A rehabilitation glove evaluation method and system based on multi-source data analysis

Abstract

This invention discloses a method and system for evaluating rehabilitation gloves based on multi-source data analysis. The method includes: collecting multi-source monitoring data through the sensor network of the rehabilitation glove; performing motion feature analysis at multiple time points; extracting multi-dimensional motion parameters from the multi-source monitoring data and mining associated parameter itemsets using the Apriori association algorithm; setting time windows based on N time points and filtering motion-related time periods through linear fitting; detecting abnormal time points according to the rehabilitation plan and establishing detection rules; filtering abnormal parameter items from the data itemset based on abnormal time points, and extracting abnormal time periods from the motion-related time periods to trace the associated motion parameters that match the abnormal time periods. This invention achieves accurate evaluation of the rehabilitation exercise process, dynamically traces abnormal associated motion characteristics, and provides data support for optimizing rehabilitation plans.

Description

Technical Field

[0001] This invention relates to the field of multi-source data analysis, and more specifically, to a method and system for evaluating rehabilitation gloves based on multi-source data analysis. Background Technology

[0002] In the field of hand rehabilitation, accurate assessment of the rehabilitation effect of patients' hand movements is crucial for developing and adjusting rehabilitation programs. Traditional rehabilitation assessment methods mainly rely on the subjective observation and experience of rehabilitation physicians, resulting in highly subjective and inaccurate assessment results. With the development of sensor and robotics technologies, wearable devices such as rehabilitation gloves are increasingly being used in rehabilitation assessments, collecting hand movement data to assist in the evaluation. However, existing assessment methods based on rehabilitation gloves mostly only perform simple analysis on single or a few movement parameters, failing to fully utilize multi-source monitoring data, leading to low data utilization and incomplete and inaccurate assessment results.

[0003] In addition, in rehabilitation glove robots, there are certain correlation characteristics among multi-source rehabilitation parameters. For users at abnormal movement time points, it is necessary to analyze and extract the correlation movement characteristics of the hand and conduct data source analysis to formulate a better rehabilitation plan and improve the accuracy of rehabilitation evaluation. However, in the existing technology, it is difficult to trace the correlation movement parameters under abnormal time periods. It often relies on manual review of data to assess the rehabilitation status, which makes it difficult to achieve an intelligent and information-based rehabilitation process.

[0004] To address the aforementioned issues, this invention proposes a rehabilitation glove evaluation method and system based on multi-source data analysis. The aim is to achieve precise evaluation of the rehabilitation exercise process and detection of abnormal correlations through in-depth analysis and mining of multi-source monitoring data, providing more objective, accurate, and comprehensive data support for rehabilitation treatment. Summary of the Invention

[0005] This invention overcomes the shortcomings of the prior art and proposes a rehabilitation glove evaluation method and system based on multi-source data analysis.

[0006] The first aspect of this invention provides a method for evaluating rehabilitation gloves based on multi-source data analysis, comprising:

[0007] S1: During the rehabilitation exercise cycle, multi-source monitoring data is collected through the sensor network of the rehabilitation glove, and multiple time nodes are set to analyze the motion characteristics of the multi-source monitoring data;

[0008] S2: Extract multi-dimensional motion parameters based on multi-source monitoring data, combine motion parameters that change in adjacent time nodes to form parameter items, collect all parameter items to form a data item set, and use the Apriori association algorithm to mine association rules from the data item set and filter out the associated parameter item set;

[0009] S3: Set time windows based on N time nodes, set variable parameters through the rehabilitation plan, and perform linear fitting calculations on multidimensional motion parameters for each time window in each time window. Filter out time windows with linear relationships and set motion-related time periods.

[0010] S4: Develop testing rules based on the rehabilitation plan, conduct conditional monitoring based on multi-source monitoring data, and screen out abnormal time points;

[0011] S5: Based on abnormal time nodes, filter out abnormal parameter items from the data item set, extract abnormal time periods from the motion-related time periods, and trace the motion parameters that match the abnormal time periods from the related item set based on the abnormal parameter items.

[0012] In this solution, S1 specifically includes:

[0013] Based on the rehabilitation plan, set a rehabilitation exercise cycle;

[0014] In the rehabilitation gloves, based on the user's rehabilitation monitoring needs, multiple sensors are set up, and the sensors are connected to the control terminal through the hardware layer to form a sensor network;

[0015] The various sensors include joint sensors, pressure sensors, and electromyography (EMG) sensors.

[0016] In this scheme, S1 further includes:

[0017] By using multi-source monitoring data, the type and frequency of exercise can be assessed.

[0018] Based on the type of exercise state and the number of exercises, preset time intervals are set to divide the rehabilitation exercise cycle into multiple time points, resulting in multiple time nodes.

[0019] In this solution, S2 specifically refers to:

[0020] Multi-dimensional motion parameters are extracted from multi-source monitoring data to obtain motion parameter data;

[0021] Based on each time node, the motion parameter data is serialized and analyzed. Taking a time node as a benchmark, multiple motion parameters under that node are analyzed. If a motion parameter changes with a motion parameter at an adjacent time node, the two motion parameters are combined to obtain the parameter item.

[0022] Analyze all time points and integrate the parameter items to obtain a data item set;

[0023] Based on the Apriori association algorithm, minimum support and minimum confidence are set;

[0024] For each parameter item, its support is calculated based on the entire data itemset, and frequent itemsets are selected using the minimum support.

[0025] Calculate association rules and their confidence levels based on frequent itemsets, filter out association rules that exceed the minimum confidence level, and form association parameter itemsets.

[0026] In this solution, S3 specifically refers to:

[0027] Set the time window based on the time step consisting of N time nodes;

[0028] During the rehabilitation exercise cycle, multiple data segments are analyzed by moving time windows, and the current time window is marked in each movement;

[0029] Within the current time window, extract various motion parameters at all time points, and select the two motion parameters with the largest rate of change as independent and dependent variables, respectively.

[0030] Within the current time window, linear fitting calculations are performed on the independent and dependent variables to obtain the motion fitting coefficients;

[0031] Determine whether the motion fitting coefficient is greater than the minimum fitting coefficient. If so, mark the current time window as the motion-related period.

[0032] For multiple movement operations, motion characteristic correlation analysis was performed, and all motion-related time periods were marked.

[0033] In this solution, S4 specifically refers to:

[0034] The testing rules are formulated based on the rehabilitation program, and the testing rules include one or more motor characteristic conditions.

[0035] For each time point, conditions are monitored and judged from multi-source monitoring data, and time points that do not meet the rules are marked as abnormal time points.

[0036] In this solution, S5 specifically refers to:

[0037] Based on abnormal time nodes, the corresponding parameter items are filtered from the data item set and marked as abnormal parameter items;

[0038] Analyze the motion-related time periods to which the abnormal time nodes belong and mark them as abnormal time periods;

[0039] If the abnormal time node does not exist in any motion-related time period, then the abnormal time period is constructed based on the adjacent time nodes;

[0040] Based on the set of associated parameter items, determine the parameter items associated with the abnormal parameter items and mark them as the first parameter items. Filter out the parameters that match the abnormal time period from the first parameter items to obtain the second parameter items.

[0041] Analyze all abnormal time points and use all the selected second parameter items as source abnormal parameters.

[0042] In this solution, S5 further includes:

[0043] Based on the abnormal parameters, the source of the abnormal parameters, and the abnormal time period, the rehabilitation gloves are assessed for auxiliary equipment abnormalities, and the analysis results are sent to the data terminal.

[0044] A second aspect of the present invention also provides a rehabilitation glove assessment system based on multi-source data analysis. The system includes a memory, a processor, and a communication interface. The communication interface is used to establish data connections between various modules within the system. The memory includes a rehabilitation glove assessment program based on multi-source data analysis. When executed by the processor, the rehabilitation glove assessment program based on multi-source data analysis performs the following steps:

[0045] S1: During the rehabilitation exercise cycle, multi-source monitoring data is collected through the sensor network of the rehabilitation glove, and multiple time nodes are set to analyze the motion characteristics of the multi-source monitoring data;

[0046] S2: Extract multi-dimensional motion parameters based on multi-source monitoring data, combine motion parameters that change in adjacent time nodes to form parameter items, collect all parameter items to form a data item set, and use the Apriori association algorithm to mine association rules from the data item set and filter out the associated parameter item set;

[0047] S3: Set time windows based on N time nodes, set variable parameters through the rehabilitation plan, and perform linear fitting calculations on multidimensional motion parameters for each time window in each time window. Filter out time windows with linear relationships and set motion-related time periods.

[0048] S4: Develop testing rules based on the rehabilitation plan, conduct conditional monitoring based on multi-source monitoring data, and screen out abnormal time points;

[0049] S5: Based on abnormal time nodes, filter out abnormal parameter items from the data item set, extract abnormal time periods from the motion-related time periods, and trace the motion parameters that match the abnormal time periods from the related item set based on the abnormal parameter items.

[0050] A third aspect of the present invention also provides a computer-readable storage medium comprising a rehabilitation glove assessment program based on multi-source data analysis, wherein when the rehabilitation glove assessment program based on multi-source data analysis is executed by a processor, it implements the steps of the rehabilitation glove assessment method based on multi-source data analysis as described in any of the preceding claims.

[0051] This invention discloses a method and system for evaluating rehabilitation gloves based on multi-source data analysis. The method includes: collecting multi-source monitoring data through the sensor network of the rehabilitation glove; performing motion feature analysis at multiple time points; extracting multi-dimensional motion parameters from the multi-source monitoring data and mining associated parameter itemsets using the Apriori association algorithm; setting time windows based on N time points and filtering motion-related time periods through linear fitting; detecting abnormal time points according to the rehabilitation plan and establishing detection rules; filtering abnormal parameter items from the data itemset based on abnormal time points, and extracting abnormal time periods from the motion-related time periods to trace the associated motion parameters that match the abnormal time periods. This invention achieves accurate evaluation of the rehabilitation exercise process, dynamically traces abnormal associated motion characteristics, and provides data support for optimizing rehabilitation plans. Attached Figure Description

[0052] Figure 1 A flowchart of the modules for evaluating rehabilitation gloves based on multi-source data analysis according to the present invention is shown.

[0053] Figure 2 The flowchart of the motion-related time period filtering process of the present invention is shown;

[0054] Figure 3 A block diagram of a rehabilitation glove assessment system based on multi-source data analysis according to the present invention is shown. Detailed Implementation

[0055] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0056] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0057] Figure 1 A flowchart of the modules for evaluating rehabilitation gloves based on multi-source data analysis according to the present invention is shown.

[0058] The first aspect of this invention provides a method for evaluating rehabilitation gloves based on multi-source data analysis, comprising:

[0059] S1: During the rehabilitation exercise cycle, multi-source monitoring data is collected through the sensor network of the rehabilitation glove, and multiple time nodes are set to analyze the motion characteristics of the multi-source monitoring data;

[0060] S2: Extract multi-dimensional motion parameters based on multi-source monitoring data, combine motion parameters that change in adjacent time nodes to form parameter items, collect all parameter items to form a data item set, and use the Apriori association algorithm to mine association rules from the data item set and filter out the associated parameter item set;

[0061] S3: Set time windows based on N time nodes, set variable parameters through the rehabilitation plan, and perform linear fitting calculations on multidimensional motion parameters for each time window in each time window. Filter out time windows with linear relationships and set motion-related time periods.

[0062] S4: Develop testing rules based on the rehabilitation plan, conduct conditional monitoring based on multi-source monitoring data, and screen out abnormal time points;

[0063] S5: Based on abnormal time nodes, filter out abnormal parameter items from the data item set, extract abnormal time periods from the motion-related time periods, and trace the motion parameters that match the abnormal time periods from the related item set based on the abnormal parameter items.

[0064] It is worth mentioning that each of the above steps is implemented through different modules: S1 corresponds to the data acquisition module; S2 corresponds to the motion parameter analysis module; S3 corresponds to the motion characteristic correlation module; S4 corresponds to the anomaly detection module; and S5 corresponds to the source analysis module. Each module corresponds to a different data processing function. In addition to the above modules, the system of this invention also includes a data terminal, which establishes a data connection with the above modules in the form of a hardware layer or a software layer. For example... Figure 1 As shown.

[0065] According to an embodiment of the present invention, S1 specifically includes:

[0066] Based on the rehabilitation plan, set a rehabilitation exercise cycle;

[0067] In the rehabilitation gloves, based on the user's rehabilitation monitoring needs, multiple sensors are set up, and the sensors are connected to the control terminal through the hardware layer to form a sensor network;

[0068] The various sensors include joint sensors, pressure sensors, and electromyography (EMG) sensors.

[0069] Here, it can be understood that the rehabilitation program specifically refers to a hand rehabilitation program. The rehabilitation exercise process involves the user's hand movements while wearing rehabilitation gloves. A rehabilitation exercise cycle is a hand activity analysis cycle, consisting of multiple hand movement processes. It is used to analyze the hand movement characteristics in different movement processes or activity states, such as joint rotation angles, range of motion, and changes in electromyographic signals. Multiple sensors are typically placed at hand muscle locations, joint locations, and fingertips, with one or more sensors arranged depending on the type of sensor and different monitoring needs. Rehabilitation gloves generally contain an assistive motor. During motion parameter analysis, relevant parameters of the assistive motor can be collected to determine the status of assisted movement; these parameters can be collected through joint sensors.

[0070] According to an embodiment of the present invention, S1 further includes:

[0071] By using multi-source monitoring data, the type and frequency of exercise can be assessed.

[0072] Based on the type of exercise state and the number of exercises, preset time intervals are set to divide the rehabilitation exercise cycle into multiple time points, resulting in multiple time nodes.

[0073] In one possible scenario, if there are multiple statistically significant movement states during a rehabilitation exercise cycle, such as repeatedly grasping objects or repeatedly opening and closing the palms, multiple time points can be set based on the number of movement states to finely analyze the changes in movement parameters at multiple consecutive time points, and subsequently implement linear fitting analysis of continuous movement characteristics.

[0074] The multi-source monitoring data includes multi-dimensional data, such as sensor data including electromyography (EMG) signal values, EMG signal intensity, pressure values, rotation angle, rotation amplitude, and movement speed. It also includes auxiliary motor data, such as power, speed, and current. Examples of recorded data are as follows: angle change: 5°-10°, amplitude change: 2cm-5cm, speed change: 0.1m / s-0.3m / s, auxiliary motor power: 2W-5W.

[0075] According to an embodiment of the present invention, step S2 specifically includes:

[0076] Multi-dimensional motion parameters are extracted from multi-source monitoring data to obtain motion parameter data;

[0077] Based on each time node, the motion parameter data is serialized and analyzed. Taking a time node as a benchmark, multiple motion parameters under that node are analyzed. If a motion parameter changes with a motion parameter at an adjacent time node, the two motion parameters are combined to obtain the parameter item.

[0078] Analyze all time points and integrate the parameter items to obtain a data item set;

[0079] Based on the Apriori association algorithm, minimum support and minimum confidence are set;

[0080] For each parameter item, its support is calculated based on the entire data itemset, and frequent itemsets are selected using the minimum support.

[0081] Calculate association rules and their confidence levels based on frequent itemsets, filter out association rules that exceed the minimum confidence level, and form association parameter itemsets.

[0082] In a preferred embodiment, the minimum support can be set to 0.3 and the minimum confidence can be set to 0.5.

[0083] It should be noted that, using a given time point as a baseline, multiple motion parameters are analyzed at that point. If a motion parameter changes in both a motion parameter at an adjacent time point, the two motion parameters are combined to obtain a parameter item. A specific example is as follows:

[0084] Parameter (change in motion amplitude:) 3cm, pressure change: +20kPa), parameter item (motion speed change) (0.1 m / s, +20 kPa), the positive and negative values ​​of the changes can be set by the actual direction of movement, or only the absolute value of the change can be recorded. Parameter items can be the same type of motion parameter. In addition, based on the needs of auxiliary analysis, motor auxiliary parameters can also be used as parameter items for combined analysis. Since motor parameters can reflect the corresponding rehabilitation assistance situation and reflect rehabilitation needs from an auxiliary perspective, they can also be used for rehabilitation exercise characteristic research and help to explore related motion data items.

[0085] Each time point includes corresponding multidimensional parameters. The motion parameter data includes electromyography (EMG) signal values, EMG signal intensity, pressure values, rotation angle, rotation amplitude, and movement speed. In addition, auxiliary motor parameters such as power, speed, and current may also be included.

[0086] The set of associated parameter items includes multiple associated items, each of which includes two or more parameter items, representing that the corresponding two or more parameter items are associated. The associated status is stored in a rule-based manner. In addition, since the participating items have time node information, the association rules also filter out the time nodes where the motion characteristics are associated.

[0087] Figure 2 The flowchart of the motion-related time period filtering process of the present invention is shown;

[0088] According to an embodiment of the present invention, S3 specifically includes:

[0089] Set the time window based on the time step consisting of N time nodes;

[0090] During the rehabilitation exercise cycle, multiple data segments are analyzed by moving time windows, and the current time window is marked in each movement;

[0091] Within the current time window, extract various motion parameters at all time points, and select the two motion parameters with the largest rate of change as independent and dependent variables, respectively.

[0092] Within the current time window, linear fitting calculations are performed on the independent and dependent variables to obtain the motion fitting coefficients;

[0093] Determine whether the motion fitting coefficient is greater than the minimum fitting coefficient. If so, mark the current time window as the motion-related period.

[0094] For multiple movement operations, motion characteristic correlation analysis was performed, and all motion-related time periods were marked.

[0095] It should be noted that each movement corresponds to the time interval within the rehabilitation exercise cycle of the time window. For example, if the rehabilitation exercise cycle is 10 time units long and the time window is 2 time units, then 5 time window movements can be performed. The rate of change can be determined by the maximum-minimum difference or the linear coefficient within the time window. The two exercise parameters with the highest rate of change can be randomly assigned as independent and dependent variables, primarily for analyzing linear changes. For example, if the rate of change of electromyography (EMG) signal and rotation amplitude is the highest within a time window, then a linear fit can be performed based on EMG signal and rotation amplitude as independent and dependent variables to analyze the correlation of exercise characteristics and determine whether to mark the associated exercise periods. Through correlation analysis of exercise characteristics, periods with correlations in exercise characteristics can be effectively marked, and associated exercise parameters can be further filtered in conjunction with abnormal time nodes.

[0096] The process for filtering exercise-related time periods is as follows: Figure 2 As shown.

[0097] According to an embodiment of the present invention, step S4 specifically includes:

[0098] The testing rules are formulated based on the rehabilitation program, and the testing rules include one or more motor characteristic conditions.

[0099] For each time point, conditions are monitored and judged from multi-source monitoring data, and time points that do not meet the rules are marked as abnormal time points.

[0100] It should be noted that the movement characteristic conditions are personalized settings based on the user's rehabilitation, used to filter out periods of rapid abnormal movement, such as movement amplitude greater than distance D, movement speed continuously less than V, and whether electromyographic signals are detected.

[0101] According to an embodiment of the present invention, step S5 specifically includes:

[0102] Based on abnormal time nodes, the corresponding parameter items are filtered from the data item set and marked as abnormal parameter items;

[0103] Analyze the motion-related time periods to which the abnormal time nodes belong and mark them as abnormal time periods;

[0104] If the abnormal time node does not exist in any motion-related time period, then the abnormal time period is constructed based on the adjacent time nodes;

[0105] Based on the set of associated parameter items, determine the parameter items associated with the abnormal parameter items and mark them as the first parameter items. Filter out the parameters that match the abnormal time period from the first parameter items to obtain the second parameter items.

[0106] Analyze all abnormal time points and use all the selected second parameter items as source abnormal parameters.

[0107] It should be noted that the motion-related time period includes multiple consecutive time nodes. Adjacent time nodes are the two nodes before and after the abnormal time node.

[0108] According to an embodiment of the present invention, step S5 further includes:

[0109] Based on the abnormal parameters, the source of the abnormal parameters, and the abnormal time period, the rehabilitation gloves are assessed for auxiliary equipment abnormalities, and the analysis results are sent to the data terminal.

[0110] It should be noted that, based on abnormal parameter items, source abnormal parameters, and abnormal time periods, the rehabilitation gloves can be analyzed for abnormal motion parameters, assess abnormal motion characteristics present during abnormal time periods, determine abnormal motion characteristics of the assistive device (assistive motor), and revise the rehabilitation plan accordingly. The rehabilitation gloves include the assistive device.

[0111] It is worth mentioning that existing technologies have difficulty capturing the time points and abnormal linkage motion parameters of abnormal linkage in hand rehabilitation devices, resulting in poor ability to trace linkage data. However, hand movements often have multiple motion characteristics and correlations of multiple motion parameters, and linkage characteristics and correlations are key to evaluating hand rehabilitation movements, which can be used to further trace and analyze the rehabilitation status and hand rehabilitation trends.

[0112] This invention effectively solves the above problems. It collects motion feature data of rehabilitation gloves through a data acquisition module. Furthermore, it uses the Apriori association algorithm to perform association matching on motion parameters under different spatiotemporal conditions. The matching process analyzes data itemsets formed by multidimensional parameters to construct association parameter itemsets with associated states. Furthermore, it uses linear fitting and time windows to set variable parameters for multidimensional motion parameters. By introducing linear fitting, it evaluates whether there are corresponding motion association characteristics in each time window, dynamically captures rehabilitation motion association time periods, quickly detects abnormal time periods using rules, and combines the association parameter itemsets with motion association time periods to filter out motion parameter items with linkage association, thereby realizing the source analysis of abnormal motion characteristics.

[0113] This invention enables dynamic tracing of related motion characteristics and screening of potential abnormal correlation parameters, effectively improving the level of rehabilitation program formulation, thereby enhancing the automation and intelligent rehabilitation assessment level of rehabilitation glove equipment, completing automated rehabilitation analysis, avoiding reliance on human experience for the analysis and monitoring of continuous movements, and improving the applicability of the equipment in multiple scenarios.

[0114] According to an embodiment of the present invention, it further includes:

[0115] The data terminal is used to evaluate the movement parameters and assistive devices during abnormal periods, optimize the rehabilitation plan, and apply it to the second rehabilitation exercise cycle.

[0116] In the second rehabilitation exercise cycle, exercise parameters and assistive device parameters were acquired based on multiple time points, and the parameters were serialized to obtain exercise parameter sequences and assistive parameter sequences, respectively.

[0117] The correlation between the motion parameter sequence and the auxiliary parameter sequence was calculated based on the grey relational analysis method.

[0118] Based on the comparison of abnormal detection data between the first and second rehabilitation exercise cycles, the changes in the number of abnormal time nodes, the rate of change of abnormal time periods, and the rate of change of correlation were obtained.

[0119] The degree of optimization of the rehabilitation program can be assessed by the changes in the number of abnormal time points, the rate of change of abnormal time periods, and the changes in correlation.

[0120] It should be noted that the rehabilitation plan includes information such as the sensor monitoring scheme for rehabilitation gloves, the operation scheme for assistive devices, and detection rules. One or more motion parameters can be selected for sequential analysis from the motion parameter sequence; generally, core motion parameters are selected for evaluation.

[0121] For the optimization of hand rehabilitation equipment solutions, existing technologies often evaluate solutions based on a single dimension, lacking optimization analysis that combines movement and assisted operation at different times, making it difficult to assess the degree of optimization of the rehabilitation solution.

[0122] This invention can effectively solve the above problems. By comparing and analyzing two rehabilitation cycles, the changes in the number of abnormal time points, the rate of change of abnormal time periods, and the changes in correlation are used as evaluation indicators to quantify the rehabilitation changes in the two cycles, thereby achieving a reasonable and accurate evaluation of the degree of optimization of the rehabilitation plan.

[0123] The change in the number of abnormal time points is specifically the difference in the number of abnormal time points between two periods; the change rate of abnormal time periods is specifically the difference in the length of abnormal time periods between two periods; and the change rate of correlation is specifically the change rate of correlation between two periods. The higher the correlation, the better the match between the exercise state and the assistive device, and the better the rehabilitation effect.

[0124] Figure 3 A block diagram of a rehabilitation glove assessment system based on multi-source data analysis according to the present invention is shown.

[0125] A second aspect of the present invention also provides a rehabilitation glove assessment system based on multi-source data analysis. The system includes a memory 13, a processor 12, and a communication interface 11. The memory includes a rehabilitation glove assessment program based on multi-source data analysis. The communication interface is used to establish data connections between various modules within the system. When the processor executes the rehabilitation glove assessment program based on multi-source data analysis, it performs the following steps:

[0126] S1: During the rehabilitation exercise cycle, multi-source monitoring data is collected through the sensor network of the rehabilitation glove, and multiple time nodes are set to analyze the motion characteristics of the multi-source monitoring data;

[0127] S2: Extract multi-dimensional motion parameters based on multi-source monitoring data, combine motion parameters that change in adjacent time nodes to form parameter items, collect all parameter items to form a data item set, and use the Apriori association algorithm to mine association rules from the data item set and filter out the associated parameter item set;

[0128] S3: Set time windows based on N time nodes, set variable parameters through the rehabilitation plan, and perform linear fitting calculations on multidimensional motion parameters for each time window in each time window. Filter out time windows with linear relationships and set motion-related time periods.

[0129] S4: Develop testing rules based on the rehabilitation plan, conduct conditional monitoring based on multi-source monitoring data, and screen out abnormal time points;

[0130] S5: Based on abnormal time nodes, filter out abnormal parameter items from the data item set, extract abnormal time periods from the motion-related time periods, and trace the motion parameters that match the abnormal time periods from the related item set based on the abnormal parameter items.

[0131] According to an embodiment of the present invention, S1 specifically includes:

[0132] Based on the rehabilitation plan, set a rehabilitation exercise cycle;

[0133] In the rehabilitation gloves, based on the user's rehabilitation monitoring needs, multiple sensors are set up, and the sensors are connected to the control terminal through the hardware layer to form a sensor network;

[0134] The various sensors include joint sensors, pressure sensors, and electromyography (EMG) sensors.

[0135] Here, it can be understood that the rehabilitation program specifically refers to a hand rehabilitation program. The rehabilitation exercise process involves the user's hand movements while wearing rehabilitation gloves. A rehabilitation exercise cycle is a hand activity analysis cycle, consisting of multiple hand movement processes. It is used to analyze the hand movement characteristics in different movement processes or activity states, such as joint rotation angles, range of motion, and changes in electromyographic signals. Multiple sensors are typically placed at hand muscle locations, joint locations, and fingertips, with one or more sensors arranged depending on the type of sensor and different monitoring needs. Rehabilitation gloves generally contain an assistive motor. During motion parameter analysis, relevant parameters of the assistive motor can be collected to determine the status of assisted movement; these parameters can be collected through joint sensors.

[0136] According to an embodiment of the present invention, S1 further includes:

[0137] By using multi-source monitoring data, the type and frequency of exercise can be assessed.

[0138] Based on the type of exercise state and the number of exercises, preset time intervals are set to divide the rehabilitation exercise cycle into multiple time points, resulting in multiple time nodes.

[0139] In one possible scenario, if there are multiple statistically significant movement states during a rehabilitation exercise cycle, such as repeatedly grasping objects or repeatedly opening and closing the palms, multiple time points can be set based on the number of movement states to finely analyze the changes in movement parameters at multiple consecutive time points, and subsequently implement linear fitting analysis of continuous movement characteristics.

[0140] The multi-source monitoring data includes multi-dimensional data, such as sensor data including electromyography (EMG) signal values, EMG signal intensity, pressure values, rotation angle, rotation amplitude, and movement speed. It also includes auxiliary motor data, such as power, speed, and current. Examples of recorded data are as follows: angle change: 5°-10°, amplitude change: 2cm-5cm, speed change: 0.1m / s-0.3m / s, auxiliary motor power: 2W-5W.

[0141] According to an embodiment of the present invention, step S2 specifically includes:

[0142] Multi-dimensional motion parameters are extracted from multi-source monitoring data to obtain motion parameter data;

[0143] Based on each time node, the motion parameter data is serialized and analyzed. Taking a time node as a benchmark, multiple motion parameters under that node are analyzed. If a motion parameter changes with a motion parameter at an adjacent time node, the two motion parameters are combined to obtain the parameter item.

[0144] Analyze all time points and integrate the parameter items to obtain a data item set;

[0145] Based on the Apriori association algorithm, minimum support and minimum confidence are set;

[0146] For each parameter item, its support is calculated based on the entire data itemset, and frequent itemsets are selected using the minimum support.

[0147] Calculate association rules and their confidence levels based on frequent itemsets, filter out association rules that exceed the minimum confidence level, and form association parameter itemsets.

[0148] In a preferred embodiment, the minimum support can be set to 0.3 and the minimum confidence can be set to 0.5.

[0149] It should be noted that, using a given time point as a baseline, multiple motion parameters are analyzed at that point. If a motion parameter changes in both a motion parameter at an adjacent time point, the two motion parameters are combined to obtain a parameter item. A specific example is as follows:

[0150] Parameter (change in motion amplitude:) 3cm, pressure change: +20kPa), parameter item (motion speed change) (0.1 m / s, +20 kPa), the positive and negative values ​​of the changes can be set by the actual direction of movement, or only the absolute value of the change can be recorded. Parameter items can be the same type of motion parameter. In addition, based on the needs of auxiliary analysis, motor auxiliary parameters can also be used as parameter items for combined analysis. Since motor parameters can reflect the corresponding rehabilitation assistance situation and reflect rehabilitation needs from an auxiliary perspective, they can also be used for rehabilitation exercise characteristic research and help to explore related motion data items.

[0151] Each time point includes corresponding multidimensional parameters. The motion parameter data includes electromyography (EMG) signal values, EMG signal intensity, pressure values, rotation angle, rotation amplitude, and movement speed. In addition, auxiliary motor parameters such as power, speed, and current may also be included.

[0152] The set of associated parameter items includes multiple associated items, each of which includes two or more parameter items, representing that the corresponding two or more parameter items are associated. The associated status is stored in a rule-based manner. In addition, since the participating items have time node information, the association rules also filter out the time nodes where the motion characteristics are associated.

[0153] Figure 2 The flowchart of the motion-related time period filtering process of the present invention is shown;

[0154] According to an embodiment of the present invention, S3 specifically includes:

[0155] Set the time window based on the time step consisting of N time nodes;

[0156] During the rehabilitation exercise cycle, multiple data segments are analyzed by moving time windows, and the current time window is marked in each movement;

[0157] Within the current time window, extract various motion parameters at all time points, and select the two motion parameters with the largest rate of change as independent and dependent variables, respectively.

[0158] Within the current time window, linear fitting calculations are performed on the independent and dependent variables to obtain the motion fitting coefficients;

[0159] Determine whether the motion fitting coefficient is greater than the minimum fitting coefficient. If so, mark the current time window as the motion-related period.

[0160] For multiple movement operations, motion characteristic correlation analysis was performed, and all motion-related time periods were marked.

[0161] It should be noted that each movement corresponds to the time interval within the rehabilitation exercise cycle of the time window. For example, if the rehabilitation exercise cycle is 10 time units long and the time window is 2 time units, then 5 time window movements can be performed. The rate of change can be determined by the maximum-minimum difference or the linear coefficient within the time window. The two exercise parameters with the highest rate of change can be randomly assigned as independent and dependent variables, primarily for analyzing linear changes. For example, if the rate of change of electromyography (EMG) signal and rotation amplitude is the highest within a time window, then a linear fit can be performed based on EMG signal and rotation amplitude as independent and dependent variables to analyze the correlation of exercise characteristics and determine whether to mark the associated exercise periods. Through correlation analysis of exercise characteristics, periods with correlations in exercise characteristics can be effectively marked, and associated exercise parameters can be further filtered in conjunction with abnormal time nodes.

[0162] The process for filtering exercise-related time periods is as follows: Figure 3 As shown.

[0163] According to an embodiment of the present invention, step S4 specifically includes:

[0164] The testing rules are formulated based on the rehabilitation program, and the testing rules include one or more motor characteristic conditions.

[0165] For each time point, conditions are monitored and judged from multi-source monitoring data, and time points that do not meet the rules are marked as abnormal time points.

[0166] It should be noted that the movement characteristic conditions are personalized settings based on the user's rehabilitation, used to filter out periods of rapid abnormal movement, such as movement amplitude greater than distance D, movement speed continuously less than V, and whether electromyographic signals are detected.

[0167] According to an embodiment of the present invention, step S5 specifically includes:

[0168] Based on abnormal time nodes, the corresponding parameter items are filtered from the data item set and marked as abnormal parameter items;

[0169] Analyze the motion-related time periods to which the abnormal time nodes belong and mark them as abnormal time periods;

[0170] If the abnormal time node does not exist in any motion-related time period, then the abnormal time period is constructed based on the adjacent time nodes;

[0171] Based on the set of associated parameter items, determine the parameter items associated with the abnormal parameter items and mark them as the first parameter items. Filter out the parameters that match the abnormal time period from the first parameter items to obtain the second parameter items.

[0172] Analyze all abnormal time points and use all the selected second parameter items as source abnormal parameters.

[0173] It should be noted that the motion-related time period includes multiple consecutive time nodes. Adjacent time nodes are the two nodes before and after the abnormal time node.

[0174] According to an embodiment of the present invention, step S5 further includes:

[0175] Based on the abnormal parameters, the source of the abnormal parameters, and the abnormal time period, the rehabilitation gloves are assessed for auxiliary equipment abnormalities, and the analysis results are sent to the data terminal.

[0176] It should be noted that, based on abnormal parameter items, source abnormal parameters, and abnormal time periods, the rehabilitation gloves can be analyzed for abnormal motion parameters, assess abnormal motion characteristics present during abnormal time periods, determine abnormal motion characteristics of the assistive device (assistive motor), and revise the rehabilitation plan accordingly. The rehabilitation gloves include the assistive device.

[0177] It is worth mentioning that existing technologies have difficulty capturing the time points and abnormal linkage motion parameters of abnormal linkage in hand rehabilitation devices, resulting in poor ability to trace linkage data. However, hand movements often have multiple motion characteristics and correlations of multiple motion parameters, and linkage characteristics and correlations are key to evaluating hand rehabilitation movements, which can be used to further trace and analyze the rehabilitation status and hand rehabilitation trends.

[0178] This invention effectively solves the above problems. It collects motion feature data of rehabilitation gloves through a data acquisition module. Furthermore, it uses the Apriori association algorithm to perform association matching on motion parameters under different spatiotemporal conditions. The matching process analyzes data itemsets formed by multidimensional parameters to construct association parameter itemsets with associated states. Furthermore, it uses linear fitting and time windows to set variable parameters for multidimensional motion parameters. By introducing linear fitting, it evaluates whether there are corresponding motion association characteristics in each time window, dynamically captures rehabilitation motion association time periods, quickly detects abnormal time periods using rules, and combines the association parameter itemsets with motion association time periods to filter out motion parameter items with linkage association, thereby realizing the source analysis of abnormal motion characteristics.

[0179] This invention enables dynamic tracing of related motion characteristics and screening of potential abnormal correlation parameters, effectively improving the level of rehabilitation program formulation, thereby enhancing the automation and intelligent rehabilitation assessment level of rehabilitation glove equipment, completing automated rehabilitation analysis, avoiding reliance on human experience for the analysis and monitoring of continuous movements, and improving the applicability of the equipment in multiple scenarios.

[0180] A third aspect of the present invention also provides a computer-readable storage medium comprising a rehabilitation glove assessment program based on multi-source data analysis, wherein when the rehabilitation glove assessment program based on multi-source data analysis is executed by a processor, it implements the steps of the rehabilitation glove assessment method based on multi-source data analysis as described in any of the preceding claims.

[0181] This invention discloses a method and system for evaluating rehabilitation gloves based on multi-source data analysis. The method includes: collecting multi-source monitoring data through the sensor network of the rehabilitation glove; performing motion feature analysis at multiple time points; extracting multi-dimensional motion parameters from the multi-source monitoring data and mining associated parameter itemsets using the Apriori association algorithm; setting time windows based on N time points and filtering motion-related time periods through linear fitting; detecting abnormal time points according to the rehabilitation plan and establishing detection rules; filtering abnormal parameter items from the data itemset based on abnormal time points, and extracting abnormal time periods from the motion-related time periods to trace the associated motion parameters that match the abnormal time periods. This invention achieves accurate evaluation of the rehabilitation exercise process, dynamically traces abnormal associated motion characteristics, and provides data support for optimizing rehabilitation plans.

[0182] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0183] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0184] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0185] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0186] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0187] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A rehabilitation glove evaluation method based on multi-source data analysis, characterized in that, include: S1. During the rehabilitation exercise cycle, multi-source monitoring data is collected through the sensor network of the rehabilitation glove, and multiple time nodes are set to analyze the motion characteristics of the multi-source monitoring data. S2. Extract multi-dimensional motion parameters based on multi-source monitoring data, combine motion parameters that change in adjacent time nodes to form parameter items, collect all parameter items to form a data item set, and use the Apriori association algorithm to mine association rules from the data item set and filter out the associated parameter item set; S3. Set time windows based on N time nodes, set variable parameters through the rehabilitation plan, and perform linear fitting calculations on the multidimensional motion parameters for each time window in each time window. Filter out time windows with linear relationships and set motion-related time periods. S4. Develop testing rules based on the rehabilitation plan, conduct conditional monitoring based on multi-source monitoring data, and screen out abnormal time points; S5. Based on the abnormal time nodes, filter out abnormal parameter items from the data item set, and extract the abnormal time period from the motion-related time period. Based on the abnormal parameter items, trace out the motion parameters that match the abnormal time period from the related item set. Specifically, S3 is: Set the time window based on the time step formed by N time nodes; During the rehabilitation exercise cycle, multiple data segments are analyzed by moving time windows, and the current time window is marked in each movement; Within the current time window, extract various motion parameters at all time points, and select the two motion parameters with the largest rate of change as independent and dependent variables, respectively. Within the current time window, linear fitting calculations are performed on the independent and dependent variables to obtain the motion fitting coefficients; Determine whether the motion fitting coefficient is greater than the minimum fitting coefficient. If so, mark the current time window as the motion-related period. For multiple movement operations, motion characteristic correlation analysis was performed, and all motion-related time periods were marked.

2. The rehabilitation glove evaluation method based on multi-source data analysis according to claim 1, characterized in that, S1 specifically includes: Based on the rehabilitation plan, set a rehabilitation exercise cycle; In the rehabilitation gloves, based on the user's rehabilitation monitoring needs, multiple sensors are set up, and the sensors are connected to the control terminal through the hardware layer to form a sensor network; The various sensors include joint sensors, pressure sensors, and electromyography (EMG) sensors.

3. The rehabilitation glove evaluation method based on multi-source data analysis according to claim 1, characterized in that, S1 further includes: By using multi-source monitoring data, the type and frequency of exercise can be assessed. Based on the type of exercise state and the number of exercises, preset time intervals are set to divide the rehabilitation exercise cycle into multiple time points, resulting in multiple time nodes.

4. The rehabilitation glove evaluation method based on multi-source data analysis of claim 1, wherein, Specifically, S2 is: Multi-dimensional motion parameters are extracted from multi-source monitoring data to obtain motion parameter data; Based on each time node, the motion parameter data is serialized and analyzed. Taking a time node as a benchmark, multiple motion parameters under that node are analyzed. If a motion parameter changes with a motion parameter at an adjacent time node, the two motion parameters are combined to obtain the parameter item. Analyze all time points and integrate the parameter items to obtain a data item set; Based on the Apriori association algorithm, minimum support and minimum confidence are set; For each parameter item, its support is calculated based on the entire data itemset, and frequent itemsets are selected using the minimum support. Calculate association rules and their confidence levels based on frequent itemsets, filter out association rules that exceed the minimum confidence level, and form association parameter itemsets.

5. The rehabilitation glove evaluation method based on multi-source data analysis of claim 1, wherein, Specifically, S4 is: The testing rules are formulated based on the rehabilitation program, and the testing rules include one or more motor characteristic conditions. For each time point, conditions are monitored and judged from multi-source monitoring data, and time points that do not meet the rules are marked as abnormal time points.

6. The rehabilitation glove evaluation method based on multi-source data analysis of claim 1, wherein, Specifically, S5 is: Based on abnormal time nodes, the corresponding parameter items are filtered from the data item set and marked as abnormal parameter items; Analyze the motion-related time periods to which the abnormal time nodes belong and mark them as abnormal time periods; If the abnormal time node does not exist in any motion-related time period, then the abnormal time period is constructed based on the adjacent time nodes; Based on the set of associated parameter items, determine the parameter items associated with the abnormal parameter items and mark them as the first parameter items. Filter out the parameters that match the abnormal time period from the first parameter items to obtain the second parameter items. Analyze all abnormal time points and use all the selected second parameter items as source abnormal parameters.

7. The rehabilitation glove evaluation method based on multi-source data analysis according to claim 6, characterized in that, The S5 also includes: Based on the abnormal parameters, the source of the abnormal parameters, and the abnormal time period, the rehabilitation gloves are assessed for auxiliary equipment abnormalities, and the analysis results are sent to the data terminal.

8. A rehabilitation glove evaluation system based on multi-source data analysis, characterized in that, The system includes: a memory, a processor, and a communication interface. The communication interface is used to establish data connections between various modules within the system. The memory includes a rehabilitation glove assessment program based on multi-source data analysis. When the processor executes the rehabilitation glove assessment program based on multi-source data analysis, it performs the following steps: S1: During the rehabilitation exercise cycle, multi-source monitoring data is collected through the sensor network of the rehabilitation glove, and multiple time nodes are set to analyze the motion characteristics of the multi-source monitoring data; S2: Extract multi-dimensional motion parameters based on multi-source monitoring data, combine motion parameters that change in adjacent time nodes to form parameter items, collect all parameter items to form a data item set, and use the Apriori association algorithm to mine association rules from the data item set and filter out the associated parameter item set; S3: Set time windows based on N time nodes, set variable parameters through the rehabilitation plan, and perform linear fitting calculations on multidimensional motion parameters for each time window in each time window. Filter out time windows with linear relationships and set motion-related time periods. S4: Develop testing rules based on the rehabilitation plan, conduct conditional monitoring based on multi-source monitoring data, and screen out abnormal time points; S5: Based on abnormal time nodes, filter out abnormal parameter items from the data item set, extract abnormal time periods from the motion-related time periods, and trace out motion parameters that match the abnormal time periods from the related item set based on the abnormal parameter items; Specifically, S3 is: Set the time window based on the time step formed by N time nodes; During the rehabilitation exercise cycle, multiple data segments are analyzed by moving time windows, and the current time window is marked in each movement; Within the current time window, extract various motion parameters at all time points, and select the two motion parameters with the largest rate of change as independent and dependent variables, respectively. Within the current time window, linear fitting calculations are performed on the independent and dependent variables to obtain the motion fitting coefficients; Determine whether the motion fitting coefficient is greater than the minimum fitting coefficient. If so, mark the current time window as the motion-related period. For multiple movement operations, motion characteristic correlation analysis was performed, and all motion-related time periods were marked.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a rehabilitation glove assessment program based on multi-source data analysis, which, when executed by a processor, implements the steps of the rehabilitation glove assessment method based on multi-source data analysis as described in any one of claims 1 to 7.

Images