Personalized rehabilitation training method and system based on state monitoring

Abstract

The invention provides a personalized rehabilitation training method and system based on state monitoring, and relates to the technical field of rehabilitation training. The method comprises the steps of obtaining myoelectricity amplitude, root mean square and frequency spectrum median, calculating change rate, monitoring muscle fiber mobilization, blood oxygen, lactic acid metabolism and heart rate variability, extracting joint angle change rate and inertia moment, and obtaining a training load reference value. According to the method, multiple physiological parameters including the electromyographic signals, the oxyhemoglobin saturation and the heart rate variability are monitored and analyzed in real time, the accurate training load reference value is established, the accuracy of neuromuscular coordination evaluation is improved, the change rate of the electromyographic signals of short and long time windows is compared, the training load is accurately adjusted, the fatigue risk is reduced, and the neuromuscular coordination evaluation accuracy is improved. By adjusting the motion trail of the rehabilitation robot in real time and optimizing the training rhythm, it is ensured that the training intensity is matched with the individual metabolic capacity, dynamic matching and personalized adjustment of the training plan are achieved, and the training safety and effect are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of rehabilitation training, and particularly to a personalized rehabilitation training method and system based on state monitoring. Background Art

[0002] The technical field of rehabilitation training includes a variety of methods and systems for intervening in patients' physical disabilities, aiming to restore patients' motor abilities, improve physiological functions and enhance the quality of life. This technical field involves multiple aspects such as biomechanical analysis, electromyography signal acquisition and processing, motion control and feedback regulation, virtual reality-assisted training, etc. Rehabilitation training systems usually combine sensor technology, computer-aided assessment and training program optimization, and through precise monitoring of patients' physiological and motion states, achieve personalized intervention. In addition, this field includes various modes such as robot-assisted rehabilitation training, neurorehabilitation training, biofeedback training, etc., to meet the needs of different patients. In recent years, with the development of artificial intelligence, big data and wearable devices, rehabilitation training technology has gradually evolved towards intelligence and adaptability, improving the accuracy and pertinence of training programs.

[0003] Among them, the personalized rehabilitation training method and system based on state monitoring refer to the method and its implementation system for formulating and dynamically adjusting rehabilitation training programs by collecting and analyzing data such as patients' motion states and physiological parameters in real time. The patent theme covers links such as physiological signal acquisition, data analysis and processing, training program generation and adjustment, and training execution feedback. Specifically, the system uses sensors to collect information such as electromyography, joint angles, gait parameters, etc., extracts key features using signal processing and pattern recognition technologies, formulates personalized training plans based on individual physiological models and rehabilitation theories, and adjusts training parameters in combination with real-time monitoring results. The training program is presented in the form of visual, tactile or force feedback, and the patient performs training actions according to the feedback, and the system continuously monitors the state changes to optimize subsequent training content.

[0004] In the aspect of training load assessment in the prior art, there is a lack of fine analysis of multiple physiological signals. The training benchmark is mainly based on fixed parameters and it is difficult to match individual physiological characteristics, resulting in insufficient adaptability in setting the training intensity. The neuromuscular response monitoring does not fully consider the time difference between the primary and auxiliary muscle groups, and cannot accurately evaluate coordination, affecting the refined adjustment of the training plan. The fatigue assessment method is relatively single and mostly relies on individual physiological parameters, making it difficult to accurately identify the state of muscle fiber mobilization, resulting in difficulty in effectively controlling fatigue accumulation during training. The method of adjusting the exercise state is relatively static, and the change of training load lacks dynamic feedback on the real-time physiological state, resulting in unstable training effects. The optimization of the training rhythm usually relies on preset time intervals and fails to adjust in combination with individual metabolic capabilities, reducing the training efficiency and possibly resulting in a mismatch between the training rhythm and metabolic requirements. The conversion of the training mode does not fully consider the matching of active and passive movements, and cannot accurately adjust the training strategy, which may affect the training effect and increase the additional physiological burden. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art and to propose a personalized rehabilitation training method and system based on state monitoring.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions:

[0007] A personalized rehabilitation training method based on state monitoring, comprising the following steps:

[0008] S1: Obtain the myoelectric amplitude, root mean square, median frequency spectrum, calculate the change rate, monitor muscle fiber mobilization, blood oxygen, lactic acid metabolism, heart rate variability, extract the change rate of joint angle, moment of inertia, and obtain the training load reference value;

[0009] S2: Call the training load reference value, calculate the time difference between myoelectric trigger and joint angle change, monitor the neuromuscular response of the primary and auxiliary muscle groups, analyze the mobilization rate, control torque, delay, screen out deviation data, and obtain the matching evaluation parameter;

[0010] S3: Call the neuromuscular matching evaluation parameter, calculate the ratio of the change rate of short and long window myoelectric signals, identify the muscle fiber mobilization rate, extract the fatigue accumulation index, analyze the fatigue inflection point threshold, calculate the stability in combination with the change rate of joint angle, screen out the training load correction parameter, monitor the motion state of the ICUAW rehabilitation robot, extract the myoelectric control parameter, analyze the matching of active and passive movements, adjust the robot motion trajectory, and generate the dynamic training load adjustment parameter;

[0011] S4: Call the dynamic training load adjustment parameter, calculate the heart rate recovery rate, analyze the glycogen metabolism trend, screen out the training rhythm optimization threshold, call the energy metabolism load index, judge the metabolic limit, and generate the training rhythm optimization parameter.

[0012] As a further solution of the present invention, the training load reference value includes an electromyogram signal reference, a metabolic state reference, and a joint movement reference. The matching evaluation parameter includes a neuromuscular response parameter, a movement control parameter, and a deviation evaluation parameter. The dynamic training load adjustment parameter includes a muscle fiber mobilization parameter, a fatigue state parameter, a joint stability parameter, and a movement matching parameter. The training rhythm optimization parameter includes a heart rate recovery parameter, a metabolic limit parameter, and an interval adjustment parameter.

[0013] As a further solution of the present invention, the specific steps for obtaining the electromyogram amplitude, root mean square, median frequency spectrum, calculating the change rate, monitoring muscle fiber mobilization, blood oxygen, lactic acid metabolism, and heart rate variability, and extracting the joint angle change rate and moment of inertia to obtain the training load reference value are as follows:

[0014] S101: Obtain the electromyogram signal of the target muscle group, extract the amplitude, root mean square value, and median frequency spectrum, compare the amplitude changes at adjacent times, calculate the change rate within the time window, and obtain the change rate of the electromyogram signal characteristics;

[0015] S102: Based on the change rate of the electromyogram signal characteristics, calculate the peak mobilization times per unit time, combine the amplitude fluctuation range, obtain the muscle fiber mobilization rate, simultaneously monitor the blood oxygen saturation, calculate the blood oxygen saturation decrease rate, and obtain the blood oxygen fluctuation range, calculate the average value of the blood oxygen level decrease per unit time, and generate the muscle fiber mobilization rate and blood oxygen metabolism rate;

[0016] S103: Call the muscle fiber mobilization rate and blood oxygen metabolism rate, calculate the lactic acid metabolism rate, monitor the heart rate variability, extract the joint angle change rate and moment of inertia, and obtain the training load reference value.

[0017] As a further solution of the present invention, the specific steps for calling the training load reference value, calculating the time difference between the electromyogram trigger and the joint angle change, monitoring the neuromuscular response of the main and auxiliary muscle groups, analyzing the mobilization rate, control torque, and delay, screening the deviation data, and obtaining the matching evaluation parameter are as follows:

[0018] S201: Call the training load reference value, extract the electromyogram signal trigger moment and the starting moment of the joint angle change, calculate the time difference, and monitor the neuromuscular response time difference between the main muscle group and the auxiliary muscle group to obtain the neuromuscular response time parameter;

[0019] S202: Based on the neuromuscular response time parameter, calculate the change interval of the muscle fiber recruitment rate, extract the muscle control torque, calculate the torque fluctuation amplitude per unit time, and simultaneously perform time series analysis on the neuromuscular response delay, extract the offset of the differential muscle group response, screen the data with deviation exceeding the set threshold on the time axis, remove outliers, and obtain the muscle fiber recruitment and nerve response analysis parameters;

[0020] S203: Invoke the muscle fiber recruitment and nerve response analysis parameters, extract the matching degrees of the muscle fiber recruitment rate, muscle control torque, and nerve response delay, and calculate the consistency offset interval on the time axis to obtain the matching evaluation parameters.

[0021] As a further solution of the present invention, the specific steps for invoking the neuromuscular matching evaluation parameters, calculating the change rate ratio of the short and long window electromyogram signals, identifying the muscle fiber recruitment rate, extracting the fatigue accumulation index, analyzing the fatigue inflection point threshold, calculating the stability in combination with the joint angle change rate, screening the training load correction parameters, monitoring the motion state of the ICUAW rehabilitation robot, extracting the electromyogram control parameters, analyzing the matching of active and passive movements, and adjusting the robot motion trajectory to generate the dynamic training load adjustment parameters are as follows:

[0022] S301: Invoke the matching evaluation parameters, calculate the change rate ratio of the electromyogram signals in the short time window and the long time window, and determine whether it exceeds the training load adjustment threshold, and simultaneously identify the muscle fiber recruitment rate to obtain the training load threshold determination parameter;

[0023] S302: Based on the training load threshold determination parameter, extract the time series of the electromyogram signal change rate, calculate the cumulative fatigue index per unit time, analyze the electromyogram signal change trend, extract the change of the muscle fiber recruitment rate in the differential time window, analyze the fatigue inflection point threshold, and simultaneously invoke the joint angle change rate to identify the joint stability parameter, and screen the data that meets the training load adjustment requirements to obtain the fatigue accumulation and joint stability parameters;

[0024] S303: Invoke the fatigue accumulation and joint stability parameters, obtain the motion state monitoring data of the ICUAW rehabilitation robot, extract the electromyogram signal control parameters, calculate the matching index of passive and active movements, and adjust the robot motion trajectory to obtain the dynamic training load adjustment parameters.

[0025] As a further solution of the present invention, the specific calculation formula of the matching index of passive and active movements is:

[0026]

[0027] Among them, Ω represents the matching index of passive movement and active movement, κ represents the total number of discrete moments within the active movement time window, and γ i represents the instantaneous amplitude of the EMG signal at the i-th moment, and λ i represents the external torque applied by the rehabilitation robot at the i-th moment, and δ i represents the neuromuscular response delay at the i-th moment, ξ represents the total number of discrete moments within the passive movement time window, and ρ j represents the restoring force generated by muscle stretching during passive movement at the j-th moment, and τ j represents the instantaneous angular acceleration of the rehabilitation robot's movement at the j-th moment.

[0028] As a further solution of the present invention, the specific steps of calling the dynamic training load adjustment parameter, calculating the heart rate recovery rate, analyzing the glycogen metabolism trend, screening the training rhythm optimization threshold, calling the energy metabolism load index, and judging the metabolic limit to generate the training rhythm optimization parameter are as follows:

[0029] S401: Call the dynamic training load adjustment parameter, extract the heart rate signal, calculate the heart rate recovery rate per unit time, analyze the trend of glycogen metabolism ratio, extract the change of glycogen metabolism ratio on the time axis, and obtain the heart rate recovery and metabolism trend parameters;

[0030] S402: Based on the heart rate recovery and metabolism trend parameters, extract the time series of glycogen metabolism ratio under different training rhythms, calculate the metabolic fluctuation range index per unit time, screen the training rhythm optimization threshold, and at the same time call the instantaneous energy metabolism load index, compare the training rhythm with the metabolic limit situation, extract the load adjustment interval of the training rhythm, and identify the change of the load increment per unit time to obtain the training rhythm metabolism matching parameter;

[0031] S403: Call the training rhythm metabolism matching parameter, calculate the adjustment interval index of the training interval time, screen the training rhythm optimization threshold, extract the rhythm adjustment requirement parameter, and generate the training rhythm optimization parameter.

[0032] As a further solution of the present invention, the specific calculation formula of the adjustment interval index of the training interval time is:

[0033]

[0034] Among them, Θ rto represents the adjustment interval index of the training interval time, N prds represents the number of training cycles considered, ξ pmk represents the value of the training rhythm metabolism matching parameter within the k-th cycle, and η adjk represents the adjustment coefficient of the k-th cycle, reflecting the sensitivity of rhythm change to the training effect.

[0035] As a further solution of the present invention, the method further includes S5: calling the training rhythm optimization parameter, combining the electromyogram change rate, matching evaluation, and load adjustment, calculating the personalized training load, adjusting the movement amplitude, matching the training plan, identifying the training conversion strategy, and generating a personalized rehabilitation plan;

[0036] The personalized rehabilitation training plan includes training load parameters, training rhythm parameters, and movement mode parameters;

[0037] S501: Call the training rhythm optimization parameter, combine the electromyogram signal change rate, neuromuscular matching evaluation parameter, and dynamic training load adjustment parameter, extract the change trend of the training index on the time axis, calculate the personalized training load configuration index, and obtain the personalized load configuration parameter;

[0038] S502: Based on the personalized load configuration parameter, extract the change range of the training movement amplitude, calculate the movement trajectory deviation index per unit time, compare with the stage training plan, extract the training movement amplitude that meets the load matching requirements, identify the active-passive training conversion strategy at the same time, analyze the load switching point between active and passive training, and screen the time interval that meets the conversion threshold to obtain the active-passive conversion and training amplitude parameters;

[0039] S503: Call the active-passive conversion and training amplitude parameters, extract the movement situation of the rehabilitation robot, analyze the deviation interval of the robot movement trajectory on the time axis, and match the personalized training load configuration to generate a personalized rehabilitation plan.

[0040] The personalized rehabilitation training system based on state monitoring includes:

[0041] The electromyogram and load monitoring module obtains the electromyogram amplitude, root mean square, and spectral median, monitors the muscle fiber mobilization, blood oxygen, lactic acid metabolism, and heart rate variability, and extracts the joint angle change rate and moment of inertia to obtain the training load reference value;

[0042] The nerve response analysis module calls the training load reference value, calculates the time difference between electromyogram trigger and joint angle change, monitors the neuromuscular response of the main and auxiliary muscle groups, analyzes the mobilization rate, control torque, and delay, and obtains the matching evaluation parameter;

[0043] The dynamic load adjustment module calls the neuromuscular matching evaluation parameter, calculates the ratio of the electromyogram signal change rate of the short and long windows, identifies the muscle fiber mobilization rate, extracts the fatigue accumulation index, analyzes the fatigue inflection point threshold, calculates the stability in combination with the joint angle change rate, screens the training load correction parameter, monitors the movement state of the ICUAW rehabilitation robot, extracts the electromyogram control parameter, analyzes the active-passive movement matching, and generates the dynamic training load adjustment parameter;

[0044] The rhythm optimization control module calls the dynamic training load adjustment parameter, calculates the heart rate recovery rate, analyzes the glycogen metabolism trend, screens the training rhythm optimization threshold, calls the energy metabolism load index, and generates the training rhythm optimization parameter;

[0045] The personalized rehabilitation planning module calls the training rhythm optimization parameter, combines the EMG change rate, matching evaluation, and load adjustment, calculates the personalized training load, adjusts the movement amplitude, matches the training plan, and generates the personalized rehabilitation plan.

[0046] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0047] In the present invention, by real-time monitoring and analyzing multiple physiological parameters including EMG signals, blood oxygen saturation, and heart rate variability, an accurate training load reference value is established, the accuracy of neuromuscular coordination evaluation is improved, the training load is accurately adjusted by comparing the EMG signal change rates in short and long time windows, the fatigue risk is reduced, and by real-time adjusting the movement trajectory of the rehabilitation robot and optimizing the training rhythm, it is ensured that the training intensity matches the individual metabolic capacity, realizing the dynamic matching and personalized adjustment of the training plan, and effectively improving the training safety and effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0049] Figure 1 It is a schematic diagram of the step flow of the present invention;

[0050] Figure 2 It is a flowchart of step S1 of the present invention;

[0051] Figure 3 It is a flowchart of step S2 of the present invention;

[0052] Figure 4 It is a flowchart of step S3 of the present invention;

[0053] Figure 5 It is a flowchart of step S4 of the present invention;

[0054] Figure 6 It is a flowchart of step S5 of the present invention;

[0055] Figure 7 It is a system module diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] The technical solutions in the present invention will be described below with reference to the accompanying drawings.

[0057] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as an "example" in the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of the word "example" is intended to present concepts in a specific manner. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one of the two can be selected.

[0058] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meanings they express are the same. "Of", "corresponding", and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meanings they express are the same.

[0059] In the embodiments of the present invention, sometimes subscripts such as W 1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meanings they express are the same.

[0060] To make the technical problems to be solved, technical solutions and advantages of the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0061] Please refer to Figure 1 , an embodiment of the present invention provides a personalized rehabilitation training method based on state monitoring, including:

[0062] S1: Obtain the amplitude, root mean square value, and spectral median of the electromyogram signal of the target muscle group, calculate the change rate within the time window, monitor the muscle fiber mobilization rate and blood oxygen saturation, calculate the lactic acid metabolism rate and heart rate variability, and extract the joint angle change rate and moment of inertia to obtain the training load reference value;

[0063] S2: Call the training load reference value, calculate the time difference between the electromyogram signal trigger moment and the starting moment of the joint angle change, monitor the neuromuscular response time gap between the main muscle group and the auxiliary muscle group, analyze the muscle fiber mobilization rate, muscle control torque, and neuromuscular response delay, screen the deviation data, and obtain the matching evaluation parameters;

[0064] S3: Call the matching evaluation parameters, calculate the ratio of the myoelectric signal change rate between the short-time window and the long-time window, determine whether it exceeds the training load adjustment threshold, identify the muscle fiber mobilization rate, extract the fatigue accumulation index, analyze the fatigue inflection point threshold, call the joint angle change rate to calculate the joint stability parameters, screen the training load correction calculation parameters, call the ICUAW rehabilitation robot motion state monitoring, extract the myoelectric signal control parameters, analyze the matching of passive and active movements, adjust the robot motion trajectory, and obtain the dynamic training load adjustment parameters;

[0065] S4: Call the dynamic training load adjustment parameters, calculate the heart rate recovery rate, analyze the trend of glycogen metabolism proportion, screen the training rhythm optimization threshold, call the instantaneous energy metabolism load index, determine the situation where the training rhythm is close to the metabolic limit, adjust the training interval time, and generate the training rhythm optimization parameters;

[0066] S5: Call the training rhythm optimization parameters, combine the myoelectric signal change rate, neuromuscular matching evaluation parameters, and dynamic training load adjustment parameters, calculate the personalized training load configuration index, adjust the training movement amplitude, match the stage training plan, identify the main passive training conversion strategy, extract the motion situation of the rehabilitation robot, and generate a personalized rehabilitation plan.

[0067] The training load reference values include the myoelectric signal reference, metabolic state reference, and joint movement reference. The matching evaluation parameters include neuromuscular response parameters, motion control parameters, and deviation evaluation parameters. The dynamic training load adjustment parameters include muscle fiber mobilization parameters, fatigue state parameters, joint stability parameters, and motion matching parameters. The training rhythm optimization parameters include heart rate recovery parameters, metabolic limit parameters, and interval adjustment parameters. The personalized rehabilitation training plan includes training load parameters, training rhythm parameters, and motion mode parameters.

[0068] Please refer to Figure 2 , the specific steps of S1 are as follows:

[0069] S101: Obtain the myoelectric signals of the target muscle group, extract the amplitude, root mean square value, and spectral median, compare the amplitude changes at adjacent times, calculate the change rate within the time window, and obtain the myoelectric signal characteristic change rate;

[0070] First, surface electromyogram (sEMG) electrodes are attached to the target muscle group, and the electrode spacing is ensured to be appropriate to reduce artifact interference. Subsequently, the original data of the myoelectric signals are collected, and the signals are collected at a sampling frequency above 1 kHz to ensure the timing accuracy of the data. After the data is obtained, pre-filtering processing is performed, including 50 Hz power frequency filtering to eliminate power supply noise, and a band-pass filter (20 Hz - 450 Hz) is used to extract the effective myoelectric signal components to obtain the processed time-domain myoelectric signals. When calculating the amplitude, the collected signals are processed by taking the absolute value to obtain the full-wave rectified signal, and its mean value is taken. When calculating the root mean square (RMS) value, the time-domain myoelectric signal is squared, averaged, and then square-rooted. When calculating the spectral median, the signal is first subjected to a fast Fourier transform (FFT) to obtain its power spectral density distribution, and the frequency value that makes the total power half is found as the median frequency. Subsequently, the amplitude change amount at adjacent times is compared. By calculating the absolute value difference of the amplitude changes at adjacent sampling points and averaging, the amplitude change trend is determined. Furthermore, the amplitude change rate within a fixed time window (such as 500 ms) is set, and the amplitude change rate within the time window is statistically calculated. This rate can be calculated by setting the mean value of all amplitude changes within the window, and finally, the characteristic change rate of the myoelectric signal is obtained.

[0071] S102: Based on the characteristic change rate of the myoelectric signal, calculate the peak mobilization times per unit time, combine the amplitude fluctuation range, and obtain the muscle fiber mobilization rate. At the same time, monitor the blood oxygen saturation, calculate the blood oxygen saturation decline rate, and obtain the blood oxygen fluctuation range, and calculate the average value of the blood oxygen level decline per unit time to generate the muscle fiber mobilization rate and the blood oxygen metabolism rate;

[0072] First, peak detection is performed on the myoelectric signals within a fixed time window. A threshold is set to judge the peak points. For example, the points where the signal within a certain time window is higher than the mean value + 2 times the standard deviation within this time period are set as peak points, and the number of peak points is statistically calculated to calculate the peak mobilization times per unit time. At the same time, the amplitude fluctuation range is set, that is, the difference between the maximum amplitude and the minimum amplitude of the myoelectric signals within this time window is calculated to obtain the amplitude fluctuation amplitude within this range. On this basis, the muscle fiber mobilization rate is calculated, that is, the ratio of the peak mobilization times per unit time to the amplitude fluctuation range. At the same time, a photoelectric sensor is used to monitor the blood oxygen saturation (SpO2), and a fixed time window is set to calculate the blood oxygen saturation decline rate. The method is to calculate the change amount of the blood oxygen saturation at the beginning and end of the window and divide it by the length of the time window to obtain the blood oxygen decline rate. In addition, by setting the blood oxygen fluctuation range, the difference between the maximum value and the minimum value of the blood oxygen saturation within the window is calculated, and further, the average value of the blood oxygen level decline per unit time is calculated. This average value can be obtained by statistically calculating the average value of the blood oxygen decline rates within multiple windows, and finally, the muscle fiber mobilization rate and the blood oxygen metabolism rate are obtained.

[0073] S103: Invoke the muscle fiber mobilization rate and blood oxygen metabolism rate, calculate the lactic acid metabolism rate, monitor the heart rate variability, extract the joint angle change rate and moment of inertia, and obtain the training load reference value;

[0074] First, calculate the lactic acid metabolism rate. The calculation of the lactic acid metabolism rate is based on the relationship between the blood oxygen metabolism rate and the muscle fiber mobilization rate. The oxygen consumption rate is calculated through the blood oxygen decline rate, and the corresponding lactic acid production rate is estimated in combination with the muscle fiber mobilization rate. The lactic acid accumulation amount at different time points is calculated using a fixed time window. The calculation formula can be expressed as lactic acid metabolism rate = (lactic acid production amount per unit time - lactic acid clearance amount per unit time) / time window. At the same time, monitor the heart rate variability (HRV), detect the RR interval of the collected heart rate signal, and calculate heart rate variability indexes such as SDNN (standard deviation), RMSSD (root mean square of successive RR intervals), etc. to evaluate the physiological load. In addition, use an inertial measurement unit (IMU) to obtain joint movement data, calculate the joint angle change rate, that is, calculate the angle change amount within the time window and divide it by the time length. At the same time, calculate the moment of inertia according to the angular velocity, mass, and lever arm length. Finally, calculate the training load reference value based on the above parameters. The training load reference value can be obtained through weighted comprehensive calculation of the muscle fiber mobilization rate, blood oxygen metabolism rate, lactic acid metabolism rate, heart rate variability, and joint movement characteristics.

[0075] Please refer to Figure 3 , the specific steps of S2 are as follows:

[0076] S201: Invoke the training load reference value, extract the trigger moment of the electromyogram signal and the starting moment of the joint angle change, calculate the time difference, and monitor the neuromuscular response time difference between the main muscle group and the auxiliary muscle group to obtain the neuromuscular response time parameter;

[0077] First, extract the trigger moment of the electromyogram signal. This trigger moment is detected by setting a threshold to detect the sudden increase point of the electromyogram signal. For example, the moment when the signal exceeds the mean + 3 times the standard deviation within a certain time window is set as the trigger point. Subsequently, extract the starting moment of the joint angle change. This moment is monitored by an inertial measurement unit (IMU) for the joint angular velocity. When the angular velocity is greater than the set static state threshold, this moment is recognized as the starting moment of the angle change. When calculating the time difference, directly perform subtraction on the timestamps of the trigger moment of the electromyogram signal and the starting moment of the joint angle change to obtain the time difference. Then, monitor the neuromuscular response time difference between the main muscle group and the auxiliary muscle group. Detect the trigger moments of the electromyogram signals of the main muscle group and the auxiliary muscle group respectively using the same threshold method, and calculate the time interval between the two. At the same time, in order to avoid the influence of individual differences on the time difference, it is necessary to compare the response time changes under different training loads, calculate the mean value of the time difference between the main muscle group and the auxiliary muscle group under different load conditions, and calculate the stability of this time difference in combination with the standard deviation. Finally, obtain the neuromuscular response time parameter.

[0078] S202: Based on the neuromuscular response time parameter, calculate the change interval of the muscle fiber recruitment rate, extract the muscle control torque, calculate the torque fluctuation amplitude per unit time, perform time series analysis on the neuromuscular response delay, extract the offset of the differential muscle group response, screen the data with deviation exceeding the set threshold on the time axis, remove the outliers, and obtain the muscle fiber recruitment and nerve response analysis parameters;

[0079] First, calculate the change interval of the muscle fiber recruitment rate. The calculation method of this change interval is to statistically calculate the maximum and minimum values of the muscle fiber recruitment rate within a fixed time window, and take the difference between the two as the fluctuation interval of the recruitment rate. Subsequently, extract the muscle control torque. The calculation method of the torque is based on the measured joint angle, angular velocity, and the force exerted by the muscle. The torque calculation formula is M = F × r, where M is the muscle control torque, F is the pulling force generated by the muscle, and r is the length of the force arm. Calculate this torque value through the measured muscle pulling force and joint biomechanics parameters. On this basis, calculate the torque fluctuation amplitude per unit time, that is, statistically calculate the difference between the maximum and minimum values of the torque within a fixed time window, and take its average value as the fluctuation amplitude per unit time. At the same time, perform time series analysis on the neuromuscular response delay. During the time series analysis process, by setting the window length, calculate the response delay within different time windows, and extract the response offset of the differential muscle group. Screen the data with deviation exceeding the set threshold on the time axis. This threshold can be set by statistically calculating the mean ± 2 times the standard deviation of historical data. The data exceeding this threshold is regarded as an outlier and excluded. Finally, obtain the muscle fiber recruitment and nerve response analysis parameters.

[0080] S203: Invoke the muscle fiber recruitment and nerve response analysis parameters, extract the matching degrees of the muscle fiber recruitment rate, muscle control torque, and neuromuscular response delay, and calculate the consistency offset interval on the time axis to obtain the matching evaluation parameters;

[0081] First, extract the matching degrees of the muscle fiber recruitment rate, muscle control torque, and neuromuscular response delay. When calculating the matching degree, calculate the correlation coefficients between the muscle fiber recruitment rate and the muscle control torque, and between the muscle fiber recruitment rate and the neuromuscular response delay within different time windows respectively. Use the Pearson correlation coefficient R to quantify the matching degree. The calculation formula is R = Cov(X,Y) / (σX × σY), where Cov(X,Y) is the covariance of X and Y, and σX and σY are the standard deviations of X and Y respectively. The closer the matching degree is to 1, the more consistent the change trends of the two are. Subsequently, calculate the consistency offset interval on the time axis, that is, statistically calculate the change range of the matching degree within different time windows, and calculate its mean value and standard deviation to obtain the matching evaluation parameters. This evaluation parameter can be used to determine the rationality of training load adjustment.

[0082] Please refer toFigure 4 , the specific steps of S3 are as follows:

[0083] S301: Invoke the matching evaluation parameters, calculate the ratio of the myoelectric signal change rates between the short time window and the long time window, determine whether it exceeds the training load adjustment threshold, simultaneously identify the muscle fiber mobilization rate, and obtain the training load threshold determination parameter;

[0084] First, calculate the ratio of the myoelectric signal change rates between the short time window and the long time window. The short time window can be set to 500 ms, and the long time window can be set to 5 s. Calculate the myoelectric signal change rate within each window, that is, by extracting the root mean square value (RMS) or the amplitude change rate of the myoelectric signal and calculating the mean value within the time window. Subsequently, calculate the ratio of the change rate value of the short time window to the change rate value of the long time window to obtain the short time window / long time window ratio, and determine whether it exceeds the training load adjustment threshold. The setting of this threshold can be based on the mean and standard deviation of the trainer's historical data. For example, a ratio exceeding the mean ± 2 times the standard deviation is set as abnormal load. Then, identify the muscle fiber mobilization rate, extract the peak mobilization times and the amplitude fluctuation range within the unit time window, calculate the change range of the muscle fiber mobilization rate, and combine it with the ratio of the short time window and the long time window to obtain the training load threshold determination parameter. This parameter is used to evaluate whether the current load is in a stable range or whether the training plan needs to be adjusted.

[0085] S302: Based on the training load threshold determination parameter, extract the time series of the myoelectric signal change rate, calculate the cumulative fatigue index per unit time, analyze the myoelectric signal change trend, extract the change of the muscle fiber mobilization rate in different time windows, analyze the fatigue inflection point threshold, simultaneously invoke the joint angle change rate, identify the joint stability parameter, and screen the data that meet the training load adjustment requirements to obtain the fatigue accumulation and joint stability parameters;

[0086] First, extract the time series of the change rate of the electromyogram (EMG) signal. Calculate the change rate at each time point according to a fixed time window to construct a time series data set. Subsequently, calculate the cumulative fatigue index per unit time. The calculation of the cumulative fatigue index is based on the decline amplitude of the median frequency of the EMG signal and the decay rate of the root mean square value. Set the change amount between the initial value and the terminal value within the time window, and accumulate and calculate the change trend within multiple windows. Then, analyze the change trend of the EMG signal. Calculate the mean and standard deviation of the change rate within different time windows, and extract the change of the muscle fiber recruitment rate within different time windows, that is, statistically analyze the change range of the muscle fiber recruitment rate under short-time windows and long-time windows. By setting the change slope for different time periods, calculate the fatigue inflection point threshold, which can be obtained by comparing the change rates before and after the maximum endurance time point. For example, if the decline of the median frequency of the EMG signal exceeds 20% within 5 minutes before and after the fatigue inflection point, it is determined as the fatigue inflection point. In addition, call the data of the change rate of joint angles, extract the mean values of the angular velocity and angular acceleration per unit time, and calculate the joint stability parameter, that is, judge whether there are abnormal changes through the fluctuation range of the angular acceleration, and screen the data that meets the requirements of training load adjustment to obtain the fatigue accumulation and joint stability parameters.

[0087] S303: Call the fatigue accumulation and joint stability parameters, obtain the motion state monitoring data of the ICUAW rehabilitation robot, extract the EMG signal control parameters, calculate the matching index between passive motion and active motion, and adjust the robot motion trajectory to obtain the dynamic training load adjustment parameters;

[0088] The specific calculation formula for the matching index between passive motion and active motion is as follows:

[0089]

[0090] Among them, Ω represents the matching index between passive motion and active motion, κ represents the total number of discrete time moments within the active motion time window, γ i represents the instantaneous amplitude of the EMG signal at the i-th moment, λ i represents the external torque applied by the rehabilitation robot at the i-th moment, δ i represents the neuromuscular response delay at the i-th moment, ξ represents the total number of discrete time moments within the passive motion time window, ρ j represents the restoring force generated by muscle stretching during passive motion at the j-th moment, τ j represents the instantaneous angular acceleration of the rehabilitation robot during motion at the j-th moment;

[0091] The purpose of the formula calculation is to determine the matching index Ω between passive movement and active movement, which evaluates the coordination degree between passive mechanical assisted movement and the patient's voluntary movement during the rehabilitation process. This calculation involves two parts: the numerator represents the correlation measurement of active movement, and the denominator represents the stability measurement of passive movement.

[0092] First, for the numerator part, γ i represents the instantaneous amplitude of the electromyogram signal at the i-th moment, which is obtained by real-time monitoring of the electromyogram sensor. If the amplitude of the electromyogram signal at a certain moment i is 0.8 mV, then record γ i = 0.8. Similarly, λ i is the external torque applied by the rehabilitation robot at the i-th moment, which is obtained from the data read by the sensor. If the torque is 15 Nm, then record λ i = 15. δ i is the neuromuscular response delay, which is calculated by analyzing the time difference between the electromyogram signal and the movement response, and is assumed to be 0.03 seconds.

[0093] Suppose there are 10 moments in a window period κ, then calculate the above terms for each moment and sum them. For example:

[0094]

[0095] For the denominator part, ρ j represents the restoring force generated by muscle stretching during passive movement at the j-th moment. If the measured data is 0.5 N, then record ρ j = 0.5. τ j represents the instantaneous angular acceleration of the rehabilitation robot during movement at the j-th moment. If the device reading is 0.02 rad / s2, then record τ j = 0.02.

[0096] If there are 10 moments in a window period ξ, then calculate:

[0097]

[0098] Substitute the above two parts into the formula to get:

[0099]

[0100] This result shows that the movement matching index between the rehabilitation robot and the patient is 0.8517, indicating a relatively high degree of coordination between the two. The patient's active participation and the robot's passive assistance are relatively consistent in terms of torque and response time, and can effectively support the patient's dynamic load adjustment during rehabilitation training.

[0101] Please refer to Figure 5 , the specific steps of S4 are:

[0102] S401: Invoke the dynamic training load adjustment parameter, extract the heart rate signal, calculate the heart rate recovery rate per unit time, analyze the trend of glycogen metabolism proportion, extract the change of glycogen metabolism proportion on the time axis, and obtain the heart rate recovery and metabolism trend parameters;

[0103] First, extract the heart rate signal, use a heart rate monitoring device to record the heart rate data per unit time, and calculate the heart rate recovery rate. Set a fixed time window, such as a 60 - second time period after the training ends, calculate the change in the starting heart rate and the terminal heart rate, and divide the change by the length of this time window to obtain the heart rate recovery rate per unit time. Subsequently, analyze the trend of glycogen metabolism proportion, extract the change in glycogen metabolism proportion per unit time. When calculating the glycogen metabolism proportion, first measure the blood lactate concentration, extract the blood lactate values at set time intervals, and calculate the ratio of glycogen and fat metabolism in combination with the respiratory exchange rate to obtain the change in glycogen metabolism on the time axis, and further calculate the average value of the glycogen metabolism rate within the time window. Finally, obtain the heart rate recovery and metabolism trend parameters.

[0104] S402: Based on the heart rate recovery and metabolism trend parameters, extract the time series of glycogen metabolism proportion under different training rhythms, calculate the metabolic fluctuation range index per unit time, and screen the training rhythm optimization threshold. At the same time, invoke the instantaneous energy metabolism load index, compare the training rhythm with the metabolic limit situation, extract the load adjustment interval of the training rhythm, and identify the change in the load increment per unit time to obtain the training rhythm metabolism matching parameters;

[0105] First, extract the time series of glycogen metabolism proportion under different training rhythms, divide the training rhythm data according to a fixed time window, and calculate the average value of glycogen metabolism proportion within each window. Subsequently, calculate the metabolic fluctuation range index per unit time, which is obtained by statistically calculating the difference between the maximum and minimum values of glycogen metabolism proportion within the time window, and further calculate the average metabolic fluctuation within multiple windows. At the same time, screen the training rhythm optimization threshold, and the setting of this threshold can be obtained through historical training data. For example, set that when the glycogen metabolism proportion fluctuates by more than 20% as the unstable interval. Subsequently, invoke the instantaneous energy metabolism load index, extract the energy load consumed per unit time, and calculate the comparison between the training rhythm and the metabolic limit situation, statistically calculate the correlation between energy consumption and glycogen metabolism within the time window, extract the load adjustment interval of the training rhythm, that is, calculate the glycogen metabolism fluctuation range under different rhythms, and screen the optimal rhythm in combination with the heart rate recovery rate. Finally, identify the change in the load increment per unit time, calculate the load increment change rate, and statistically calculate the change trend within multiple windows to obtain the training rhythm metabolism matching parameters.

[0106] S403: Call the training rhythm metabolism matching parameters, calculate the adjustment interval index of the training interval time, screen the training rhythm optimization threshold, extract the rhythm adjustment requirement parameters, and generate the training rhythm optimization parameters;

[0107] The calculation formula of the adjustment interval index of the training interval time is specifically:

[0108]

[0109] Among them, Θ rto represents the adjustment interval index of the training interval time, N prds represents the number of training cycles considered, ξ pmk represents the value of the training rhythm metabolism matching parameter in the k-th cycle, η adjk represents the adjustment coefficient in the k-th cycle, reflecting the sensitivity of the rhythm change to the training effect;

[0110] The formula aims to calculate the adjustment interval index Θ rto of the training interval time, and adjust the future training plan by analyzing the training data of a certain period N prds . This formula combines the training rhythm metabolism matching parameter ξ pmk of each cycle and the adjustment coefficient η adjk of the corresponding cycle to obtain a weighted average value, representing the optimization threshold.

[0111] Set a practical example: Assume that the number of cycles N prds considered is 5, indicating that the analysis of five training cycles has been carried out. For the training rhythm metabolism matching parameter ξ pmk of each cycle, assume that it is obtained through electromyogram signal analysis and energy consumption calculation, and the specific values are:

[0112] 0.92, 0.85, 0.88, 0.90, 0.87;

[0113] The adjustment coefficient η adjk represents the dynamic adjustment for individual adaptability. Assume that these values are obtained based on the previous training response and the coach's evaluation, and the values are set as:

[0114] 1.05, 0.95, 1.00, 1.03, 0.98;

[0115] The specific calculation process is as follows: 1. Calculation for the first cycle: 0.92×1.05 = 0.966 2. Calculation for the second cycle: 0.85×0.95 = 0.8075 3. Calculation for the third cycle: 0.88×1.00 = 0.88 4. Calculation for the fourth cycle: 0.90×1.03 = 0.927 5. Calculation for the fifth cycle: 0.87×0.98 = 0.8526

[0116] The sum of these results is as follows:

[0117] 0.966 + 0.8075 + 0.88 + 0.927 + 0.8526 = 4.4331;

[0118] Then divide by the number of cycles N prds :

[0119]

[0120] This numerical result indicates that after comprehensive dynamic adjustment and matching analysis of five training cycles, the adjustment interval index of the training interval time is 0.88662. This means that the subsequent adjustment of the training rhythm can be optimized around this threshold to improve the adaptability and effectiveness of training. This threshold is comprehensively calculated based on actual training data and dynamic adjustment coefficients, ensuring the personalization and scientific nature of the training plan.

[0121] Please refer to Figure 6 , and the specific steps of S5 are as follows:

[0122] S501: Invoke the training rhythm optimization parameters, combine the myoelectric signal change rate, neuromuscular matching evaluation parameters, and dynamic training load adjustment parameters, extract the change trend of training indicators on the time axis, calculate the personalized training load configuration index, and obtain the personalized load configuration parameters;

[0123] First, combine the myoelectric signal change rate, extract the root mean square value (RMS) and amplitude fluctuation range of the myoelectric signal within a fixed time window, calculate the myoelectric signal change rate, and extract the change trend on the time axis. Subsequently, combine the neuromuscular matching evaluation parameters, extract the myoelectric signal trigger time of different muscle groups, calculate the time difference between the main muscle group and the auxiliary muscle group, and analyze the consistency of neural responses. Further, combine the dynamic training load adjustment parameters, extract the training load change value per unit time, calculate the motion state matching degree at different load levels, and calculate the change trend of each parameter during training by setting a time series window. At the same time, set the personalized load configuration index. First, calculate the mean value of the training load in different time periods, then analyze the standard deviation of the training load change, and set the personalized load configuration parameter threshold. This threshold can be set through historical training data. For example, when the training load fluctuation range does not exceed 20% of the mean value, it is in a stable state. Finally, obtain the personalized load configuration parameters.

[0124] S502: Based on the personalized load configuration parameters, extract the variation range of the training action amplitude, calculate the motion trajectory deviation index per unit time, compare with the stage training plan, extract the training action amplitude that meets the load matching requirements, identify the active-passive training conversion strategy, analyze the load switching point between active and passive training, and screen the time interval that meets the conversion threshold to obtain the active-passive conversion and training amplitude parameters;

[0125] First, extract the variation range of the training action amplitude, set the unit time window, extract the maximum and minimum joint angles within each window, and calculate their variation amplitude. At the same time, calculate the motion trajectory deviation index per unit time, set the angular deviation between the target trajectory and the actual trajectory, and calculate the root mean square error (RMSE) within each time window. Subsequently, compare with the stage training plan, extract the training action amplitude that meets the load matching requirements, count the target amplitude range of different stage trainings, and set the deviation threshold of the training action. If the deviation is less than 5%, the matching is successful. At the same time, identify the active-passive training conversion strategy, extract the electromyogram signal triggering situation in different training stages, analyze the time point of the conversion from active training to passive training, calculate the load switching point between active and passive training, that is, count the time point when the peak value of the electromyogram signal drops by more than 50% during the training process, and screen the time interval that meets the conversion threshold. This threshold can be set to convert when the training load drops by more than 15% of the mean value, and finally obtain the active-passive conversion and training amplitude parameters.

[0126] S503: Call the active-passive conversion and training amplitude parameters, extract the motion situation of the rehabilitation robot, analyze the deviation interval of the robot motion trajectory on the time axis, and match the personalized training load configuration to generate a personalized rehabilitation plan;

[0127] First, extract the motion situation of the rehabilitation robot, record the robot motion trajectory data on the time axis, and calculate the trajectory variation range within different time windows. At the same time, analyze the deviation interval of the robot motion trajectory on the time axis, calculate the trajectory deviation per unit time, and set the deviation threshold. For example, set that the trajectory deviation exceeding 10% of the target trajectory is abnormal data. Subsequently, match the personalized training load configuration, extract the personalized load configuration parameters, calculate the matching degree between the robot motion trajectory and the training load, count the motion load variation within the unit time window, and calculate its correlation coefficient, and finally generate a personalized rehabilitation plan.

[0128] Please refer to Figure 7 , the personalized rehabilitation training system based on state monitoring, including:

[0129] The electromyogram and load monitoring module obtains the electromyogram amplitude, root mean square, and spectral median, monitors the muscle fiber recruitment, blood oxygen, lactic acid metabolism, and heart rate variability, extracts the joint angle change rate and moment of inertia, and obtains the training load reference value;

[0130] The neural response analysis module calls the training load reference value, calculates the time difference between the myoelectric trigger and the joint angle change, monitors the neuromuscular responses of the primary and auxiliary muscle groups, analyzes the mobilization rate, control torque, and delay, and obtains the matching evaluation parameters;

[0131] The dynamic load adjustment module calls the neuromuscular matching evaluation parameters, calculates the change rate ratio of the myoelectric signals in the short and long windows, identifies the muscle fiber mobilization rate, extracts the fatigue accumulation index, analyzes the fatigue inflection point threshold, calculates the stability in combination with the joint angle change rate, screens the training load correction parameters, monitors the motion state of the ICUAW rehabilitation robot, extracts the myoelectric control parameters, analyzes the active and passive motion matching, and generates the dynamic training load adjustment parameters;

[0132] The rhythm optimization control module calls the dynamic training load adjustment parameters, calculates the heart rate recovery rate, analyzes the glycogen metabolism trend, screens the training rhythm optimization threshold, calls the energy metabolism load index, and generates the training rhythm optimization parameters;

[0133] The personalized rehabilitation planning module calls the training rhythm optimization parameters, combines the myoelectric change rate, matching evaluation, and load adjustment, calculates the personalized training load, adjusts the movement amplitude, matches the training plan, and generates the personalized rehabilitation plan.

[0134] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claimed rights.

Claims

1. A personalized rehabilitation training method based on state monitoring, characterized in that: include: S1: Obtain the EMG amplitude, RMS, and spectrum median, calculate the rate of change, monitor muscle fiber mobilization, blood oxygen, lactate metabolism, and heart rate variability, extract the rate of change of joint angles and moment of inertia, and obtain the training load baseline value; S2: calling the training load baseline value, calculating the time difference between myoelectric triggering and joint angle change, monitoring the neuromuscular response of the main and auxiliary muscle groups, analyzing the mobilization rate, control torque, delay, screening deviation data, and obtaining matching evaluation parameters; S3: calling the neuromuscular matching evaluation parameters, calculating the ratio of the change rate of the short- and long-window electromyographic signals, identifying the muscle fiber mobilization rate, extracting the fatigue accumulation index, analyzing the fatigue inflection point threshold, calculating the stability in combination with the joint angle change rate, screening the training load correction parameters, monitoring the motion state of the ICUAW rehabilitation robot, extracting the electromyographic control parameters, analyzing the matching of active and passive motions, adjusting the robot motion trajectory, and generating dynamic training load adjustment parameters; S4: calling the dynamic training load adjustment parameters, calculating the heart rate recovery rate, analyzing the glycogen metabolism trend, screening the training rhythm optimization threshold, calling the energy metabolism load index, judging the metabolic limit, and generating the training rhythm optimization parameters.

2. The personalized rehabilitation training method based on state monitoring according to claim 1, characterized in that: The training load benchmark value includes an electromyographic signal benchmark, a metabolic state benchmark, and a joint movement benchmark; the matching evaluation parameters include a neuromuscular response parameter, a motion control parameter, and a deviation evaluation parameter; the dynamic training load adjustment parameters include a muscle fiber mobilization parameter, a fatigue state parameter, a joint stability parameter, and a motion matching parameter; the training rhythm optimization parameters include a heart rate recovery parameter, a metabolic limit parameter, and an interval adjustment parameter.

3. The personalized rehabilitation training method based on state monitoring according to claim 1, characterized in that: The specific steps to obtain the EMG amplitude, RMS, and spectrum median, calculate the rate of change, monitor muscle fiber mobilization, blood oxygen, lactate metabolism, and heart rate variability, extract the rate of change of joint angles and moment of inertia, and obtain the training load baseline value are as follows: S101: Acquire the electromyographic signal of the target muscle group, extract the amplitude, root mean square value, and spectrum median, compare the amplitude changes at adjacent moments, calculate the change rate within the time window, and obtain the characteristic change rate of the electromyographic signal; S102: Based on the characteristic change rate of the electromyographic signal, the number of peak mobilization times per unit time is calculated, and the muscle fiber mobilization rate is obtained by combining the amplitude fluctuation range. At the same time, the blood oxygen saturation is monitored, the blood oxygen saturation decrease rate is calculated, and the blood oxygen fluctuation range is obtained. The mean of the decrease in blood oxygen level per unit time is calculated to generate the muscle fiber mobilization rate and the blood oxygen metabolism rate; S103: calling the muscle fiber mobilization rate and blood oxygen metabolism rate, calculating the lactate metabolism rate, monitoring the heart rate variability, extracting the joint angle change rate and moment of inertia, and obtaining a training load baseline value.

4. The personalized rehabilitation training method based on state monitoring according to claim 1, characterized in that: The specific steps of calling the training load baseline value, calculating the time difference between myoelectric triggering and joint angle change, monitoring the neuromuscular response of the main and auxiliary muscle groups, analyzing the mobilization rate, control torque, delay, screening deviation data, and obtaining matching evaluation parameters are as follows: S201: calling the training load reference value, extracting the triggering time of the electromyographic signal and the starting time of the joint angle change, calculating the time difference, and monitoring the neuromuscular response time difference between the main muscle group and the auxiliary muscle group to obtain the neuromuscular response time parameter; S202: Based on the neuromuscular response time parameters, the variation interval of the muscle fiber mobilization rate is calculated, and the muscle control torque is extracted, and the torque fluctuation amplitude per unit time is calculated. At the same time, a time series analysis is performed on the neuromuscular response delay, and the offset of the differentiated muscle group response is extracted. The data with deviations exceeding the set threshold on the time axis are screened, and outliers are removed to obtain muscle fiber mobilization and neural response analysis parameters; S203: calling the muscle fiber mobilization and neural response analysis parameters, extracting the matching degree of the muscle fiber mobilization rate, the muscle control torque, and the neural response delay, and calculating the consistency offset interval on the time axis to obtain the matching evaluation parameters.

5. The personalized rehabilitation training method based on state monitoring according to claim 1, characterized in that: The specific steps of calling the neuromuscular matching evaluation parameters, calculating the ratio of the change rate of short- and long-window electromyographic signals, identifying the muscle fiber mobilization rate, extracting the fatigue accumulation index, analyzing the fatigue inflection point threshold, calculating the stability in combination with the joint angle change rate, screening the training load correction parameters, monitoring the motion state of the ICUAW rehabilitation robot, extracting the electromyographic control parameters, analyzing the matching of active and passive motions, adjusting the robot motion trajectory, and generating the dynamic training load adjustment parameters are as follows: S301: calling the matching evaluation parameter, calculating the ratio of the electromyographic signal change rate in the short time window to the long time window, and judging whether it exceeds the training load adjustment threshold, identifying the muscle fiber mobilization rate, and obtaining the training load threshold determination parameter; S302: Based on the training load threshold determination parameter, extract the time series of the electromyographic signal change rate, calculate the cumulative fatigue index per unit time, analyze the electromyographic signal change trend, extract the change of the muscle fiber mobilization rate in the differentiated time window, analyze the fatigue inflection point threshold, call the joint angle change rate at the same time, identify the joint stability parameter, and screen the data that meets the training load adjustment requirements to obtain fatigue accumulation and joint stability parameters; S303: Call the fatigue accumulation and joint stability parameters, obtain the ICUAW rehabilitation robot motion state monitoring data, extract the electromyographic signal control parameters, calculate the matching index of passive motion and active motion, and adjust the robot motion trajectory to obtain the dynamic training load adjustment parameters.

6. The personalized rehabilitation training method based on state monitoring according to claim 5, characterized in that: The calculation formula of the matching index of passive movement and active movement is specifically: Among them, Ω represents the matching index between passive movement and active movement, κ represents the total number of discrete moments in the active movement time window, and γ i represents the instantaneous amplitude of the electromyographic signal at the i-th moment, λ i represents the external torque applied by the rehabilitation robot at the i-th moment, δ i represents the neuromuscular response delay at the i-th moment, ξ represents the total number of discrete moments in the passive movement time window, and ρ j represents the restoring force generated by muscle stretching during passive movement at the jth moment, τ j Represents the instantaneous angular acceleration of the rehabilitation robot performing motion at the jth moment.

7. The personalized rehabilitation training method based on state monitoring according to claim 1, characterized in that: The specific steps of calling the dynamic training load adjustment parameters, calculating the heart rate recovery rate, analyzing the glycogen metabolism trend, screening the training rhythm optimization threshold, calling the energy metabolism load index, judging the metabolic limit, and generating the training rhythm optimization parameters are as follows: S401: calling the dynamic training load adjustment parameter, extracting the heart rate signal, calculating the heart rate recovery rate per unit time, analyzing the glycogen metabolism ratio trend, extracting the change of the glycogen metabolism ratio on the time axis, and obtaining the heart rate recovery and metabolic trend parameters; S402: Based on the heart rate recovery and metabolic trend parameters, extract the time series of glycogen metabolism proportion under the differentiated training rhythm, calculate the metabolic fluctuation range index within a unit time, and screen the training rhythm optimization threshold, and at the same time call the instantaneous energy metabolic load index, compare the training rhythm with the metabolic limit, extract the load adjustment interval of the training rhythm, and identify the change of the load increment within a unit time, so as to obtain the training rhythm metabolic matching parameter; S403: calling the training rhythm metabolism matching parameter, calculating the adjustment interval index of the training interval time, and screening the training rhythm optimization threshold, extracting the rhythm adjustment demand parameter, and generating the training rhythm optimization parameter.

8. The personalized rehabilitation training method based on state monitoring according to claim 7, characterized in that: The specific calculation formula of the adjustment interval index of the training interval time is: Among them, Θ rto Represents the adjustment interval index of the training interval, N prds represents the number of training cycles considered, ξ pmk represents the value of the training rhythm metabolism matching parameter in the kth cycle, η adjk Represents the adjustment coefficient of the kth cycle, reflecting the sensitivity of rhythm changes to training effects.

9. The personalized rehabilitation training method based on state monitoring according to claim 1, characterized in that: The method also Including, S5: calling the training rhythm optimization parameters, combining the electromyographic change rate, matching evaluation, and load adjustment, calculating the personalized training load, adjusting the movement range, matching the training plan, identifying the training conversion strategy, and generating a personalized rehabilitation plan; The personalized rehabilitation training program includes training load parameters, training rhythm parameters, and movement pattern parameters; S501: calling the training rhythm optimization parameters, combining the electromyographic signal change rate, the neuromuscular matching evaluation parameters, and the dynamic training load adjustment parameters, extracting the change trend of the training indicators on the time axis, calculating the personalized training load configuration indicators, and obtaining the personalized load configuration parameters; S502: Based on the personalized load configuration parameters, extract the training movement amplitude variation range, calculate the motion trajectory deviation index per unit time, and compare the stage training plan to extract the training movement amplitude that meets the load matching requirements, and at the same time identify the active and passive training conversion strategy, analyze the load switching point between active and passive training, and select the time interval that meets the conversion threshold to obtain the active and passive conversion and training amplitude parameters; S503: calling the active-passive conversion and training amplitude parameters, extracting the movement of the rehabilitation robot, analyzing the offset interval of the robot's movement trajectory on the time axis, matching the personalized training load configuration, and generating a personalized rehabilitation plan.

10. A personalized rehabilitation training system based on state monitoring, characterized in that: According to any one of claims 19, the personalized rehabilitation training method based on state monitoring comprises: The electromyography and load monitoring module obtains the electromyography amplitude, root mean square, and spectrum median, monitors muscle fiber mobilization, blood oxygen, lactate metabolism, and heart rate variability, extracts the rate of change of joint angles and moment of inertia, and obtains the training load baseline value; The neural response analysis module calls the training load baseline value, calculates the time difference between myoelectric triggering and joint angle change, monitors the neuromuscular response of the main and auxiliary muscle groups, analyzes the mobilization rate, control torque, and delay, and obtains matching evaluation parameters; The dynamic load adjustment module calls the neuromuscular matching evaluation parameters, calculates the ratio of the change rate of the short- and long-window electromyographic signals, identifies the muscle fiber mobilization rate, extracts the fatigue accumulation index, analyzes the fatigue inflection point threshold, calculates the stability in combination with the joint angle change rate, screens the training load correction parameters, monitors the motion state of the ICUAW rehabilitation robot, extracts the electromyographic control parameters, analyzes the active and passive motion matching, and generates the dynamic training load adjustment parameters; The rhythm optimization control module calls the dynamic training load adjustment parameter, calculates the heart rate recovery rate, analyzes the glycogen metabolism trend, screens the training rhythm optimization threshold, calls the energy metabolism load index, and generates the training rhythm optimization parameter; The personalized rehabilitation planning module calls the training rhythm optimization parameters, combines the electromyography change rate, matching evaluation, and load adjustment, calculates the personalized training load, adjusts the movement range, matches the training plan, and generates a personalized rehabilitation plan.

Images