Intelligent control method of hand rehabilitation training device

Abstract

The application provides an intelligent control method for a hand rehabilitation training device, and relates to the technical field of device control. The method comprises the following steps: collecting pressure information and bending information, then acquiring image information, performing offset analysis, acquiring data collection offset categories, then analyzing a real-time training stage, performing assistance analysis, and then controlling the device, calculating force information array, and controlling the device through an impedance analyzer. The application mainly solves the problem that the existing hand rehabilitation training device lacks intelligent control, resulting in inaccurate device control and unstable training effect, and cannot be individually adjusted according to individual differences of users. Sensor technology is used to monitor the hand movement and strength of users, and the monitoring data is transmitted to the control system. The control system automatically adjusts the movement and strength of the training device according to the monitoring data and the preset training plan, thereby providing personalized rehabilitation training for users.

Description

Technical Field

[0001] This application relates to the field of equipment control technology, specifically to an intelligent control method for hand rehabilitation training equipment. Background Technology

[0002] Rehabilitation medicine is an applied medical discipline that studies the rehabilitation of people with disabilities and their families. It aims to help users restore physical function and improve their quality of life through physical therapy, exercise therapy, drug therapy, speech therapy, and other methods. Hand rehabilitation training is an important research direction within rehabilitation medicine. Because the hand has delicate and complex movements and functions, hand rehabilitation training requires precise control of the movement and force of training equipment to ensure effective rehabilitation for users. With continuous technological advancements, intelligent control technology has been widely applied in various fields. In rehabilitation medicine, intelligent control technology has also been applied to the control methods of hand rehabilitation training equipment. By adopting intelligent control methods, hand rehabilitation training equipment can more precisely control movement and force, thereby providing users with more effective rehabilitation training.

[0003] However, in the process of implementing the technical solution of the invention in the embodiments of this application, it was found that the above-mentioned technology has at least the following technical problems:

[0004] Existing hand rehabilitation training equipment lacks intelligent control, resulting in inaccurate equipment control and unstable training effects, and it cannot be personalized according to the individual differences of users. Summary of the Invention

[0005] This application primarily addresses the lack of intelligent control in existing hand rehabilitation training equipment, which leads to inaccurate equipment control, unstable training effects, and the inability to personalize adjustments based on individual user differences. It employs sensor technology to monitor the user's hand movements and force, transmitting the monitoring data to the control system. Based on the monitoring data and a preset training plan, the control system automatically adjusts the movement and force of the training equipment, thereby providing personalized rehabilitation training for the user.

[0006] In view of the above problems, in a first aspect, this application provides an intelligent control method for a hand rehabilitation training device, the method comprising: when a user uses the hand rehabilitation training device, acquiring pressure information and bending information at multiple locations within each designated finger training unit through a pressure sensor array and a strain-bending sensor, obtaining at least one pressure information array and at least one bending information; acquiring images of the user using the hand rehabilitation training device, obtaining usage image information, performing data acquisition offset analysis, obtaining a data acquisition offset category, correcting the at least one pressure information array and at least one bending information, obtaining at least one corrected pressure information array and at least one corrected bending information; and analyzing the user's hair loss based on the at least one corrected pressure information array and at least one corrected bending information. The system obtains at least one force intention information. When the at least one force intention information is greater than the force intention threshold, it analyzes and obtains a real-time training phase based on the at least one corrected bending information. Combining the at least one corrected pressure information array, it performs assistance analysis through an assistance training analyzer to obtain at least one assistance information. According to the at least one assistance information, it performs equipment control and obtains at least one corrected assistance pressure information array after control. Combining the at least one assistance information, it calculates and obtains at least one force information array. According to the at least one force information array, it performs resistance analysis through an impedance training analyzer to obtain at least one resistance information. It performs equipment control, and when the at least one corrected bending information reaches the next training phase, it continues to perform assistance analysis and resistance analysis and perform equipment control.

[0007] Secondly, this application provides an intelligent control system for a hand rehabilitation training device. The system includes: an information acquisition module, used to acquire pressure and bending information at multiple locations within each designated finger training unit via a pressure sensor array and a strain-bending sensor when the user uses the hand rehabilitation training device, obtaining at least one pressure information array and at least one bending information; an image information acquisition module, used to acquire images of the user using the hand rehabilitation training device, obtain image information, perform data acquisition offset analysis, obtain data acquisition offset categories, and correct the at least one pressure information array and at least one bending information, obtaining at least one corrected pressure information array and at least one corrected bending information; and a force intention information acquisition module, which analyzes the user's force application based on the at least one corrected pressure information array and at least one corrected bending information. The system comprises: an intention module for obtaining at least one force exertion intention information; an assistance analysis module for analyzing and obtaining a real-time training phase based on at least one corrected bending information when at least one force exertion intention information exceeds a force exertion intention threshold, and combining the at least one corrected pressure information array with an assistance training analyzer to obtain at least one assistance information; a force information array acquisition module for controlling the equipment according to the at least one assistance information and acquiring at least one corrected assistance pressure information array after control, and calculating at least one force information array based on the at least one assistance information; and an equipment control module for performing resistance analysis based on the at least one force information array with an impedance training analyzer to obtain at least one resistance information, controlling the equipment, and continuing to perform assistance and resistance analysis and equipment control when at least one corrected bending information reaches the next training phase.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] This application provides an intelligent control method for a hand rehabilitation training device, relating to the field of device control technology. The method includes: collecting pressure information and bending information, then acquiring usage image information, performing offset analysis, obtaining data acquisition offset category, then analyzing the real-time training stage, performing assistance analysis, and then controlling the device, calculating and obtaining a force information array, and controlling the device through an impedance analyzer.

[0010] This application primarily addresses the lack of intelligent control in existing hand rehabilitation training equipment, which leads to inaccurate equipment control, unstable training effects, and the inability to personalize adjustments based on individual user differences. Typically, sensor technology is used to monitor the user's hand movements and force, transmitting the data to the control system. Based on the monitoring data and a pre-set training plan, the control system automatically adjusts the movement and force of the training equipment, thereby providing personalized rehabilitation training for the user.

[0011] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0013] Figure 1 This application provides a schematic flowchart of an intelligent control method for a hand rehabilitation training device.

[0014] Figure 2 This application provides a flowchart illustrating a method for obtaining the data acquisition offset category in an intelligent control method for a hand rehabilitation training device.

[0015] Figure 3 This application provides a schematic flowchart of a method for obtaining at least one force exertion intention information in an intelligent control method for a hand rehabilitation training device.

[0016] Figure 4 This application provides a schematic diagram of the structure of an intelligent control system for a hand rehabilitation training device.

[0017] Explanation of reference numerals in the attached diagram: Information acquisition module 10, Image information acquisition module 20, Force intention information acquisition module 30, Assist analysis module 40, Force information array acquisition module 50, Equipment control module 60. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] This application primarily addresses the lack of intelligent control in existing hand rehabilitation training equipment, which leads to inaccurate equipment control, unstable training effects, and the inability to personalize adjustments based on individual user differences. Typically, sensor technology is used to monitor the user's hand movements and force, transmitting the data to the control system. Based on the monitoring data and a pre-set training plan, the control system automatically adjusts the movement and force of the training equipment, thereby providing personalized rehabilitation training for the user.

[0020] To better understand the above technical solution, the following will provide a detailed description of the solution in conjunction with the accompanying drawings and specific implementation methods:

[0021] Example 1

[0022] like Figure 1 The present invention discloses an intelligent control method for a hand rehabilitation training device. The method is applied to a hand rehabilitation training device comprising five finger training units, each finger training unit being equipped with a pressure sensor array and a strain bending sensor. The method includes:

[0023] When a user uses the hand rehabilitation training device, pressure information and bending information at multiple locations within each designated finger training unit are collected through a pressure sensor array and a strain bending sensor, resulting in at least one pressure information array and at least one bending information.

[0024] Specifically, when a user uses the hand rehabilitation training device, pressure and bending information at multiple locations within each designated finger training unit are collected via a pressure sensor array and a strain-bending sensor, resulting in at least one pressure information array and at least one bending information array. This information can be used to monitor the user's hand movements and strength in real time, as well as to process and analyze training data. The hand rehabilitation training device is used for rehabilitation training of movements such as gripping after a finger injury. It includes five finger training units, for example, composed of airbags, which can apply pressure to the back or front of each finger, providing assistance or resistance to gripping and other movements. The pressure sensor array can monitor pressure information at multiple locations within each finger training unit, including pressure at the joints and fingertips of each finger. This pressure information reflects finger muscle strength and motor coordination. The strain-bending sensor can monitor bending information at multiple locations within each finger training unit, including bending angles at the joints and fingertips of each finger. This bending information reflects finger joint range of motion and flexibility. By collecting pressure and bending information from multiple locations within each designated finger training unit, at least one pressure information array and at least one bending information array can be obtained, providing more comprehensive and accurate hand rehabilitation training data.

[0025] Images of the user using the hand rehabilitation training device are acquired to obtain usage image information. Data acquisition offset analysis is performed to obtain data acquisition offset categories. The at least one pressure information array and at least one bending information are corrected to obtain at least one corrected pressure information array and at least one corrected bending information.

[0026] Specifically, when a user uses the hand rehabilitation training device, images of the user using the device can be captured to obtain usage image information. This image information may include data such as the user's hand movements and strength, as well as data such as the position and posture of the device and finger training units. By processing and analyzing the captured images, data acquisition offset analysis can be performed to obtain data acquisition offset categories. Data acquisition offset analysis involves preprocessing the captured usage image information, such as removing noise and performing image enhancement to improve image quality and clarity; extracting features from the preprocessed image information, such as extracting hand contours and finger joint positions; and classifying and identifying data acquisition offset categories based on the extracted features. Data acquisition offset categories may include categories such as inaccurate sensor position and incorrect finger joint position; based on the classification and identification results, the data acquisition offset categories are processed and corrected accordingly. For example, for the problem of inaccurate sensor position, the sensor position can be adjusted or the captured data can be corrected; for the problem of incorrect finger joint position, the user can be reminded to adjust the position of the finger joints or correct the captured data. The data collected may be offset due to changes in the device's position and posture, or the position and posture of the finger training unit. Since users may wear the hand rehabilitation training device improperly, such as with misalignment, or because the user's hand is too large or too small, causing discrepancies between the pressure sensor and the corresponding joint position, it is necessary to correct the data based on the type of offset (e.g., upward or downward, left or right, different offset distances) identified through image analysis. During the user's use of the hand rehabilitation training device, image information can be collected, and data acquisition offset analysis can be performed. Through data acquisition offset analysis, the data acquisition offset category can be obtained, and the collected pressure information array and bending information can be corrected. The corrected information can more accurately reflect the user's hand muscle strength, motor coordination, and flexibility, obtaining at least one corrected pressure information array and at least one corrected bending information. Correction methods may include operations such as correcting and calibrating the raw data according to the data acquisition offset category, thereby obtaining more accurate and reliable corrected pressure information arrays and corrected bending information. The corrected pressure information array and bending information can more accurately reflect the user's hand muscle strength, motor coordination, joint range of motion, and flexibility.

[0027] Based on the at least one correction pressure information array and at least one correction bending information, the user's force application intention is analyzed to obtain at least one force application intention information;

[0028] Specifically, based on the at least one corrected pressure information array and at least one corrected bending information, the user's force application intention can be further analyzed to obtain at least one force application intention information. Force application intention refers to the user's control of the training equipment's force and movement during training through the contraction and relaxation of hand muscles, thereby achieving the specified training goal. By analyzing the corrected pressure information array and bending information, the user's force application intention information during training can be extracted. Analysis of the corrected pressure information array can extract information such as the user's finger joint force changes and force application patterns during training. Simultaneously, analysis of the corrected bending information can extract information such as the user's finger joint mobility and flexibility during training. By comprehensively analyzing this information, at least one force application intention information can be obtained. This force application intention information may include the user's force application pattern, finger joint force changes, and mobility. Furthermore, force application intention information can also be used to evaluate and optimize the performance of the training equipment, improving its accuracy and stability. For example, by comparing and analyzing the force application intention information of multiple users, differences and commonalities among different users during training can be identified, providing a reference for equipment improvement and optimization.

[0029] When at least one force intention information is greater than the force intention threshold, the real-time training phase is obtained based on the at least one corrected bending information. Combined with the at least one corrected pressure information array, the assist training analyzer is used to perform assist analysis to obtain at least one assist information.

[0030] Specifically, when at least one force intention information exceeds a force intention threshold, the real-time training stage can be determined based on the at least one corrected bending information. Combining the at least one corrected pressure information array, an assistive training analyzer is used to perform assistive analysis and obtain at least one assistive information. When a user generates a force intention during training, the corresponding force intention information will exceed the force intention threshold. Based on these force intention information exceeding the threshold, the corrected bending information can be analyzed to determine the user's current real-time training stage. After obtaining the real-time training stage, the corrected pressure information array can be combined with the assistive training analyzer for assistive analysis. The assistive training analyzer can be trained based on machine learning or deep learning algorithms to classify and identify the user's assistive behavior based on the combination of five sample pressure information arrays and five sample assistive information sets from the real-time training stage. Through assistive analysis, at least one assistive information can be obtained, which can reflect the user's finger joint assist patterns, force changes, and range of motion during training. In summary, by monitoring the user's force intention information, it can be determined whether the user is exerting force, and the real-time training stage the user is in can be determined based on the corrected bending information. By combining the calibration pressure information array with the assist training analyzer, assist information reflecting the user's current assist status can be obtained.

[0031] Based on the at least one assist information, the equipment is controlled, and at least one corrected assist pressure information array after control is obtained. Combined with the at least one assist information, at least one force information array is calculated.

[0032] Specifically, based on the at least one assist information, device control can be performed, and at least one corrected assist pressure information array after control can be obtained. Combining the at least one assist information, at least one force information array can be calculated. Device control refers to controlling the auxiliary force or resistance of the training device according to the assist information to help or hinder the user from achieving the specified training goal. Device control can be achieved by adjusting the mechanical structure, electronic control system, or software algorithm of the training device. An assist training analyzer can control the magnitude, direction, and trend of the auxiliary force or resistance of the training device based on the assist information. The corrected assist pressure information array after control can reflect the changes in the auxiliary force or resistance of the training device at different positions of the finger joints. The force information array refers to the information matrix of the user's force application calculated based on the assist information and the corrected pressure information array. By analyzing and calculating the assist information and the corrected pressure information array, the force application of each finger joint at different positions can be obtained, thus forming at least one force information array. The force information array can reflect the user's finger joint force application pattern, force changes, and range of motion during training. Simultaneously, the force information array can also be used to evaluate and optimize the performance of the training device, improving its accuracy and stability. By collecting and analyzing images, pressure information arrays, bending information, and force exertion intentions of users using hand rehabilitation training equipment, more comprehensive and accurate hand rehabilitation training data can be obtained.

[0033] Based on the at least one force information array, resistance analysis is performed through an impedance training analyzer to obtain at least one resistance information, and equipment control is performed. When at least one corrected bending information reaches the next training stage, assist analysis and resistance analysis are performed again for equipment control.

[0034] Specifically, based on the at least one force information array, resistance analysis can be performed using an impedance training analyzer to obtain at least one resistance information, and then the device can be controlled. When at least one corrected bending information reaches the next training stage, assistance analysis and resistance analysis continue, and device control is maintained. Impedance training refers to adjusting the resistance of the training device to help or hinder the user from achieving a specified training goal. Impedance training can reflect the performance of the training device and the user's training state. The force information array can be analyzed using an impedance training analyzer to obtain the resistance at different positions of each finger joint, thus obtaining at least one resistance information. This resistance information reflects the changes in the resistance of the training device at different finger joint positions. Based on the resistance information, device control can be performed by adjusting the mechanical structure, electronic control system, or software algorithm of the training device to change its resistance. When at least one corrected bending information reaches the next training stage, assistance analysis and resistance analysis continue, and device control is maintained. This helps the user gradually adapt to different training stages and resistance levels, thereby improving the effectiveness and quality of hand rehabilitation training. Based on the user's force exertion, resistance analysis is performed to conduct resistance training for grasping exercises, increasing training intensity. Data analysis also improves the accuracy of device control and enhances training effectiveness. By collecting and analyzing images, pressure information arrays, bending information, and force exertion intentions of users using hand rehabilitation training equipment, more comprehensive and accurate hand rehabilitation training data can be obtained.

[0035] Furthermore, the method of this application, when a user uses the hand rehabilitation training device, collects pressure and bending information at multiple locations within each designated finger training unit through a pressure sensor array and a strain-bending sensor, obtaining at least one pressure information array and at least one bending information, including:

[0036] When a user uses the hand rehabilitation training device, at least one finger training unit is determined and activated based on the finger that the user needs to train.

[0037] By using a pressure sensor array and a strain bending sensor within at least one finger training unit, pressure and bending information at multiple locations within the finger training unit are collected to obtain at least one pressure information array and at least one bending information.

[0038] Specifically, when a user uses the hand rehabilitation training device, at least one finger training unit can be determined and activated based on the finger the user needs to train. The finger training unit can be a mechanical, electronic, or software system targeting a single finger, multiple fingers, or the entire palm, used to simulate and train finger joint movements, muscle strength, and sensory feedback. Pressure and bending information at multiple locations within the finger training unit can be collected using a pressure sensor array and a strain-bending sensor. The pressure sensor array monitors pressure at different locations on the finger, including pressure at joints and fingertips. The strain-bending sensor monitors the bending angle and force at different locations on the finger, including bending angles and forces at joints and fingertips. By collecting pressure and bending information from multiple locations within the finger training unit, at least one pressure information array and at least one bending information can be obtained. This information reflects indicators such as muscle strength, motor coordination, and flexibility of the finger at different locations. After obtaining at least one pressure information array and at least one bending information, it can be transmitted to an assistive training analyzer and an impedance training analyzer for analysis and processing. The assisted training analyzer can calculate the user's force application intention and assistance information based on pressure and bending information, which is used to control the auxiliary force or resistance of the training equipment. The impedance training analyzer can calculate the user's resistance information based on pressure information, which is used to control the resistance of the training equipment. Based on the assistance and resistance information, equipment control can be implemented by adjusting the mechanical structure, electronic control system, or software algorithm of the training equipment to change the auxiliary force or resistance and resistance level. In summary, by collecting and analyzing data such as pressure and bending information from users using hand rehabilitation training equipment, more comprehensive and accurate hand rehabilitation training data can be obtained.

[0039] Furthermore, the method of this application, which involves acquiring images of the user using the hand rehabilitation training device, obtaining usage image information, performing data acquisition offset analysis to obtain data acquisition offset categories, and correcting the at least one pressure information array and at least one bending information, includes:

[0040] Images of the user using the hand rehabilitation training device are captured to obtain usage image information;

[0041] Capture an image of the user's hand;

[0042] Based on the image information and hand image, an analysis of data acquisition offset caused by wearing offset is performed to obtain the data acquisition offset category;

[0043] Based on the data acquisition offset category, a mapping analysis is performed within the data calibration table to obtain at least one pressure calibration coefficient array and at least one bending calibration coefficient. The data calibration table includes the mapping relationship between the data acquisition category and the pressure calibration coefficients and full lesson coefficients of bending information and pressure parameters at different positions within different training units.

[0044] Using the at least one pressure calibration coefficient array and the at least one bending calibration coefficient, the at least one pressure information array and the at least one bending information are corrected to obtain the at least one corrected pressure information array and the at least one corrected bending information.

[0045] Specifically, during the user's use of the hand rehabilitation training device, images of the user using the device can be captured to obtain usage image information. Simultaneously, images of the user's hand can also be captured to understand the user's initial hand state and changes in the hand during training. Based on the usage image information and hand images, data acquisition offset analysis due to wearing offset can be performed to obtain data acquisition offset categories. Wearing offset refers to the offset in acquired data caused by inaccurate wearing method and position of the training device, such as inaccurate sensor position or incorrect finger joint position. Based on the data acquisition offset categories, mapping analysis is performed within a data calibration table to obtain at least one pressure calibration coefficient array and at least one bending calibration coefficient. The data calibration table includes the mapping relationship between data acquisition categories and pressure calibration coefficients and full lesson coefficients for bending information and pressure parameters at different positions within different training units. These mapping relationships are calculated using a large amount of pre-collected data and corresponding calibration coefficients and can be used to correct data acquisition offsets caused by wearing offset. Using the at least one pressure calibration coefficient array and at least one bending calibration coefficient, correction calculations can be performed on the at least one pressure information array and at least one bending information to obtain at least one corrected pressure information array and at least one corrected bending information. These corrected information can more accurately reflect indicators such as the user's hand muscle strength, motor coordination, and flexibility. In summary, by collecting and analyzing information such as images, pressure information arrays, bending information, and force exertion intentions of users using hand rehabilitation training equipment, more comprehensive and accurate hand rehabilitation training data can be obtained.

[0046] Furthermore, such as Figure 2 As shown, the method of this application, based on the image information and hand image, performs data acquisition offset analysis due to wearing offset to obtain data acquisition offset categories, including:

[0047] Based on the usage data records of the hand rehabilitation training device, a sample usage image information set, a sample hand image set, and a sample data acquisition offset category set are obtained;

[0048] Based on a convolutional neural network, the sample image information set and the sample hand image set are used as inputs, and the sample data acquisition offset category set is used as output to construct an offset classifier. The offset classifier includes a first offset classification branch and a second offset classification branch.

[0049] The offset classifier is used to identify the image information and hand image to obtain a first offset category and a second offset category. When the first offset category and the second offset category are consistent, the data acquisition offset category is obtained.

[0050] Specifically, based on the usage data records of the hand rehabilitation training device, a set of sample usage image information, a set of sample hand images, and a set of sample data acquisition offset categories can be obtained. This sample data can be used to train and use an offset classifier. Based on a convolutional neural network, the set of sample usage image information and the set of sample hand images can be used as input, combined with the set of sample data acquisition offset categories as output, to construct a usage offset classifier. This usage offset classifier includes a first usage offset classification branch and a second usage offset classification branch. Each branch can be a convolutional neural network used to extract features from the input image information and classify them based on these features. Using the offset classifier, the usage image information and hand images can be identified to obtain a first offset category and a second offset category. When the first offset category and the second offset category are consistent, the data acquisition offset category can be obtained. This data acquisition offset category can be used to correct the acquired pressure and bending information to improve the accuracy and reliability of the data. In summary, by constructing and using an offset classifier, data acquisition offsets caused by wearing offsets can be identified more accurately, and the acquired data can be corrected. This helps to improve the effectiveness and quality of hand rehabilitation training.

[0051] Furthermore, such as Figure 3 As shown, the method of this application, based on the at least one correction pressure information array and at least one correction bending information, analyzes the user's force application intention to obtain at least one force application intention information, including:

[0052] Based on the user's usage data records of the five finger training units, a set of five sample pressure information arrays, a set of five sample bending information sets, and a set of force intention information sets are obtained.

[0053] By using the five sample pressure information array sets and the five sample bending information sets respectively, and combining them with the force intention information set, a force intention analyzer including five force intention analysis branches is constructed.

[0054] The at least one corrected pressure information array and the at least one corrected bending information are input into the corresponding at least one force application intention analysis branch to obtain the at least one force application intention information.

[0055] Specifically, based on the user's usage data records of the five finger training units, five sets of sample pressure information arrays, five sets of sample bending information arrays, and a set of force intention information can be obtained. This sample data can be used to train and construct a force intention analyzer. The five sets of sample pressure information arrays and five sets of sample bending information arrays can be combined with the force intention information set to construct a force intention analyzer including five force intention analysis branches. Each force intention analysis branch can be a neural network or algorithm model used to extract features from the input pressure and bending information, and to classify and identify force intentions based on these features. By inputting at least one corrected pressure information array and at least one corrected bending information array into the corresponding at least one force intention analysis branch, at least one force intention information can be obtained. This force intention information can reflect the user's finger joint force patterns, strength changes, and range of motion during training. In summary, by collecting and analyzing the pressure information, bending information, and force intention information of users using hand rehabilitation training equipment, more comprehensive and accurate force intention information can be obtained.

[0056] Furthermore, the method of this application, based on the at least one corrected bending information, analyzes and obtains a real-time training phase, and combines the at least one corrected pressure information array with an assist training analyzer to perform assist analysis and obtain at least one assist information, including:

[0057] Based on the five finger training units, five sets of sample bending information and five sets of training stages are collected to construct a training stage classifier.

[0058] Based on the training stage classifier, the at least one corrected bending information is classified to obtain at least one training stage, and the training stage with the highest occurrence rate is selected as the real-time training stage.

[0059] Based on the historical training data corresponding to the real-time training phase, an auxiliary training analyzer is constructed. The auxiliary training analyzer is trained based on the combination of five sample pressure information arrays and five sample auxiliary information sets in the real-time training phase, and includes five auxiliary training analysis branches.

[0060] The assist training analyzer is used to perform assist analysis on the at least one correction pressure information array to obtain the at least one assist information.

[0061] Specifically, based on the five finger training units, five sets of sample bending information and five sets of training stages can be collected to construct a training stage classifier. The training stage classifier can be based on machine learning or deep learning methods, learning how to classify different training stages through training sample data. Based on the training stage classifier, the at least one corrected bending information can be classified to obtain at least one training stage. The training stage with the highest occurrence rate is selected as the real-time training stage. The real-time training stage refers to the training stage the user is currently in. Based on the historical assist training data corresponding to the real-time training stage, an assist training analyzer can be constructed. The historical assist training data includes the combination of five sample pressure information arrays and five sample assist information sets from the real-time training stage. The assist training analyzer can adopt a similar structure and method to the force intention analyzer, constructed and trained based on deep learning or machine learning algorithms. Using the assist training analyzer, assist analysis can be performed on the at least one corrected pressure information array to obtain the at least one assist information. This assistive information can reflect the assistive patterns, force changes, and range of motion of the user's finger joints during training. In summary, by collecting and analyzing information such as the bending information, assistive information, and force exertion intention of the user when using hand rehabilitation training equipment, more comprehensive and accurate assistive information can be obtained.

[0062] Furthermore, the method of this application, based on the at least one force information array, performs resistance analysis through an impedance training analyzer to obtain at least one resistance information for equipment control, including:

[0063] Based on the historical resistance training data corresponding to the real-time training phase, obtain a set of five sample force information arrays and a set of five sample resistance information arrays.

[0064] A resistance training analyzer is constructed using the five sets of force information arrays and the five sets of resistance information arrays, respectively, wherein the resistance training analyzer includes five resistance training analysis branches;

[0065] The resistance training analyzer is used to perform resistance analysis on the at least one force information array to obtain at least one resistance information.

[0066] Specifically, based on the historical resistance training data corresponding to the real-time training phase, five sets of sample force information arrays and five sets of sample resistance information can be obtained. This sample data can be used to train and construct a resistance training analyzer. The resistance training analyzer can be constructed using the five sets of sample force information arrays and five sets of sample resistance information. Each resistance training analysis branch can be a neural network or algorithm model used to extract features from the input force and resistance information, and to classify and identify resistance conditions based on these features. Using the resistance training analyzer, resistance analysis can be performed on at least one force information array to obtain at least one resistance information. This resistance information can reflect the resistance pattern, force changes, and range of motion of the user's finger joints during training. In summary, by collecting and analyzing the force, resistance, and bending information of users using hand rehabilitation training equipment, more comprehensive and accurate resistance information can be obtained.

[0067] Example 2

[0068] Based on the same inventive concept as the intelligent control method for a hand rehabilitation training device described in the foregoing embodiments, such as Figure 4 As shown, this application provides an intelligent control system for a hand rehabilitation training device, the system comprising:

[0069] Information acquisition module 10 is used to acquire pressure information and bending information at multiple locations within each specified finger training unit by means of a pressure sensor array and a strain bending sensor when the user uses the hand rehabilitation training device, thereby obtaining at least one pressure information array and at least one bending information.

[0070] The image information acquisition module 20 is used to acquire images of the user using the hand rehabilitation training device, obtain image information, perform data acquisition offset analysis, obtain data acquisition offset category, and correct the at least one pressure information array and at least one bending information to obtain at least one corrected pressure information array and at least one corrected bending information.

[0071] The force intention information acquisition module 30 analyzes the user's force intention based on the at least one correction pressure information array and at least one correction bending information to obtain at least one force intention information.

[0072] The assist analysis module 40 is used to analyze and obtain the real-time training stage based on the at least one corrected bending information when at least one force intention information is greater than the force intention threshold, and to perform assist analysis through the assist training analyzer in combination with the at least one corrected pressure information array to obtain at least one assist information.

[0073] The force information array acquisition module 50 is used to control the equipment according to the at least one assist information, and acquire at least one corrected assist pressure information array after control, and calculate at least one force information array in combination with the at least one assist information.

[0074] The equipment control module 60 is used to perform resistance analysis based on the at least one force information array through an impedance training analyzer to obtain at least one resistance information, perform equipment control, and continue to perform assist analysis and resistance analysis when at least one correction bending information reaches the next training stage, and perform equipment control.

[0075] Furthermore, the system also includes:

[0076] The pressure information and bending information acquisition module is used to determine and activate at least one finger training unit based on the finger that the user needs to train when using the hand rehabilitation training device; and to collect pressure information and bending information at multiple locations within the finger training unit through a pressure sensor array and a strain bending sensor within the at least one finger training unit, thereby obtaining at least one pressure information array and at least one bending information.

[0077] Furthermore, the system also includes:

[0078] The calibration information acquisition module is used to acquire images of the user using the hand rehabilitation training device to obtain usage image information; acquire images of the user's hand; perform data acquisition offset analysis due to wearing offset based on the usage image information and hand images to obtain data acquisition offset categories; perform mapping analysis in a data calibration table based on the data acquisition offset categories to obtain at least one pressure calibration coefficient array and at least one bending calibration coefficient, wherein the data calibration table includes the mapping relationship between the data acquisition categories and the pressure calibration coefficients and complete lesson coefficients of bending information and pressure parameters at different positions in different training units; and use the at least one pressure calibration coefficient array and at least one bending calibration coefficient to perform correction calculations on the at least one pressure information array and at least one bending information to obtain the at least one corrected pressure information array and at least one corrected bending information.

[0079] Furthermore, the system also includes:

[0080] The offset category acquisition module acquires a set of sample usage image information, a set of sample hand images, and a set of sample data acquisition offset categories based on the usage data records of the hand rehabilitation training device. Using a convolutional neural network, it constructs a usage offset classifier by taking the set of sample usage image information and the set of sample hand images as inputs, and combining them with the set of sample data acquisition offset categories as outputs. The usage offset classifier includes a first usage offset classification branch and a second usage offset classification branch. Using the offset classifier, it identifies the usage image information and hand images to obtain a first offset category and a second offset category. When the first offset category and the second offset category are consistent, the data acquisition offset category is obtained.

[0081] Furthermore, the system also includes:

[0082] The force intention information acquisition module acquires five sets of sample pressure information arrays, five sets of sample bending information arrays, and a set of force intention information based on the user's usage data records of the five finger training units. It then uses the five sets of sample pressure information arrays and five sets of sample bending information arrays, combined with the set of force intention information, to construct a force intention analyzer comprising five force intention analysis branches. Finally, it inputs at least one corrected pressure information array and at least one corrected bending information array into the corresponding at least one force intention analysis branch to obtain at least one set of force intention information.

[0083] Furthermore, the system also includes:

[0084] The assist information acquisition module, based on the five finger training units, collects five sets of sample bending information and five sets of training stages to construct a training stage classifier. Based on the training stage classifier, it classifies the at least one corrected bending information to obtain at least one training stage, and selects the training stage with the highest occurrence rate as the real-time training stage. Based on the historical assist training data corresponding to the real-time training stage, it constructs an assist training analyzer, which is trained based on the combination of the five sample pressure information arrays and the five sample assist information sets of the real-time training stage, and includes five assist training analysis branches. Using the assist training analyzer, it performs assist analysis on the at least one corrected pressure information array to obtain the at least one assist information.

[0085] Furthermore, the system also includes:

[0086] The auxiliary training analyzer construction module obtains five sample force information array sets and five sample resistance information sets based on the historical resistance training data corresponding to the real-time training phase; it constructs a resistance training analyzer using the five sample force information array sets and the five sample resistance information sets respectively, wherein the resistance training analyzer includes five resistance training analysis branches; and it uses the resistance training analyzer to perform resistance analysis on at least one force information array to obtain at least one resistance information.

[0087] Through the detailed description of the intelligent control method for a hand rehabilitation training device described above, those skilled in the art can clearly understand the intelligent control system of the hand rehabilitation training device in this embodiment. As the system disclosed in the embodiment corresponds to the device disclosed in the embodiment, the description is relatively simple, and relevant parts can be referred to the method section.

[0088] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent control system for a hand rehabilitation training device, characterized in that, The system is applied to a hand rehabilitation training device, which includes five finger training units. Each finger training unit is equipped with a pressure sensor array and a strain bending sensor. The system includes: The information acquisition module is used to acquire pressure information and bending information at multiple locations within each specified finger training unit by means of a pressure sensor array and a strain bending sensor when the user uses the hand rehabilitation training device, thereby obtaining at least one pressure information array and at least one bending information. The image information acquisition module is used to acquire images of the user using the hand rehabilitation training device, obtain image information, perform data acquisition offset analysis, obtain data acquisition offset category, and correct the at least one pressure information array and at least one bending information to obtain at least one corrected pressure information array and at least one corrected bending information. The force intention information acquisition module analyzes the user's force intention based on the at least one correction pressure information array and at least one correction bending information to obtain at least one force intention information. The assist analysis module is used to analyze and obtain the real-time training stage based on the at least one corrected bending information when at least one force intention information is greater than the force intention threshold, and to perform assist analysis through the assist training analyzer in combination with the at least one corrected pressure information array to obtain at least one assist information. The force information array acquisition module is used to control the equipment according to the at least one assist information, and acquire at least one corrected assist pressure information array after control, and calculate at least one force information array in combination with the at least one assist information. The equipment control module is used to perform resistance analysis based on the at least one force information array through an impedance training analyzer to obtain at least one resistance information, perform equipment control, and continue to perform assist analysis and resistance analysis when at least one correction bending information reaches the next training stage, and perform equipment control.

2. The intelligent control system of the hand rehabilitation training device according to claim 1, characterized in that, The system includes: The pressure information and bending information acquisition module is used to determine and activate at least one finger training unit based on the finger that the user needs to train when using the hand rehabilitation training device; and to collect pressure information and bending information at multiple locations within the finger training unit through a pressure sensor array and a strain bending sensor within the at least one finger training unit, thereby obtaining at least one pressure information array and at least one bending information.

3. The intelligent control system of the hand rehabilitation training device according to claim 1, characterized in that, The system includes: The calibration information acquisition module is used to acquire images of the user using the hand rehabilitation training device to obtain usage image information; acquire images of the user's hand; perform data acquisition offset analysis due to wearing offset based on the usage image information and hand images to obtain data acquisition offset categories; perform mapping analysis in a data calibration table based on the data acquisition offset categories to obtain at least one pressure calibration coefficient array and at least one bending calibration coefficient, wherein the data calibration table includes the mapping relationship between the data acquisition categories and the pressure calibration coefficients and bending calibrations of bending information and pressure parameters at different positions in different training units; and use the at least one pressure calibration coefficient array and at least one bending calibration coefficient to perform correction calculations on the at least one pressure information array and at least one bending information to obtain the at least one corrected pressure information array and at least one corrected bending information.

4. The intelligent control system of the hand rehabilitation training device according to claim 1, characterized in that, The system includes: The offset category acquisition module acquires a set of sample usage image information, a set of sample hand images, and a set of sample data acquisition offset categories based on the usage data records of the hand rehabilitation training device. Using a convolutional neural network, it constructs a usage offset classifier by taking the set of sample usage image information and the set of sample hand images as inputs, and combining them with the set of sample data acquisition offset categories as outputs. The usage offset classifier includes a first usage offset classification branch and a second usage offset classification branch. Using the offset classifier, it identifies the usage image information and hand images to obtain a first offset category and a second offset category. When the first offset category and the second offset category are consistent, the data acquisition offset category is obtained.

5. The intelligent control system of the hand rehabilitation training device according to claim 1, characterized in that, The system includes: The force intention information acquisition module acquires five sets of sample pressure information arrays, five sets of sample bending information arrays, and a set of force intention information based on the user's usage data records of the five finger training units. It then uses the five sets of sample pressure information arrays and five sets of sample bending information arrays, combined with the set of force intention information, to construct a force intention analyzer comprising five force intention analysis branches. Finally, it inputs at least one corrected pressure information array and at least one corrected bending information array into the corresponding at least one force intention analysis branch to obtain at least one set of force intention information.

6. The intelligent control system of the hand rehabilitation training device according to claim 1, characterized in that, The system includes: The assist information acquisition module, based on the five finger training units, collects five sets of sample bending information and five sets of training stages to construct a training stage classifier. Based on the training stage classifier, it classifies the at least one corrected bending information to obtain at least one training stage, and selects the training stage with the highest occurrence rate as the real-time training stage. Based on the historical assist training data corresponding to the real-time training stage, it constructs an assist training analyzer, which is trained based on the combination of the five sample pressure information arrays and the five sample assist information sets of the real-time training stage, and includes five assist training analysis branches. Using the assist training analyzer, it performs assist analysis on the at least one corrected pressure information array to obtain the at least one assist information.

7. The intelligent control system of the hand rehabilitation training device according to claim 1, characterized in that, The system includes: The auxiliary training analyzer construction module obtains five sample force information array sets and five sample resistance information sets based on the historical resistance training data corresponding to the real-time training phase; it constructs a resistance training analyzer using the five sample force information array sets and the five sample resistance information sets respectively, wherein the resistance training analyzer includes five resistance training analysis branches; and it uses the resistance training analyzer to perform resistance analysis on at least one force information array to obtain at least one resistance information.

Images