An intelligent orthopedic rehabilitation training adaptive adjustment method and system

Abstract

This invention relates to the field of medical rehabilitation engineering technology and discloses an intelligent orthopedic rehabilitation training adaptive adjustment method and system, including an adaptive adjustment system. The adaptive adjustment system includes a rehabilitation movement acquisition unit, a cloud-based intelligent management platform, and a medical consultation interaction unit. By incorporating the adaptive adjustment system, this invention facilitates the objective and comprehensive monitoring of the rehabilitation training process: by simultaneously acquiring motion state data Q1 and biomechanical data Q2 through multimodal devices such as wearable inertial sensors and surface electromyography sensors, the subjective "whether the movement is standard" is transformed into objective time-series data such as joint angle We, trajectory Wr, and electromyography (EMG). This achieves digital recording and quantitative evaluation of the entire training process, realizing a leap from static image evaluation to dynamic recovery process tracking. Combined with the comprehensive analysis of training load state Sl, abnormal trends can be identified earlier than traditional re-examinations.

Description

Technical Field

[0001] This invention relates to the field of medical rehabilitation engineering technology, and more specifically to an intelligent orthopedic rehabilitation training adaptive adjustment method and system. Background Technology

[0002] Post-orthopedic surgery rehabilitation training is crucial for functional recovery, but its effectiveness highly depends on the standardization, continuity, and personalized adjustments of the training. Current mainstream rehabilitation models have significant limitations: Lack of training supervision: When patients train at home, there is a lack of objective and continuous monitoring and real-time feedback on whether the movements are correct and whether the training volume is sufficient, which can easily lead to incorrect movement patterns or secondary injuries. Subjective lag in assessment: Recovery status mainly relies on periodic outpatient follow-ups. Doctors make judgments through visual observation, questioning, and limited imaging examinations. There is a lack of continuous and quantitative functional status data, making it difficult to detect recovery stagnation or abnormalities in a timely manner. While some wearable devices or applications attempt to record motion data in the current technology, they are mostly limited to single data recording or simple counting, failing to achieve deep fusion analysis of multi-source heterogeneous data (movement, electromyography, imaging), and have not built a complete technical closed loop from "real-time monitoring - intelligent assessment - early warning - adaptive adjustment - remote medical confirmation", which means they cannot truly replace professional guidance and have limited clinical practical value. Therefore, there is an urgent need for an intelligent orthopedic rehabilitation training adaptive adjustment method and system to solve the aforementioned technical problems. Summary of the Invention

[0003] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide an intelligent orthopedic rehabilitation training adaptive adjustment method and system to solve the technical problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an intelligent orthopedic rehabilitation training adaptive adjustment system, comprising an adaptive adjustment system, wherein the adaptive adjustment system includes a rehabilitation movement acquisition unit, a cloud-based intelligent management platform, and a medical consultation interaction unit; The rehabilitation movement acquisition unit is used to acquire rehabilitation training movement data Q of the target object in real time through an external posture acquisition device; The cloud-based intelligent management platform is communicatively connected to the rehabilitation motion acquisition unit, and includes: The data storage module is used to receive and store the rehabilitation training action data Q and the orthopedic examination results of the target object; The core analysis engine evaluates whether the rehabilitation training movements meet the standards and calculates the training volume based on the rehabilitation training movement data. At the same time, it analyzes whether the recovery speed is normal based on the orthopedic examination results and generates markers and warnings for abnormal situations. An adaptive decision engine is used to parametrically adjust the rehabilitation training program based on the analysis results of the core analysis engine, and generate an adjusted training program. The medical consultation interaction unit is communicatively connected to the cloud-based intelligent management platform and is used to receive and display the markers and warnings, the adjusted training scheme, and related analysis data for medical personnel to review and confirm remotely.

[0005] Preferably, the data storage module, the core analysis engine, and the adaptive decision engine communicate and transmit data through a predefined application programming interface (API).

[0006] Preferably, the rehabilitation training movement data Q includes movement state data Q1 and biomechanical data Q2; The motion status data Q1 includes joint angle time-series data W collected continuously within one or more complete rehabilitation training cycles. e Motion trajectory time series data W r Angular velocity timing data W t And acceleration time series data W y ; The biomechanical data Q2 includes the following data collected synchronously during the corresponding training cycle: electromyography (EMG) signal time series data and pressure distribution time series data F.

[0007] Preferably, the analysis process of the core analysis engine is as follows: S100: Acquire and fuse rehabilitation training movement data Q from the data storage module with the orthopedic examination results of the target object; perform timestamp alignment, filtering and noise reduction, and feature extraction on the movement state data Q1 and biomechanical data Q2 to generate a standardized temporal feature set U. S200. Based on the orthopedic examination results, extract key imaging indicators and quantify their degree of recovery R within the current rehabilitation cycle; S300. The standardized temporal feature set U is compared with the pre-stored standard action feature template, and the action quality score Sc is output. At the same time, the actual training volume is calculated based on the feature set U and compared with the preset standard training volume threshold, and the training load status Sl is output. S400. The recovery degree R, action quality score Sc, and training load status Sl are compared with the preset corresponding threshold ranges, and a comprehensive analysis is performed. If the abnormal conditions are met, an early warning mark containing the abnormal type and level is generated and output.

[0008] Preferably, the process of timestamp alignment, filtering and denoising, and feature extraction for motion state data Q1 and biomechanical data Q2 includes: Based on a unified system clock source, the joint angle timing data We, motion trajectory timing data Wr, angular velocity timing data Wt, acceleration timing data Wy, electromyographic signal timing data EMG, and pressure distribution timing data F are synchronized and aligned in time. The Butterworth low-pass filter or the Kalman filter are used to filter and denoise the various types of time-series data after alignment. From the denoised time-series data, multidimensional features including time-domain mean, variance, frequency-domain dominant frequency, signal amplitude integral, and cross-modal correlation coefficient are extracted. The extracted multidimensional features are normalized to eliminate dimensional differences and transform them into a unified numerical range, thereby generating the standardized time-series feature set U.

[0009] Preferably, the step S200 of "extracting key imaging indicators and quantifying their degree of recovery R" includes: extracting at least one quantifiable structural healing indicator based on the temporal imaging data in the orthopedic examination results; The structural healing indicators include fracture line ambiguity score, callus area percentage, and joint space width; The degree of recovery R is obtained by calculating the percentage change in the structural healing index between the current period and the previous period.

[0010] Preferably, the step of comparing the standardized temporal feature set U with the pre-stored standard movement feature templates specifically involves: performing item-by-item similarity matching between the standardized features representing joint angles and movement trajectories in the feature set U and the standard angle feature sequences and standard trajectory feature sequences corresponding to the rehabilitation movements in the template library; and the movement quality score Sc is derived based on the matching results of the joint angles and movement trajectories. The calculation of the actual training volume based on the feature set U is specifically as follows: the overall training load value is determined by weighted summation based on the intensity characteristics of angular velocity and acceleration and the characteristics of effective exercise duration in the feature set U; the training load state S1 is determined by comparing the overall training load value with the load threshold range preset for the current rehabilitation stage.

[0011] Preferably, the "comprehensive analysis" mentioned in step S400 refers to: performing a preset correlation logic judgment based on the comparison results of the recovery degree R, the action quality score Sc, and the training load state Sl with their respective threshold ranges; The correlation logic judgment includes at least one of the following abnormal pattern determinations: Recovery Abnormal Mode: When the recovery level R is judged to be below the normal range, and the training load state Sl is judged to be within the normal or acceptable range, the determination is triggered. Action-load imbalance mode: When the action quality score Sc is judged to be below the normal range and the training load state Sl is judged to be above the normal range, the judgment is triggered; Composite Abnormal Mode: When both the recovery degree R and the action quality score Sc are judged to be below the normal range, a judgment is triggered.

[0012] Preferably, the adaptive decision engine adjusts and generates a corresponding training scheme based on the abnormal pattern generated in step S400, specifically including: If the abnormal pattern is "slow recovery", then adjust the phase plan of rehabilitation training or reduce the load parameters of strength training; If the abnormal mode is "abnormal movement quality", then adjust the difficulty level of the training movement, break it down into simpler steps, or add auxiliary training parameters to enhance stability. If the abnormal mode is "abnormal training load", then adjust the training frequency, the duration of a single training session, or the rest time between sets.

[0013] An intelligent orthopedic rehabilitation training adaptive adjustment method, applied to an adaptive adjustment system, is characterized by comprising the following steps: Step 1: In the cloud-based intelligent management platform, based on the target object's injury diagnosis information and initial functional assessment results, configure a personalized initial rehabilitation training plan, pre-store corresponding standard movement feature templates, and threshold ranges for various assessment indicators. Step 2: Guide and monitor the target object to perform rehabilitation training through the rehabilitation movement acquisition unit, synchronously collect the rehabilitation training movement data Q and transmit it to the cloud intelligent management platform, and periodically or in response to event triggers, obtain and upload the target object's orthopedic examination results; Step 3: The core analysis engine executes the analysis process as described in claim 4, and generates a quantitative recovery degree R, movement quality score Sc, training load status Sl, and comprehensive early warning mark based on the rehabilitation training movement data Q and orthopedic examination results. Step 4: The adaptive decision engine adjusts the current rehabilitation training program according to the abnormal pattern indicated by the comprehensive early warning marker, or according to the quantitative results of the recovery degree R, movement quality score Sc, and training load state Sl, and generates the adjusted training program. Step 5: Push the comprehensive early warning marker, quantitative evaluation results, and the adjusted training plan to the medical consultation interaction unit. After review, medical personnel communicate remotely with the target through the unit and confirm, modify, or reject the adjusted training plan to form the final execution plan. Step 6: Synchronously update the final execution plan to the rehabilitation action acquisition unit to guide the rehabilitation training in the next cycle; The system returns to step two, forming a closed-loop adaptive management process that combines continuous data collection, intelligent analysis, dynamic adjustment, and manual review.

[0014] The technical effects and advantages of this invention are as follows: This invention, by incorporating an adaptive adjustment system, facilitates the objective and comprehensive monitoring of the rehabilitation training process: by simultaneously collecting motion state data Q1 and biomechanical data Q2 through multimodal devices such as wearable inertial sensors and surface electromyography sensors, the subjective "whether the movement is standard" is transformed into objective time-series data such as joint angles We, trajectory Wr, and electromyography (EMG), thereby realizing the digital recording and quantitative evaluation of the entire training process. This invention, by incorporating an adaptive adjustment system, facilitates in-depth analysis such as dynamic time warping by aligning time-series data with timestamps, extracting features, and generating a standardized time-series feature set U. It also calculates the action quality score Sc. Furthermore, by quantifying the recovery degree R of time-series radiographic indicators, it achieves a leap from static image evaluation to dynamic recovery process tracking. Combined with the comprehensive analysis of training load state Sl, it can identify abnormal trends earlier than traditional review. This invention uses an adaptive decision engine to automatically invoke preset rules to parametrically adjust the training program (such as adjusting the load, decomposing the movements, and changing the frequency) based on early warning markers or quantitative evaluation results. This breaks the limitations of traditional static rehabilitation programs, enabling the training program to dynamically evolve with the patient's recovery status and always remain within the safe and effective "optimal challenge range". Attached Figure Description

[0015] Figure 1 This is an overall flowchart of the adaptive adjustment system of the present invention.

[0016] Figure 2 This is an overall flowchart of the adaptive adjustment method of the present invention. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The intelligent orthopedic rehabilitation training adaptive adjustment method and system involved in the present invention are not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Reference Figure 1 As shown, the present invention discloses an intelligent orthopedic rehabilitation training adaptive adjustment system, including an adaptive adjustment system, which includes a rehabilitation movement acquisition unit, a cloud-based intelligent management platform, and a medical consultation interaction unit. The rehabilitation movement acquisition unit is used to acquire rehabilitation training movement data Q of the target object in real time through an external posture acquisition device; The cloud-based intelligent management platform communicates with the rehabilitation motion acquisition unit and includes: The data storage module is used to receive and store rehabilitation training movement data Q and orthopedic examination results of the target subject; The core analysis engine evaluates whether the rehabilitation training movements meet the standards and calculates the training volume based on rehabilitation training movement data. At the same time, it analyzes whether the recovery speed is normal based on orthopedic examination results and generates markers and warnings for abnormal situations. The adaptive decision engine is used to parametrically adjust the rehabilitation training program based on the analysis results of the core analysis engine, and generate the adjusted training program. The medical consultation interaction unit communicates with the cloud-based intelligent management platform to receive and display tags and warnings, adjusted training plans, and related analysis data for medical personnel to review and confirm remotely.

[0019] In this embodiment, the "external posture acquisition device" is not a single device, but a multimodal sensor integrated system that can be flexibly configured according to the accuracy of rehabilitation assessment, application scenarios, and cost considerations. Its core purpose is to capture the limb movement and biomechanical signals of the target object in three-dimensional space in a non-contact or low-invasive manner with high precision. The device mainly includes implementation equipment based on one or more of the following technical paths: Wearable Inertial Measurement Unit Sensor Array: Specific device: It consists of multiple (usually no less than 2) nine-axis inertial measurement unit sensor nodes, each node integrating a three-axis accelerometer, a three-axis gyroscope and a three-axis magnetometer; Its core principle is: Attitude calculation: By using algorithms within the sensors (such as complementary filtering based on quaternions or Kalman filtering), the specific force measured by the accelerometer, the angular velocity measured by the gyroscope, and the heading reference provided by the magnetometer are fused in real time to calculate the three-dimensional spatial attitude (i.e., pitch angle, roll angle, and yaw angle relative to the Earth coordinate system) of each sensor node. Skeletal model construction: Sensor nodes are fixedly worn near specific anatomical bony landmarks of the limb to be evaluated (for example, when used for knee joint evaluation, they are worn on the distal lateral femur of the thigh and the proximal lateral tibia of the lower leg, respectively). The system is pre-input with parameters such as bone length and joint center position of each limb segment to form a simplified human skeletal model. Motion data derivation: By calculating the relative posture change between two adjacent sensor nodes (i.e., representing two adjacent limb segments), the joint angle time series data We can be calculated in real time. Further, by integrating the angular velocity time series data Wt and combining it with the initial position, the motion trajectory time series data Wr of the limb end point can be estimated. The acceleration data Wy is directly provided by the accelerometer.

[0020] Optical motion capture system Specific equipment: It consists of multiple (usually 4-8) high-speed infrared cameras, reflective markers, and a data processing unit; Working principle: Marker tracking: Reflective markers are attached to bony landmarks on the surface of the target object. Multiple infrared cameras simultaneously emit infrared light from different angles and capture the reflected light from the markers. Through the principle of triangulation, the precise coordinates (X, Y, Z) of each marker in three-dimensional space are calculated. Motion reconstruction: The system continuously captures the coordinates of all marker points at extremely high frequencies (such as above 100Hz) to form the motion trajectory time series data Wr of the marker points. By connecting specific marker points, the joint angle time series data We can be calculated. By differentiating the position data, the velocity and acceleration data can be obtained.

[0021] Reference Figure 1 As shown, this invention discloses an intelligent orthopedic rehabilitation training adaptive adjustment system, in which the data storage module, core analysis engine and adaptive decision engine communicate and transmit data through a predefined application programming interface (API). The rehabilitation training movement data Q includes motion state data Q1 and biomechanical data Q2; The motion status data Q1 includes joint angle time-series data W collected continuously within one or more complete rehabilitation training cycles. e Motion trajectory time series data W r Angular velocity timing data W t And acceleration time series data W y ; Biomechanical data Q2 includes: electromyography (EMG) temporal data and pressure distribution temporal data F, which were collected synchronously during the corresponding training cycle. The core analysis engine's analysis workflow is as follows: S100: Acquire and fuse rehabilitation training movement data Q from the data storage module with the orthopedic examination results of the target object; perform timestamp alignment, filtering and noise reduction, and feature extraction on the movement state data Q1 and biomechanical data Q2 to generate a standardized temporal feature set U. S200. Based on orthopedic examination results, extract key imaging indicators and quantify their recovery degree R within the current rehabilitation cycle; S300: Compare the standardized temporal feature set U with the pre-stored standard action feature template, and output the action quality score Sc. At the same time, calculate the actual training volume based on the feature set U, and compare it with the preset standard training volume threshold, and output the training load status Sl. S400: Compare the recovery level R, motion quality score Sc, and training load status Sl with the preset corresponding threshold ranges respectively, perform comprehensive analysis, and if the abnormal conditions are met, generate and output an early warning mark containing the abnormal type and level. The timestamp alignment, filtering, denoising, and feature extraction of motion state data Q1 and biomechanical data Q2 include: Based on a unified system clock source, the timing data of joint angles We, motion trajectory Wr, angular velocity Wt, acceleration Wy, electromyography signal EMG, and pressure distribution F are synchronized and aligned in time. Filters are used to perform filtering and noise reduction processing on the various types of aligned time-series data. From the denoised time-series data, multidimensional features including time-domain mean, variance, frequency-domain dominant frequency, signal amplitude integral, and cross-modal correlation coefficient are extracted. The extracted multidimensional features are normalized to eliminate dimensional differences and transform them into a unified numerical range, generating a standardized time-series feature set U.

[0022] In this part of the application embodiments, the filter is a Butterworth low-pass filter or a Kalman filter; The purpose of “normalizing the extracted multidimensional features” is to convert the previously extracted multidimensional original feature vectors with different dimensions into a standardized feature set U in which all components are on the same numerical scale, so as to facilitate subsequent fair comparison and analysis. The specific implementation is as follows: Input and Objective: The input to this step is the original multidimensional feature vector containing multiple feature components such as time-domain mean, variance, frequency-domain dominant frequency, signal amplitude integral, and cross-modal correlation coefficient. The processing objective is to eliminate the order-of-magnitude differences caused by the different physical meanings of each feature. Core processing method: To achieve this goal, the system uses the "min-max normalization" method to process each feature component of the vector independently. For the ii-th feature component, its normalization value Xi is calculated as follows: Where X is a general variable, representing any feature taken from the multidimensional feature vector, where X... i For the i-th feature component, X min and X max To obtain the minimum and maximum values ​​of this feature from the normalized reference set in advance, this calculation linearly maps the value of each feature to the interval [0,1]. Output: After all feature components have been normalized, the result is obtained from these X1 components. 归 X2 归 X i 归 The new vector formed by sequential combination is the standardized temporal feature set U.

[0023] Step S200, “extracting key imaging indicators and quantifying their degree of recovery R”, includes: extracting at least one quantifiable structural healing indicator based on temporal imaging data from orthopedic examination results. Structural healing indicators include fracture line ambiguity score, callus area percentage, and joint space width; The degree of recovery, R, is obtained by calculating the percentage change in the structural healing index between the current cycle and the previous cycle. The standardized temporal feature set U is compared with the pre-stored standard movement feature templates. Specifically, the standardized features representing joint angles and movement trajectories in the feature set U are matched item by item with the standard angle feature sequences and standard trajectory feature sequences of the corresponding rehabilitation movements in the template library. The movement quality score Sc is derived based on the matching results of joint angles and movement trajectories. The actual training load is calculated based on the feature set U. Specifically, the overall training load value is determined by weighted summation based on the intensity characteristics of angular velocity and acceleration and the characteristics of effective exercise duration in the feature set U. The training load state Sl is determined by comparing the overall training load value with the load threshold range preset for the current rehabilitation stage. In this embodiment of the application, the process of matching the motion quality score Sc with the joint angle and the motion trajectory is as follows: The system will automatically match and analyze the real-time generated, standardized joint angle and motion trajectory temporal feature set with the corresponding standard sequence in the pre-stored standard action template. First, the system performs time alignment and difference quantization. It employs a sequence alignment algorithm capable of handling time scale scaling to perform non-linear time-axis alignment between the real-time sequence and the standard sequence. This eliminates temporal phase differences caused by varying motion speeds. Based on this alignment, the system calculates the distance between sequences and quantizes the overall difference D between the real-time motion and the standard sequence in terms of joint angle change patterns. angle And the overall deviation Dt from the standard path in the three-dimensional spatial motion trajectory. raj ; Secondly, the system performs score conversion and synthesis, using a preset mapping relationship to convert the aforementioned difference degree D. angle and D traj Each is converted into an independent, comparable numerical sub-rating S. angle and S traj This mapping relationship is configured to ensure that the lower the difference, the higher the corresponding sub-score, thus directly representing the quality of the action in that dimension. The scale range of the sub-scores is predefined and consistent (e.g., 0-100 points). The final comprehensive action quality score Sc is calculated by sub-scores S. angle and S traj It is generated through a fusion calculation based on pre-set comprehensive rules that reflect the clinical focus of different rehabilitation stages. For example, in the early rehabilitation phase, the sub-score S of joint angle accuracy angle It may carry a higher weight in the overall calculation; however, in the later functional training stage, the sub-score S of motion trajectory stability... traj The proportion of that will increase accordingly.

[0024] In this part of the application embodiments, the method for setting the "preset load threshold range for the current rehabilitation stage" is as follows: The system is based on a pre-built load baseline database established according to the consensus of rehabilitation medicine, providing initial threshold references for different injuries and rehabilitation stages. On this basis, the system performs personalized calibration according to the individual functional baseline of the target object (such as initial muscle strength, range of motion, and weight) to determine the load range that matches their individual capabilities. This threshold range is not a fixed value; its lower bound L min To ensure the minimum requirements for training effectiveness, the upper bound L max The highest limit to ensure safety and prevent overtraining is the intermediate target value L. 中This is the ideal load. During the rehabilitation process, the system will dynamically and slightly adjust this threshold range based on the patient's recent training completion and physiological feedback, thereby achieving gradual and individualized load management. The training load state Sl is determined by comparing the calculated overall training load value with the dynamically set threshold range [L]. min L 中 L max It is determined by comparison.

[0025] In step S400, “perform comprehensive analysis” means: based on the comparison results of the recovery degree R, the movement quality score Sc, and the training load state Sl with their respective threshold ranges, perform a preset correlation logic judgment. The correlation logic judgment includes at least one of the following abnormal patterns: Recovery Abnormal Mode: When the recovery level R is judged to be below the normal range, and the training load status Sl is judged to be within the normal or acceptable range, the judgment is triggered. Action-load imbalance mode: When the action quality score Sc is judged to be below the normal range and the training load state Sl is judged to be above the normal range, the judgment is triggered. Composite Abnormal Mode: A judgment is triggered when both the recovery level R and the action quality score Sc are judged to be below the normal range; The adaptive decision engine adjusts and generates corresponding training schemes based on the abnormal patterns generated in step S400, specifically including: If the abnormal pattern is "slow recovery", then adjust the phase plan of rehabilitation training or reduce the load parameters of strength training. If the abnormal mode is "abnormal movement quality", then adjust the difficulty level of the training movements, break them down into simpler steps, or add auxiliary training parameters to enhance stability. If the abnormal mode is "abnormal training load", then adjust the training frequency, the duration of a single training session, or the rest time between sets.

[0026] Reference Figure 2 As shown, this invention provides an intelligent orthopedic rehabilitation training adaptive adjustment method, applied to an adaptive adjustment system, characterized by comprising the following steps: Step 1: In the cloud-based intelligent management platform, based on the target individual's injury diagnosis information and initial functional assessment results, configure a personalized initial rehabilitation training plan, pre-store corresponding standard movement feature templates, and threshold ranges for various assessment indicators. Step 2: Guide and monitor the target subject to perform rehabilitation training through the rehabilitation movement acquisition unit, synchronously collect rehabilitation training movement data Q and transmit it to the cloud intelligent management platform, and periodically or in response to event triggers, obtain and upload the target subject's orthopedic examination results; Step 3: The core analysis engine executes the analysis process as described in claim 4, and generates quantitative recovery degree R, movement quality score Sc, training load status Sl and comprehensive early warning mark based on rehabilitation training movement data Q and orthopedic examination results; Step 4: The adaptive decision engine adjusts the current rehabilitation training program based on the abnormal patterns indicated by the comprehensive early warning markers, or based on the quantitative results of the recovery degree R, movement quality score Sc, and training load status Sl, and generates the adjusted training program. Step 5: Push the comprehensive early warning markers, quantitative assessment results, and adjusted training plan to the medical consultation interaction unit. After review, medical personnel will communicate remotely with the target through this unit and confirm, modify, or reject the adjusted training plan to form the final execution plan. Step 6: Update the final implementation plan to the rehabilitation movement acquisition unit to guide the rehabilitation training in the next cycle; The system returns to step two, forming a closed-loop adaptive management process that combines continuous data collection, intelligent analysis, dynamic adjustment, and manual review.

[0027] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0028] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0029] In conclusion, the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent orthopedic rehabilitation training adaptive adjustment system, comprising an adaptive adjustment system, characterized in that: The adaptive adjustment system includes a rehabilitation motion acquisition unit, a cloud-based intelligent management platform, and a medical consultation interaction unit. The rehabilitation movement acquisition unit is used to acquire rehabilitation training movement data Q of the target object in real time through an external posture acquisition device; The cloud-based intelligent management platform is communicatively connected to the rehabilitation motion acquisition unit, and includes: The data storage module is used to receive and store the rehabilitation training action data Q and the orthopedic examination results of the target object; The core analysis engine evaluates whether the rehabilitation training movements meet the standards and calculates the training volume based on the rehabilitation training movement data. At the same time, it analyzes whether the recovery speed is normal based on the orthopedic examination results and generates markers and warnings for abnormal situations. An adaptive decision engine is used to parametrically adjust the rehabilitation training program based on the analysis results of the core analysis engine, and generate an adjusted training program. The medical consultation interaction unit is communicatively connected to the cloud-based intelligent management platform and is used to receive and display the markers and warnings, the adjusted training scheme, and related analysis data for medical personnel to review and confirm remotely.

2. The intelligent orthopedic rehabilitation training adaptive adjustment system according to claim 1, characterized in that: The data storage module, core analysis engine, and adaptive decision engine communicate and transfer data through a predefined application programming interface (API).

3. The intelligent orthopedic rehabilitation training adaptive adjustment system according to claim 1, characterized in that: The rehabilitation training motion data Q includes motion state data Q1 and biomechanical data Q2; in Motion status data Q1 includes joint angle time-series data W collected continuously within one or more complete rehabilitation training cycles. e Motion trajectory time series data W r Angular velocity timing data W t And acceleration time series data W y ; The biomechanical data Q2 includes the following data collected synchronously during the corresponding training cycle: electromyography (EMG) signal time series data and pressure distribution time series data F.

4. The intelligent orthopedic rehabilitation training adaptive adjustment system according to claim 3, characterized in that: The analysis process of the core analysis engine is as follows: S100: Acquire and fuse rehabilitation training movement data Q from the data storage module with the orthopedic examination results of the target object; perform timestamp alignment, filtering and noise reduction, and feature extraction on the movement state data Q1 and biomechanical data Q2 to generate a standardized temporal feature set U. S200. Based on the orthopedic examination results, extract key imaging indicators and quantify their degree of recovery R within the current rehabilitation cycle; S300. The standardized temporal feature set U is compared with the pre-stored standard action feature template, and the action quality score Sc is output. At the same time, the actual training volume is calculated based on the feature set U and compared with the preset standard training volume threshold, and the training load status Sl is output. S400. The recovery degree R, action quality score Sc, and training load status Sl are compared with the preset corresponding threshold ranges, and a comprehensive analysis is performed. If the abnormal conditions are met, an early warning mark containing the abnormal type and level is generated and output.

5. The intelligent orthopedic rehabilitation training adaptive adjustment system according to claim 4, characterized in that: The process of timestamp alignment, filtering and denoising, and feature extraction for motion state data Q1 and biomechanical data Q2 includes: Based on a unified system clock source, the joint angle timing data We, motion trajectory timing data Wr, angular velocity timing data Wt, acceleration timing data Wy, electromyographic signal timing data EMG, and pressure distribution timing data F are synchronized and aligned in time. Filters are used to perform filtering and noise reduction processing on the various types of aligned time-series data. From the denoised time-series data, multidimensional features including time-domain mean, variance, frequency-domain dominant frequency, signal amplitude integral, and cross-modal correlation coefficient are extracted. The extracted multidimensional features are normalized to eliminate dimensional differences and transform them into a unified numerical range, thereby generating the standardized time-series feature set U.

6. The intelligent orthopedic rehabilitation training adaptive adjustment system according to claim 4, characterized in that: The step S200, "extracting key imaging indicators and quantifying their degree of recovery R", includes: extracting at least one quantifiable structural healing indicator based on the temporal imaging data in the orthopedic examination results. The structural healing indicators include fracture line ambiguity score, callus area percentage, and joint space width; The degree of recovery R is obtained by calculating the percentage change in the structural healing index between the current period and the previous period.

7. The intelligent orthopedic rehabilitation training adaptive adjustment system according to claim 5, characterized in that: The step of comparing the standardized temporal feature set U with the pre-stored standard movement feature templates is as follows: the standardized features in the feature set U that represent joint angles and movement trajectories are matched item by item with the standard angle feature sequences and standard trajectory feature sequences of the corresponding rehabilitation movements in the template library. The movement quality score Sc is obtained by comprehensively considering the matching results of the joint angles and movement trajectories. The calculation of the actual training volume based on the feature set U is specifically as follows: the overall training load value is determined by weighted summation based on the intensity characteristics of angular velocity and acceleration and the characteristics of effective exercise duration in the feature set U; the training load state S1 is determined by comparing the overall training load value with the load threshold range preset for the current rehabilitation stage.

8. The intelligent orthopedic rehabilitation training adaptive adjustment system according to claim 7, characterized in that: The "comprehensive analysis" mentioned in step S400 refers to: performing a preset correlation logic judgment based on the comparison results of the recovery degree R, the action quality score Sc, and the training load state Sl with their respective threshold ranges; The correlation logic judgment includes at least one of the following abnormal pattern determinations: Recovery Abnormal Mode: When the recovery level R is judged to be below the normal range, and the training load state Sl is judged to be within the normal or acceptable range, the determination is triggered. Action-load imbalance mode: When the action quality score Sc is judged to be below the normal range and the training load state Sl is judged to be above the normal range, the judgment is triggered; Composite Abnormal Mode: When both the recovery degree R and the action quality score Sc are judged to be below the normal range, a judgment is triggered.

9. The intelligent orthopedic rehabilitation training adaptive adjustment system according to claim 8, characterized in that: The adaptive decision engine adjusts and generates a corresponding training scheme based on the abnormal pattern generated in step S400, specifically including: If the abnormal pattern is "slow recovery", then adjust the phase plan of rehabilitation training or reduce the load parameters of strength training. If the abnormal mode is "abnormal movement quality", then adjust the difficulty level of the training movement, break it down into simpler steps, or add auxiliary training parameters to enhance stability. If the abnormal mode is "abnormal training load", then adjust the training frequency, the duration of a single training session, or the rest time between sets.

10. An intelligent orthopedic rehabilitation training adaptive adjustment method, applied to the adaptive adjustment system described in any one of claims 1-9, characterized in that, Includes the following steps: Step 1: In the cloud-based intelligent management platform, based on the target object's injury diagnosis information and initial functional assessment results, configure a personalized initial rehabilitation training plan, pre-store corresponding standard movement feature templates, and threshold ranges for various assessment indicators. Step 2: Guide and monitor the target object to perform rehabilitation training through the rehabilitation movement acquisition unit, synchronously collect the rehabilitation training movement data Q and transmit it to the cloud intelligent management platform, and periodically or in response to event triggers, obtain and upload the target object's orthopedic examination results; Step 3: The core analysis engine executes the analysis process as described in claim 4, and generates a quantitative recovery degree R, movement quality score Sc, training load status Sl, and comprehensive early warning mark based on the rehabilitation training movement data Q and orthopedic examination results. Step 4: The adaptive decision engine adjusts the current rehabilitation training program according to the abnormal pattern indicated by the comprehensive early warning marker, or according to the quantitative results of the recovery degree R, movement quality score Sc, and training load state Sl, and generates the adjusted training program. Step 5: Push the comprehensive early warning marker, quantitative evaluation results, and the adjusted training plan to the medical consultation interaction unit. After review, medical personnel communicate remotely with the target through the unit and confirm, modify, or reject the adjusted training plan to form the final execution plan. Step 6: Synchronously update the final execution plan to the rehabilitation action acquisition unit to guide the rehabilitation training in the next cycle; The system returns to step two, forming a closed-loop adaptive management process that combines continuous data collection, intelligent analysis, dynamic adjustment, and manual review.

Images