A rehabilitation nursing risk early warning system, a fatigue risk early warning method, equipment and a medium

Abstract

The application provides a rehabilitation nursing risk early warning system, a fatigue risk early warning method, equipment and a medium. The three-axis acceleration signals and the surface electromyography signals of a target muscle of a patient during rehabilitation nursing are synchronously collected. The surface electromyography signals are filtered according to the frequency spectrum characteristics of the three-axis acceleration signals to suppress artifacts generated by limb macro-movement, and a pre-processed electromyography signal is obtained. The median frequency sequence and the approximate entropy sequence of the amplitude envelope of the pre-processed electromyography signal are determined in a continuous sliding time window. A risk index of a muscle fatigue state is constructed according to the cooperative change relationship between the median frequency sequence and the approximate entropy sequence. The risk index is combined with a pre-set fatigue risk threshold to generate a fatigue risk early warning report of the patient during rehabilitation nursing. The scheme of the application can distinguish between normal physiological fatigue and abnormal neuromuscular decompensation.

Description

Technical Field

[0001] This application relates to the field of rehabilitation nursing technology, and more specifically, to a risk warning system, fatigue risk warning method, equipment and medium for rehabilitation nursing. Background Technology

[0002] Rehabilitation nursing, through an engineered process involving standardized assessment, modular intervention, data-driven regulation, and closed-loop management, transforms a patient's functional impairment into a quantifiable, traceable, and optimizable functional output system, thereby achieving a paradigm shift from experience-based decision-making to precise regulation. In this process, effectively managing training-related risks such as muscle overexertion, secondary injury, and cardiovascular overload is a core challenge in ensuring rehabilitation safety and effectiveness. Among these, accurate early warning of muscle fatigue risk is particularly crucial.

[0003] Traditional fatigue monitoring methods often rely on isolated judgments or linear combinations of single or a few features such as electromyography (EMG) spectra and amplitudes. This makes it difficult to analyze the complex coupling relationships and dynamic patterns of multidimensional physiological signals in the evolution of fatigue. For example, traditional methods often simply attribute single changes such as decreased muscle activation and leftward shift of the spectrum to routine physical exertion, but fail to identify the neuromuscular control strategy decompensation or potential damage risk that these changes indicate when they are accompanied by abnormal patterns such as disordered muscle coordination timing and decoupling of EMG-mechanical output efficiency. Therefore, how to distinguish between normal physiological fatigue processes and abnormal neuromuscular decompensation states has become a challenge for the industry. Summary of the Invention

[0004] This application provides a risk warning system, fatigue risk warning method, equipment and medium for rehabilitation nursing, which can distinguish between normal physiological fatigue process and abnormal neuromuscular decompensation state.

[0005] Firstly, this application provides a fatigue risk early warning method for providing early warning of fatigue risk to users during rehabilitation care. The method includes the following steps:

[0006] Simultaneously collect triaxial acceleration signals and surface electromyography signals of the target muscles during rehabilitation care;

[0007] The surface electromyography signal is filtered based on the spectral characteristics of the triaxial acceleration signal to suppress artifacts generated by macroscopic limb movements, resulting in a preprocessed electromyography signal.

[0008] Within a continuously sliding time window, the median frequency sequence and the approximate entropy sequence of the amplitude envelope of the preprocessed electromyographic signal are determined;

[0009] A risk index for muscle fatigue state is constructed based on the synergistic relationship between the median frequency sequence and the approximate entropy sequence.

[0010] The risk index is combined with a preset fatigue risk threshold to generate a fatigue risk warning report for patients during rehabilitation care.

[0011] In some embodiments, the surface electromyography (EMG) signal is filtered based on the spectral characteristics of the triaxial acceleration signal to suppress artifacts generated by macroscopic limb movements, resulting in a preprocessed EMG signal. Specifically, this preprocessing includes:

[0012] Extract reference noise features characterizing the macroscopic movement patterns of the limbs from the triaxial acceleration signals;

[0013] An adaptive filter for filtering motion artifacts is constructed based on the reference noise characteristics;

[0014] The surface electromyography (EMG) signal is processed by the adaptive filter to suppress artifacts generated by macroscopic limb movements, resulting in a preprocessed EMG signal.

[0015] In some embodiments, determining the median frequency sequence and the approximate entropy sequence of the amplitude envelope of the preprocessed electromyographic signal within a continuously sliding time window specifically includes:

[0016] Preset the time window for continuous sliding;

[0017] The preprocessed electromyographic signal is divided into a sliding window using the time window to obtain continuous signal segments.

[0018] The median frequency of each signal segment is determined, thereby obtaining the median frequency sequence of the preprocessed electromyographic signal;

[0019] The amplitude envelope of each signal segment is extracted and the approximate entropy of the corresponding amplitude envelope is calculated, thereby obtaining the approximate entropy sequence of the amplitude envelope.

[0020] In some embodiments, constructing a risk index for muscle fatigue state based on the synergistic relationship between the median frequency sequence and the approximate entropy sequence specifically includes:

[0021] Set the time window length for collaborative change analysis, and then extract the data segment within the most recent time window from the median frequency sequence and the approximate entropy sequence;

[0022] A quantitative index for determining the cooperative change relationship between the median frequency sequence and the approximate entropy sequence based on the extracted data segments;

[0023] The quantitative indicators are mapped to a risk index for muscle fatigue.

[0024] In some embodiments, generating a fatigue risk warning report for patients during rehabilitation care by combining the risk index with a preset fatigue risk threshold specifically includes:

[0025] Preset a fatigue risk threshold for patients during rehabilitation care;

[0026] The patient's fatigue risk level is determined by combining the risk index with a preset fatigue risk threshold.

[0027] A structured report containing early warning information is generated based on the fatigue risk level as the fatigue risk early warning report.

[0028] In some embodiments, the target muscle refers to the primary muscle group selected as the subject of fatigue monitoring based on the rehabilitation nursing movement pattern.

[0029] In some embodiments, the surface electromyography signal refers to the electrophysiological signal generated by the activity of the target muscle motor unit.

[0030] Secondly, this application provides a risk warning system for rehabilitation nursing, which includes a fatigue risk warning unit, the fatigue risk warning unit comprising:

[0031] Thirdly, this application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the fatigue risk warning method described above.

[0032] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the fatigue risk warning method described above.

[0033] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0034] The rehabilitation nursing risk warning system, fatigue risk warning method, equipment, and media provided in this application firstly involve synchronously acquiring triaxial acceleration and surface electromyography (EMG) signals, and filtering the original EMG signals using the spectral characteristics of the acceleration signals. The aim is to suppress artifact interference introduced by macroscopic limb movement at the source, providing a high-quality electrophysiological data foundation reflecting pure neuromuscular activity for subsequent precise feature extraction. This step is a physical prerequisite for ensuring the signal-to-noise ratio of the extracted features is physiologically relevant. Subsequently, within consecutive sliding time windows, the median frequency sequence and the approximate entropy sequence of the amplitude envelope of the EMG signals are determined in parallel. The amplitude fluctuation pattern of the surface EMG signals is transformed from a time-domain waveform into a quantifiable nonlinear complexity index. This process calculates the approximate entropy of the amplitude envelope within each sliding time window to obtain a time-evolving complexity change sequence, thereby transforming the abstract degree of pattern disorder into a concrete and traceable mathematical sequence. This provides a key dimension for characterizing the stability and predictability of neuromuscular control strategies in this scheme: in normal life... In rational fatigue, due to the synchronization of motor unit recruitment, the amplitude envelope often exhibits a regular increase and a decrease in complexity. However, in abnormal decompensation states (such as due to pain suppression or compensatory muscle interference), control commands become chaotic, which may lead to unexpected fluctuations or abnormal increases in the complexity of the amplitude envelope. Therefore, a risk index is constructed based on the synergistic relationship between the median frequency sequence and the approximate entropy sequence. The core of this approach is to design a mathematical model (such as a fusion function based on correlation coefficient, phase synchronization, or trend divergence measurement) to quantify the dynamic correlation pattern of these two key features on the time axis. This process can mathematically distinguish and quantify two fundamentally different states—a decrease in median frequency accompanied by a synchronous decrease in approximate entropy (typical physiological fatigue) and a decrease in median frequency but abnormal fluctuations or increases in approximate entropy (potential decompensation)—through a comprehensive index. Finally, by comparing the calculated dynamic risk index with a preset threshold, an early warning report is generated. In summary, this scheme can distinguish between normal physiological fatigue processes and abnormal neuromuscular decompensation states. Attached Figure Description

[0035] Figure 1 This is a flowchart illustrating a fatigue risk warning method according to some embodiments of this application;

[0036] Figure 2 This is a schematic flowchart illustrating the determination of preprocessed electromyographic signals according to some embodiments of this application;

[0037] Figure 3 This is a schematic flowchart illustrating the determination of median frequency according to some embodiments of this application;

[0038] Figure 4 This is a schematic diagram of the structure of a fatigue risk warning unit according to some embodiments of this application;

[0039] Figure 5 This is an internal structural diagram of a computer device for implementing a fatigue risk warning method according to some embodiments of this application. Detailed Implementation

[0040] To better understand the technical solutions in this embodiment, the technical solutions in this embodiment will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0041] refer to Figure 1 The figure is a flowchart illustrating a fatigue risk warning method according to some embodiments of this application. The fatigue risk warning method mainly includes the following steps:

[0042] In step 101, triaxial acceleration signals and surface electromyography signals of the target muscles of the patient are collected simultaneously during rehabilitation care.

[0043] In practice, an integrated biosignal sensor combining a triaxial accelerometer and surface electromyography (EMG) electrodes is used for data acquisition. First, based on the major muscle groups involved in the rehabilitation movements, the muscle belly of the target muscle is determined according to clinical EMG measurement standards, and the skin is cleaned. Then, the EMG electrodes of the sensor are placed along the muscle fiber direction at the target muscle belly, and the reference electrode is placed at a nearby bony landmark. Simultaneously, the triaxial accelerometer module of the sensor is fixed to the limb, with its three sensitive axes roughly aligned anatomically with the anterior-posterior, medial-lateral, and superior-inferior directions. During data acquisition, a unified clock within the same hardware device controls the recording at a synchronized sampling frequency. The sampling frequency of the EMG signal must be no less than 1000 Hz, and the sampling frequency of the triaxial accelerometer signal is typically no less than 100 Hz. The signals are transmitted to the signal processing unit via wired or wireless means, forming a dual-channel raw data stream with strictly aligned timestamps.

[0044] It should be noted that the target muscle mentioned in this application refers to the main muscle group selected as the subject of fatigue monitoring based on the rehabilitation nursing movement pattern; the triaxial acceleration signal refers to the signal reflecting the change of limb movement acceleration over time in three-dimensional space, with its physical dimension being gravitational acceleration, mainly used to quantify macroscopic limb movement; the surface electromyography signal refers to the electrophysiological signal generated by the activity of the target muscle motor unit, with its amplitude unit being millivolts, and is the core raw signal for assessing muscle fatigue; synchronous acquisition refers to the simultaneous recording of electromyography signal and acceleration signal under the same time reference, the key technology of which is that the two signals share the same sampling clock to ensure time alignment, which is a prerequisite for subsequent motion artifact filtering; the integrated biosignal sensor is a special device that encapsulates the electromyography acquisition front end and the miniature triaxial accelerometer in the same medical-grade patch.

[0045] In step 102, the surface electromyography signal is filtered according to the spectral characteristics of the triaxial acceleration signal to suppress artifacts generated by macroscopic limb movements, thereby obtaining a preprocessed electromyography signal.

[0046] In some embodiments, reference Figure 2 As shown in the figure, this is a flowchart illustrating the process of determining the preprocessed electromyographic (EMG) signal according to some embodiments of this application. The surface EMG signal is filtered based on the spectral characteristics of the triaxial acceleration signal to suppress artifacts generated by macroscopic limb movements. The preprocessed EMG signal can be obtained through the following steps:

[0047] First, in step 1021, reference noise features characterizing the macroscopic movement pattern of the limbs are extracted from the triaxial acceleration signal;

[0048] Then, in step 1022, an adaptive filter for filtering out motion artifacts is constructed based on the reference noise characteristics;

[0049] Finally, in step 1023, the surface electromyography signal is processed by the adaptive filter to suppress artifacts generated by macroscopic limb movements, resulting in a preprocessed electromyography signal.

[0050] In specific implementation, the reference noise features characterizing the macroscopic movement pattern of the limbs can be extracted from the triaxial acceleration signal in the following way: for example, the time-stamp aligned triaxial acceleration signal is first preprocessed to remove the DC bias and smoothed; then the first principal component time sequence obtained by principal component analysis of the smoothed triaxial acceleration signal is used as the reference noise feature. In other embodiments, other methods can also be used, which are not limited here.

[0051] It should be noted that the reference noise feature mentioned in this application refers to the signal feature extracted from the triaxial acceleration signal to characterize the macroscopic movement pattern of the limb. Its function is to serve as the reference input of the adaptive filter to guide the filtering out artifacts generated by movement in the surface electromyography signal.

[0052] In specific implementation, the adaptive filter for filtering motion artifacts based on the reference noise features can be constructed in the following ways, for example: First, the filter structure is determined to be a finite-length unit impulse response transverse filter; the reference noise features are used as the reference input sequence of the filter; the synchronously acquired raw surface electromyography signal is used as the desired signal; the normalized least mean square algorithm is used as the core adaptive algorithm of the filter, which iteratively adjusts the filter weight vector to minimize the mean square error between the filter output and the desired signal; during the construction process, an adaptive step size parameter to ensure convergence is set for the normalized least mean square algorithm, the value of which is normalized according to the power of the reference noise features or set to an empirically fixed small value; at the same time, an order related to the possible duration of motion artifacts is set for the transverse filter, which is preset according to the sampling frequency and the typical time delay of the artifacts, and this application does not limit this.

[0053] It should be noted that the adaptive filter described in this application refers to a signal processing component that can dynamically adjust its parameters according to the reference noise characteristics. Its function is to estimate and subtract motion artifact components from the surface electromyography signal to obtain the artifact-suppressed electromyography signal.

[0054] In specific implementation, the surface electromyography (EMG) signal is processed by the adaptive filter to suppress artifacts generated by macroscopic limb movements. The preprocessed EMG signal can be obtained in the following way: at each processing moment, the real-time acquired reference noise feature sample is input to the reference input channel of the adaptive filter; the original surface EMG signal sample at the same moment is input to the desired input channel of the filter; the adaptive filter performs linear convolution on the input reference noise sample according to its currently updated weight vector to generate an estimate of the motion artifact component in the current EMG signal; then, the artifact estimate is subtracted from the input original EMG signal sample in real time, and the difference is output as the error signal at the current moment. This error signal is specifically used as the preprocessed EMG signal sample after filtering out motion artifacts; at the same time, the error signal and the reference input signal work together to apply the update rule of the normalized least mean square algorithm to calculate and update the filter weight vector at the next moment, thereby realizing online adaptive adjustment of the filter coefficients to continuously track changes in limb movement patterns; finally, the continuously output error signal sequence constitutes the preprocessed EMG signal.

[0055] It should be noted that the preprocessed electromyographic signal mentioned in this application refers to the surface electromyographic signal after being processed by an adaptive filter, in which artifacts generated by macroscopic limb movement are effectively suppressed. Its role is to retain the core components reflecting muscle electrophysiological activity and provide a clean signal basis for subsequent fatigue feature extraction.

[0056] In step 103, the median frequency sequence and the approximate entropy sequence of the amplitude envelope of the preprocessed electromyographic signal are determined within a continuously sliding time window.

[0057] In some embodiments, determining the median frequency sequence and the approximate entropy sequence of the amplitude envelope of the preprocessed electromyographic signal within a continuously sliding time window can be achieved by the following steps:

[0058] Preset the time window for continuous sliding;

[0059] The preprocessed electromyographic signal is divided into a sliding window using the time window to obtain continuous signal segments.

[0060] The median frequency of each signal segment is determined, thereby obtaining the median frequency sequence of the preprocessed electromyographic signal;

[0061] The amplitude envelope of each signal segment is extracted and the approximate entropy of the corresponding amplitude envelope is calculated, thereby obtaining the approximate entropy sequence of the amplitude envelope.

[0062] In practical implementation, a time window length and a sliding step size can be preset based on the non-stationary characteristics of electromyographic signals and the real-time requirements of fatigue monitoring. The window length needs to be long enough to include several muscle activation cycles to stably estimate the signal spectrum, while also being short enough to ensure timely response to changes in fatigue state. Its specific value can be set within a reasonable range based on the main frequency components of the preprocessed electromyographic signal and the actual application scenario. The sliding step size determines the temporal resolution of the feature sequence update, and its value is less than or equal to the window length. It can be preset after balancing computational resources and real-time monitoring requirements. This application does not limit the specific values ​​of the window length and the sliding step size.

[0063] In specific implementation, the preprocessed electromyography (EMG) signal is divided into continuous signal segments by a sliding window through the time window. This can be achieved in the following way: starting from the starting point of the preprocessed EMG signal, signal segments with a length equal to the window length are sequentially extracted at intervals of the sliding step size. Each extraction operation extracts the data points of the corresponding time segment in the preprocessed EMG signal to form a signal segment. Since the sliding step size is less than the window length, there is partial overlap in time between adjacent signal segments. This extraction process is repeated until the signal ends, thereby obtaining a series of temporally continuous and partially overlapping signal segments. In other embodiments, other methods can also be used to determine this, which are not limited here.

[0064] It should be noted that the signal segment mentioned in this application refers to a block of signal data within a continuous time period extracted from the preprocessed electromyographic signal through a sliding time window.

[0065] For specific implementation, refer to Figure 3 As shown in the figure, this is a schematic flowchart illustrating the process of determining the median frequency in some embodiments of this application. Determining the median frequency of each signal segment, and thus obtaining the median frequency sequence of the preprocessed electromyographic signal, can be achieved in the following manner: For each segment, a short-time Fourier transform is first performed on each segment to estimate its power spectral density; then, in the frequency domain, the power spectral density is integrated from low to high frequencies to calculate the cumulative power; the frequency point corresponding to when the cumulative power reaches 50% of the total power of the corresponding signal segment is taken as the median frequency of the corresponding signal segment; the above short-time Fourier transform and median frequency lookup process is repeated for the continuous signal segments in chronological order, and all calculated median frequency values ​​are arranged according to the center time order of their corresponding signal segments, thus forming the median frequency sequence of the preprocessed electromyographic signal.

[0066] It should be noted that the median frequency sequence mentioned in this application refers to time-series data formed by arranging the median frequencies of various signal segments in chronological order. Its function is to characterize the trend of the centroid of the power spectrum of electromyographic signals changing over time, so as to reflect the spectral migration phenomenon during muscle fatigue.

[0067] In specific implementation, extracting the amplitude envelope of each signal segment and calculating the approximate entropy of the corresponding amplitude envelope to obtain the approximate entropy sequence of the amplitude envelope can be achieved in the following way: For each signal segment, firstly, perform a Hilbert transform on the signal segment and calculate the modulus of its analytic signal, using the modulus sequence as the amplitude envelope of the current signal segment; then, calculate the approximate entropy of the amplitude envelope sequence, the process of which includes: preset the mode dimension and tolerance parameter, wherein the tolerance parameter is usually set as a scaling factor of the standard deviation of the amplitude envelope sequence; then, statistically analyze the conditional probability of matching between a vector with a length equal to the mode dimension and a vector with a length equal to the mode dimension plus one in the amplitude envelope sequence; finally, obtain the approximate entropy value of the current amplitude envelope by taking the logarithm of the conditional probability and performing the operation; repeat the above amplitude envelope extraction and approximate entropy calculation process for each signal segment, and arrange all the obtained approximate entropy values ​​according to the center time order of their corresponding signal segments, thus forming the approximate entropy sequence of the amplitude envelope.

[0068] It should be noted that the approximate entropy sequence of amplitude envelopes mentioned in this application refers to time-series data formed by arranging the approximate entropy of the amplitude envelopes of each signal segment in chronological order. Its function is to characterize the change in the complexity of the electromyographic signal amplitude envelope over time, so as to reflect the evolution of neural control strategies during muscle fatigue.

[0069] In step 104, a risk index for muscle fatigue state is constructed based on the synergistic change relationship between the median frequency sequence and the approximate entropy sequence.

[0070] In some embodiments, constructing a risk index for muscle fatigue state based on the synergistic relationship between the median frequency sequence and the approximate entropy sequence can be achieved through the following steps:

[0071] Set the time window length for collaborative change analysis, and then extract the data segment within the most recent time window from the median frequency sequence and the approximate entropy sequence;

[0072] A quantitative index for determining the cooperative change relationship between the median frequency sequence and the approximate entropy sequence based on the extracted data segments;

[0073] The quantitative indicators are mapped to a risk index for muscle fatigue.

[0074] In specific implementation, setting the time window length for collaborative change analysis and then extracting the most recent data segment from the median frequency sequence and the approximate entropy sequence can be achieved in the following way: First, based on the swiftness of changes in muscle fatigue state and the real-time requirements of risk warning, a collaborative change analysis time window length is preset; this time window length needs to be sufficient to cover the duration required for observable collaborative changes between the median frequency and the approximate entropy, usually on the order of tens of seconds, and its specific value can be set within a reasonable range based on empirical data from actual rehabilitation and nursing scenarios; then, based on the current moment, the latest data segment with a time length equal to the collaborative change analysis time window length is extracted from the median frequency sequence and the approximate entropy sequence, respectively, thereby obtaining two time-aligned data segments for collaborative change analysis, which are denoted as the current median frequency analysis segment and the current approximate entropy analysis segment, i.e., the data segment within the most recent time window.

[0075] It should be noted that the data segment within the most recent time window mentioned in this application refers to a subset of data extracted from the median frequency sequence and the approximate entropy sequence within the most recent time period. Its purpose is to provide input data for analyzing the cooperative relationship between the two feature sequences within the most recent time window, so as to support real-time risk assessment.

[0076] In specific implementation, the quantitative index for determining the cooperative change relationship between the median frequency sequence and the approximate entropy sequence based on the extracted data segment can be implemented in the following ways. For example, one implementation is to calculate the Pearson correlation coefficient between the current median frequency analysis segment and the current approximate entropy analysis segment, and use the absolute value of the correlation coefficient or the value obtained after specific symbol processing as a quantitative index characterizing the degree of cooperative change between the two. The process of calculating the Pearson correlation coefficient includes: firstly, calculating the mean and standard deviation of each of the two analysis segments respectively; then calculating the covariance of the corresponding data points of the two analysis segments; and finally dividing the covariance by the product of the two standard deviations to obtain the correlation coefficient. Another implementation is to perform linear trend fitting on the current median frequency analysis segment and the current approximate entropy analysis segment respectively to obtain the slope of their changes, and then using the ratio of the two slopes or the normalized difference value as the quantitative index. This application provides specific mathematical methods for calculating the quantitative index. The model is not limited, but its core lies in its ability to numerically represent the consistency of the changing trends of two sequences within a recent time window. As a preferred embodiment, the current median frequency analysis segment and the current approximate entropy analysis segment can be combined into a two-dimensional feature sequence, where each data point consists of the median frequency value and the approximate entropy value at the same time. Then, the actual joint probability distribution of this two-dimensional sequence in the feature space and a theoretical joint probability distribution assuming that the two are completely independent are calculated respectively. Subsequently, the statistical distance between these two probability distributions, such as the Jensen-Shannon divergence or Wasserstein distance, is calculated, and the specific value of this statistical distance is used as a quantitative indicator to characterize the strength of the statistical dependence between the two. The larger the value of this indicator, the more the actual distribution of the two sequences deviates from the independence assumption, that is, the stronger their cooperative change relationship. This method, starting from the essence of statistics, measures the fatigue correlation information implied by the joint feature, which goes beyond the independent information of each sequence.

[0077] It should be noted that the quantitative index of the cooperative change relationship described in this application refers to a statistical quantity or measure used to numerically characterize the degree of cooperative change between the median frequency sequence and the approximate entropy sequence within a recent time window. Its function is to quantify the correlation strength or detuning state of the two features in the fatigue process.

[0078] In specific implementation, mapping the quantitative index to a risk index of muscle fatigue can be achieved in the following ways: First, define a mapping function from the value range of the quantitative index to the value range of a standard risk index (e.g., between 0 and 100). This mapping function can be a linear function, a piecewise linear function, or a nonlinear function, aiming to convert the degree of synergy or disharmony reflected by the quantitative index into a single-dimensional scalar where a larger value represents a higher risk of muscle fatigue. For example, when using the correlation coefficient as the quantitative index, the mapping function can be designed such that the closer the correlation coefficient is to -1 or +1 (indicating high synergy or high negative synergy), the lower the risk index is mapped, while the closer the correlation coefficient is to 0 (indicating disharmony), the higher the risk index is mapped. Finally, input the calculated quantitative index into the mapping function, and its output value is determined as the risk index of the muscle fatigue state at the current moment. In other embodiments, other methods can also be used, which are not limited here.

[0079] It should be noted that the risk index of muscle fatigue state mentioned in this application refers to a standardized value obtained by mapping the synergistic change relationship. Its function is to comprehensively characterize the current risk level of muscle fatigue. The larger the value, the higher the fatigue risk.

[0080] In step 105, a fatigue risk warning report for patients during rehabilitation care is generated by combining the risk index with a preset fatigue risk threshold.

[0081] In some embodiments, generating a fatigue risk warning report for patients during rehabilitation care by combining the risk index with a preset fatigue risk threshold can be achieved through the following steps:

[0082] Preset a fatigue risk threshold for patients during rehabilitation care;

[0083] The patient's fatigue risk level is determined by combining the risk index with a preset fatigue risk threshold.

[0084] A structured report containing early warning information is generated based on the fatigue risk level as the fatigue risk early warning report.

[0085] In specific implementation, based on the clinical needs of muscle fatigue risk warning and the numerical range characteristics of the risk index, one or more ordered values ​​obtained in advance through historical data analysis or clinical experience calibration can be used as the fatigue risk threshold. The process of determining the threshold through historical data analysis includes: collecting risk index data calculated from a large number of rehabilitation nursing cases and their corresponding clinical fatigue assessment results; and using statistical methods, such as percentile or receiver operating characteristic curve analysis, to determine one or more numerical points that best distinguish different fatigue states. Clinical experience calibration can be set by rehabilitation therapists based on the general performance of the patient group or the specific characteristics of individual patients. For example, a lower threshold with higher sensitivity can be set for early-stage rehabilitation patients, while a higher threshold with higher tolerance can be set for later-stage rehabilitation patients. This application does not limit the specific value of the fatigue risk threshold; its core purpose is to provide an objective and adjustable benchmark for subsequent risk level determination.

[0086] In specific implementation, determining a patient's fatigue risk level by combining the risk index with preset fatigue risk thresholds can be achieved in the following way: for example, comparing the risk index with a plurality of preset ordered fatigue risk thresholds in sequence; and using the discrete risk label corresponding to the comparison result as the patient's fatigue risk level; the specific process is as follows: comparing the risk index with preset low-risk thresholds, medium-risk thresholds, and high-risk thresholds, etc. If the risk index is lower than the low-risk threshold, it is determined to be a low-risk level; if the risk index is between the low-risk threshold and the medium-risk threshold, it is determined to be a medium-risk level; if the risk index exceeds the high-risk threshold, it is determined to be a high-risk level; to enhance the stability of the determination, a hysteresis comparison logic can be used, that is, when the risk index rises above a certain threshold, the level is immediately upgraded, and the level is only downgraded when it falls below the threshold minus a preset hysteresis amount.

[0087] It should be noted that the fatigue risk level mentioned in this application refers to a discrete risk classification label determined by comparing the risk index of muscle fatigue state with a preset threshold.

[0088] In specific implementation, generating a structured report containing early warning information based on the fatigue risk level as the fatigue risk early warning report can be achieved in the following way: First, construct a report template containing fields such as timestamp, target muscle information, risk index, fatigue risk level, and corresponding early warning measures; then, fill the current time, the target muscle information, the risk index, and the fatigue risk level into the template; next, according to the fatigue risk level, retrieve and fill the corresponding early warning prompt text and nursing suggestion text from a preset early warning information mapping table; finally, output the completed template data as the fatigue risk early warning report; wherein, the early warning information mapping table is a predefined data structure that associates different fatigue risk levels with specific early warning prompts, such as "normal," "attention," "warning," and nursing suggestions, such as "continue training," "rest recommended," and "stop immediately"; the output form of the report can be a screen visualization interface, a voice broadcast signal, or a standardized format electronic document, which is not limited in this application.

[0089] It should be noted that the fatigue risk warning report mentioned in this application refers to a structured output document containing timestamps, target muscle information, risk index, fatigue risk level and corresponding warning measures. Its purpose is to provide rehabilitation nursing staff with clear fatigue status assessment results and actionable nursing suggestions.

[0090] In another aspect, in some embodiments, this application provides a risk warning system for rehabilitation care, including a fatigue risk warning unit, as referenced. Figure 4 The figure is a schematic diagram of the structure of a fatigue risk warning unit according to some embodiments of this application. The fatigue risk warning unit 200 includes: a data acquisition module 201, a processing module 202, and an execution module 203, which are described below:

[0091] The acquisition module 201 in this application is mainly used to synchronously acquire the triaxial acceleration signal and surface electromyography signal of the target muscle of the patient during rehabilitation care;

[0092] Processing module 202, in this application, is mainly used to filter the surface electromyography signal according to the spectral characteristics of the triaxial acceleration signal in order to suppress artifacts generated by macroscopic limb movement and obtain a preprocessed electromyography signal.

[0093] In addition, the processing module 202 in this application is also used to determine the median frequency sequence and the approximate entropy sequence of the amplitude envelope of the preprocessed electromyographic signal within a continuously sliding time window;

[0094] In addition, the processing module 202 in this application is also used to construct a risk index of muscle fatigue state based on the synergistic change relationship between the median frequency sequence and the approximate entropy sequence;

[0095] The execution module 203 in this application is mainly used to generate a fatigue risk warning report for patients during rehabilitation care by combining the risk index with a preset fatigue risk threshold.

[0096] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the fatigue risk warning method described above.

[0097] In some embodiments, reference Figure 5 This figure is an internal structural diagram of a computer device implementing a fatigue risk warning method according to some embodiments of this application. The fatigue risk warning method in the above embodiments can be implemented through... Figure 5 The computer device shown is used to implement this, and the computer device 300 includes at least one processor 301, a communication bus 302, a memory 303, and at least one communication interface 304.

[0098] The processor 301 may be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more devices used to control the execution of the fatigue risk warning method in this application.

[0099] The communication bus 302 is used to transmit information between the aforementioned components.

[0100] Memory 303 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory 303 may exist independently and be connected to processor 301 via communication bus 302. Memory 303 may also be integrated with processor 301.

[0101] The memory 303 stores program code for executing the solution of this application, and its execution is controlled by the processor 301. The processor 301 executes the program code stored in the memory 303. The program code may include one or more software modules. In the above embodiments, the fatigue risk warning method can be implemented by the processor 301 and one or more software modules in the program code in the memory 303.

[0102] Communication interface 304 uses any transceiver-like device to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0103] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0104] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device may be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0105] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described fatigue risk warning method.

[0106] In summary, the rehabilitation nursing risk warning system, fatigue risk warning method, equipment, and medium disclosed in this application firstly, synchronously collect the triaxial acceleration signal and surface electromyography (EMG) signal of the target muscle during rehabilitation nursing; filter the surface EMG signal according to the spectral characteristics of the triaxial acceleration signal to suppress artifacts generated by macroscopic limb movement, obtaining a pre-processed EMG signal; determine the median frequency sequence and the approximate entropy sequence of the amplitude envelope of the pre-processed EMG signal within a continuously sliding time window; construct a risk index of muscle fatigue state based on the synergistic change relationship between the median frequency sequence and the approximate entropy sequence; and generate a fatigue risk warning report for the patient during rehabilitation nursing by combining the risk index with a preset fatigue risk threshold; thus, it can distinguish between normal physiological fatigue processes and abnormal neuromuscular decompensation states.

[0107] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0108] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A fatigue risk warning method for warning a fatigue risk when a user is given rehabilitation care, characterized by, The method includes the following steps: Simultaneously collect triaxial acceleration signals and surface electromyography signals of the target muscles during rehabilitation care; The surface electromyography signal is filtered based on the spectral characteristics of the triaxial acceleration signal to suppress artifacts generated by macroscopic limb movements, resulting in a preprocessed electromyography signal. Within a continuously sliding time window, the median frequency sequence and the approximate entropy sequence of the amplitude envelope of the preprocessed electromyographic signal are determined; A risk index for muscle fatigue state is constructed based on the synergistic relationship between the median frequency sequence and the approximate entropy sequence. The fatigue risk warning report for patients during rehabilitation care is generated by combining the risk index with a preset fatigue risk threshold. Specifically, determining the median frequency sequence and the approximate entropy sequence of the amplitude envelope of the preprocessed electromyographic signal within a continuously sliding time window includes: Preset the time window for continuous sliding; The preprocessed electromyographic signal is divided into a sliding window using the time window to obtain continuous signal segments. The median frequency of each signal segment is determined, thereby obtaining the median frequency sequence of the preprocessed electromyographic signal; Extract the amplitude envelope of each signal segment and calculate the approximate entropy of the corresponding amplitude envelope to obtain the approximate entropy sequence of the amplitude envelope; Specifically, the risk index for muscle fatigue state constructed based on the synergistic relationship between the median frequency sequence and the approximate entropy sequence includes: Set the time window length for collaborative change analysis, and then extract the data segment within the most recent time window from the median frequency sequence and the approximate entropy sequence; A quantitative index for determining the cooperative change relationship between the median frequency sequence and the approximate entropy sequence based on the extracted data segments; The quantitative indicators are mapped to a risk index for muscle fatigue.

2. The method of claim 1, wherein, The surface electromyography (EMG) signal is filtered based on the spectral characteristics of the triaxial acceleration signal to suppress artifacts generated by macroscopic limb movements, resulting in a preprocessed EMG signal. Specifically, this preprocessed EMG signal includes: Extract reference noise features characterizing the macroscopic movement patterns of the limbs from the triaxial acceleration signals; An adaptive filter for filtering motion artifacts is constructed based on the reference noise characteristics; The surface electromyography (EMG) signal is processed by the adaptive filter to suppress artifacts generated by macroscopic limb movements, resulting in a preprocessed EMG signal.

3. The method of claim 1, wherein, The fatigue risk warning report for patients during rehabilitation care is generated by combining the risk index with a preset fatigue risk threshold, specifically including: Preset a fatigue risk threshold for patients during rehabilitation care; The patient's fatigue risk level is determined by combining the risk index with a preset fatigue risk threshold. A structured report containing early warning information is generated based on the fatigue risk level as the fatigue risk early warning report.

4. The method of claim 1, wherein, The target muscles refer to the main muscle groups selected as the subjects of fatigue monitoring based on the rehabilitation nursing movement patterns.

5. The method of claim 1, wherein, The surface electromyography signal refers to the electrophysiological signal generated by the activity of the target muscle motor unit.

6. A risk early warning system for rehabilitation care, which performs risk early warning for rehabilitation care by the method according to any one of claims 1 to 5, the system comprising a fatigue risk early warning unit, characterized in that, The fatigue risk early warning unit includes: The acquisition module is used to simultaneously acquire triaxial acceleration signals and surface electromyography signals of the target muscles during rehabilitation care. The processing module is used to filter the surface electromyography signal according to the spectral characteristics of the triaxial acceleration signal to suppress artifacts generated by macroscopic limb movement and obtain a preprocessed electromyography signal. The processing module is also used to determine the median frequency sequence and the approximate entropy sequence of the amplitude envelope of the preprocessed electromyographic signal within a continuously sliding time window. The processing module is also used to construct a risk index of muscle fatigue state based on the synergistic change relationship between the median frequency sequence and the approximate entropy sequence; The execution module is used to generate a fatigue risk warning report for patients during rehabilitation care by combining the risk index with a preset fatigue risk threshold. 7.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-6 when the computer program is executed by the processor. When the processor executes the computer program, it implements the steps of the fatigue risk warning method according to any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, the computer-readable storage medium comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 7. When the computer program is executed by the processor, it implements the steps of the fatigue risk warning method as described in any one of claims 1 to 5.

Images