Wrist joint rehabilitation training method and system based on motion perception

Abstract

The invention discloses a wrist joint rehabilitation training method and system based on motion perception, and relates to the technical field of wrist joint rehabilitation, and the method comprises the steps: collecting wrist joint perception parameters of a target user in real time according to a motion perception unit; inquiring whether historical wrist joint data exists in a wrist joint sensing record library or not; if not, using the user wrist joint data as wrist joint feature sensing data; obtaining a wrist joint rehabilitation training scheme based on the wrist joint feature sensing data; performing rehabilitation training on a target user based on the wrist joint rehabilitation training scheme to obtain training monitoring data; and based on the training monitoring data, adjusting the wrist joint rehabilitation training scheme, and generating a wrist joint training optimization scheme. The technical problems that in the prior art, wrist joint rehabilitation sensing precision is insufficient, a personalized training scheme is lacked, and clinical rehabilitation requirements are difficult to meet are solved, and the technical effects that the personalized rehabilitation training scheme is provided, and the rehabilitation effect and safety are improved are achieved.

Description

Technical Field

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

[0002] Rehabilitation after wrist injury is crucial but the process is complicated. Traditional rehabilitation methods rely on manual and simple equipment, and there are problems such as difficulty in accurately controlling the intensity of training and strong subjectivity in effect evaluation. With the development of intelligent technology, rehabilitation systems based on motion perception have attracted attention, but existing technologies have shortcomings. In terms of sensor technology, the perception accuracy is not enough, such as errors in measuring wrist angles and muscle activity. In the generation of training plans, there is a lack of personalization, and individual differences of patients are not fully considered, which makes it difficult to meet clinical rehabilitation needs.

[0003] The existing technology has technical problems such as insufficient wrist joint rehabilitation perception accuracy and lack of personalized training programs, making it difficult to meet clinical rehabilitation needs. Summary of the invention

[0004] The present application provides a wrist joint rehabilitation training method and system based on motion perception, which is used to solve the technical problems in the prior art of insufficient wrist joint rehabilitation perception accuracy, lack of personalized training programs, and difficulty in meeting clinical rehabilitation needs.

[0005] In view of the above problems, the present application provides a wrist joint rehabilitation training method and system based on motion perception.

[0006] The first aspect of the present application provides a wrist joint rehabilitation training method based on motion perception, the method comprising: According to the motion perception unit, the wrist joint perception parameters of the target user are collected in real time to obtain the user's wrist joint data; the wrist joint perception record library is queried to see whether there is historical wrist joint data of the target user; if the historical wrist joint data of the target user does not exist in the wrist joint perception record library, the user's wrist joint data is used as wrist joint feature perception data; based on the wrist joint feature perception data, according to the passive training mode constraints, the healthy side driving the affected side training mode constraints and the wrist joint rehabilitation training channel, a wrist joint rehabilitation training plan is obtained; based on the wrist joint rehabilitation training plan, rehabilitation training is performed on the target user to obtain training monitoring data; based on the training monitoring data, the wrist joint rehabilitation training plan is adjusted according to the passive training conditions and the healthy side driving the affected side training conditions to generate a wrist joint training optimization plan.

[0007] The second aspect of the present application provides a wrist joint rehabilitation training system based on motion perception, the system comprising: A user wrist joint data acquisition module, which collects the wrist joint perception parameters of the target user in real time according to the action perception unit to obtain the user's wrist joint data; a historical wrist joint data query module, which is used to query whether there is historical wrist joint data of the target user in the wrist joint perception record library; a wrist joint feature perception data acquisition module, which is used to use the user's wrist joint data as wrist joint feature perception data if there is no historical wrist joint data of the target user in the wrist joint perception record library; a wrist joint rehabilitation training program acquisition module, which is used to query whether there is historical wrist joint data of the target user in the wrist joint perception record library; a wrist joint feature perception data acquisition module, which is used to obtain the wrist joint data of the target user as the wrist joint feature perception data if there is no historical wrist joint data of the target user in the wrist joint perception record library; a wrist joint rehabilitation training program acquisition module, which is used to obtain the wrist joint rehabilitation training program The joint rehabilitation training program acquisition module obtains the wrist joint rehabilitation training program based on the wrist joint feature perception data, according to the passive training mode constraints, the healthy side driving the affected side training mode constraints and the wrist joint rehabilitation training channel; the training monitoring data acquisition module, the training monitoring data acquisition module performs rehabilitation training on the target user based on the wrist joint rehabilitation training program to obtain training monitoring data; the wrist joint training optimization program generation module, the wrist joint training optimization program generation module adjusts the wrist joint rehabilitation training program based on the training monitoring data according to the passive training conditions and the healthy side driving the affected side training conditions to generate a wrist joint training optimization program.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: According to the motion perception unit, the wrist joint perception parameters of the target user are collected in real time to obtain the user's wrist joint data; the wrist joint perception record library is queried to see whether there is historical wrist joint data; if the wrist joint perception record library does not exist, the user's wrist joint data is used as the wrist joint feature perception data; based on the wrist joint feature perception data, a wrist joint rehabilitation training program is obtained; based on the wrist joint rehabilitation training program, rehabilitation training is performed on the target user to obtain training monitoring data; based on the training monitoring data, the wrist joint rehabilitation training program is adjusted according to the passive training conditions and the healthy side driving the affected side training conditions to generate a wrist joint training optimization program. The technical effect of providing a personalized rehabilitation training program and improving the rehabilitation effect and safety is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic diagram of a flow chart of a wrist joint rehabilitation training method based on motion perception provided in an embodiment of the present application; Figure 2 A schematic structural diagram of a wrist joint rehabilitation training system based on motion perception provided in an embodiment of the present application.

[0011] Explanation of the reference numerals: user wrist joint data acquisition module 10, historical wrist joint data query module 20, wrist joint feature perception data acquisition module 30, wrist joint rehabilitation training program acquisition module 40, training monitoring data acquisition module 50, wrist joint training optimization program generation module 60. DETAILED DESCRIPTION

[0012] The present application provides a wrist joint rehabilitation training method and system based on motion perception, which is used to solve the technical problems in the prior art of insufficient wrist joint rehabilitation perception accuracy, lack of personalized training programs, and difficulty in meeting clinical rehabilitation needs.

[0013] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0014] Embodiment 1, as Figure 1 As shown, the present application provides a wrist joint rehabilitation training method based on motion perception, the method comprising: Step S100: According to the motion sensing unit, wrist joint sensing parameters of the target user are collected in real time to obtain wrist joint data of the user.

[0015] Specifically, the motion sensing unit uses a flexible film sensor to accurately measure the flexion and extension angles of the wrist joint. This sensor is tightly attached to the user's skin and can monitor the bending state of the wrist joint at any time and capture the range of motion of the joint in real time. At the same time, the motion sensing unit uses a gyroscope to measure the movement direction and rotation angle of the wrist joint to ensure comprehensive acquisition of the user's wrist joint movement data in multiple dimensions. In addition, the motion sensing unit also records the muscle activity of the wrist flexors and wrist extensors during exercise through the surface electromyography signal acquisition circuit. As a specific example, taking patient Zhang San as an example, the user's wrist joint data obtained is shown in Table 1 below: serial number 1 2 3 4 5 6 7 8 parameter Active flexion angle of left wrist Passive flexion angle of left wrist Active extension angle of left wrist Left wrist passive extension angle Right wrist active flexion angle Passive flexion angle of right wrist Right wrist active extension angle Passive extension angle of right wrist data 15 55 7 40 15 55 7 40 serial number 9 10 11 12 13 14 15 16 parameter The left wrist extension direction (horizontal, vertical, flip) represents the spatial coordinates of the left wrist The left wrist flexion direction (lateral, longitudinal, rotational) represents the spatial coordinates of the left wrist The right wrist extension direction (horizontal, vertical, flip) represents the spatial coordinates of the left wrist The flexion direction of the right wrist (lateral, longitudinal, rotational) represents the spatial coordinates of the left wrist Surface electromyography value of the active wrist flexor muscle of the left wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) Surface electromyography value of the passive wrist flexor muscle of the left wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) The surface electromyography value of the active wrist extensor muscle of the left wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) The surface electromyography value of the passive wrist extensor muscle of the left wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) data （0，7，0） （0，15，0） （0，7，0） （0，15，0） （15，35，55，75，95，105，125，145） （15，55，95，135，175，215，255，295） （15，35，55，75，95，105，125，145） （15，55，95，135，175，215，255，295） serial number 17 18 19 20 21 22 23 24 parameter Surface electromyography value of the active wrist flexor muscle of the right wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) Surface electromyography value of passive wrist flexor muscle of right wrist (after FFT filtering, collect the RMS value of electromyography, and divide the maximum angle into bins to obtain the corresponding electromyography value) Surface electromyography value of the active wrist extensor muscle of the right wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) The surface electromyography value of the passive wrist extensor muscle of the right wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) Left wrist passive flexion time (S) Left wrist passive extension time (S) Right wrist passive flexion time (S) Right wrist passive extension time (S) data （25，45，65，85，105，115，135，155） （25，65，95，135，185，225，265，305） （25，45，55，85，105，115，135，155） （25，65，95，135，185，225，265，305） 5 5 5 5 Through these multi-dimensional data collection, key parameters such as the user's wrist joint movement status, muscle strength, etc. can be obtained in real time, providing data support for the formulation of subsequent rehabilitation training programs. The high precision and comprehensiveness of real-time data collection can provide more scientific and effective support for the wrist joint rehabilitation process.

[0016] Step S200: Query the wrist joint perception record library to see whether there is historical wrist joint data of the target user.

[0017] Specifically, the wrist joint perception record library is automatically queried to determine whether the target user has previous rehabilitation training or evaluation data. The wrist joint perception record library stores the user's previous wrist joint movement data, electromyographic signals, training plans, and evaluation results. By querying the database, the data change trend of the user at different training stages is identified. The historical evaluation parameters are shown in Table 2 below: serial number 1 2 3 4 5 6 7 8 parameter Active flexion angle of left wrist Passive flexion angle of left wrist Active extension angle of left wrist Left wrist passive extension angle Right wrist active flexion angle Passive flexion angle of right wrist Right wrist active extension angle Passive extension angle of right wrist data 13 55 5 40 13 55 5 40 serial number 9 10 11 12 13 14 15 16 parameter The left wrist extension direction (horizontal, vertical, flip) represents the spatial coordinates of the left wrist The left wrist flexion direction (lateral, longitudinal, rotational) represents the spatial coordinates of the left wrist The right wrist extension direction (horizontal, vertical, flip) represents the spatial coordinates of the left wrist The flexion direction of the right wrist (lateral, longitudinal, rotational) represents the spatial coordinates of the left wrist Surface electromyography value of the active wrist flexor muscle of the left wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) Surface electromyography value of the passive wrist flexor muscle of the left wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) The surface electromyography value of the active wrist extensor muscle of the left wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) The surface electromyography value of the passive wrist extensor muscle of the left wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) data （0，7，0） （0，15，0） （0，7，0） （0，15，0） （10，30，50，70，90，100，120，140） （10，50，90，130，170，210，250，290） （10，30，50，70，90，100，120，140） （10，50，90，130，170，210，250，290） serial number 17 18 19 20 21 22 23 24 parameter Surface electromyography value of the active wrist flexor muscle of the right wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) Surface electromyography value of passive wrist flexor muscle of right wrist (after FFT filtering, collect the RMS value of electromyography, and divide the maximum angle into bins to obtain the corresponding electromyography value) Surface electromyography value of the active wrist extensor muscle of the right wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) The surface electromyography value of the passive wrist extensor muscle of the right wrist (after FFT filtering, the RMS value of the electromyography is collected, and the electromyography value corresponding to the maximum angle is divided into bins) Left wrist passive flexion time (S) Left wrist passive extension time (S) Right wrist passive flexion time (S) Right wrist passive extension time (S) data （20，40，60，80，100，110，130，150） （20，60，90，130，180，220，260，300） （20，40，50，80，100，110，130，150） （20，60，90，130，180，220，260，300） 4 4 4 4 If the user's historical data is found, it will be combined with the current real-time collected data to generate a more accurate and personalized rehabilitation training plan. If the user's historical data does not exist in the database, it will be considered as the user's first training record, and the currently collected data will be used as the basis for subsequent evaluation and training plan generation. This query process ensures that the user's historical information can be fully utilized to improve the scientificity and pertinence of the rehabilitation plan, especially when it is necessary to compare the user's rehabilitation progress or personalize the training plan. This step is particularly important.

[0018] Step S300: If the historical wrist joint data of the target user does not exist in the wrist joint perception record library, the wrist joint data of the user is used as the wrist joint feature perception data.

[0019] Specifically, when the wrist joint perception record library is queried and it is confirmed that the target user has no historical wrist joint data, the currently collected wrist joint data will be automatically used as the characteristic perception data of the wrist joint. These data include the user's wrist flexion angle, extension angle, movement direction, rotation angle, and wrist electromyographic signal. These real-time collected data will become the user's initial rehabilitation assessment basic data for the generation of subsequent rehabilitation training programs. Due to the lack of historical data for comparison, the initial intensity and goals of the training will be set based on these first collected data. This not only ensures the personalization and accuracy of rehabilitation training, but also accumulates basic data for future training records, which facilitates the optimization and adjustment of subsequent training programs. This step is especially important for users who are undergoing rehabilitation training for the first time. It can ensure that the system starts from scratch and provides users with training programs that suit their physical condition.

[0020] Step S400: Based on the wrist joint feature perception data, a wrist joint rehabilitation training program is obtained according to the passive training mode constraints, the healthy side driving the affected side training mode constraints, and the wrist joint rehabilitation training channel.

[0021] Specifically, based on the wrist joint feature perception data collected in real time and combined with the preset training mode constraints, a personalized rehabilitation training program is generated. First, based on the range of motion and electromyographic signals of the wrist joint, it is determined whether the user meets the requirements of the passive training mode. If the user's wrist joint range of motion is less than a certain set threshold, it indicates that his autonomous movement ability is weak. The passive training mode will be selected, relying on external equipment to assist the wrist joint to move in order to promote joint movement. The passive training mode can ensure that when the patient's active movement ability is insufficient, the joint still receives the necessary movement stimulation with the help of external equipment. If it is detected that the user's healthy wrist joint activity is normal, the healthy side will be recommended to drive the affected side training mode. This mode collects the motion data of the user's healthy hand (such as motion angle, direction and electromyographic signals), and uses these data to control the low-frequency electrical stimulation of the affected side, thereby achieving synchronous training of both arms. This method can not only improve the range of motion of the affected arm, but also help the affected arm recover faster through the guidance of the healthy hand. According to the current feature data of the wrist joint and the selection of the above training mode, the final rehabilitation training program will be generated through the wrist joint rehabilitation training channel. The plan will include detailed parameters such as specific training intensity, range of motion, training mode, etc. to ensure the personalization, scientificity and effectiveness of the training process.

[0022] Step S500: performing rehabilitation training on the target user based on the wrist joint rehabilitation training program to obtain training monitoring data.

[0023] Specifically, first, the rehabilitation training program includes parameters such as training mode, training site, flexion and extension angles of each joint, and corresponding time. For example, if the training mode is a passive training mode, the device will control the movement of the wrist joint of the target user according to the set parameters, such as controlling the left and right wrists to move according to the set flexion angles (such as 55 degrees for the left wrist flexion angle and 55 degrees for the right wrist flexion angle) and extension angles (such as 40 degrees for the left wrist extension angle and 40 degrees for the right wrist extension angle) and corresponding time (such as 5S for the left wrist flexion time, 5S for the left wrist extension time, 5S for the right wrist flexion time, and 5S for the right wrist extension time). During the training process, the training situation is monitored in real time through each module, and the motion perception unit continuously collects the motion data and surface electromyography data of the wrist joint, which are processed by corresponding filtering as part of the training monitoring data. For example, the wrist joint motion data is filtered by average filtering to obtain relevant perception data, and the wrist joint surface electromyography data is filtered by Fourier transform to obtain corresponding data. At the same time, it is judged whether the training monitoring data meets the abnormal conditions of wrist joint training. For example, if the collected data shows that the wrist joint movement angle exceeds the set reasonable range, or the surface electromyography data fluctuates abnormally, it means that an abnormal situation has occurred. If the wrist joint training abnormal conditions are met, a training stop command will be generated to ensure the safety of the user and the effectiveness of the rehabilitation training.

[0024] Step S600: Based on the training monitoring data, the wrist joint rehabilitation training program is adjusted according to the passive training conditions and the healthy side driving the affected side training conditions to generate a wrist joint training optimization program.

[0025] Specifically, the wrist joint rehabilitation training program is adjusted according to the training monitoring data to generate a more optimized program. When the rehabilitation training program is in the training mode of the healthy side driving the affected side: first, it is necessary to determine whether the training monitoring data meets the training conditions of the healthy side driving the affected side. If not, a passive mode switching instruction will be generated. Then, a weighted average will be performed based on the training monitoring data and the wrist joint feature perception data to obtain the wrist joint feature update data. Finally, based on the passive mode switching instruction, combined with the wrist joint feature update data and the relevant parameters of the wrist joint rehabilitation training channel, the original rehabilitation training program is adjusted to generate a wrist joint training optimization program. For example, if the training monitoring data shows that the movement angle of the healthy side driving the affected side is inaccurate, the angle parameters can be adjusted through the above steps to optimize the training program. When the rehabilitation training program is in the passive training mode: it is also necessary to first determine whether the training monitoring data meets the passive training conditions. If not, a driving mode switching instruction will be generated. Then, based on the driving mode switching instruction and the wrist joint feature update data (data obtained and updated during the previous training process) and the setting of the wrist joint rehabilitation training channel, the training program is adjusted to generate a wrist joint training optimization program. For example, if the force feedback data during passive training does not meet the requirements, the force parameters can be adjusted through this process to make the training plan more in line with actual needs.

[0026] In a possible implementation, step S100 further includes: Step S110: Obtain wrist joint motion data and wrist joint surface electromyography data of the target user according to the motion sensing unit.

[0027] Step S120: Perform average filtering according to the wrist joint motion data to obtain first wrist joint perception data.

[0028] Step S130: Perform Fourier transform filtering on the wrist joint surface electromyography data to obtain second wrist joint perception data.

[0029] Step S140: Add the first wrist joint perception data and the second wrist joint perception data to the user wrist joint data.

[0030] Specifically, in this step, the collection of wrist-related data of the target user is realized through the motion sensing unit. The motion sensing unit monitors the movement direction and angle information of the wrist joint of the target user in real time, as well as the surface electromyography values ​​of the wrist flexors and wrist extensors during wrist movement through high-precision flexible film sensors, gyroscopes, and electromyographic signal acquisition circuits. For wrist joint movement data, the flexible film sensor in the motion sensing unit calculates the bending angles of the left and right wrist joints based on the corresponding relationship between its resistance and curvature. At the same time, the gyroscope collects the movement directions of the wrist joints in different dimensions, including the angle a of the left and right translation with the wrist as the center when the palm is placed flat, the angle b of the upside-down flip, and the angle c of the elbow joint rotation. These angle data and possible movement speed, acceleration and other information together constitute the wrist joint movement data. For the wrist joint surface electromyographic data, the electromyographic signal acquisition circuit collects the original electromyographic signals at a sampling frequency of 1Khz. These signals reflect the electrical activity of the muscles during contraction and relaxation, providing a basis for the subsequent analysis of the force state of the wrist joint muscles.

[0031] Average filtering based on wrist joint motion data is a key data processing process, which aims to improve the accuracy and reliability of the data, so as to better reflect the actual motion state of the wrist joint and finally obtain the first perception data of the wrist joint. First of all, the wrist joint motion data contains multiple aspects of information, such as the left and right wrist joint bending angles obtained by the flexible film sensor, and the wrist joint motion direction data in different dimensions collected by the gyroscope (including the angle a of the left and right translation with the wrist as the center when the palm is flat, the angle b of the upside-down flip, and the angle c of the elbow joint rotation). These raw data have noise and fluctuations, which will affect the judgment of the real motion of the wrist joint.

[0032] The bending angle data calculated by the flexible film sensor is processed according to the median filter average algorithm. The specific steps are as follows: Data sorting: sort all the angle data in the data set according to the formula P(n, m)=n(n−1)(n−2)...(n−m+1). This step is to accurately find the median position in the future and prepare for further data processing. Determine the median range: determine the median according to the parity of the number of data sets. If the number of data sets is an odd number, the median is the number in the middle; if it is an even number, the median is the average of the two middle numbers. Calculate the average value: select a range (such as a window size) near the median position, and then calculate the average value of all values ​​within this range. This average value is used as the filtered angle data, which can effectively reduce the impact of noise and outliers in the original data. For other motion data, the moving average algorithm is used for processing. The moving average algorithm smoothes the data by calculating the average value of the data within a certain time window, thereby reducing data fluctuations. The wrist joint motion data is comprehensively processed through these different filtering processing methods, and finally the first perception data of the wrist joint is obtained. This data can more accurately reflect the movement state of the wrist joint and provide more reliable data support for the formulation of subsequent rehabilitation training programs.

[0033] Fourier transform filtering of wrist surface electromyography data is a key operation, which aims to remove interference components and more clearly reflect the characteristics of muscle activity, thereby obtaining the second sensory data of the wrist joint. First of all, the wrist surface electromyography data is the original signal collected by the electromyography signal acquisition circuit at a sampling frequency of 1Khz. These original signals contain the electrical activity information generated during muscle contraction and relaxation, but also contain some interference components, such as power frequency and frequency multiplication interference, which will affect the judgment of the actual activity state of the muscle. Then, use the FFT formula X(k)=∑n=0N-1X(n) ,in =exp(-j2πkn / N), j is an imaginary number, k represents the number of movements, N represents the size of the data to be converted, X(k): represents the kth value in the frequency domain after Fourier transform. n is the index variable when summing the original signal, traversing from 0 to N-1. N represents the size of the data to be converted, that is, the length of the original signal. X(n) is the nth value of the original signal in the time domain. It is called the rotation factor, which is a complex number. Its function is to convert the time domain signal to the frequency domain. k represents the number of movements or the frequency index. As k changes, the amplitude and phase information of different frequency components can be obtained.

[0034] =exp(-j2πkn / N), exp represents the natural exponential function. The exponential part -j2πkn / N here determines the specific value of the rotation factor. j is the imaginary unit, and the part 2πkn / N determines the angle of rotation. In the frequency domain transformation, different k values ​​correspond to different frequency components. Through this formula, the complex coefficients corresponding to each frequency component can be calculated.

[0035] The function of FFT is to convert the original time domain signal to the frequency domain. In the time domain, the signal is a function that changes with time, and it is difficult to directly see the frequency composition of the signal. After being converted to the frequency domain through Fourier transform, the components of different frequencies can be separated, and it is clear which frequency components are contained in the signal and their relative strengths. For example, for wrist surface electromyography data, the original signal contains electrical signals generated by muscle activity and various interference components. Through FFT, these interference components (such as power frequency and multiplier interference, etc.) are identified in the frequency domain, because these interferences usually have a specific frequency range. Once the frequency position of the interference component is determined, it can be removed through subsequent processing steps, thereby obtaining a signal that more clearly reflects the characteristics of muscle activity.

[0036] The collected original electromyographic signal is filtered from time domain to frequency. This step converts the original time domain signal to the frequency domain. In the frequency domain, components of different frequencies are separated, making it easier to identify and remove unnecessary frequency components, that is, power frequency and frequency multiplication interference. Then, after removing the interference components, median average filtering is performed. Median average filtering further optimizes the data and reduces the impact of residual noise. The specific operation is similar to the filtering of angle data in wrist joint motion data, that is, first determine the median range, and then select a range near the median to calculate the average value. Through the above-mentioned Fourier transform filtering and subsequent median average filtering operations, the surface electromyographic data of the wrist joint is comprehensively processed, and finally the second perception data of the wrist joint is obtained. This data can more accurately reflect the activity characteristics of the muscle and provide more reliable data support for the formulation of subsequent rehabilitation training programs.

[0037] Adding the first wrist joint perception data and the second wrist joint perception data to the user wrist joint data is an integration of the previous data processing results, so that the data set can more comprehensively reflect the actual situation of the target user's wrist joint. The user's wrist joint data is a comprehensive data set, which is constantly updated and enriched with the rehabilitation training process. The first wrist joint perception data is obtained after the wrist joint motion data is averaged and filtered, and contains information such as the processed left and right wrist joint bending angles and other dimensional motion direction data. These data can more accurately reflect the motion state of the wrist joint. The second wrist joint perception data is obtained after Fourier transform filtering and median average filtering of the wrist joint surface electromyography data, which can more clearly reflect the activity characteristics of the muscle. Adding these two perception data to the user's wrist joint data enables the user's wrist joint data to integrate information on both the motion state of the wrist joint and the muscle activity state. In this way, the user's wrist joint data provides a more sufficient and comprehensive basis for the subsequent formulation and adjustment of rehabilitation training programs based on these data, which helps to more accurately understand the actual situation of the target user's wrist joint, so as to formulate a rehabilitation training program that better meets individual needs.

[0038] In a possible implementation, step S200 further includes: Step S210: If the wrist joint perception record library contains historical wrist joint data of the target user, generate neighborhood historical wrist joint data.

[0039] Step S220: performing weighted averaging based on the neighborhood historical wrist joint data and the user wrist joint data to generate the wrist joint feature perception data.

[0040] Specifically, when it is confirmed that there is historical wrist data of the target user in the wrist joint perception record library, it is necessary to generate neighborhood historical wrist data. Here, it is particularly mentioned that the current collection data and the most recent training collection data are weighted averaged to achieve this. First of all, it is necessary to clarify that the current collection data is the relevant data of the target user's wrist joint obtained in real time by the motion perception unit. It reflects the actual state of the user's current wrist joint, including the comprehensive information obtained after the motion data and surface electromyography data are processed, and the most recent training collection data is the wrist joint data collected during the last training process selected from the historical records. This part of the data records the user's wrist joint situation in the previous training stage. In order to generate neighborhood historical wrist data, corresponding weights will be given according to the importance of these two groups of data. The current collection data will be given a relatively high weight because it can better reflect the current instant state, such as a weight of α (0<α<1); although the most recent training collection data is historical data, it also plays an important role in understanding the changing trend of the user's wrist joint, and is given a weight of β (0<β<1), and α+β=1. Then, through the weighted average calculation method, that is, neighborhood historical wrist joint data = α × current collected data + β × the most recent training collected data, the wrist joint conditions at the current time and the most recent training are comprehensively considered to generate neighborhood historical wrist joint data that can better reflect the changing trend of the user's wrist joint status and is more valuable for reference for subsequent analysis and plan formulation.

[0041] The neighborhood historical wrist data is obtained in step S210, which integrates the historical data information related to the current situation. This information is of great significance for understanding the historical state and change trend of the target user's wrist data. The user wrist data is obtained through a series of collection and processing steps, reflecting the actual situation of the user's current wrist, including information on movement status and muscle activity. When performing weighted averaging, the importance and relevance of the two sets of data need to be considered. For the neighborhood historical wrist data, the weight setting is based on multiple factors, such as the time interval between the historical data and the current data, the degree of similarity of the data, etc. If the time interval between the historical data and the current data is short and the data similarity is high, a higher weight will be given; otherwise, the weight is lower. For the user wrist data, since it directly reflects the current actual situation, it is usually also given a certain weight. Assuming that the weight of the neighborhood historical wrist data is α, the weight of the user wrist data is β, and α+β=1, then the calculation method of the wrist feature perception data is: wrist feature perception data=α×neighborhood historical wrist data+β×user wrist data. Through this weighted averaging method, historical experience and current actual conditions are organically combined, and the generated wrist joint feature perception data can more accurately reflect the characteristics of the target user's wrist joint, providing a more reliable data basis for the subsequent formulation of rehabilitation training programs.

[0042] In a possible implementation, step S400 further includes: Step S410: Determine whether the wrist joint feature perception data satisfies the passive training mode constraint condition.

[0043] Step S420: If the wrist joint feature perception data meets the passive training mode constraint condition, activate the wrist joint passive training model in the wrist joint rehabilitation training channel.

[0044] Step S430: generating the wrist joint rehabilitation training program based on the wrist joint feature perception data and the wrist joint passive training model.

[0045] Preferably, the passive training mode constraint is that the active flexion / extension range of motion of both wrist joints is less than 30 degrees. The significance of this judgment is to determine whether the patient is suitable for the passive training mode. When the range of motion is small, it indicates that the patient's own ability to actively move is limited and high-intensity active training cannot be performed. At this time, the passive training mode can help the patient to move the wrist joint through external equipment or auxiliary means to promote rehabilitation. If the wrist joint feature perception data meets this condition, the wrist joint passive training model in the wrist joint rehabilitation training channel is further activated to formulate a suitable passive training plan for the patient. If not met, it is necessary to continue to determine whether the healthy side drives the affected side training mode constraint condition to determine the rehabilitation training mode that best suits the patient's current state.

[0046] When the wrist joint feature perception data meets the passive training mode constraint conditions, the wrist joint passive training model in the wrist joint rehabilitation training channel needs to be activated. The following is a detailed description of the process: a special mechanism is set up internally to determine whether the wrist joint feature perception data meets the passive training mode constraint conditions. When the conditions are confirmed to be met, an activation instruction is sent to the wrist joint rehabilitation training channel. This instruction contains the relevant identification information of the target user and a specific identification indicating that it is activated based on the passive training mode. After receiving the instruction, the wrist joint rehabilitation training channel will quickly locate the wrist joint passive training model among the multiple training models stored in it. The channel adopts an index or classification storage method to quickly and accurately find the required model. Wrist joint passive training model structure, input layer: The input of the model is the wrist joint feature perception data. This data contains the wrist joint motion information (such as joint angle, angular velocity, angular acceleration, etc.) and muscle activity information (such as electromyographic signal strength, muscle force frequency, etc.) processed by the previous steps. These data are input to the input layer of the model in a specific format and order, providing the model with a comprehensive description of the current state of the wrist joint. Hidden layer: The model contains one or more hidden layers for complex feature extraction and nonlinear transformation of the input data. The number of neurons and the connection mode in the hidden layer are determined based on a large number of experiments and data analysis. Each neuron receives input from the previous layer and processes it through a specific activation function ReLU, and then passes the processed results to the next layer of neurons. Through the layer-by-layer processing of the hidden layer, the model can mine the potential features and rules in the input data, so as to better understand the state and needs of the wrist joint. Output layer: The function of the output layer is to generate the relevant parameters of the rehabilitation training program. These parameters include but are not limited to training intensity, training frequency, training time, training angle range (such as flexion angle, extension angle, etc.) and specific settings of the training mode (such as whether auxiliary force is required, the size and direction of the auxiliary force, etc.). The output layer converts the results of the hidden layer processing into rehabilitation training program parameters with practical application value. When activating the model, the parameters of the model need to be initialized. These parameters are crucial to the performance and output results of the model. The initialization process takes into account the individual characteristics of the user. For example, the initial value of the training intensity is determined based on factors such as the user's age, gender, degree of injury, and the current state of the wrist joint. For older users with more serious injuries, the training intensity will be initialized to a lower value to avoid further damage to the wrist joint due to excessive training. Similarly, the training frequency and training time will be appropriately adjusted according to the user's daily activity habits and rehabilitation needs. Through the above steps, the wrist joint passive training model in the wrist joint rehabilitation training channel is successfully activated and is ready to generate a suitable rehabilitation training plan based on the wrist joint feature perception data.

[0047] The wrist joint feature perception data contains many aspects of information, including the range of motion of the wrist joint, such as the wrist joint bending angle and the angle of motion direction in different dimensions collected by flexible film sensors and gyroscopes; it also includes the muscle force, which is reflected by the relevant data obtained after the electromyographic signal is collected and processed. Then, the passive training model of the wrist joint is a model that comprehensively considers multiple factors. It analyzes according to the input wrist joint feature perception data. For example, the model determines the starting angle and ending angle of the training based on the wrist joint activity data. If the current active flexion angle of the wrist joint is small, the starting angle of flexion in the training plan will be set based on the current angle, and the ending angle will also be determined according to the algorithm and rules in the model, taking into account factors such as rehabilitation goals and the physiological limits of the wrist joint. For muscle force, the model determines the training intensity based on it. If the muscle force is weak, the training intensity will be reduced accordingly to avoid damage to the wrist joint caused by excessive training. At the same time, parameters such as training frequency and training time will also be set in the model according to the wrist joint feature perception data and some preset rules. For example, according to the user's age, degree of injury and other factors, combined with the actual state of the current wrist joint, determine a suitable training frequency and training time. The wrist joint rehabilitation training program finally generated includes specific parameters such as training mode (passive training mode in this case), training site (such as left wrist or right wrist or bilateral wrist joints), flexion and extension angles of each joint and corresponding time. These parameters will guide the subsequent rehabilitation training process to help the target user restore the function of the wrist joint.

[0048] In a possible implementation, step S410 further includes: Step S411: If the wrist joint feature perception data does not satisfy the passive training mode constraint condition, determine whether the wrist joint feature perception data satisfies the healthy side driving the affected side training mode constraint condition.

[0049] Step S412: If the wrist joint feature perception data meets the healthy side driving the affected side training mode constraint condition, activate the healthy side driving the affected side training model in the wrist joint rehabilitation training channel.

[0050] Step S413: Based on the wrist joint feature perception data and according to the healthy side driving the affected side training model, the wrist joint rehabilitation training program is output.

[0051] Specifically, when the wrist joint feature perception data does not meet the passive training mode constraint, it is necessary to further determine whether the wrist joint feature perception data meets the healthy side driving the affected side training mode constraint. Preferably, the healthy side driving the affected side training mode constraint is that the range of motion of one hand is within the normal range of motion. The normal range of motion includes palmar flexion 50° to 60° and dorsiflexion 30° to 60°. If the wrist joint feature perception data shows that the palmar flexion angle of one hand is within the range of 50° to 60°, and the dorsiflexion angle is within the range of 30° to 60°, then the healthy side driving the affected side training mode constraint is met.

[0052] The wrist rehabilitation training channel contains multiple training models. When it is determined that the wrist feature perception data meets the constraints of the healthy side driving the affected side training mode, the healthy side driving the affected side training model is located in this channel. The healthy side driving the affected side training model uses multiple algorithms to construct the healthy side driving the affected side training model. First, the data is preprocessed. The median average filtering algorithm is used for the wrist joint activity data. The electromyographic signal data is first filtered by the FFT formula and then the median average filtering is performed. When determining the input parameters, if there is historical data, the weighted average algorithm is used. The model is based on the neural network algorithm. Its input layer receives the preprocessed wrist joint activity and muscle activity information. The number of neurons is the same as the input data dimension; there are multiple hidden layers, and the connection weight matrix uses a specific initialization method. The activation function uses ReLU to mine data features through the hidden layer; the output layer outputs the parameters related to the rehabilitation training program. During the training process, the mean square error is defined as the loss function, and the stochastic gradient descent or its variant optimization algorithm is used to calculate the gradient and update the weight according to the loss function to make the model better fit the data. The positioning process is based on the identification or indexing mechanism of the model to ensure that the corresponding model is found accurately. When activating the model, the parameters of the model need to be initialized, which are crucial to the performance and output results of the model. According to the principle of adjusting parameters according to the individual characteristics of the user, the initialization process will consider factors such as the user's age, gender, degree of injury, and the current state of the wrist joint. For example, for older users, the training intensity may be initialized to a lower value to avoid further damage to the wrist joint caused by excessive training. Similarly, parameters such as training frequency and training time will be appropriately adjusted according to the user's individual situation and rehabilitation needs. If the user's injury is more serious, the training frequency will be reduced and the training time will be appropriately extended, but it is also necessary to avoid long-term training that causes user fatigue. Through the above process, the healthy side driving the affected side training model in the wrist joint rehabilitation training channel is successfully activated, preparing for the subsequent generation of rehabilitation training plans based on wrist joint feature perception data.

[0053] The wrist joint feature perception data is input into the healthy side driving the affected side training model. These data contain information such as wrist joint range of motion and muscle force. The model processes the input data. For the range of motion data, the training angle range of the healthy and affected wrist joints is determined by combining the normal range of motion of the healthy side and the relationship between the affected and healthy sides. For example, the current range of motion is gradually expanded or adjusted to a suitable rehabilitation angle. For muscle force, if the force is weak, a lower training intensity is output to avoid overtraining. At the same time, the training frequency and training time are determined by comprehensively considering individual characteristics such as user age, gender, degree of injury, and the preset rules of the model. The final output includes a wrist joint rehabilitation training program that includes parameters such as the healthy side driving the affected side training mode, training site, flexion and extension angles of the healthy and affected wrist joints, and corresponding training time, training frequency, and training intensity.

[0054] In a possible implementation, step S600 further includes: Step S610: when the wrist joint rehabilitation training program is in the healthy side driving the affected side training mode, it is determined whether the training monitoring data meets the healthy side driving the affected side training condition.

[0055] Step S620: If the training monitoring data does not meet the condition that the healthy side drives the affected side training, a passive mode switching instruction is generated.

[0056] Step S630: performing weighted averaging based on the training monitoring data and the wrist joint feature perception data to obtain wrist joint feature update data.

[0057] Step S640: Based on the passive mode switching instruction, the wrist joint feature update data and the wrist joint rehabilitation training channel are used to obtain the wrist joint training optimization plan.

[0058] Specifically, when the wrist rehabilitation training program is in the healthy side driving the affected side training mode, the training monitoring data is comprehensively judged to determine whether the healthy side driving the affected side training conditions are met. The training monitoring data covers multiple key aspects of information, including the range of motion of the healthy and affected wrist joints (such as flexion angle, extension angle and range of motion), muscle force (such as muscle contraction force, contraction frequency and force duration), and coordination of the healthy and affected wrist joint movements (such as movement synchronization, relative movement deviation), etc.

[0059] Preferably, the condition for the healthy side to drive the affected side training is that during the healthy side driving the affected side training, the activity of the target user's healthy side for three consecutive activity cycles is greater than / equal to 30 degrees. Among them, one activity cycle is 10S. If during the healthy side driving the affected side training, the activity of the target user's healthy side for three consecutive activity cycles is less than 30 degrees, it can be judged that the patient's subjective initiative is insufficient and the healthy side driving the affected side training condition is not met.

[0060] If it is determined in step S610 that the training monitoring data does not meet the condition of healthy side driving the affected side training, a passive mode switching instruction will be generated. This instruction contains two important contents: a mode switching mark, and the instruction clearly includes an identifier for switching from the healthy side driving the affected side training mode to the passive training mode. This is a key guide for subsequent operations to ensure that the system can accurately identify the mode conversion tasks that need to be performed. Related data recording, at the same time, the instruction will also record some key data related to the current training status, including the current healthy and affected side wrist joint activity data, muscle force data, and movement coordination related data. These data will provide an important reference for generating wrist joint training optimization plans in subsequent steps, so that the system can better adjust the training parameters according to the current actual situation.

[0061] The wrist joint feature update data is obtained by weighted averaging the training monitoring data and the wrist joint feature perception data. The training monitoring data reflects the actual situation during the training process, while the wrist joint feature perception data is a comprehensive perception of the wrist joint state before and during training. The weighted averaging process is to effectively integrate the two types of data to obtain feature update data that can more accurately reflect the current wrist joint state. The weight distribution of the weighted average will be reasonably set according to the importance and reliability of the data. For example, the training monitoring data is given a higher weight because of its real-time and direct reflection of the current training situation, while the wrist joint feature perception data has a corresponding weight due to its comprehensive evaluation of the overall state of the wrist joint.

[0062] Based on the passive mode switching instruction, the passive wrist training model is located in the wrist rehabilitation training channel, and the wrist feature update data is used as input. The model operates based on the input data and mines data features through its internal structure (such as hidden layer, connection weight matrix and activation function). According to the muscle force, activity and other information in the wrist feature update data, the training intensity, frequency and time are adjusted respectively, and the training position and angle range are determined. The final output is a wrist training optimization plan with a clear training mode as the passive training mode, which contains various parameters adapted to the current wrist state.

[0063] In a possible implementation, step S600 further includes: Step S650: When the wrist joint rehabilitation training program is in the passive training mode, determining whether the training monitoring data meets the passive training conditions.

[0064] Step S660: If the training monitoring data does not meet the passive training condition, a driving mode switching instruction is generated.

[0065] Step S670: Based on the driving mode switching instruction, the wrist joint feature update data and the wrist joint rehabilitation training channel are used to generate the wrist joint training optimization plan.

[0066] Specifically, when the wrist joint rehabilitation training program is in the passive training mode, it is necessary to determine whether the training monitoring data meets the passive training conditions. Preferably, the passive training conditions include monitoring that the range of motion of the healthy hand flexion / extension is less than 80% of the maximum range of motion for three consecutive cycles. The passive flexion / extension angle of the wrist joint is used as the maximum range of motion.

[0067] If the flexion / extension range of motion of the healthy hand is monitored to be ≥ 80% of the maximum range of motion for three consecutive cycles, the training monitoring data does not meet the passive training conditions, and a driving mode switching instruction is generated. This instruction contains a clear mode switching mark, switching from the passive training mode (the healthy side drives the affected side mode), and also records the current wrist joint range of motion, muscle force and external force related data to provide a reference for subsequent optimization of the training plan, so as to adjust the training parameters to adapt to the new training mode and wrist joint state. According to the driving mode switching instruction, the healthy side driving the affected side training model is located in the wrist joint rehabilitation training channel.

[0068] The model adjusts the training intensity based on the muscle strength in the wrist joint feature update data. If the muscle strength is weak, the training intensity will be appropriately reduced to avoid damage to the wrist joint caused by excessive training. If the muscle strength is strong, the training intensity may be appropriately increased based on the actual situation, but the patient's tolerance and rehabilitation goals must also be considered.

[0069] At the same time, the model will also adjust the training frequency according to the wrist joint activity in the wrist joint feature update data. If the wrist joint activity is poor and the range of motion is small, the training frequency will be appropriately reduced to give the wrist joint more time to recover; if the wrist joint activity is good, the training frequency will be appropriately increased, but it is also necessary to avoid too frequent training to cause patient fatigue. For the adjustment of training time, the model will also comprehensively consider various factors in the wrist joint feature update data. If the wrist joint activity is poor, the training time needs to be appropriately extended, but it is also necessary to avoid too long training to cause patient fatigue; if the wrist joint activity is good, the training time can be appropriately shortened, but the training effect must be ensured. The output training optimization plan clearly states that the training mode is the healthy side driving the affected side training mode, which is determined by the driving mode switching instruction. The training site is determined according to the wrist joint feature update data, which is a single wrist joint (such as the left wrist or the right wrist) or a double wrist joint, depending on the actual situation and rehabilitation needs of the patient. According to the wrist joint activity information in the wrist joint feature update data, the training angle range is determined. For example, if the current flexion angle of the wrist joint is small, the training angle range may start from the current angle and gradually expand to a reasonable target angle; if the current extension angle of the wrist joint is small, there will also be corresponding completion. The output training optimization plan contains the adjusted training frequency and training time, which are determined based on the model's calculation results of the wrist joint feature update data. The training intensity is also determined based on the model calculation results. It is a reasonable value suitable for the current wrist joint state, which can avoid excessive injury while ensuring the rehabilitation effect. Through the above process, based on the driving mode switching instructions, wrist joint feature update data and wrist joint rehabilitation training channel (using the healthy side to drive the affected side training model), a wrist joint training optimization plan is generated, which can better adapt to the patient's current wrist joint state and rehabilitation needs.

[0070] In addition, if 80% of the maximum range of motion is not reached in three consecutive cycles, or if active force can be exerted to exceed 120% of the maximum range of motion of flexion / extension in three consecutive cycles, the updated wrist joint feature data is input into the passive training model of the wrist joint to obtain the wrist joint training optimization plan.

[0071] In a possible implementation, step S500 further includes: Step S510: Determine whether the training monitoring data meets the wrist joint training abnormality condition.

[0072] Step S520: If the training monitoring data meets the wrist joint training abnormality condition, a training stop instruction is generated.

[0073] Specifically, during the wrist joint rehabilitation training process, the training monitoring data is continuously received and analyzed to determine whether the wrist joint training abnormal conditions are met. Preferably, the wrist joint training abnormal conditions include the wrist joint reaching or exceeding the maximum range of motion for 3 seconds or more in the flexion / extension direction. If it is determined in step S510 that the training monitoring data meets the wrist joint training abnormal conditions, that is, it is judged that the patient has abnormal conditions such as spasms, and a training stop instruction is immediately generated. The function of this training stop instruction is to quickly interrupt the current training process to prevent more serious injuries to the patient. At the same time, the current training data is recorded, which includes but is not limited to the activity angle of the wrist joint, muscle force, training time and other information. These data are very important for subsequent analysis of the causes of abnormal conditions and for re-formulating rehabilitation training programs. After recording the current data, the rehabilitation training program is regenerated based on these data and the patient's initial condition. The new program will take into account this abnormal situation and adjust parameters such as training intensity, training mode, and training time to be more suitable for the patient's current physical condition to ensure that rehabilitation training can be carried out safely and effectively.

[0074] Embodiment 2 is based on the same inventive concept as the wrist joint rehabilitation training method based on motion perception in the above embodiment. Figure 2 As shown, the present application provides a wrist joint rehabilitation training system based on motion perception, and the system and method embodiments in the present application are based on the same inventive concept. The system includes: The user wrist joint data acquisition module 10 collects wrist joint perception parameters of the target user in real time according to the motion perception unit to obtain the user's wrist joint data.

[0075] The historical wrist joint data query module 20 is used to query whether there is historical wrist joint data of the target user in the wrist joint perception record library.

[0076] The wrist joint feature perception data acquisition module 30 is used to use the user's wrist joint data as wrist joint feature perception data if there is no historical wrist joint data of the target user in the wrist joint perception record library.

[0077] The wrist joint rehabilitation training program acquisition module 40 obtains the wrist joint rehabilitation training program based on the wrist joint feature perception data, according to the passive training mode constraints, the healthy side driving the affected side training mode constraints and the wrist joint rehabilitation training channel.

[0078] The training monitoring data acquisition module 50 performs rehabilitation training on the target user based on the wrist joint rehabilitation training program to obtain training monitoring data.

[0079] The wrist joint training optimization program generation module 60 adjusts the wrist joint rehabilitation training program based on the training monitoring data and the passive training conditions and the healthy side driving the affected side training conditions to generate a wrist joint training optimization program.

[0080] Furthermore, the user wrist joint data acquisition module 10 also includes: A wrist joint data acquisition unit, the wrist joint data acquisition unit is used to obtain the wrist joint motion data and wrist joint surface electromyography data of the target user according to the motion perception unit. A first perception data acquisition unit, the first perception data acquisition unit is used to perform average filtering according to the wrist joint motion data to obtain first wrist joint perception data. A second perception data acquisition unit, the second perception data acquisition unit is used to perform Fourier transform filtering according to the wrist joint surface electromyography data to obtain second wrist joint perception data. A wrist joint data adding unit, the wrist joint data adding unit is used to add the first wrist joint perception data and the second wrist joint perception data to the user's wrist joint data.

[0081] Furthermore, the historical wrist joint data query module 20 also includes: A neighborhood historical wrist data generation unit, the neighborhood historical wrist data generation unit is used to generate neighborhood historical wrist data if the wrist perception record library contains the target user's historical wrist data. A wrist feature perception data generation unit, the wrist feature perception data generation unit is used to generate the wrist feature perception data by weighted averaging the neighborhood historical wrist data and the user's wrist data.

[0082] Furthermore, the wrist joint rehabilitation training program acquisition module 40 also includes: A wrist joint feature perception data judgment unit, the wrist joint feature perception data judgment unit is used to judge whether the wrist joint feature perception data meets the passive training mode constraint condition. A wrist joint passive training model activation unit, the wrist joint passive training model activation unit is used to activate the wrist joint passive training model in the wrist joint rehabilitation training channel if the wrist joint feature perception data meets the passive training mode constraint condition. A wrist joint rehabilitation training program generation unit, the wrist joint rehabilitation training program generation unit generates the wrist joint rehabilitation training program based on the wrist joint feature perception data and the wrist joint passive training model.

[0083] Furthermore, the wrist joint feature perception data judgment unit also includes: A healthy side drives the affected side training mode constraint judgment unit, the healthy side drives the affected side training mode constraint judgment unit is used to judge whether the wrist joint feature perception data meets the healthy side drives the affected side training mode constraint if the wrist joint feature perception data does not meet the passive training mode constraint. A healthy side drives the affected side training model activation unit, the healthy side drives the affected side training model activation unit is used to activate the healthy side drives the affected side training model in the wrist joint rehabilitation training channel if the wrist joint feature perception data meets the healthy side drives the affected side training mode constraint. A joint rehabilitation training program output unit, the joint rehabilitation training program output unit outputs the wrist joint rehabilitation training program based on the wrist joint feature perception data and the healthy side drives the affected side training model.

[0084] Furthermore, the wrist joint training optimization program generation module 60 also includes: A training monitoring data judgment unit, the training monitoring data judgment unit is used to judge whether the training monitoring data meets the healthy side driving the affected side training condition when the wrist rehabilitation training program is in the healthy side driving the affected side training mode. A passive mode switching instruction generation unit, the passive mode switching instruction generation unit is used to generate a passive mode switching instruction if the training monitoring data does not meet the healthy side driving the affected side training condition. A wrist joint feature update data acquisition unit, the wrist joint feature update data acquisition unit is used to obtain wrist joint feature update data by weighted averaging the training monitoring data and the wrist joint feature perception data. A wrist joint training optimization program acquisition unit, the wrist joint training optimization program acquisition unit obtains the wrist joint training optimization program based on the passive mode switching instruction, the wrist joint feature update data and the wrist joint rehabilitation training channel.

[0085] Furthermore, the wrist joint training optimization program generation module 60 also includes: A passive training condition judgment unit, the passive training condition judgment unit is used to judge whether the training monitoring data meets the passive training condition when the wrist rehabilitation training program is in the passive training mode. A driving mode switching instruction generation unit, the driving mode switching instruction generation unit is used to generate a driving mode switching instruction if the training monitoring data does not meet the passive training condition. A wrist joint training optimization program generation unit, the wrist joint training optimization program generation unit generates the wrist joint training optimization program based on the driving mode switching instruction, the wrist joint feature update data and the wrist joint rehabilitation training channel.

[0086] Furthermore, the training monitoring data acquisition module 50 also includes: A wrist joint training abnormality condition judgment unit, the wrist joint training abnormality condition judgment unit is used to judge whether the training monitoring data meets the wrist joint training abnormality condition. A training stop instruction generation unit, the training stop instruction generation unit is used to generate a training stop instruction if the training monitoring data meets the wrist joint training abnormality condition.

[0087] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0088] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0089] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A wrist joint rehabilitation training method based on motion perception, characterized in that: The method comprises: According to the motion sensing unit, wrist joint sensing parameters of the target user are collected in real time to obtain wrist joint data of the user; Querying the wrist joint perception record library to see whether there is historical wrist joint data of the target user; If the historical wrist joint data of the target user does not exist in the wrist joint perception record library, the wrist joint data of the user is used as the wrist joint feature perception data; Based on the wrist joint feature perception data, a wrist joint rehabilitation training program is obtained according to the passive training mode constraints, the healthy side driving the affected side training mode constraints, and the wrist joint rehabilitation training channel; Performing rehabilitation training on the target user based on the wrist joint rehabilitation training program to obtain training monitoring data; Based on the training monitoring data, the wrist joint rehabilitation training program is adjusted according to passive training conditions and healthy side driving affected side training conditions to generate a wrist joint training optimization program.

2. A wrist joint rehabilitation training method based on motion perception as claimed in claim 1, characterized in that: According to the motion perception unit, the wrist joint perception parameters of the target user are collected in real time to obtain the user's wrist joint data, including: Obtaining wrist joint motion data and wrist joint surface electromyography data of the target user according to the motion sensing unit; Performing average filtering on the wrist joint motion data to obtain first wrist joint perception data; Perform Fourier transform filtering on the wrist joint surface electromyography data to obtain second wrist joint perception data; The first wrist joint perception data and the second wrist joint perception data are added to the user wrist joint data.

3. A wrist joint rehabilitation training method based on motion perception as claimed in claim 1, characterized in that: Query the wrist joint perception record library to see whether there is historical wrist joint data of the target user, including: If the wrist joint perception record library contains historical wrist joint data of the target user, generating neighborhood historical wrist joint data; The wrist joint feature perception data is generated by performing weighted averaging on the neighborhood historical wrist joint data and the user wrist joint data.

4. A wrist joint rehabilitation training method based on motion perception as claimed in claim 1, characterized in that: Based on the wrist joint feature perception data, according to the passive training mode constraints, the healthy side driving the affected side training mode constraints and the wrist joint rehabilitation training channel, a wrist joint rehabilitation training program is obtained, including: Determining whether the wrist joint feature perception data satisfies the passive training mode constraint condition; If the wrist joint feature perception data meets the passive training mode constraint condition, activating the wrist joint passive training model in the wrist joint rehabilitation training channel; Based on the wrist joint feature perception data and according to the wrist joint passive training model, the wrist joint rehabilitation training program is generated.

5. A wrist joint rehabilitation training method based on motion perception as claimed in claim 4, characterized in that: Determining whether the wrist joint feature perception data satisfies the passive training mode constraint condition includes: If the wrist joint feature sensing data does not satisfy the passive training mode constraint condition, determining whether the wrist joint feature sensing data satisfies the healthy side driving the affected side training mode constraint condition; If the wrist joint feature perception data satisfies the healthy side driving the affected side training mode constraint condition, activating the healthy side driving the affected side training model in the wrist joint rehabilitation training channel; Based on the wrist joint feature perception data and according to the healthy side driving the affected side training model, the wrist joint rehabilitation training program is output.

6. A wrist joint rehabilitation training method based on motion perception as claimed in claim 1, characterized in that: Based on the training monitoring data, the wrist joint rehabilitation training program is adjusted according to the passive training conditions and the healthy side driving the affected side training conditions to generate a wrist joint training optimization program, including: When the wrist joint rehabilitation training program is in the healthy side driving the affected side training mode, determining whether the training monitoring data meets the healthy side driving the affected side training condition; If the training monitoring data does not meet the condition that the healthy side drives the affected side to train, a passive mode switching instruction is generated; Performing weighted averaging on the training monitoring data and the wrist joint feature perception data to obtain wrist joint feature update data; Based on the passive mode switching instruction, the wrist joint training optimization plan is obtained according to the wrist joint feature update data and the wrist joint rehabilitation training channel.

7. A wrist joint rehabilitation training method based on motion perception as claimed in claim 6, characterized in that: Based on the training monitoring data, the wrist joint rehabilitation training program is adjusted according to the passive training conditions and the healthy side driving the affected side training conditions to generate a wrist joint training optimization program, including: When the wrist joint rehabilitation training program is in a passive training mode, determining whether the training monitoring data meets the passive training conditions; If the training monitoring data does not meet the passive training condition, generating a driving mode switching instruction; Based on the driving mode switching instruction, the wrist joint training optimization plan is generated according to the wrist joint feature update data and the wrist joint rehabilitation training channel.

8. The wrist joint rehabilitation training method based on motion perception as claimed in claim 1, characterized in that: Performing rehabilitation training on the target user based on the wrist joint rehabilitation training program to obtain training monitoring data, including: Determining whether the training monitoring data meets the wrist joint training abnormality condition; If the training monitoring data meets the wrist joint training abnormality condition, a training stop instruction is generated.

9. A wrist joint rehabilitation training system based on motion perception, characterized in that: The system is used to implement a wrist joint rehabilitation training method based on motion perception according to any one of claims 1 to 8, and the system comprises: A user wrist joint data acquisition module, which collects wrist joint perception parameters of the target user in real time according to the action perception unit to obtain the user's wrist joint data; A historical wrist joint data query module, which is used to query whether there is historical wrist joint data of the target user in the wrist joint perception record library; A wrist joint feature perception data acquisition module, wherein if the wrist joint perception record library does not contain the historical wrist joint data of the target user, the wrist joint data of the user is used as the wrist joint feature perception data; A wrist joint rehabilitation training program acquisition module, which obtains a wrist joint rehabilitation training program based on the wrist joint feature perception data, according to the passive training mode constraint conditions, the healthy side driving the affected side training mode constraint conditions and the wrist joint rehabilitation training channel; A training monitoring data acquisition module, wherein the training monitoring data acquisition module performs rehabilitation training on the target user based on the wrist joint rehabilitation training program to obtain training monitoring data; A wrist joint training optimization program generation module is provided. The wrist joint training optimization program generation module adjusts the wrist joint rehabilitation training program based on the training monitoring data according to passive training conditions and healthy side driving affected side training conditions to generate a wrist joint training optimization program.

Images