Balance exercise ability training system for people with walking obstacles

Abstract

The invention discloses a balance exercise ability training system for people with walking obstacles, which comprises an input module, a gravity monitoring module, a position monitoring module, an upper arm monitoring module, a storage module, a processing module, a comparison module and an acousto-optic reminding module. The position monitoring module accurately recognizes the position change of the user in the space and evaluates whether the walking action of the user is standard or not and the movement amplitude of the specified action is completed, and the upper arm monitoring module records the force data of the upper limbs of the user in the movement process in real time and analyzes the compensation phenomenon. By recording the balance ability, position detection and upper limb compensation conditions and comprehensively considering the recovery condition of the user, powerful support is provided for a rehabilitator to formulate and adjust the training intensity and scheme, and the rehabilitation effect of the user is improved.

Description

Technical Field

[0001] The invention relates to the technical field of training systems, in particular to a balance movement ability training system for people with walking disabilities. Background Art

[0002] As the global population ages, especially after a stroke, the need for rehabilitation training for the elderly is becoming increasingly urgent. Traditional rehabilitation methods often lack personalization and systematization, and are difficult to meet the needs of the elderly in terms of strength, balance, and control. This system aims to provide comprehensive rehabilitation training support through the combination of gravity detection, position detection, and upper limb assistive devices.

[0003] After searching, the existing patent (publication number: CN114558302B) discloses a system and method for exercising sports ability. The sports ability training system includes a sports device placed on the ground for trainees to exercise according to a predetermined guidance program, and the sports device is at least arrayed with a plurality of light-emitting elements for generating guidance according to the predetermined guidance program. The system guides the trainees to exercise their sports ability by the following method: guiding the trainees to perform sports training, and correcting the trainees' movements and / or adjusting the training program for guiding the trainees based on the system's monitoring of the trainees, so that the corresponding training program matches the trainees' balance level, thereby ensuring the training effect and training enthusiasm of the trainees during the exercise.

[0004] It receives the physiological parameters of the trainee who is currently in the training mode and the accuracy parameters of the training movements, and then adjusts the training program. However, it has its shortcomings. It does not have the function of recognizing the training movements of the trainee, so the training movements of the trainee are not in place, resulting in poor training results. Summary of the invention

[0005] In view of the deficiencies of the prior art, the present invention provides a balance movement ability training system for people with walking disabilities, which solves the problems raised by the above-mentioned background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a balance movement ability training system for people with walking disabilities, including an input module, a gravity monitoring module, a position monitoring module, an upper arm monitoring module, a storage module, a processing module, a comparison module and an audio-visual reminder module, wherein the gravity monitoring module includes a gravity sensor; the position monitoring module includes a position switch, an accelerometer and an ultrasonic sensor; the upper arm monitoring module includes a grip pressure sensor and a support pressure sensor;

[0007] Step 1: inputting the upper limb force compensation operation weight, the correct standing center of gravity operation weight and the correct walking posture operation weight into the storage module through the input module;

[0008] Step 2, the gravity monitoring module obtains the center of gravity transfer data of the user when standing with both feet upright and standing on one foot through the gravity sensor, and sends the data to the processing module, the processing module starts the comparison module and sends the data to the comparison module, the comparison module mines the data inside the storage module for comparison, finds out whether the user's balance ability is standard, and sends the result to the display module for display;

[0009] Step 3, the position monitoring module accurately identifies the data of the position change of the user's legs in space through the position switch, accelerometer and ultrasonic sensor, and sends the data to the processing module, the processing module starts the comparison module and sends the data to the comparison module, the comparison module mines the data inside the storage module for comparison, and finds out whether the user's walking action completes the specified action in a standard manner, and sends the result to the display module for display;

[0010] Step 4. The upper arm monitoring module obtains the data of the palm grip and the arm support force through the grip pressure sensor and the support pressure sensor, and sends the data to the processing module. The processing module starts the comparison module and sends the data to the comparison module. The comparison module mines the data inside the storage module for comparison, and adjusts the training intensity and content according to whether there is a compensation phenomenon based on the comparison results, so as to improve the upper limb strength and balance ability in a targeted manner.

[0011] Preferably, the step of mining the internal data of the storage module by the comparison module is:

[0012] Step 21, extract a feature from the storage data inside the storage module, and normalize the storage data inside the storage module through the feature, so that the data value range corresponding to each feature of the storage data inside the storage module is consistent, forming compressed code training data:

[0013] Using the compressed code training data, encoder training is performed according to the following formula that minimizes the objective function of classification error:

[0014]

[0015] Among them, a ik is the weight coefficient, l(y ik .f(X i )) is the loss function, and the expression of the loss function is l(a)=(0,1-a), y ik ∈(-1,1) is X i The category identifier of the kth category, X i is the i-th compressed code training data, is the classification parameter corresponding to the mth feature in the kth category, bk is the bias parameter, M is the number of features, A is the projection matrix, and A (m) is the projection matrix corresponding to the mth feature, φ(A (m) X i (m) ) is a hash function, N is the number of the compressed code training data, K is the number of categories of the compressed code training data, g(W) and h(A) are two normalization functions, which are used to adjust the classification parameter matrix W and the projection matrix A respectively, λ1 and λ2 are two real numbers, which are used to adjust the normalization functions g(W) and h(A) respectively;

[0016] After training, the projection matrix A, the classification parameter matrix W and the bias matrix B are obtained, where B = {b1, b2, ...b k}; and the hash function: φ(A (m) X i (m) ) as a binary compression code encoder;

[0017] Step 22: using a classifier based on binary compression code to classify the binary compression code of the internal data of the storage module, obtaining the storage data category inside the storage module, and classifying the binary compression code of the internal data of the storage module by the following function:

[0018]

[0019] Step 23: Use the cosine method to calculate the similarity between the user data obtained by various sensors and the data of each category in turn, and select the category with the highest similarity. The formula for calculating the similarity using the cosine method is as follows: where x1, x2…x n is the user data obtained by various sensors, y1, y2…y n For each category of data, the smaller cosθ is, the higher the similarity is.

[0020] Preferably, the feature extraction method is:

[0021] The iterative formula for constructing the sample matrix according to the stored data in the storage module

[0022]

[0023] The sample matrix is ​​iteratively updated according to the iterative formula to obtain a feature matrix, wherein β0 is a constant, a random initial value of iteration β0=1 is given, Y0=H0, and it is iteratively updated according to the iterative formula. When the number of iterations reaches the set threshold k, the iteration is terminated to obtain H kThat is the characteristic matrix, k represents the product of the matrix, Y is the intermediate parameter, P() represents the projection gradient algorithm, and P(Z) means projecting all negative numbers in the matrix Z to 0.

[0024] Preferably, the comparison calculation method of the comparison module is:

[0025] The user data at each time end obtained by various sensors is converted into a two-dimensional matrix. The comparison module adds the values ​​in various two-dimensional matrices respectively to obtain the sum of various two-dimensional matrices, and then calculates the sum of the sums of each two-dimensional matrix respectively. Then, each sum is divided by the number n of each two-dimensional matrix to obtain various operation weights. Various operation weights represent the average value of the sum of each two-dimensional matrix. Then, this operation weight is compared with various correct data operation weights input in advance. If the absolute value of the difference between the two is greater than the preset parameter, it is determined that the user's action is not standard, otherwise it is standard.

[0026] Beneficial Effects

[0027] The present invention provides a system for training the balance ability of people with walking disabilities. Compared with the prior art, it has the following beneficial effects:

[0028] 1. The balance training system for people with walking disabilities automatically records the center of gravity data of the user during each training session through the gravity monitoring module, accurately obtains the center of gravity position change data, and evaluates the user's balance ability. The position monitoring module accurately identifies the user's position changes in space and evaluates whether the user's walking movements complete the specified movements in a standard manner. The upper arm monitoring module records the force data of the user's upper limbs in real time and analyzes the compensation phenomenon to adjust the training intensity and method to improve the effect. The sound and light reminder module reminds the user of training movements, training market, and the output of various appreciation languages ​​through the sound and light module during training.

[0029] 2. The balance motor ability training system for people with walking disabilities iteratively updates the sample matrix according to the iterative formula to obtain the optimal feature matrix and the optimal feature vector. By classifying the stored data, not only the accuracy of data mining is improved, but also the efficiency of data mining is improved, thereby improving the speed of device comparison calculation, and the user's training status can be quickly understood. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 Schematic diagram of the workflow of the present invention.

[0031] In the figure: 1. Gravity monitoring module; 2. Position monitoring module; 3. Upper arm monitoring module; 4. Processing module; 5. Comparison module; 6. Gravity sensor; 7. Input module; 8. Accelerometer; 9. Ultrasonic sensor; 10. Grip pressure sensor; 11. Support pressure sensor; 12. Storage module; 13. Display module; 14. Position switch; 15. Sound and light reminder module. DETAILED DESCRIPTION

[0032] 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.

[0033] See also Figure 1The present invention provides a technical solution: a balance movement ability training system for people with walking disabilities, including an input module 7, a gravity monitoring module 1, a position monitoring module 2, an upper arm monitoring module 3, a storage module 12, a processing module 4, a comparison module 5 and an audio-visual reminder module 15, wherein the gravity monitoring module 1 includes a gravity sensor 6; the position monitoring module 2 includes a position switch 14, an accelerometer 8 and an ultrasonic sensor 9; the upper arm monitoring module 3 includes a grip pressure sensor 10 and a support pressure sensor 11; step 1, inputting the upper limb force compensation operation weight, the correct standing center of gravity operation weight and the correct walking posture operation weight into the storage module 12 through the input module 7; step 2 1. The gravity monitoring module 1 obtains the center of gravity transfer data of the user when standing with both feet upright or standing on one foot through the gravity sensor 6, and sends the data to the processing module 4. The processing module 4 starts the comparison module 5 and sends the data to the comparison module 5. The comparison module 5 mines the data inside the storage module 12 for comparison to find out whether the user's balance ability is standard, and sends the result to the display module 13 for display; Step 3. The position monitoring module 2 accurately identifies the data of the position change of the user's legs in space through the position switch 14, the accelerometer 8 and the ultrasonic sensor 9, and sends the data to the processing module 4. The processing module 4 starts the comparison module 5 and sends the data to the display module 13. Comparison module 5, comparison module 5 mines the data inside storage module 12 for comparison, finds out whether the user's walking action completes the specified action in a standard manner, and sends the result to display module 13 for display; Step 4, upper arm monitoring module 3 obtains the data of palm grip and arm support force through grip pressure sensor 10 and support pressure sensor 11, and sends the data to processing module 4, processing module 4 starts comparison module 5 and sends the data to comparison module 5, comparison module 5 mines the data inside storage module 12 for comparison, and adjusts the training intensity and content according to whether there is compensation phenomenon according to the comparison result, so as to improve upper limb strength and balance ability in a targeted manner. ; The comparison calculation method of the comparison module 5 is: converting the user data at each time end obtained by various sensors into a two-dimensional matrix, the comparison module 5 respectively adds the values ​​in various two-dimensional matrices to obtain the sum of various two-dimensional matrices, and then respectively calculates the sum of the sums of various two-dimensional matrices, and then divides each sum obtained by the number n of each two-dimensional matrix to obtain various operation weights, and various operation weights represent the average value of the sum of each two-dimensional matrix, and then respectively compares this operation weight with various correct data operation weights input in advance; if the absolute value of the difference between the two is greater than the preset parameter, it is determined that the user's action is not standard, otherwise it is standard;

[0034] The gravity monitoring module 1 automatically records the center of gravity data of the user's training each time, accurately obtains the center of gravity position change data, and evaluates the user's balance ability. The position monitoring module 2 accurately identifies the user's position changes in space, and evaluates whether the user's walking movements complete the specified movements in a standard manner. The upper arm monitoring module 3 records the user's upper limb force data in real time, analyzes the compensation phenomenon, and adjusts the training intensity and method to improve the effect. The sound and light reminder module reminds the user of training movements, training market, and the output of various appreciation languages ​​through the sound and light module during training.

[0035] Furthermore, the step of mining the internal data of the storage module 12 of the comparison module 5 is as follows: Step 21, extracting a feature from the storage data inside the storage module 12, and normalizing the storage data inside the storage module 12 by using the feature, so that the data value ranges corresponding to each feature of the storage data inside the storage module 12 are consistent, forming compressed code training data:

[0036] Using the compressed code training data, the encoder is trained according to the following formula that minimizes the objective function of classification error: Among them, a ik is the weight coefficient, l(y ik .f(X i )) is the loss function, and the expression of the loss function is l(a)=(0,1-a), y ik ∈(-1,1) is X i The category identifier of the kth category, X i is the i-th compressed code training data, W k m is the classification parameter corresponding to the mth feature in the kth category, b k is the bias parameter, M is the number of features, A is the projection matrix, and A (m) is the projection matrix corresponding to the mth feature, φ(A (m) X i (m) ) is a hash function, N is the number of compressed code training data, K is the number of categories of compressed code training data, g(W) and h(A) are two normalization functions, which are used to adjust the classification parameter matrix W and the projection matrix A respectively, λ1 and λ2 are two real numbers, which are used to adjust the normalization function g(W) and h(A) respectively; after training, the projection matrix A, classification parameter matrix W and bias matrix B are obtained, B={b1,b2,...b k}; and the hash function: φ(A (m) X i (m)) as a binary compression code encoder; step 22, using a classifier based on the binary compression code to classify the binary compression code of the internal data of the storage module 12, to obtain the storage data category inside the storage module 12, and classify the binary compression code of the internal data of the storage module 12 through the following function:

[0037] Step 23: Use the cosine method to calculate the similarity between the user data obtained by various sensors and the data of each category in turn, and select the category with the highest similarity. The formula for calculating the similarity using the cosine method is as follows: where x1, x2…x n is the user data obtained by various sensors, y1, y2…y n For each category of data, the smaller the cosθ, the higher the similarity. The feature extraction method is: construct an iterative formula of the sample matrix based on the stored data inside the storage module 12 The sample matrix is ​​iteratively updated according to the iterative formula to obtain the feature matrix, where β0 is a constant, a random initial value of β0 = 1 is given, Y0 = H0, and it is iteratively updated according to the iterative formula. When the number of iterations reaches the set threshold k, the iteration is terminated and H is obtained. k That is, the feature matrix, k represents the product of the matrix, Y is the intermediate parameter, P() represents the projection gradient algorithm, and P(Z) means projecting all negative numbers in the matrix Z to 0;

[0038] By iteratively updating the sample matrix according to the iterative formula, the optimal feature matrix and the optimal feature vector are obtained, and by classifying the stored data, not only the accuracy of data mining is improved, but also the efficiency of data mining is improved, thereby improving the speed of device comparison calculation, and quickly understanding the user's training situation.

[0039] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0040] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions. The sentence "includes an element defined by ... does not exclude the existence of other identical elements in the process, method, article or device including the element".

[0041] 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 system for training the balance ability of people with walking disabilities, characterized in that: The device comprises an input module (7), a gravity monitoring module (1), a position monitoring module (2), an upper arm monitoring module (3), a storage module (12), a processing module (4), a comparison module (5) and an audio-visual reminder module (15), wherein the gravity monitoring module (1) comprises a gravity sensor (6); the position monitoring module (2) comprises a position switch (14), an accelerometer (8) and an ultrasonic sensor (9); and the upper arm monitoring module (3) comprises a grip pressure sensor (10) and a support pressure sensor (11); Step 1: inputting the upper limb force compensation operation weight, the correct standing center of gravity operation weight and the correct walking posture operation weight into the storage module (12) through the input module (7); Step 2, the gravity monitoring module (1) obtains the center of gravity transfer data of the user when standing with both feet upright and standing on one foot through the gravity sensor (6), and sends the data to the processing module (4); the processing module (4) starts the comparison module (5) and sends the data to the comparison module (5); the comparison module (5) mines the data inside the storage module (12) for comparison, and finds out whether the user's balance ability is standard, and sends the result to the display module (13) for display; Step 3, the position monitoring module (2) accurately identifies the data of the position change of the user's legs in space through the position switch (14), the accelerometer (8) and the ultrasonic sensor (9), and sends the data to the processing module (4); the processing module (4) starts the comparison module (5) and sends the data to the comparison module (5); the comparison module (5) mines the data inside the storage module (12) for comparison, and finds out whether the user's walking action completes the specified action in a standard manner, and sends the result to the display module (13) for display; Step 4, the upper arm monitoring module (3) obtains data on the grip force of the palm and the support force of the arm through the grip pressure sensor (10) and the support pressure sensor (11), and sends the data to the processing module (4); the processing module (4) starts the comparison module (5) and sends the data to the comparison module (5); the comparison module (5) mines the data inside the storage module (12) for comparison, and adjusts the training intensity and content according to whether there is a compensation phenomenon based on the comparison results, so as to improve the upper limb strength and balance ability in a targeted manner.

2. The balance training system for people with walking disabilities according to claim 1, characterized in that: The steps of mining the internal data of the storage module (12) of the comparison module (5) are as follows: Step 21, extracting a feature from the stored data in the storage module (12), and normalizing the stored data in the storage module (12) by using the feature, so that the data value ranges corresponding to the various features of the stored data in the storage module (12) are consistent, thereby forming compressed code training data: Using the compressed code training data, encoder training is performed according to the following formula that minimizes the objective function of classification error: Among them, a ik is the weight coefficient, l(y ik .f(X i )) is the loss function, and the expression of the loss function is l(a)=(0,1-a), y ik ∈(-1,1) is X i The category identifier of the kth category, X i is the i-th compressed code training data, is the classification parameter corresponding to the mth feature in the kth category, b k is the bias parameter, M is the number of features, A is the projection matrix, and A (m) is the projection matrix corresponding to the mth feature, φ(A (m) X i (m) ) is a hash function, N is the number of the compressed code training data, K is the number of categories of the compressed code training data, g(W) and h(A) are two normalization functions, which are used to adjust the classification parameter matrix W and the projection matrix A respectively, λ1 and λ2 are two real numbers, which are used to adjust the normalization functions g(W) and h(A) respectively; After training, the projection matrix A, the classification parameter matrix W and the bias matrix B are obtained, where B = {b1, b2, ...b k }; and the hash function: φ(A (m) X i (m) ) as a binary compression code encoder; Step 22: using a classifier based on binary compression codes to classify the binary compression codes of the internal data of the storage module (12), obtaining the category of the stored data in the storage module (12), and classifying the binary compression codes of the internal data of the storage module (12) by the following function: Step 23: Use the cosine method to calculate the similarity between the user data obtained by various sensors and the data of each category in turn, and select the category with the highest similarity. The formula for calculating the similarity using the cosine method is as follows: where x1, x2…x n is the user data obtained by various sensors, y1, y2…y n For each category of data, the smaller cosθ is, the higher the similarity is.

3. The balance training system for people with walking disabilities according to claim 2, characterized in that: The feature extraction method is: An iterative formula for constructing a sample matrix based on the stored data in the storage module (12) The sample matrix is ​​iteratively updated according to the iterative formula to obtain a feature matrix, wherein β0 is a constant, a random initial value of iteration β0=1 is given, Y0=H0, and it is iteratively updated according to the iterative formula. When the number of iterations reaches the set threshold k, the iteration is terminated to obtain H k That is, k represents the product of the matrices, Y is the intermediate parameter, P( ) represents the projected gradient algorithm, and P(Z) represents projecting all negative numbers in the matrix Z to 0.

4. The balance training system for people with walking disabilities according to claim 3, characterized in that: The comparison calculation method of the comparison module (5) is: The user data at each time end obtained by various sensors is converted into a two-dimensional matrix. The comparison module adds the values ​​in various two-dimensional matrices respectively to obtain the sum of various two-dimensional matrices, and then calculates the sum of the sums of each two-dimensional matrix respectively. Then, each sum is divided by the number n of each two-dimensional matrix to obtain various operation weights. Various operation weights represent the average value of the sum of each two-dimensional matrix. Then, this operation weight is compared with various correct data operation weights input in advance. If the absolute value of the difference between the two is greater than the preset parameter, it is determined that the user's action is not standard, otherwise it is standard.

Images