阅读说明：本技术 假肢接受腔体的舒适度评估方法、装置及存储介质 (Method and device for evaluating comfort of artificial limb receiving cavity and storage medium ) 是由 刘宁 刘薛勤 李锋 袁超杰 李羚 苏中 李擎 于 2021-09-13 设计创作，主要内容包括：本发明公开了一种假肢接受腔体的舒适度评估方法、装置及存储介质。其中,该方法包括：采集假肢穿戴者的历史运动状态数据,对所采集到的历史运动状态数据进行数据预处理以建立数据集；设定CNN网络的网络参数,基于所述数据集进行卷积计算和池化计算,生成CNN网络模型；将所述数据集中的一部分作为训练集,另一部分作为测试集,利用所述训练集对所述CNN网络模型进行样本训练,并利用所述测试集对所述CNN网络模型进行测试；基于训练和测试后的所述CNN网络模型来评估被穿戴的假肢接受腔体的舒适度。本发明解决了由于假肢接受腔体结构固定导致患者穿戴舒适度不足,严重影响患者体验的技术问题。(The invention discloses a method and a device for evaluating the comfort level of an artificial limb receiving cavity and a storage medium. Wherein, the method comprises the following steps: collecting historical motion state data of a prosthesis wearer, and performing data preprocessing on the collected historical motion state data to establish a data set; setting network parameters of the CNN network, and performing convolution calculation and pooling calculation based on the data set to generate a CNN network model; taking one part of the data set as a training set and the other part of the data set as a test set, carrying out sample training on the CNN network model by using the training set, and testing the CNN network model by using the test set; evaluating comfort of a worn prosthetic socket cavity based on the trained and tested CNN network model. The invention solves the technical problem that the wearing comfort of the patient is insufficient and the experience of the patient is seriously influenced because the artificial limb receiving cavity structure is fixed.)

1. A method for assessing the comfort of a prosthetic socket comprising:

collecting historical motion state data of a prosthesis wearer, and performing data preprocessing on the collected historical motion state data to establish a data set;

setting network parameters of the CNN network, and performing convolution calculation and pooling calculation based on the data set to generate a CNN network model;

taking one part of the data set as a training set and the other part of the data set as a test set, carrying out sample training on the CNN network model by using the training set, and testing the CNN network model by using the test set;

evaluating comfort of a worn prosthetic socket cavity based on the trained and tested CNN network model.

2. The method of claim 1, wherein collecting historical motion state data for a prosthesis wearer comprises:

collecting walking, upstairs, downstairs, sitting and standing posture data of the prosthesis wearer through an inertia sensing unit arranged on the prosthesis receiving cavity;

acquiring gait data of the prosthesis wearer in a walking process through a plantar pressure sensor arranged on a sole of the prosthesis receiving cavity, wherein the gait data comprises data of two phases, namely a leg supporting phase and a swinging phase, the supporting phase is a walking phase which starts from heel landing of the prosthesis wearer and ends at tiptoe off, and the swinging phase is a walking phase which starts from tiptoe off and ends at heel landing;

acquiring pressure data of all directions of the inner wall of the artificial limb receiving cavity through a cavity pressure sensor arranged in the artificial limb receiving cavity;

wherein the historical motion state data comprises the posture data, the gait data and the pressure data.

3. The method of claim 1, wherein data pre-processing the collected historical motion state data to create a data set comprises:

dividing the collected historical motion state data by adopting a sliding window folding method, moving the collected historical motion state data and extracting data segments by setting the window length and the step length in the sliding window;

and transforming the extracted data segments through sliding window iteration to obtain a data set retaining time information, wherein the relationship between each data and the surrounding several adjacent data is retained in the data set.

4. The method of claim 3, wherein the sliding window is folded as follows:

T＝(x₁,x₂,...,x_n-1,x_n)

where T represents the data set, n represents the length of the data set, I represents the window length in a sliding window, s represents the step size in a sliding window, x represents the data segment, and T' represents the matrix formed by the segmented data set, where there is a 50% overlap between each sliding window.

5. The method of claim 1, wherein performing convolution and pooling calculations based on the data set comprises:

selecting a convolution kernel, expanding the convolution matrix into a to-be-convolved matrix with preset dimensionality, wherein 0 is used as a filling number in an expansion part, then performing convolution calculation on the to-be-convolved matrix from left to right through the convolution kernel from top to bottom to obtain a convolution output matrix;

and the pooling layer performs pooling calculation on the convolution output matrix so as to perform dimensionality reduction screening on the convolution output matrix.

6. The method of claim 5, wherein the convolution calculation is performed by the following equation:

wherein, a is the matrix to be convolved, K is the convolution kernel in convolution operation, B is a convolution output matrix, i is the row number of the convolution output matrix B, j is the column number of the convolution output matrix B, m is the row number of the matrix of the convolution kernel, and n is the column number of the matrix of the convolution kernel.

7. The method of claim 5, wherein the pooling calculation is performed by the following equation:

wherein x is_il，jl、x_il，jl+1、x_il+1，jl、x_il+1，il+l、Andrespectively, different elements, y, of the corner marks in the convolution output matrix_i，jFor the pooled matrix, i is the row number of the convolution output matrix, j is the column number of the convolution output matrix, p_wAnd p_hRespectively representing width and height in the pooled kernel, l being the step size.

8. The method of claim 1, wherein evaluating the comfort of a worn prosthetic socket cavity based on the trained and tested CNN network model comprises:

acquiring mechanical information of a prosthesis wearer in real time;

and inputting the mechanical information into the trained and tested CNN network model to evaluate the comfort degree of the artificial limb receiving cavity worn by the artificial limb wearer.

9. A comfort assessment device for a prosthetic socket comprising:

the acquisition processing module is configured to acquire historical motion state data of a prosthesis wearer and perform data preprocessing on the acquired historical motion state data to establish a data set;

the model generation module is configured to set network parameters of the CNN network, and carry out convolution calculation and pooling calculation based on the data set to generate a CNN network model;

a training test module configured to use one part of the data set as a training set and another part of the data set as a test set, perform sample training on the CNN network model by using the training set, and perform testing on the CNN network model by using the test set;

an evaluation module configured to evaluate comfort of a worn prosthetic socket cavity based on the trained and tested CNN network model.

10. A computer-readable storage medium having stored thereon a program which, when executed, causes the computer to perform the method of any one of claims 1 to 8.

Technical Field

The invention relates to the field of intelligent wearing, in particular to a method and a device for evaluating the comfort level of an artificial limb receiving cavity and a storage medium.

Background

In recent years, the number of patients with lower leg amputation has been increasing, and due to the inconvenience of movement, not only has a serious burden on the family, but also the patients gradually lose confidence in their lives.

However, with the development of science and technology, the level of manufacturing artificial limbs is also continuously improved, the walking gait of amputees can be obviously improved by the current commercialized intelligent bionic artificial limb, and the amputees can be greatly helped to recover the daily activities, but the wearing comfort of the patients is insufficient due to the fact that the artificial limb receiving cavity structure is fixed at present, and the experience of the patients is seriously influenced.

In view of the above problems, no effective solution has been proposed.

Disclosure of Invention

The embodiment of the invention provides a comfort evaluation method, a comfort evaluation device and a storage medium for an artificial limb receiving cavity, which are used for at least solving the technical problems that the wearing comfort of a patient is insufficient and the experience of the patient is seriously influenced due to the fixed structure of the artificial limb receiving cavity.

According to an aspect of an embodiment of the present invention, there is provided a method for evaluating comfort of a prosthetic socket, the method comprising: collecting historical motion state data of a prosthesis wearer, and performing data preprocessing on the collected historical motion state data to establish a data set; setting network parameters of the CNN network, and performing convolution calculation and pooling calculation based on the data set to generate a CNN network model; taking one part of the data set as a training set and the other part of the data set as a test set, carrying out sample training on the CNN network model by using the training set, and testing the CNN network model by using the test set; evaluating comfort of a worn prosthetic socket cavity based on the trained and tested CNN network model.

According to another aspect of the embodiments of the present invention, there is also provided a comfort evaluation device for a prosthetic socket, including: the acquisition processing module is configured to acquire historical motion state data of a prosthesis wearer and perform data preprocessing on the acquired historical motion state data to establish a data set; the model generation module is configured to set network parameters of the CNN network, and carry out convolution calculation and pooling calculation based on the data set to generate a CNN network model; a training test module configured to use one part of the data set as a training set and another part of the data set as a test set, perform sample training on the CNN network model by using the training set, and perform testing on the CNN network model by using the test set; an evaluation module configured to evaluate comfort of a worn prosthetic socket cavity based on the trained and tested CNN network model.

According to another aspect of embodiments of the present invention, there is also provided a computer readable storage device having a program stored thereon, which when executed, causes a computer to perform the method for assessing the comfort of a prosthetic socket as described above.

In the embodiment of the invention, the inertial measurement unit and the pressure sensor are utilized to collect leg inertial information of a wearer and pressure data inside the receiving cavity and on the sole of the foot to identify different motion states of the patient and abnormal pressure information of the receiving cavity to the residual limb under different motion states, so that the wearing comfort of the patient is intelligently evaluated, and the technical problem that the wearing comfort of the patient is insufficient and the experience of the patient is seriously influenced due to the fixed structure of the receiving cavity of the artificial limb is solved.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:

FIG. 1 is a first flowchart of a method for evaluating the comfort of a prosthetic socket according to an embodiment of the present invention;

FIG. 2 is a second flowchart of a method for evaluating the comfort of a prosthetic socket according to an embodiment of the invention;

figure 3 is a diagram of a human walking gait in the sagittal plane according to an embodiment of the invention.

FIG. 4 is a data acquisition profile according to an embodiment of the present invention;

FIG. 5 is a flow chart of a comfort assessment method for a prosthetic socket according to an embodiment of the invention;

FIG. 6 is a schematic structural diagram of a comfort assessment device for a prosthetic socket according to an embodiment of the present invention;

fig. 7 is a schematic structural diagram of a comfort assessment system for a prosthetic socket according to an embodiment of the invention.

Fig. 8 is a schematic structural diagram of an inertial sensing unit according to an embodiment of the present invention.

Detailed Description

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

Example 1

According to an embodiment of the invention, there is provided a method for evaluating comfort of a prosthetic socket, as shown in fig. 1, the method including:

step S102, collecting historical motion state data of a prosthesis wearer, and performing data preprocessing on the collected historical motion state data to establish a data set.

In one exemplary embodiment, the posture data of walking, going upstairs, going downstairs, sitting and standing of the prosthesis wearer is collected by an inertia sensing unit arranged on the prosthesis receiving cavity; acquiring gait data of the prosthesis wearer in a walking process through a plantar pressure sensor arranged on a sole of the prosthesis receiving cavity, wherein the gait data comprises data of two phases, namely a leg supporting phase and a swinging phase, the supporting phase is a walking phase which starts from heel landing of the prosthesis wearer and ends at tiptoe off, and the swinging phase is a walking phase which starts from tiptoe off and ends at heel landing; acquiring pressure data of all directions of the inner wall of the artificial limb receiving cavity through a cavity pressure sensor arranged in the artificial limb receiving cavity; wherein the historical motion state data comprises the posture data, the gait data and the pressure data.

In an exemplary embodiment, the collected historical motion state data is segmented by adopting a sliding window folding method, and the collected historical motion state data is moved and data segments are extracted by setting the window length and the step length in a sliding window; and transforming the extracted data segments through sliding window iteration to obtain a data set retaining time information, wherein the relationship between each data and the surrounding several adjacent data is retained in the data set.

And step S104, setting network parameters of the CNN network, and performing convolution calculation and pooling calculation based on the data set to generate a CNN network model.

In an exemplary embodiment, a convolution kernel is selected, a convolution matrix is expanded into a to-be-convolved matrix with a preset dimensionality, wherein 0 is used as a filling number in an expansion part, then the to-be-convolved matrix is subjected to convolution calculation from left to right through the convolution kernel from top to bottom, and a convolution output matrix is obtained; and the pooling layer performs pooling calculation on the convolution output matrix so as to perform dimensionality reduction screening on the convolution output matrix.

And step S106, taking one part of the data set as a training set and the other part of the data set as a test set, carrying out sample training on the CNN network model by using the training set, and testing the CNN network model by using the test set.

Step S108, evaluating the comfort degree of the worn artificial limb accepting cavity based on the trained and tested CNN network model.

In one exemplary embodiment, the mechanical information of the prosthesis wearer is obtained in real time; and inputting the mechanical information into the trained and tested CNN network model to evaluate the comfort degree of the artificial limb receiving cavity worn by the artificial limb wearer.

The application utilizes the deep neural network to identify the movement gait of an artificial limb wearer in real time, combines the artificial limb receiving cavity, the plantar pressure identification technology and the wearing feedback of the patient, and intelligently evaluates the comfort condition of the patient. Data support is provided for the artificial limb to improve the coordination and comfort of all parts of the artificial limb.

Example 2

According to an embodiment of the invention, another method for evaluating the comfort of a prosthetic socket is provided. As shown in fig. 2, the method comprises the steps of:

step S201, data is collected.

A large number of prosthesis wearers are tested in different movement states respectively.

The posture data of walking, going upstairs, going downstairs, sitting and standing of the prosthesis wearer are collected through an inertia sensing unit arranged on the prosthesis receiving cavity.

The gait data in the walking process is collected through a sole pressure sensor arranged at the sole position of the prosthetic limb receiving cavity, wherein the gait comprises two phases of a leg supporting phase and a swinging phase. As shown in FIG. 3, the support phase is a walking phase beginning at heel strike and ending at toe-off, during which the support leg remains in contact with the ground, supporting the forward movement of the person's body center of gravity. The swing phase is a walking phase that begins at the toe-off and ends at the heel-strike, in which the position of the foot is gradually raised and gradually moved forward of the center of gravity, ready for the start of the next support phase. The supporting phase can be further divided into three stages of heel supporting, heel sole simultaneous supporting and sole supporting according to the contact condition of the lower limbs and the ground.

Pressure data of each direction of the inner wall of the artificial limb accepting cavity of a patient in the movement process are collected through a cavity pressure sensor in the artificial limb accepting cavity. In the process, abnormal pressure data under different conditions are acquired by changing the coverage range of the cavity pressure sensor in the prosthetic socket.

FIG. 4 shows a profile of data acquisition according to one embodiment of the present application. As shown in fig. 4, in the present embodiment, 424400 pieces of data of a walking state, 122869 pieces of data of an upstairs state, 100427 pieces of data of a downstairs state, 59939 pieces of data of a sitting state, and 48395 pieces of data of a standing state are collected.

And performing data preprocessing, characteristic value selection and other operations on the acquired data, such as posture data, gait data and pressure data, so as to establish a data set.

Step S202, data preprocessing.

The data in the database is segmented by adopting a sliding window folding method, data segments are extracted by moving on the data by setting the window length and the step length in the sliding window, and in order to keep the time relationship between data points in activities, the overlapping rate of 50 percent exists between every two windows. After labeling the data, 70% of the data were used as the training set and 30% as the test set.

Assuming that there is a data set T of length n, the window length in the sliding window is l, and the step size in the sliding window is s. The folding process of the sliding window is as follows:

T＝(x₁,x₂,...,x_n-1,x_n)

where T represents the data set, n represents the length of the data set, I represents the window length in a sliding window, s represents the step size in a sliding window, x represents the data segment, and T' represents the matrix formed by the segmented data set, where there is a 50% overlap between each sliding window.

The transformed T' is obtained through sliding window iteration, and the relation between the x data and the surrounding several adjacent data is reserved in the matrix. The three-axis sensor data can be processed by a sliding window folding method, and sliding windows on a time sequence are performed by taking a time window as a unit by the sliding window folding method, so that a data field with reserved time information is finally obtained. The x, y and z axes in the three-axis sensor are converted into a data format according to a time sliding window folding method, and the representation form is as follows:

the method comprises the following steps of obtaining a matrix corresponding to an X axis, obtaining a matrix corresponding to a Y axis, obtaining a matrix corresponding to a Z axis, obtaining data points on the X axis, the Y axis and the Z axis in a three-axis sensor, and obtaining a step length.

The data processed by the method ensures that the data at a certain time point has close relation with the surrounding data like pixel point data.

Step S203 sets network parameters of the CNN network.

Setting the number of training layers and training parameters of the CNN network, two convolutional layers and pooling layers, and determining the size of a convolution kernel, the convolution moving step length and the number of the convolution kernels.

Step S204, convolution and pooling process.

And performing convolution operation on the selected convolution kernel and a matrix with the same size in the training set to obtain a data matrix after convolution. The pooling layer performs dimension reduction on the convolution layer output.

Suppose a is the matrix to be convolved, K is the convolution kernel in the convolution operation, and B is the final convolution result matrix. The convolution formula for the a matrix is as follows:

wherein, a is the matrix to be convolved, K is the convolution kernel in convolution operation, B is a convolution output matrix, i is the row number of the convolution output matrix B, j is the column number of the convolution output matrix B, m is the row number of the matrix of the convolution kernel, and n is the column number of the matrix of the convolution kernel.

Firstly, expanding the convolution matrix A into a matrix to be convolved, wherein the matrix to be convolved is (m + n) × (m + n), 0 is used as a filling number in an expansion part, and then convolution calculation is carried out on the matrix A from left to right and from top to bottom through a convolution kernel K, so that a convolution output matrix B is finally obtained.

The pooling layer mainly performs dimension reduction screening on the output value of the convolutional layer, the pooling method comprises average pooling and median pooling, and the pooling process in the embodiment adopts a maximum pooling strategy. CNN-based networkThe motion state model of the network adopts 2 x 2 pooling kernels, and the width and the height in the pooling kernels are respectively p by assuming that the step length is l_wAnd p_hAnd (4) showing. The operation formula in the maximum pooling process is as follows:

wherein x is_il，jl、x_il，jl+1、x_il-1，jl、x_il+1jl+1 Andrespectively, different elements, y, of the corner marks in the convolution output matrix_i，jFor the pooled matrix, i is the row number of the convolution output matrix, j is the column number of the convolution output matrix, p_wAnd p_hRespectively, width and height in the pooled kernel, l being the step size, il representing i x l, jl representing j x l.

The data dimension reduction after the pooling reduces the size of original sample data by half, which is beneficial to reducing training parameters and improving the running speed of the algorithm model.

And step S205, training a sample.

And (3) carrying out convolution operation on the sample data, calculating according to the convolution parameters in the previous step, introducing a padding (padding) process for ensuring that the input size is the same as the output size, and carrying out data expansion by adding 0. And (4) performing dimensionality reduction treatment on the pooling layer to shorten the model training time, wherein 2 x 2 pooling kernels enable the matrix scale obtained by convolution to be changed into a half of the original matrix scale. After several convolution and pooling processes, a deep neural network structure will be formed.

And step S206, evaluating a comfort result.

And collecting mechanical information of the prosthesis wearer in real time, inputting the mechanical information into the trained and tested CNN network model, and evaluating the comfort level of the prosthesis receiving cavity worn by the prosthesis wearer.

Furthermore, the comfort level evaluated by the CNN network model can be compared with the real feedback of the patient, and the comfort level evaluated by the CNN network model is used for evaluating the precision of the CNN network model and summarizing the evaluation result.

And step S207, manufacturing an optimization scheme.

And (4) analyzing the gait difference of the normal person and the amputee patient by combining a human body dynamics model according to the actual evaluation result of the model, and guiding the next-stage model optimization.

It should be noted that, for simplicity of description, the above-mentioned method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present invention is not limited by the order of acts, as some steps may occur in other orders or concurrently in accordance with the invention. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required by the invention.

Through the above description of the embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but the former is a better implementation mode in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, or a network device) to execute the method according to the embodiments of the present invention.

Example 3

The present application also provides a method for evaluating the comfort of a prosthetic socket, as shown in fig. 5, the method comprising the steps of:

and step S501, data acquisition.

The posture data of walking, going upstairs, going downstairs, sitting and standing of the prosthesis wearer are collected through an inertia sensing unit arranged on the prosthesis receiving cavity. Gait data in the walking process are collected through a sole pressure sensor arranged at the sole position of the artificial limb receiving cavity. Pressure data of each direction of the inner wall of the artificial limb accepting cavity of a patient in the movement process are collected through a cavity pressure sensor in the artificial limb accepting cavity.

Step S502, a data set is established.

And performing data preprocessing, characteristic value selection and other operations on the acquired data, such as posture data, gait data and pressure data, so as to establish a data set.

Step S503, data preprocessing.

The data in the database is segmented by adopting a sliding window folding method, data segments are extracted by moving on the data by setting the window length and the step length in the sliding window, and in order to keep the time relationship between data points in activities, the overlapping rate of each window is 30-40%. After labeling the data, 80% of the data were used as the training set and 20% as the test set.

By processing the data set by sliding window folding, the processed data ensures that the data at a certain point in time has close relation with the surrounding data like pixel point data.

Step S504, setting CNN network parameters.

Setting the number of training layers and training parameters of the CNN network, two convolutional layers and pooling layers, and determining the size of a convolution kernel, the convolution moving step length and the number of the convolution kernels.

Step S505, convolution calculation.

And expanding the convolution matrix into a to-be-convolved matrix with a preset dimensionality, using 0 as a filling number in an expansion part, and performing convolution calculation on the to-be-convolved matrix from left to right and from top to bottom through a convolution kernel K to finally obtain a convolution output matrix.

Step S506, pooling calculation.

The pooling layer mainly performs dimension reduction screening on the output value of the convolutional layer, and the pooling method can use an average pooling method, a median pooling method, a maximum pooling method and the like. The data dimension reduction after the pooling reduces the size of original sample data by half, which is beneficial to reducing training parameters and improving the running speed of the algorithm model.

And step S507, establishing a network model.

And (3) carrying out convolution operation on the sample data, calculating according to the convolution parameters in the previous step, introducing a padding (padding) process for ensuring that the input size is the same as the output size, and carrying out data expansion by adding 0. And (4) performing dimensionality reduction treatment on the pooling layer to shorten the model training time, wherein 2 x 2 pooling kernels enable the matrix scale obtained by convolution to be changed into a half of the original matrix scale. After several convolution and pooling processes, a deep neural network structure will be formed.

In step S508, the result is evaluated.

And collecting mechanical information of the prosthesis wearer in real time, inputting the mechanical information into the trained and tested CNN network model, and evaluating the comfort level of the prosthesis receiving cavity worn by the prosthesis wearer.

Example 4

The present application also provides a comfort assessment device for a prosthetic socket cavity, as shown in fig. 6, comprising: an acquisition processing module 62, a model generation module 64, a training test module 66, and an evaluation module 68.

The acquisition and processing module 62 is configured to acquire historical motion state data of the prosthesis wearer, and to perform data preprocessing on the acquired historical motion state data to create a data set.

The model generation module 64 is configured to set network parameters of the CNN network, perform convolution and pooling calculations based on the data set, and generate a CNN network model.

The training test module 66 is configured to use one part of the data set as a training set and another part as a test set, perform sample training on the CNN network model using the training set, and perform testing on the CNN network model using the test set.

The evaluation module 68 is configured to evaluate the comfort of the worn prosthetic socket cavity based on the trained and tested CNN network model.

In one exemplary embodiment, the posture data of walking, going upstairs, going downstairs, sitting and standing of the prosthesis wearer is collected by an inertia sensing unit arranged on the prosthesis receiving cavity; acquiring gait data of the prosthesis wearer in a walking process through a plantar pressure sensor arranged on a sole of the prosthesis receiving cavity, wherein the gait data comprises data of two phases, namely a leg supporting phase and a swinging phase, the supporting phase is a walking phase which starts from heel landing of the prosthesis wearer and ends at tiptoe off, and the swinging phase is a walking phase which starts from tiptoe off and ends at heel landing; acquiring pressure data of all directions of the inner wall of the artificial limb receiving cavity through a cavity pressure sensor arranged in the artificial limb receiving cavity; wherein the historical motion state data comprises the posture data, the gait data and the pressure data.

The acquisition processing module 62 receives data acquired by the inertia sensing unit, the sole pressure sensor and the cavity pressure sensor, and performs data preprocessing. For example, the collected historical motion state data is segmented by adopting a sliding window folding method, and the collected historical motion state data is moved and data segments are extracted by setting the window length and the step length in the sliding window; and transforming the extracted data segments through sliding window iteration to obtain a data set retaining time information, wherein the relationship between each data and the surrounding several adjacent data is retained in the data set.

In an exemplary embodiment, the model generation module 64 performs convolution and pooling calculations based on the data set. For example, selecting a convolution kernel, expanding the convolution matrix into a to-be-convolved matrix with preset dimensionality, wherein 0 is used as a filling number in an expansion part, then performing convolution calculation on the to-be-convolved matrix from left to right through the convolution kernel from top to bottom, and obtaining a convolution output matrix; the pooling layer performs pooling calculation on the convolution output matrix so as to perform dimensionality reduction screening on the convolution output matrix

In an exemplary embodiment, the training test module 66 performs a convolution operation on the sample data. To ensure that the input and output sizes are the same, a padding (padding) process is introduced, and data expansion is performed by adding 0. And (4) performing dimensionality reduction treatment on the pooling layer to shorten the model training time, wherein 2 x 2 pooling kernels enable the matrix scale obtained by convolution to be changed into a half of the original matrix scale. After several convolution and pooling processes, a deep neural network structure will be formed.

In an exemplary embodiment, the evaluation module 68 inputs the prosthesis wearer's mechanical information collected in real time into the CNN network model after training and testing to evaluate the comfort of the prosthesis accepting cavity worn by the prosthesis wearer.

The comfort evaluation device for a prosthetic socket provided in this embodiment can implement the comfort evaluation method for a prosthetic socket in embodiments 1, 2, and 3, which is not described herein again.

Example 5

Referring to fig. 7, which is a schematic structural diagram of a comfort evaluation system of a prosthetic socket according to an embodiment of the present invention, as shown in fig. 7, the system includes a prosthetic socket 72 and a comfort evaluation device 74, wherein an inertia sensing unit 722, a plantar pressure sensor 724 and a cavity sensor 726 are disposed on the prosthetic socket 72. The appearance structure of the inertial sensing unit 722 is shown in fig. 8.

The cavity sensor 726 and the plantar pressure sensor 724 arranged on the sole of the foot in the prosthesis receiving cavity 72 and the inertia sensing unit 722 collect gait data, cavity pressure data and posture data of a prosthesis wearer, mark a timestamp on the collected data information and transmit the data information to the comfort evaluation device 74.

The comfort assessment device 74 is configured to receive the acquired data, pre-process the acquired data to create a data set; setting network parameters of the CNN network, and performing convolution calculation and pooling calculation based on the data set to generate a CNN network model; taking one part of the data set as a training set and the other part as a test set, carrying out sample training on the CNN network model by using the training set, and testing the CNN network model by using the test set; the comfort of the worn prosthetic socket cavity is then evaluated based on the trained and tested CNN network model.

In one embodiment, the comfort assessment device 74 can be the comfort assessment device of the prosthetic socket of embodiment 4, which is not described herein.

Example 6

Embodiments of the present disclosure also provide a storage medium. Alternatively, in the present embodiment, the storage medium may implement the method described in embodiments 1 to 4 described above.

Optionally, in this embodiment, the storage medium may be located in at least one network device of a plurality of network devices in a network of the inertial navigation system.

Optionally, in this embodiment, the storage medium may include, but is not limited to: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.

Alternatively, in the present embodiment, the processor executes the methods in embodiments 1 to 3 described above according to the program code stored in the storage medium.

Optionally, for a specific example in this embodiment, reference may be made to the examples described in embodiments 1 to 3, which is not described herein again.

The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.

The integrated unit in the above embodiments, if implemented in the form of a software functional unit and sold or used as a separate product, may be stored in the above computer-readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing one or more computer devices (which may be personal computers, servers, network devices, etc.) to execute all or part of the steps of the method according to the embodiments of the present invention.

In the above embodiments of the present invention, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

In the several embodiments provided in the present application, it should be understood that the disclosed client may be implemented in other manners. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one type of division of logical functions, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, units or modules, and may be in an electrical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.