An active stroke rehabilitation auxiliary training system

Abstract

This invention belongs to the field of rehabilitation training technology, specifically an active stroke rehabilitation assistive training system. It generates a personalized training baseline based on patient clinical assessment data, guides patients to collaboratively perform training movements using rehabilitation equipment, and analyzes training execution efficiency scores, patient active movement characteristics, and equipment-assisted performance characteristics in real time. Based on the real-time interactive status of these two characteristics and the synchronous training execution efficiency score, it dynamically adjusts the assistance intensity of the rehabilitation equipment and the difficulty of the training tasks during the training period. Finally, it updates the patient's rehabilitation ability status based on the multi-dimensional temporal characteristics of the entire training session and outputs baseline correction instructions for the next training cycle based on the updated rehabilitation ability status. This not only breaks through the limitations of previous methods that solely focused on patient movements and clearly distinguishes the boundary between patient's autonomous effort and equipment assistance during rehabilitation, but also effectively improves the dynamic adaptability and accuracy of stroke rehabilitation training.

Description

Technical Field

[0001] This invention belongs to the field of rehabilitation training technology, specifically an active stroke rehabilitation auxiliary training system. Background Technology

[0002] Post-stroke motor dysfunction severely impacts patients' quality of life. Clinical studies have shown that active rehabilitation training can significantly improve functional recovery through neural remodeling mechanisms. However, in traditional rehabilitation training, the degree of patient's voluntary effort is difficult to objectively quantify and assess, leading to insufficient adaptation of training programs to the patient's actual rehabilitation status. Therefore, integrating active assessment into the training process is crucial for accurately guiding rehabilitation and improving training efficiency.

[0003] Several stroke rehabilitation training programs that integrate active assessment have been proposed in the prior art. For example, a method for estimating patient active participation in robot-assisted rehabilitation training, which is published in Chinese Patent Publication No. CN111161835A, calculates the baseline work done by the robot in its individual movement state and compares it with the total work done by the human-machine system when the patient participates in the training. This method can estimate the patient's real-time work done to assess active participation, avoids complex dynamic modeling, and is more suitable for clinical application.

[0004] Another Chinese patent publication, CN120340759A, describes a rehabilitation training method and device based on AAN challenge level adjustment to prevent slackening. It determines the challenge level based on the active force of the human and the driving force of the preceding human-machine hybrid system, constrains the range of auxiliary force through a virtual force field, and generates non-model robust auxiliary force by combining time delay estimation. It provides reasonable machine assistance to prevent training slackening, encourages active participation, and improves the effectiveness of rehabilitation training.

[0005] Although existing technologies attempt to address the problem of initiative assessment, they still have the following limitations: 1. Existing technologies already have various methods for assessing the initiative of patients' rehabilitation training, such as fuzzy algorithms, work volume estimation, and trajectory deviation detection. However, they all focus on the unilateral motor output characteristics of patients and do not remove the interference of rehabilitation equipment on the data. As a result, the assessment results cannot accurately distinguish the boundary between the patient's voluntary efforts and equipment assistance, and the accuracy of the evaluation is significantly limited.

[0006] 2. Existing technologies often refer to the surface effects of training execution and dynamically adjust the training, neglecting the dual interactive factors of patient initiative and equipment assistance involved in the training execution effect. In addition, although some existing solutions focus on adjusting the equipment assistance force based on the patient's active effort, such as increasing or decreasing virtual assistance force in the above-mentioned solutions, they do not coordinate the optimization of training task difficulty parameters. When the patient's muscle strength improves but motor control does not improve, reducing the assistance force while maintaining high precision requirements may increase training frustration and lead to insufficient training fit. Summary of the Invention

[0007] To overcome the shortcomings of the prior art, embodiments of the present invention provide an active stroke rehabilitation auxiliary training system, which can effectively solve the problems involved in the prior art.

[0008] The objective of this invention can be achieved through the following technical solution: an active stroke rehabilitation auxiliary training system, comprising: a baseline generation module, a training execution module, a dynamic adjustment module, and an iterative optimization module.

[0009] The baseline generation module is connected to the training execution module, the training execution module is connected to the dynamic adjustment module, and the dynamic adjustment module is connected to the iterative optimization module.

[0010] The baseline generation module generates a personalized training baseline based on the patient's clinical assessment data. The training baseline includes the training objectives and training plan for the current training cycle.

[0011] The training execution module guides patients to collaboratively perform training movements using rehabilitation equipment, referencing the training baseline, and analyzes the training execution effectiveness score, patient's active movement characteristics, and equipment-assisted performance characteristics in real time.

[0012] The dynamic adjustment module, based on the real-time interaction between active movement characteristics and auxiliary performance characteristics, and in conjunction with the training execution efficiency score, dynamically adjusts the auxiliary force of the rehabilitation equipment and the difficulty of the training task during the training period, with the adjustment direction contributing to the continuous improvement of the patient's autonomous movement.

[0013] The iterative optimization module updates the patient's rehabilitation ability status based on the multidimensional temporal characteristics of the entire training process, and outputs the baseline correction instruction for the next training cycle based on the updated rehabilitation ability status.

[0014] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention analyzes the patient’s active movement characteristics and equipment-assisted performance characteristics simultaneously for the patient’s training performance, realizes the comprehensive capture and linkage analysis of key information in the rehabilitation process, breaks the limitation of the previous single focus on the patient’s movements, and thus clearly distinguishes the boundaries between the patient’s autonomous efforts and equipment assistance in the rehabilitation process, providing a precise basis for dynamic adjustment.

[0015] (2) This invention utilizes the dual-feature contribution ratio to clearly reflect the dynamic balance between the patient's active movement and the equipment assistance, and the efficacy score to quantify the quality of the training effect. Based on the combination strategy of dual-feature contribution ratio and efficacy score, it accurately captures the changes in the patient's ability and actual needs at different training stages, thereby adjusting the assistance intensity and task difficulty in a targeted manner, and improving the dynamic adaptability and scientific nature of the rehabilitation training process.

[0016] (3) This invention updates the patient’s rehabilitation ability status through full-process time-series features and directly outputs the baseline correction instruction for the next cycle, realizing clinical-level adaptive training without manual intervention and solving the problem of disconnect between evaluation and execution in existing technologies. Attached Figure Description

[0017] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the module connection of the present invention.

[0019] Figure 2 This is a schematic diagram illustrating the analytical logic of the patient's active movement features in the training execution module of this invention.

[0020] Figure 3 This is a schematic diagram illustrating the analytical logic of device-assisted performance features in the training execution module of this invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Reference Figure 1 As shown, the present invention provides an active stroke rehabilitation auxiliary training system, including: a baseline generation module, a training execution module, a dynamic adjustment module, and an iterative optimization module.

[0023] The baseline generation module is connected to the training execution module, the training execution module is connected to the dynamic adjustment module, and the dynamic adjustment module is connected to the iterative optimization module.

[0024] The baseline generation module generates a personalized training baseline based on the patient's clinical assessment data. The training baseline includes the training objectives and training plan for the current training cycle.

[0025] In a preferred embodiment of the present invention, the training baseline generation process includes: extracting stroke pathological parameters from clinical assessment data and determining the severity level of the patient's condition by comparing them with preset stroke symptom assessment standards.

[0026] It should be noted that the above-mentioned pre-set stroke symptom assessment criteria are based on evidence-based medicine consensus. The data comes from the integrated analysis of neurological deficit scales, medical imaging parameters, and physiological and biochemical indicators. The criteria include two core contents: first, the numerical range division criteria for each stroke pathological parameter, which clarifies the quantitative boundaries of different indicators in pathological assessment; and second, the correspondence between each stroke symptom severity level and the numerical range of pathological parameter indicators, that is, the numerical range that each indicator must meet for a specific severity level.

[0027] The severity level of the above patients' conditions needs to be determined by searching the actual numerical range of each index of their stroke pathology parameters and taking the highest severity level reached as the final determination result.

[0028] The motor function parameters of the clinical assessment data are compared with the preset clinical norm values ​​of healthy people of the same age group to identify the patient's functional deficits.

[0029] It should be noted that the above-mentioned preset clinical norm values ​​are statistical reference ranges for healthy individuals of the same age in standardized functional tests, which can be obtained from relevant medical guidelines for stroke rehabilitation, including joint range of motion, duration of movement completion, and peak muscle strength parameters.

[0030] The process of identifying patient functional deficits includes the following steps: matching the patient's age with a norm dataset corresponding to the age group, wherein the norm dataset contains the mean and standard deviation of various motor function indicators of healthy individuals.

[0031] For each functional indicator in the motor function parameters, a standardized deviation value is calculated. Specifically, the difference between the measured value of the indicator and the mean value of the norm for that function is calculated, and the difference is further compared with the standard deviation of the norm.

[0032] If the standardized deviation of a certain functional indicator is greater than the preset value, then that functional indicator is identified as a functional defect.

[0033] Based on the severity level of the condition and the current recovery stage of the training cycle, the training objectives for the functional deficit items are determined using stage adaptation rules.

[0034] It should be noted that the above-mentioned disease recovery stages include the acute phase, the recovery phase, and the sequelae phase. For example, the acute phase can be defined as within 2 weeks after the onset of the disease, the recovery phase as from the 2nd to the 12th week after the onset of the disease, and the sequelae phase as more than 12 weeks after the onset of the disease.

[0035] The phase adaptation rules are as follows: When the patient is in the acute phase: if the severity level is severe, the goal is to inhibit abnormal movement patterns; if the severity level is moderate or mild, the goal is to control symptom progression and inhibit abnormal movement patterns.

[0036] When a patient is in the recovery phase: if the severity level is severe, the goal is to establish basic joint range of motion; if the severity level is moderate, the goal is to increase the range of voluntary movement; and if the severity level is mild, the goal is to restore functional coordination.

[0037] When patients are in the sequelae stage: if the severity level is severe, the goal is to achieve compensation for basic living functions; if the severity level is moderate, the goal is to establish compensation for fine motor functions; and if the severity level is mild, the goal is to optimize motor efficiency and adaptive compensation.

[0038] The training scheme bound to the training objective is invoked from the preset clinical rule base. The training scheme includes the following elements: I. Rehabilitation equipment type and its initial parameter configuration.

[0039] II. The standard path trajectory of training movements and the allowable deviation distance limit of the trajectory.

[0040] III. Critical path node locations and standard pose tolerance limits for training movements.

[0041] IV. Requirements for the complete training action execution cycle and the transition window time limit for adjacent critical path nodes.

[0042] The training execution module guides patients to perform training actions collaboratively through rehabilitation equipment, referring to the training baseline, and analyzes the training execution efficiency score, patient's active movement characteristics, and equipment-assisted performance characteristics in real time.

[0043] In a preferred embodiment of the present invention, the training performance evaluation process includes: real-time acquisition of limb kinematic data of the patient performing training actions.

[0044] A full score for the performance of the current training action is assigned when the following conditions are met simultaneously: a) The real-time movement trajectory of the patient's limb end is spatially aligned with the standardized path trajectory, and the deviation of the action trajectory is quantified by dynamic time warping distance, wherein the deviation meets the permissible trajectory deviation distance limit.

[0045] It should be noted that the above dynamic time warping process aims to perform non-equal time sequence matching on two spatially aligned trajectories, and calculate the minimum warped path distance using the standard dynamic time warping distance formula, which is then used as the execution deviation.

[0046] b) The deviation between the patient's actual pose and the standard pose of the critical path node when the patient's limb reaches the critical path node is within the tolerance limit.

[0047] It should be noted that the above-mentioned deviation quantification process of the actual pose relative to the standard pose of the node is as follows: the standard pose of the node is imported into professional drawing software as the bottom layer, and the actual pose is imported as the second layer. The spatial position relationship between each unit point of the actual pose and the corresponding point of the standard pose is located by comparing the layers. The nearest distance of each unit point relative to the standard pose is retrieved, and the average value is taken to obtain the pose distance deviation. At the same time, the angle difference between the actual pose and the standard pose at each corresponding joint is extracted, and the maximum value is taken as the pose angle deviation.

[0048] c) The total execution time of the complete training movement meets the execution cycle requirements.

[0049] d) The arrival time interval between adjacent critical nodes meets the transition window time limit.

[0050] If any conditions are not met, the performance score will be reduced accordingly based on the type of unmet condition to generate a training execution performance score.

[0051] It should be noted that when the above-mentioned deduction is made to the full performance score based on the type of condition not met, the deduction score is determined by the influence weight of each condition on the training execution performance, specifically the product of the influence weight and the full performance score.

[0052] Regarding the weighting of the impact of various conditions on training execution efficiency, it can be pre-set based on industry experience or derived through a limited number of experimental data. For example, first collect the artificial pre-scoring values ​​of training execution efficiency under different types of missing conditions, quantify the correlation coefficients of the impact of different types of missing conditions on training execution efficiency, then determine the contribution of each condition through regression analysis or logistic regression analysis, and finally normalize the contribution and convert it into an impact weight, where the sum of the impact weights is required to be 1.

[0053] Reference Figure 2 As shown in a preferred embodiment of the present invention, the patient's active movement feature analysis process includes: capturing the patient's original electromyographic signals of muscle groups through a surface electrode array, and generating an electromyographic amplitude envelope through rectification and filtering.

[0054] The movement trajectory and speed of the patient's limb extremities are collected simultaneously, and the kinetic energy of the limb extremities is calculated accordingly.

[0055] It should be noted that the above calculation process of limb end motion kinetic energy is as follows: the weight of the limb is obtained by a gravity sensor attached to the patient's limb, and combined with the limb end motion speed, and then substituted into the existing standard kinetic energy formula to obtain the limb end motion kinetic energy.

[0056] The integral results of the electromyographic amplitude envelope are analyzed in the time domain with the kinetic energy of the limb end, and the autonomous drive contribution index, which characterizes the degree of coupling between muscle activation and limb movement, is output.

[0057] It should be noted that the above time-domain correlation analysis can be exemplified as the Pearson correlation coefficient analysis between the integral result of the electromyographic amplitude envelope and the rate of change of kinetic energy at the limb end.

[0058] Based on the detection and identification of the coordinated activation period of multi-muscle group phase synchronization, the theoretical physiological trajectory is constructed by fusing joint angle data within the coordinated activation period. The autonomous drive contribution index is calibrated according to the difference between the patient's limb end movement trajectory and the theoretical physiological trajectory, and the patient's active movement characteristics are output.

[0059] It should be noted that the above phase synchronization detection involves performing a Hilbert transform on the electromyographic signal of each muscle group to obtain the instantaneous phase, quantifying the phase difference variance between muscle groups, and when the variance is lower than the preset coordination target variance threshold, it indicates a multi-muscle group coordinated activation state.

[0060] It should also be noted that the above-mentioned calibration process for the autonomous drive contribution index based on the difference between the patient's limb end movement trajectory and the theoretical physiological trajectory is as follows: both the patient's limb end movement trajectory and the theoretical physiological trajectory are converted into coordinate point sets. The distance from each trajectory point in the limb end movement trajectory coordinate point set to the nearest trajectory point in the theoretical physiological trajectory coordinate point set is obtained, and the maximum value of the distance is recorded as the one-way reference distance 1. Similarly, the distance from each trajectory point in the theoretical physiological trajectory coordinate point set to the nearest trajectory point in the limb end movement trajectory coordinate point set is obtained, and the maximum value of the distance is recorded as the one-way reference distance 2. The maximum value of the two one-way reference distances is taken as the Hausdorff distance between the two trajectories.

[0061] If the Hausdorf distance is less than the first preset distance, the autonomous driving contribution index is numerically increased and calibrated according to the reduction ratio of the Hausdorf distance relative to the first preset distance.

[0062] If the Hausdorff distance is greater than the second preset distance, the autonomous driving contribution index is numerically reduced and calibrated according to the excess ratio of the Hausdorff distance relative to the second preset distance.

[0063] If the Hausdorff distance is within the closed interval defined by the first and second preset distances, then the autonomous driving contribution index will only be marked and not calibrated.

[0064] The first preset distance is the critical distance that represents the ideal state of trajectory matching. When the calculated Hausdorff distance is less than the first preset distance, it indicates that the patient's limb end movement trajectory matches the theoretical physiological trajectory to a high degree, which plays a positive role in promoting the autonomous driving process, that is, it makes a positive contribution.

[0065] The second preset distance is the critical distance that characterizes the trajectory deviation to the upper limit of the acceptable range. When the Hausdorff distance is greater than the second preset distance, it indicates that the deviation between the patient's limb end movement trajectory and the theoretical physiological trajectory is too large, which has an adverse effect on the autonomous driving process, that is, it produces a reverse contribution.

[0066] The first preset distance is less than the second preset distance, and both were obtained through extensive experimental testing, data statistics and analysis before system development. They can objectively reflect the correlation between the degree of trajectory matching and the contribution of autonomous driving.

[0067] Reference Figure 3 As shown in a preferred embodiment of the present invention, the device auxiliary performance feature analysis process includes: real-time acquisition of the joint space auxiliary torque vector output by the rehabilitation device through an embedded torque sensor, and calculation of instantaneous auxiliary power based on the motor current and rotor angular velocity of the rehabilitation device.

[0068] Based on the permissible trajectory deviation distance limit of the standardized training movement path in the training program, the target joint torque is derived, and the ratio of the vector norm of the real-time collected auxiliary torque to the target joint torque is used as the proportion of the auxiliary force of the rehabilitation equipment.

[0069] It should be noted that the above derivation process of the target joint torque is as follows: the allowable trajectory deviation distance limit is decomposed into a three-dimensional Cartesian space to generate a position deviation vector. The task space stiffness matrix is ​​constructed by combining the auxiliary torque of the joint space of the rehabilitation equipment with the position deviation mapping relationship in the task space. The Jacobian matrix is ​​introduced to convert the torque requirement obtained in the task space based on the position deviation vector and the stiffness matrix into the joint space to generate the target joint torque.

[0070] Establish the correlation between auxiliary torque energy consumption and patient trajectory execution deviation, and determine the power density required for the device to maintain the trajectory.

[0071] It should be noted that the power density mentioned above is specifically the result of a ratio calculation with the cumulative energy consumption of the auxiliary torque as the numerator and the integral of the patient trajectory execution deviation over time as the denominator.

[0072] The auxiliary force ratio is subjected to correlation processing based on the power density, and the auxiliary performance characteristics of the output device are obtained.

[0073] It should be noted that the above-mentioned correlation processing can be exemplified as linear correlation correction. For example, the power density converted from the preset power density influence coefficient can be superimposed on the proportion of assist force, and piecewise nonlinear correlation correction can also be exemplified. Its logical principle is the same as the logical principle of calibrating the autonomous drive contribution index by the difference between the patient's limb end movement trajectory and the theoretical physiological trajectory, that is, judging the positive and negative influence relationship of the power density value on the proportion of assist force. It will not be repeated here.

[0074] This invention provides a method for simultaneously analyzing the patient's active movement characteristics and equipment-assisted performance characteristics during training. This enables comprehensive capture and coordinated analysis of key information during the rehabilitation process, breaking away from the previous limitation of focusing solely on the patient's movements. It clearly distinguishes the boundaries between the patient's voluntary efforts and equipment assistance during the rehabilitation process, providing a precise basis for dynamic adjustment.

[0075] The dynamic adjustment module, based on the real-time interaction between active movement characteristics and auxiliary performance characteristics, and in conjunction with the training execution efficiency score, dynamically adjusts the auxiliary force of the rehabilitation equipment and the difficulty of the training task during the training period, with the adjustment direction contributing to the continuous improvement of the patient's autonomous movement.

[0076] In a preferred embodiment of the present invention, the dynamic adjustment process of the assistive force and task difficulty of the rehabilitation device includes: during the training period, periodically monitoring the training execution efficiency score, the patient's active movement characteristics and the device's assistive performance characteristics according to preset adjustment response time points, and recording the time period defined by adjacent adjustment response time points as the monitoring period.

[0077] Based on the expected synergy ratio range defined by the clinical rehabilitation strategy, the relative contribution relationship between the patient's active movement characteristics and the device-assisted performance characteristics during the monitoring period was determined.

[0078] Based on the combination of the relative contribution relationship and the training execution efficiency achievement status, an adjustment strategy is implemented. The adjustment strategy includes: based on the training execution efficiency achievement status, if the patient's active movement characteristics are dominant, the task difficulty is increased; if the device-assisted performance characteristics are dominant, the assistance level is reduced.

[0079] Based on the state of unsatisfactory training performance, if the patient's active movement characteristics are dominant, the task difficulty should be reduced; if the device-assisted performance characteristics are dominant, the assistance level should be increased.

[0080] In a preferred embodiment of the present invention, the relative contribution relationship between the patient's active movement characteristics and the device-assisted performance characteristics is determined by the following method: the patient's active movement characteristics and the device-assisted performance characteristics during the monitoring period are normalized and the feature contribution ratio is calculated.

[0081] It should be noted that the above feature contribution ratio refers to the ratio of the patient's active movement features to the device-assisted performance features after normalization.

[0082] If the feature contribution ratio is within the expected synergy ratio range, it is determined that the patient's active participation and the device assistance have reached the expected synergy state, with no dominant subject.

[0083] If the feature contribution ratio exceeds the expected synergistic ratio range, and the patient's active movement features are significantly higher than the device-assisted performance features, then the patient's active movement features are determined to be dominant.

[0084] If the feature contribution ratio exceeds the expected synergistic ratio range, and the device-assisted performance feature is significantly higher than the patient's active movement feature, then the device-assisted performance feature is determined to be dominant.

[0085] It should be noted that the trigger condition mentioned above, which is significantly higher than the preset reasonable value, is that the absolute difference between the two features is greater than the preset reasonable value.

[0086] In a preferred embodiment of the present invention, the adjustment strategy execution process further includes: the auxiliary force adjustment is achieved by updating the initial parameter configuration of the rehabilitation device, changing the joint space auxiliary torque parameter and the motor operating parameter.

[0087] Task difficulty adjustment is achieved by updating at least one of the following elements in the training scheme, including the permissible trajectory deviation distance limit, the standard pose tolerance limit for critical path nodes, the execution cycle requirement for the complete training action, and the transition window time limit for adjacent critical path nodes.

[0088] The adjustment amplitude is positively correlated with the degree to which the feature contribution ratio deviates from the expected synergistic ratio range.

[0089] This invention utilizes a dual-feature contribution ratio to clearly reflect the dynamic balance between the patient's active movement and equipment assistance, and an efficacy score to quantify the quality of training effects. Based on a combination strategy of dual-feature contribution ratio and efficacy score, it accurately captures the changes in the patient's abilities and actual needs at different training stages, thereby adjusting the assistance intensity and task difficulty in a targeted manner, and improving the dynamic adaptability and scientific nature of the rehabilitation training process.

[0090] The iterative optimization module updates the patient's rehabilitation ability status based on the multidimensional temporal characteristics of the entire training process, and outputs the baseline correction instruction for the next training cycle based on the updated rehabilitation ability status.

[0091] In a preferred embodiment of the present invention, the patient rehabilitation ability status update process includes: extracting the training execution efficiency score sequence of a single training session, introducing a time decay factor for weighted fusion to output a functional progress score benchmark.

[0092] The fluctuation variance of the patient's active movement characteristics and the evolution trend of the device-assisted performance characteristics were obtained throughout a single training session.

[0093] It should be noted that the evolution trend of the above-mentioned equipment-assisted performance characteristics depends on the positive or negative slope of the linear fitting time series curve. If the slope is positive, it indicates that the rehabilitation equipment-assisted performance characteristics show a positive trend; if the slope is negative, it indicates that the rehabilitation equipment-assisted performance characteristics show a negative trend.

[0094] When the functional progress scoring benchmark reaches the preset clinical advancement standard, the fluctuation of the patient's active movement characteristics meets the preset stability requirements, and the rehabilitation equipment's assisted performance characteristics show a negative trend, the patient's rehabilitation ability status is updated to the preparatory advancement status.

[0095] When the functional progress score is lower than the preset clinical maintenance standard, the fluctuation of the patient's active movement characteristics exceeds the preset sensitivity requirement, and the rehabilitation equipment assistive performance characteristics show a positive trend, the patient's rehabilitation ability status will be updated to a capability decline warning status.

[0096] When the functional progress scoring benchmark is between the preset clinical advancement standard and the maintenance standard, and the fluctuation of the patient's active motor characteristics is between the preset stability requirement and the sensitivity requirement, the patient's rehabilitation ability status is updated to the maintenance training status.

[0097] In a preferred embodiment of the present invention, the baseline correction instruction output process for the next training cycle includes: when the patient's rehabilitation ability status is in the advanced preparation state, selecting the motor function that meets the advanced requirements of the current rehabilitation stage from the functional deficiency items as the new training target, and feeding it back to the next training cycle.

[0098] When the state of capability degradation warning is reached, the current training objective is broken down into low-order sub-action objectives and fed back to the next training cycle.

[0099] To maintain the training status, keep the current training baseline parameters unchanged and apply them to the next training cycle.

[0100] This invention updates the patient's rehabilitation status through full-process temporal features and directly outputs the baseline correction instruction for the next cycle, realizing clinical-level adaptive training without manual intervention and solving the problem of disconnect between evaluation and execution in existing technologies.

[0101] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. An active stroke rehabilitation auxiliary training system, characterized in that, include: The baseline generation module generates a personalized training baseline based on the patient's clinical assessment data. The training baseline includes the training objectives and training plan for the current training cycle. The training execution module guides patients to collaboratively perform training movements using rehabilitation equipment, referencing the training baseline, and analyzes the training execution effectiveness score, patient's active movement characteristics, and equipment-assisted performance characteristics in real time. The dynamic adjustment module dynamically adjusts the auxiliary force of the rehabilitation equipment and the difficulty of the training task during the training period based on the real-time interaction between active movement characteristics and auxiliary performance characteristics, combined with the training execution efficiency score. The adjustment direction contributes to continuously improving the patient's autonomous movement. The iterative optimization module updates the patient's rehabilitation ability status based on the multidimensional temporal characteristics of the entire training process, and outputs the baseline correction instruction for the next training cycle based on the updated rehabilitation ability status. The dynamic adjustment process of the rehabilitation equipment's assistive force and task difficulty includes: During the training period, the training performance score, patient active movement characteristics and device-assisted performance characteristics were periodically monitored according to the preset adjustment response time points, and the time period defined by adjacent adjustment response time points was recorded as the monitoring period. Based on the expected synergy ratio range defined by the clinical rehabilitation strategy, the relative contribution relationship between the patient's active movement characteristics and the device-assisted performance characteristics during the monitoring period was determined. Based on the combination of the relative contribution relationship and the training execution performance achievement status, an adjustment strategy is implemented, the adjustment strategy including: Based on the training performance target status, if the patient's active movement characteristics are dominant, the task difficulty should be increased; if the device-assisted performance characteristics are dominant, the assistance level should be reduced. Based on the state of unsatisfactory training performance, if the patient's active movement characteristics are dominant, the task difficulty should be reduced; if the device-assisted performance characteristics are dominant, the assistance level should be increased.

2. The active stroke rehabilitation assistive training system according to claim 1, characterized in that: The training baseline generation process includes: Extract stroke pathology parameters from clinical assessment data and determine the severity level of the patient's condition by comparing them with preset stroke symptom assessment standards; The motor function parameters of the clinical assessment data are compared with the preset clinical norm values ​​of the same age group of healthy people to identify the patient's functional deficits. Based on the severity level of the condition and the recovery stage of the current training cycle, the training objectives for the functional deficit items are determined through stage adaptation rules. The training scheme bound to the training objective is invoked from the preset clinical rule base. The training scheme includes the following elements: I. Rehabilitation equipment type and its initial parameter configuration; II. The standard path trajectory of training movements and the permissible deviation distance limits; III. Critical path node locations and standard pose tolerance limits for training movements; IV. Requirements for the complete training action execution cycle and the transition window time limit for adjacent critical path nodes.

3. The active stroke rehabilitation assistive training system according to claim 2, characterized in that: The training performance scoring and analysis process includes: Real-time acquisition of limb kinematic data of patients performing training movements; A full score is assigned to the performance of the current training action when all of the following conditions are met: a) Spatially align the real-time movement trajectory of the patient's limb extremities with the standardized path trajectory, and quantify the movement trajectory execution deviation through dynamic time warping distance, wherein the deviation meets the permissible trajectory deviation distance limit; b) The actual pose of the patient's limb end when it reaches the critical path node conforms to the tolerance limit of the node's standard pose; c) The total execution time of the complete training movement meets the execution cycle requirements; d) The arrival time interval between adjacent critical nodes meets the transition window time limit; If any conditions are not met, the performance score will be reduced accordingly based on the type of unmet condition to generate a training execution performance score.

4. The active stroke rehabilitation assistive training system according to claim 1, characterized in that: The process of analyzing the patient's active movement features includes: Raw electromyographic signals of the patient's muscle groups are captured by a surface electrode array and then rectified and filtered to generate an electromyographic amplitude envelope. Simultaneously collect the movement trajectory and speed of the patient's limb extremities, and calculate the kinetic energy of the limb extremities accordingly; The integral results of the electromyographic amplitude envelope are analyzed in the time domain with the kinetic energy of the limb end, and the autonomous drive contribution index, which characterizes the degree of coupling between muscle activation and limb movement, is output. Based on the detection and identification of the coordinated activation period of multi-muscle group phase synchronization, the theoretical physiological trajectory is constructed by fusing joint angle data within the coordinated activation period. The autonomous drive contribution index is calibrated according to the difference between the patient's limb end movement trajectory and the theoretical physiological trajectory, and the patient's active movement characteristics are output.

5. The active stroke rehabilitation assistive training system according to claim 2, characterized in that: The device-assisted performance feature analysis process includes: The joint space auxiliary torque vector output by the rehabilitation equipment is collected in real time by an embedded torque sensor, and the instantaneous auxiliary power is calculated based on the motor current and rotor angular velocity of the rehabilitation equipment. Based on the permissible trajectory deviation distance limit of the standardized training movement path in the training program, the target joint torque is derived, and the ratio of the vector norm of the real-time collected auxiliary torque to the target joint torque is used as the proportion of the auxiliary force of the rehabilitation equipment. Establish the correlation between auxiliary torque energy consumption and patient trajectory execution deviation, and determine the power density required for the device to maintain the trajectory; The auxiliary force ratio is subjected to correlation processing based on the power density, and the auxiliary performance characteristics of the output device are obtained.

6. The active stroke rehabilitation assistive training system according to claim 1, characterized in that: The relative contribution relationship between the patient's active motor characteristics and the device-assisted performance characteristics was determined in the following way: The patient's active movement characteristics and equipment-assisted performance characteristics during the monitoring period were normalized and the feature contribution ratio was calculated. If the feature contribution ratio is within the expected synergy ratio range, it is determined that the patient's active participation and the device assistance have reached the expected synergy state, and there is no dominant subject. If the feature contribution ratio exceeds the expected synergistic ratio range, and the patient's active movement features are significantly higher than the device-assisted performance features, then the patient's active movement features are determined to be dominant. If the feature contribution ratio exceeds the expected synergistic ratio range, and the device-assisted performance feature is significantly higher than the patient's active movement feature, then the device-assisted performance feature is determined to be dominant.

7. The active stroke rehabilitation assistive training system according to claim 6, characterized in that: The adjustment strategy execution process also includes: The adjustment of the assistive force is achieved by updating the initial parameter configuration of the rehabilitation equipment and changing the joint space assistive torque parameters and motor operating parameters; Task difficulty adjustment is achieved by updating at least one of the following elements in the training scheme, including the permissible trajectory deviation distance limit, the standard pose tolerance limit of critical path nodes, the execution cycle requirement of the complete training action, and the transition window time limit of adjacent critical path nodes; The adjustment amplitude is positively correlated with the degree to which the feature contribution ratio deviates from the expected synergistic ratio range.

8. The active stroke rehabilitation assistive training system according to claim 2, characterized in that: The process of updating the patient's rehabilitation capacity status includes: Extract the training execution efficiency score sequence of the entire training process in a single training session, introduce a time decay factor for weighted fusion to output a functional progress score benchmark; The fluctuation variance of the patient's active movement characteristics and the evolution trend of the device-assisted performance characteristics were obtained throughout a single training session. When the functional progress scoring benchmark reaches the preset clinical advancement standard, the fluctuation of the patient's active movement characteristics meets the preset stability requirements, and the rehabilitation equipment assistive performance characteristics show a negative trend, the patient's rehabilitation ability status is updated to the preparatory advancement status. When the functional progress score is lower than the preset clinical maintenance standard, the fluctuation of the patient's active movement characteristics exceeds the preset sensitivity requirement, and the rehabilitation equipment assistive performance characteristics show a positive trend, the patient's rehabilitation ability status will be updated to a capability decline warning status. When the functional progress scoring benchmark is between the preset clinical advancement standard and the maintenance standard, and the fluctuation of the patient's active motor characteristics is between the preset stability requirement and the sensitivity requirement, the patient's rehabilitation ability status is updated to the maintenance training status.

9. The active stroke rehabilitation auxiliary training system according to claim 8, characterized in that: The baseline correction instruction output process for the next training cycle includes: When the patient's rehabilitation ability is in the advanced preparation stage, select the motor function that meets the advanced requirements of the current rehabilitation stage from the functional deficit items as the new training target and feed it back to the next training cycle. When the ability degradation warning state is in effect, the current training objective is broken down into low-order sub-action objectives and fed back to the next training cycle; To maintain the training status, keep the current training baseline parameters unchanged and apply them to the next training cycle.

Images