjunction with specific embodiments and accompanying drawings.

[0041] This invention provides an interactive design method and device for ankle-foot orthotics based on digital twins, which will be described below in conjunction with... Figures 1 to 5 The present invention is illustrated by the schematic diagram.

[0042] Example 1

[0043] like Figure 1 The flowchart shown illustrates an interactive design method and apparatus for ankle-foot orthoses based on digital twins, provided by this invention, comprising the following steps:

[0044] Obtain structural data of the ankle and foot and ankle-foot orthotics.

[0045] A three-dimensional model was built based on the structural data of the ankle and foot and ankle-foot orthotics.

[0046] Real-time pressure distribution data of the soles of the feet is acquired and mapped onto a 3D model to form a digital twin model containing dynamic behavioral information.

[0047] Plantar pressure features are extracted from the digital twin model, and these features are used to construct a gait evaluation model y and a gait anomaly detection model p. c The gait evaluation model y is used to reflect the stability and stress distribution of gait, and the gait anomaly detection model p c Used for real-time detection of abnormal gait. When the gait evaluation model y is lower than a preset threshold or the gait abnormality detection model detects gait abnormality, feedback information is generated and sent to the ankle-foot orthosis for parameter adjustment.

[0048] In the interactive design of ankle-foot orthoses, each component has a physical entity and a corresponding virtual model. The physical entity is the actual hardware part of the component, while the virtual model is a digital image created through pressure sensors, data acquisition, and modeling tools. It consists of two parts: a geometric model and a dynamic behavior model, and is used to simulate, analyze, and adjust the performance of the barefoot ankle-foot orthosis.

[0049] Combination Figures 1 to 4 The flowchart shown illustrates the key steps of the interactive design method for ankle-foot orthoses, and the specific implementation methods are as follows:

[0050] 1. Obtain structural data of the user's ankle, foot, and ankle-foot orthosis.

[0051] Specifically, a 3D scanner can be used to scan the user's ankle and foot and ankle-foot orthosis separately, thereby collecting structural data on the user's ankle and foot and ankle-foot orthosis.

[0052] 2. Based on the structural data of the ankle and foot and ankle-foot orthotics, a three-dimensional model was established.

[0053] The point cloud data of the ankle and foot and the point cloud data of the ankle-foot orthosis from the 3D scanner are acquired as raw data. Then, the point cloud data of the ankle and foot and the point cloud data of the ankle-foot orthosis are imported into Rhino software. The Rhino software is used to perform 3D reconstruction to generate a geometric model of the ankle and foot and the ankle-foot orthosis. The geometric model of the ankle and foot and the ankle-foot orthosis is a high-precision 3D model created based on the actual size and structure of the ankle and foot and the ankle-foot orthosis, which corresponds precisely to the physical entity in shape and size.

[0054] 3. Acquire real-time plantar pressure distribution data and map it onto a 3D model to form a digital twin model containing dynamic behavioral information.

[0055] Multiple pressure sensors placed on the sole of the foot are used to collect real-time pressure distribution data of the sole under different gait conditions.

[0056] The pressure sensor used is a flexible pressure sensor, such as a row a and column b flexible pressure sensor. Figure 5 The arrangement shown indicates that the pressure data from the pressure sensor at time t is distributed as a vector F(t).

[0057]

[0058] Where i takes the value of an integer between 1 and a, j takes the value of an integer between 1 and b, a and b are the number of rows and columns of the pressure sensor, respectively. Since the sensing range and measurement accuracy of each model of flexible pressure sensor are different, a and b are determined based on the model of the flexible pressure sensor used.

[0059] To facilitate data processing using Kalman filtering, F(t) is converted into an a*b row, 1 column matrix, represented as a vector P(t).

[0060]

[0061] Where, p i,j (t) represents the plantar pressure value of the pressure sensor in the i-th row and j-th column at time t.

[0062] Let the state equation of the state-space model be P(t) = A·P(t-1) + w(t).

[0063] Where P(t) is the plantar pressure distribution vector at time t, P(t-1) is the plantar pressure distribution vector at time t-1, A is the state transition matrix, which is the identity matrix, and w(t) is the Gaussian process noise with a mean of 0 and a covariance of Q.

[0064] The measurement equation for the state-space model is z(t) = H·P(t) + v(t).

[0065] Where z(t) is the noisy measurement value at time t, which is the real-time plantar pressure distribution data collected by the pressure sensor, P(t) is the plantar pressure distribution vector at time t, H is the measurement matrix, which is the identity matrix, and v(t) is the Gaussian measurement noise with a mean of 0 and a covariance of R.

[0066] To eliminate noise, the real-time plantar pressure distribution data needs to be preprocessed. The preprocessing of real-time plantar pressure distribution data mainly includes noise reduction and standardization.

[0067] The noise reduction method uses Kalman filtering, and the filtering process is as follows;

[0068] (1) Initialization:

[0069] X(0|0)=I, Q=0.1·I, R=0.5·I,

[0070] In this matrix, the number of elements in P′(0|0), X(0|0), Q, and R is a*b, and I is the identity matrix.

[0071] (2) Prediction:

[0072] 1) Prediction of plantar pressure status:

[0073] P′(t|t-1)=A·P′(t-1|t-1),

[0074] Where P′(t|t-1) is the predicted state at time t, A is the state transition matrix, which is the identity matrix, and P′(t-1|t-1) is the pressure state estimate at time t-1, that is, the denoised plantar pressure distribution estimate at time t-1.

[0075] 2) Error covariance prediction:

[0076] X(t|t-1)=A·X(t-1|t-1)·A T +Q,

[0077] Where X(t|t-1) is the prediction covariance matrix, A is the state transition matrix, which is the identity matrix, X(t-1|t-1) is the covariance matrix at time t-1 after the update, and Q is the covariance matrix of the process noise.

[0078] (3) Update:

[0079] 1) Kalman gain calculation:

[0080] K(t)=X(t|t-1)·H T ·[H·X(t|t-1)·H T +R] -1 ,

[0081] Where K(t) is the Kalman gain, reflecting the influence of the measurement on the state estimation, R is the covariance matrix of the measurement noise, H is the measurement matrix, and X(t|t-1) is the prediction covariance matrix.

[0082] 2) Plantar pressure status update:

[0083] P′(t|t)=P′(t|t-1)+K(t)·[z(t)-H·P′(t|t-1)],

[0084] Where z(t) is the noisy measurement value, i.e. the real-time plantar pressure distribution data collected by the pressure sensor, P′(t|t) is the updated state estimate, i.e. the denoised plantar pressure distribution estimate, P′(t|t-1) is the pressure prediction state at time t, H is the measurement matrix, and K(t) is the Kalman gain.

[0085] 3) Error covariance update:

[0086] X(t|t)=(IK(t)·H)·X(t|t-1),

[0087] Where X(t|t) is the covariance matrix at time t after the update, I is the identity matrix, H is the measurement matrix, K(t) is the Kalman gain, and X(t|t-1) is the prediction covariance matrix.

[0088] The standardization process is as follows:

[0089]

[0090] in, Let p′ be the pressure value of the pressure sensor in the i-th row and j-th column at time t, after standardization. i,j (t|t) represents the pressure value of the pressure sensor in the i-th row and j-th column after noise reduction at time t, P max (t) represents the maximum value of the pressure after noise reduction of all pressure sensors at time t, P min (t) represents the minimum pressure value after noise reduction of all pressure sensors at time t.

[0091] P max (t) and P min The formula for calculating (t) is as follows:

[0092] P max (t) = max{P′(t|t)},

[0093] P min (t)=min{P′(t|t)}.

[0094] To map pressure sensor data to a digital twin model and establish the same layout as the physical pressure sensor, a linear mapping is used to directly map the data. The data is projected onto the corresponding points of the digital twin model, enabling the digital twin model to receive and process pressure sensor data, reflecting the dynamic behavior of each component in real time, such as the rate of change of pressure during gait cycles and pressure distribution.

[0095] The physical entity of the pressure sensor and its corresponding virtual model constitute a digital twin unit. Through the interaction between the physical entity and the virtual model, the digital twin unit enables intelligent feedback and adaptive adjustment of the barefoot ankle-foot orthosis.

[0096] 4. Extract plantar pressure features from the digital twin model, and use these features to construct a gait evaluation model y and a gait anomaly detection model p. c The gait evaluation model y is used to reflect the stability and stress distribution of gait, and the gait anomaly detection model p c Used for real-time detection of abnormal gait. When the gait evaluation model y is lower than a preset threshold or the gait abnormality detection model detects gait abnormality, feedback information is generated and sent to the ankle-foot orthosis for parameter adjustment and wearing suggestions are synchronized on the mobile terminal.

[0097] Extracting plantar pressure features mainly includes the pressure center point (x c (t),y c (t) and maximum pressure value P max (t), pressure distribution area A(t), gait cycle pressure change rate ΔP(t), and pressure distribution variance The calculation methods for each feature are as follows:

[0098] (1) Pressure center point (x) c (t),y c (t)):

[0099] After noise reduction, the position of the pressure sensor in the i-th row and j-th column of the plantar pressure sensor array is (x i ,y j The pressure value at time t is Then the pressure center point (x) c (t),y c The calculation method for (t) is as follows:

[0100]

[0101] in, Let y be the pressure value of the pressure sensor in the i-th row and j-th column at time t, after standardization. j Let x be the column coordinate of the pressure sensor in the i-th row and j-th column, i.e., its position in the horizontal direction. iLet be the row coordinate of the pressure sensor in the i-th row and j-th column, i.e., its position in the vertical direction.

[0102] (2) Maximum pressure value P max (t):

[0103] Maximum pressure value

[0104] in, This represents the plantar pressure value after denoising at time t.

[0105] (3) Pressure distribution area A(t):

[0106] The pressure distribution area A(t) is used to analyze the distribution of pressure concentration areas, given a plantar pressure threshold P. threshhold Calculate the area A(t) of the region exceeding this threshold:

[0107]

[0108] in, For the plantar pressure in the i-th row and j-th column after noise reduction at time t, α i,j For the area covered by each pressure sensor, Π() is an indicator function, when... If the value is 1, then the value is 0; otherwise, the value is 0.

[0109] (4) Gait cycle pressure change rate ΔP(t):

[0110] The gait cycle pressure change rate ΔP(t) represents the pressure change between two time points. The starting point t of the gait cycle is the time when the heel of the left or right foot first contacts the ground, detected by a pressure sensor. s The end point t of the gait cycle is the time when the heel touches the ground again. e Then the gait period is Δt = t e -t s The method for calculating the rate of change of pressure during the gait cycle is as follows:

[0111] in,

[0112]

[0113] P s (t e (t) represents time t e Total pressure on the sole of the foot

[0114] P s (t s (t) represents time t s Total pressure on the sole of the foot.

[0115] (5) Pressure distribution variance

[0116] Pressure distribution variance Assess the dispersion of pressure values ​​across different parts of the sole to identify areas of concentrated pressure and their distribution.

[0117] in,

[0118]

[0119] Let be the average pressure at time t. Let be the plantar pressure in the i-th row and j-th column at time t after noise reduction.

[0120] The extracted plantar pressure features are used to construct a vector X, which is represented as follows:

[0121]

[0122] The vector X is preprocessed, including feature Z-score standardization, to eliminate dimensional differences. After preprocessing, the plantar pressure feature vector Z is obtained. The i-th element in the plantar pressure feature vector Z is denoted by z[i], and the preprocessing formula for z[i] is:

[0123]

[0124] Where X[i] represents the i-th element in vector X, and within a given time period S, μ[i] and σ[i] represent the mean and standard deviation of the i-th plantar pressure feature, respectively, and their calculation formulas are as follows:

[0125]

[0126] Where j is an integer between 0 and S.

[0127] Neural network architecture design includes an input layer, a first hidden layer, a second hidden layer, and an output layer.

[0128] The input layer only transmits data. The input is the preprocessed plantar pressure feature vector Z. The input layer includes 6 neurons, each corresponding to one of the 6 plantar pressure features.

[0129] The first hidden layer is a fully connected layer with 128 neurons. The activation function is the modified linear unit ReLU, and the calculation formula is: h1=ReLU(W1Z+b1), where W1 is a 128-row, 6-column weight matrix, b1 is a 128-row, 1-column bias vector, and h1 is the output of the first hidden layer of the neural network model.

[0130] The second hidden layer is a fully connected layer with 64 neurons. The activation function is the same as that of the first hidden layer, which is ReLU. The calculation formula is: h2 = ReLU(W2h1 + b2), where W2 is a 64-row, 128-column weight matrix, b2 is a 64-row, 1-column bias vector, h1 is the output of the first hidden layer of the neural network model, and h2 is the output of the second hidden layer of the neural network model.

[0131] The output layer is used for gait evaluation and anomaly detection. When used for gait evaluation, the output layer has 1 neuron, the activation function is linear, and the formula for calculating the gait evaluation model y is: y = W0h2 + b0, where W0 is a 1-row, 64-column weight matrix, b0 is a 1-row, 1-column bias matrix, h2 is the output of the second hidden layer of the neural network model, and y is the gait evaluation model, representing the evaluation score.

[0132] For gait anomaly detection, the output layer has 2 neurons, the activation function is Softmax, and the gait anomaly detection model is p. c The calculation formula is p c =Softmax(W c h2+b c ), W c The weight matrix is ​​2 rows and 64 columns, b c Let h1 be a 2x1 bias vector, h2 be the output of the second hidden layer of the neural network model, and p be the bias vector. c This is a gait anomaly detection model, and it represents the gait detection result.

[0133] Model Training: Gait evaluation uses Huber loss as the loss function to quantify prediction error and guide model parameter updates. The formula for Huber loss is:

[0134]

[0135] Among them, L e Let y be the loss function for gait evaluation, and y be the gait evaluation model, representing the evaluation score, which ranges from 0 to 100. δ is the predicted value of the gait model, and δ is the threshold parameter that controls the sensitivity to outliers, with a value of 1.

[0136] Gait anomaly detection uses weighted cross-entropy loss as the loss function to guide model parameter updates, as shown in the formula:

[0137]

[0138] Among them, L cla Let C be the loss function for gait anomaly detection, where C is the total number of categories, set to 2 here to represent both abnormal and normal cases. Ti Here, p represents the true label of the sample. Ti∈{0,1}. p ci w represents the predicted probability of class i by the model. i Let N be the weight of the i-th class. toal N represents the total number of samples. i Let be the number of samples in the i-th class.

[0139] Before real-time detection, 30 minutes of plantar pressure distribution data is collected from the user. The collected data is then sliced ​​according to the smallest time unit supported by the pressure sensor model used, forming a model training dataset. 70% of the data in this dataset is used to train the constructed neural network model. After training, another 30% of the data in the training dataset is used to test the neural network model. Once the test results meet the requirements, the real-time collected plantar pressure distribution data is input into the neural network model for gait evaluation and abnormal gait detection. In gait evaluation, a feedback mechanism is activated when the predicted value is less than 70.

[0140] Gait assessment and abnormality detection results are used to guide the adjustment of pressure distribution in ankle-foot orthoses. The plantar pressure characteristics, gait assessment results, and abnormal gait detection results measured above are fed back to the software in a timely manner. When the gait assessment is below 70 or abnormal gait is present, the software provides feedback to the user through vibration, ringing, etc., and gives ankle-foot orthotic adjustment suggestions, so that the user can make adjustments according to their actual needs.

[0141] The digital twin model has a real-time feedback function. When the gait assessment value is below 70 or gait abnormality is detected, it generates feedback information to be sent to the ankle-foot orthosis for parameter adjustment and wear suggestions to be synchronized on the mobile terminal. This allows the ankle-foot orthosis to increase support in the low-pressure area and decrease support in the high-pressure area, thereby adjusting the support of the ankle-foot orthosis and changing the pressure distribution.

[0142] Feedback information is used to determine the level of abnormality, thus providing a basis for adjusting the orthosis. The abnormality level score is based on the maximum pressure P. max The maximum pressure reference value P is calculated using pressure distribution area A(t) and pressure distribution area A(t). base and pressure distribution area reference value A base The value of P is related to the individual user. base Take the average pressure during a normal gait, A base The effective contact area of ​​the sole under normal gait is taken, and the abnormality level is calculated using a two-parameter joint scoring model, as shown in the following formula:

[0143]

[0144] Here, score(t) represents the abnormality level at time t, and α and β are adjustment coefficients. Based on the clinical hazards of pressure overload, α and β are set to 0.7 and 0.3, respectively. If score(t) < 0.5, it indicates a normal range and no adjustment is needed; if 0.5 ≤ score(t) < 1.0, it indicates a level 1 abnormality and requires slight adjustment; if score(t) ≥ 1.0, it indicates a level 2 abnormality and requires severe adjustment.

[0145] After the user completes the ankle-foot orthosis support adjustment based on feedback and wearing recommendations, the gait at time t+1 is recalculated using a gait assessment model, and the current pressure distribution is evaluated to verify the adjustment effect. Continuous monitoring of pressure distribution changes after adjustment forms a closed-loop feedback, enabling the ankle-foot orthosis to adaptively adjust support and pressure distribution with each gait change.

[0146] The above-mentioned interactive design method for ankle and foot orthotics based on digital twins can be implemented with the help of electronic devices. Specifically, the electronic devices include a storage device, a processor, and a computer program stored on the storage device and capable of running on the processor. The processor executes the computer program to implement the above-mentioned interactive design method for ankle and foot orthotics based on digital twins.

[0147] This embodiment also provides a computer-readable storage medium storing a computer program that is executed by a processor to implement the above-described interactive design method for ankle-foot orthotics based on digital twins.

[0148] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods.

[0149] Any references to memory, storage, database, or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM may take many forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0150] Based on the modeling and simulation analysis results of the digital twin model, feedback can be provided on the user's actual usage, allowing for timely intervention and adjustment of the ankle-foot orthosis's working status, providing personalized rehabilitation support. Furthermore, through a mobile application, users can view the ankle-foot orthosis's working status in real time, adjust parameters, set personalized usage modes, and receive rehabilitation suggestions based on the modeling and simulation analysis results of the digital twin model, enabling user interaction and control.

[0151] This invention provides an interactive design method and device for ankle-foot orthoses based on digital twins. This allows the orthosis to better adapt to individual user differences, providing personalized rehabilitation support. It also ensures good comfort even during prolonged use, resulting in excellent corrective effects and a superior user experience, thus enhancing the functionality and user experience of the ankle-foot orthosis. Specifically, the solution of this invention uses precise modeling to accurately match individual user differences, optimizing the design of the ankle-foot orthosis. It fully utilizes the real-time monitoring and feedback functions of digital twin technology to meet personalized rehabilitation needs, improve user experience, and demonstrates strong practicality, making it worthy of widespread adoption.

[0152] The above-disclosed embodiments are merely preferred embodiments of the present invention. However, the embodiments of the present invention are not limited thereto, and any variations that can be conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. An interactive design method for ankle-foot orthoses based on digital twins, characterized in that, Includes the following steps: Obtain structural data of the ankle and foot and ankle-foot orthoses; A three-dimensional model was built based on the structural data of the ankle and foot and ankle-foot orthotics. Real-time pressure distribution data of the soles of the feet is acquired and mapped onto a 3D model to form a digital twin model containing dynamic behavioral information; Plantar pressure features are extracted from digital twin models, and gait assessment models are constructed using these features. and gait anomaly detection model The gait evaluation model The gait anomaly detection model is used to reflect gait stability and pressure distribution. Used for real-time detection of abnormal gait; When gait evaluation model When the gait is below a preset threshold or when the gait abnormality detection model detects a gait abnormality, feedback information is generated and sent to the ankle-foot orthosis for parameter adjustment. The plantar pressure characteristics include the pressure center point. Maximum pressure value Pressure distribution area gait cycle pressure change rate and pressure distribution variance ; Gait assessment model constructed using plantar pressure characteristics and gait anomaly detection model This includes the following steps: Construct a vector from plantar pressure features ; For vectors Preprocessing is performed to obtain the plantar pressure feature vector Z; A neural network is constructed, comprising an input layer, a first hidden layer, a second hidden layer, and an output layer. The input layer is used to input the plantar pressure feature vector Z. The first hidden layer is a fully connected layer with 128 neurons, and the activation function is the Corrected Linear Unit (ReLU), calculated as follows: , It is a weight matrix with 128 rows and 6 columns. It is a bias vector with 128 rows and 1 column. This is the output of the first hidden layer of the neural network model; the second hidden layer is a fully connected layer with 64 neurons, and the activation function is the Corrected Linear Unit (ReLU), calculated as follows: , The weight matrix is ​​64 rows and 128 columns. It is a 64-row, 1-column bias vector. This is the output of the first hidden layer of the neural network model. This is the output of the second hidden layer of the neural network model; when the output layer is used for gait evaluation, the number of neurons in the output layer is 1, and the activation function is linear. (Gait evaluation model) , Gait assessment model, representing the assessment score. Weight matrix, It is the output of the second hidden layer of the neural network model. This is the bias matrix; when the output layer is used for gait anomaly detection, the number of neurons in the output layer is 2, and the activation function is Softmax. (Gait anomaly detection model) , This is a gait anomaly detection model, and it represents the gait detection results. This is the weight matrix. It is the output of the second hidden layer of the neural network model. For bias vectors, The activation function for the output layer of a neural network model; The feedback information is determined based on the following formula. in, It is the anomaly level at time t. and It is the adjustment coefficient. This is the maximum pressure reference value. It is the baseline value of the pressure distribution area. It is the maximum pressure value. It is the area of ​​pressure distribution.

2. The interactive design method for ankle-foot orthoses based on digital twins according to claim 1, characterized in that, Based on the structural data of the ankle and foot and ankle-foot orthoses, a three-dimensional model is established, including the following steps: Acquire point cloud data of the ankle and foot from a 3D scanner and point cloud data of the ankle-foot orthosis; A three-dimensional model is obtained by performing three-dimensional reconstruction based on the point cloud data of the ankle and foot and the point cloud data of the ankle and foot orthosis.

3. The interactive design method for ankle-foot orthoses based on digital twins according to claim 1, characterized in that, Acquire real-time plantar pressure distribution data and map it onto a 3D model to form a digital twin model containing dynamic behavioral information, including the following steps: Preprocess real-time plantar pressure distribution data; The preprocessed real-time plantar pressure distribution data is mapped onto a three-dimensional model to form a digital twin model.

4. The interactive design method for ankle-foot orthoses based on digital twins according to claim 3, characterized in that, The preprocessing includes the following steps: Kalman filtering was used to denoise the real-time plantar pressure distribution data. The real-time plantar pressure distribution data after noise reduction is standardized.

5. An electronic device, characterized in that, include: The storage device, the processor, and the computer program stored on the storage device and executable on the processor, wherein the processor executes the computer program to implement the interactive design method for ankle-foot orthotics based on digital twins as described in claim 1.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program is executed by a processor to implement the interactive design method for ankle-foot orthotics based on digital twins as described in claim 1.

Images