Wheelchair user limb straightening guidance system based on medical health big data sharing

Abstract

The invention relates to the technical field of medical health, in particular to a wheelchair user limb straightening guidance system based on medical health big data sharing, which comprises a limb state sensing module for acquiring limb joint angles, limb pressure distribution and trunk posture data; the medical big data analysis center is used for identifying potential health risks caused by improper sitting postures and positioning limb abnormal stress points and posture deviation sources; the intelligent guidance decision engine is used for dynamically generating a personalized limb straightening guidance scheme and rehabilitation assistance suggestions in combination with a preset rehabilitation rule and user historical data; the limb posture dynamic adjusting unit drives the wheelchair adjusting mechanism to synchronously provide straightening guide feedback; and the medical health data sharing module carries out cross-mechanism sharing on the limb straightening guidance scheme, the rehabilitation assistance suggestion, the patient medical record data and the potential health risk, synchronously updates the user state information and carries out access authority management and control. Therefore, the problems of poor adaptability, sharing difficulty and the like in the prior art are solved.

Description

Technical Field

[0001] This invention relates to the field of medical and health technology, specifically to a limb alignment guidance system for wheelchair users based on the sharing of medical and health big data. Background Technology

[0002] Wheelchair limb alignment assistive systems are crucial support tools for the mobility and rehabilitation of patients with limb dysfunction, the elderly, and other groups. The scientific accuracy and precision of their limb alignment guidance directly determine the user's health status, rehabilitation progress, and quality of life. During long-term use, factors such as differences in user body shape, types of limb impairment, and dynamic changes in the rehabilitation stage (e.g., progress in rehabilitation, evolution of complications) can easily lead to improper limb alignment, resulting in complications such as joint contractures, muscle atrophy, pressure sores, and scoliosis. This severely reduces the user's quality of life and increases the medical burden on families and society. As the healthcare field develops towards personalization, precision, and intelligence, higher demands are placed on wheelchair limb alignment assistive systems in areas such as medical data integration and sharing, personalized guidance plan generation, dynamic adjustment and adaptation, and security and privacy protection. There is an urgent need for a system that can deeply integrate medical and health big data with limb alignment guidance scenarios, dynamically match the user's personalized medical characteristics, and intelligently drive precise guidance decisions to ensure the standardization and safety of limb alignment during daily mobility and rehabilitation.

[0003] However, traditional wheelchair limb alignment technologies and systems have significant shortcomings: poor adaptability, often relying on fixed mechanical limiting structures or simple intelligent detection modules for universal correction, failing to fully consider differences in body shape and limb disability type among users, and lacking personalized adaptation and dynamic adjustment capabilities; lack of data linkage, with existing intelligent wheelchair posture detection and correction functions mostly operating independently, failing to integrate personalized medical data such as medical records, rehabilitation training records, and history of complications, resulting in guidance plans lacking clinical data support and insufficient accuracy and scientific rigor; hindered data sharing, with medical and health data being scattered, difficult to share, and challenging to protect privacy, lacking an effective linkage mechanism with limb alignment guidance scenarios, and failing to fully leverage the supporting value of big data; and the lack of a precise guidance system based on medical big data, making it difficult for operators and users to obtain personalized plans tailored to individual rehabilitation needs, and lagging in the ability to predict and avoid complication risks. Overall, the technology faces multiple challenges, including insufficient adaptability, lack of data linkage, and poor guidance accuracy. Summary of the Invention

[0004] This application provides a limb alignment guidance system for wheelchair users based on the sharing of medical and health big data, in order to solve the problems of poor adaptability and difficulty in sharing in the existing technology.

[0005] The first aspect of this application provides a limb alignment guidance system for wheelchair users based on medical and health big data sharing, comprising: a limb state perception module, a medical big data analysis center, an intelligent guidance decision engine, a limb posture dynamic adjustment unit, and a medical and health data sharing module; wherein, the limb state perception module is used to collect data on the limb joint angles, limb pressure distribution, and trunk posture of the wheelchair user; the medical big data analysis center is used to fuse and process the data on the limb joint angles, limb pressure distribution, and trunk posture of the wheelchair user, identify potential health risks such as pressure sores, aggravation of scoliosis, and muscle spasms caused by improper sitting posture, and locate abnormal force points and postural deviations of the limbs. The intelligent guidance decision engine is used to dynamically generate personalized limb alignment guidance plans and rehabilitation assistance suggestions based on the potential health risks, abnormal limb force points, and postural deviation sources, combined with preset rehabilitation rules and user historical data; the limb posture dynamic adjustment unit is used to drive the wheelchair adjustment mechanism according to the personalized limb alignment guidance plan and rehabilitation assistance suggestions, and simultaneously provide alignment guidance feedback; the medical health data sharing module is used to share the personalized limb alignment guidance plan and rehabilitation assistance suggestions, patient medical record data, and the potential health risks, abnormal limb force points, and postural deviation sources across institutions, synchronously update user status information, and manage access permissions.

[0006] Preferably, the limb state perception module includes a joint angle sensor group, a pressure sensor array, and a trunk posture acquisition unit. The joint angle sensor group is used to collect rotation angle data of the shoulder joint, elbow joint, hip joint, and knee joint in real time. The pressure sensor array is installed in the wheelchair seat and backrest area to collect pressure distribution data between the limb and the contact surface. The trunk posture acquisition unit is used to capture trunk pitch angle, tilt angle, and rotation angle data.

[0007] Preferably, the medical big data analysis center includes a data fusion processing unit, a health risk identification unit, and an anomaly localization unit. The data fusion processing unit employs a weighted fusion algorithm to perform spatiotemporal alignment and redundancy removal on limb joint angle, limb pressure distribution, and trunk posture data, generating a unified limb state dataset. The health risk identification unit, based on preset thresholds and machine learning models, identifies pressure ulcer risk levels, scoliosis progression trends, and the probability of muscle spasms. The anomaly localization unit analyzes pressure distribution hotspots and posture deviation data, using correlation analysis algorithms to locate the specific positions of abnormal force points on the limbs and the sources of posture deviations in the trunk and limbs.

[0008] Preferably, the intelligent guidance decision engine includes a guidance plan generation unit, a rehabilitation suggestion adaptation unit, and a medical data matching unit. The guidance plan generation unit generates a personalized limb alignment guidance plan, including limb adjustment angles, timing, and duration, based on potential health risk levels, abnormal stress points, and sources of postural deviations, combined with preset rehabilitation rules and user historical data. The rehabilitation suggestion adaptation unit generates targeted muscle relaxation training and joint mobility assistance suggestions using a user feature adaptation algorithm, combined with the user's age, medical history, and rehabilitation stage. The medical data matching unit accesses a multi-center medical and health database and uses a similarity calculation algorithm to match intervention plans for similar cases, optimizing the current guidance plan and rehabilitation suggestions.

[0009] Preferably, the limb posture dynamic adjustment unit includes an adjustment mechanism drive unit, a guidance feedback providing unit, and a posture calibration unit. The adjustment mechanism drive unit receives personalized alignment guidance and drives the wheelchair seat lifting, backrest angle adjustment, and tilt compensation actuators via a trajectory planning algorithm. The guidance feedback providing unit guides the user to actively cooperate with the alignment movements through voice prompts, flashing visual indicator lights, and tactile vibration feedback. The posture calibration unit collects real-time limb state data after adjustment, compares it with target posture parameters, and calibrates the adjustment accuracy.

[0010] Preferably, the medical and health data sharing module includes a data sharing unit, an access control unit, and a status synchronization unit. The data sharing unit is used to share personalized limb alignment guidance plans, rehabilitation assistance suggestions, core patient medical record data, and risk assessment results across hospitals and rehabilitation institutions via an encrypted transmission protocol, ensuring privacy and security during transmission. The access control unit is used to set different data access permissions for medical staff, rehabilitation therapists, and patients' families based on a role-based access control model. The data synchronization unit is used to synchronize user posture adjustment effect data, health risk change data, and rehabilitation progress data in real time, ensuring data consistency during cross-institutional collaboration.

[0011] The second aspect of this application provides a method for guiding wheelchair users to limb alignment based on shared medical and health big data. The method includes: acquiring data on the joint angles, pressure distribution, and trunk posture of the wheelchair user's limbs; fusing the data on joint angles, pressure distribution, and trunk posture; assessing the potential health risks of pressure sores, abnormal spinal load, and muscle fatigue caused by improper sitting posture based on preset thresholds and a machine learning model; locating abnormal force points and postural deviation sources in the limbs using a correlation analysis algorithm; dynamically generating personalized limb alignment guidance plans and rehabilitation assistance suggestions based on the potential health risks, abnormal force points, and postural deviation sources, combined with a multi-center medical and health database; under user confirmation and safety monitoring, driving the wheelchair adjustment mechanism to adjust and providing alignment guidance feedback based on the personalized limb alignment guidance plans and rehabilitation assistance suggestions; and simultaneously sharing the personalized limb alignment guidance plans and rehabilitation assistance suggestions, patient medical record data, and the potential health risks, abnormal force points, and postural deviation sources across institutions securely, and updating user status information synchronously.

[0012] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the program to implement a method for guiding wheelchair users to limb alignment based on the sharing of medical and health big data, as described in the above embodiments.

[0013] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement a method for guiding wheelchair users to limb alignment based on the sharing of medical and health big data, as described in the above embodiments.

[0014] The fifth aspect of this application provides a computer program product, including a computer program or instructions, for implementing a method for guiding wheelchair users to limb alignment based on the sharing of medical and health big data as described in the above embodiments.

[0015] Therefore, this application has the following beneficial effects: This application embodiment uses a limb state perception module to accurately collect multi-dimensional data on limb joint angles, pressure distribution, and trunk posture, providing complete basic data for subsequent analysis and judgment. The medical big data analysis center, leveraging a weighted fusion algorithm and machine learning model, performs deep fusion processing on multi-source data, accurately identifying potential health risks such as pressure sores and worsening scoliosis, locating abnormal stress points and postural deviation sources, and significantly improving the accuracy of risk prediction and abnormal location. The intelligent guidance decision engine deeply integrates preset rehabilitation rules, user historical data, and multi-center similar case data to dynamically generate personalized alignment plans and adaptive rehabilitation suggestions, including details such as adjustment angles and timing, achieving a comprehensive approach from general... This represents an upgrade from orthodontic correction to personalized, precise guidance. The dynamic limb posture adjustment unit, through multi-actuator collaborative drive and multi-modal guidance feedback including voice, vision, and touch, ensures precise implementation of alignment movements and active user cooperation. Combined with a real-time calibration mechanism, it further enhances adjustment accuracy. The medical and health data sharing module, through encrypted transmission protocols and role-based access control models, enables secure data sharing and hierarchical access management across hospitals and rehabilitation institutions. It synchronously updates user rehabilitation progress and risk changes in real time, effectively protecting data privacy and security while improving cross-institutional collaboration efficiency. This enhances the standardization of wheelchair users' limb positioning, reduces the risk of complications, accelerates the rehabilitation process, and optimizes the quality of collaborative allocation of medical resources. Thus, it solves the problems of poor adaptability and sharing difficulties in existing technologies.

[0016] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0017] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a schematic diagram of a wheelchair user limb alignment guidance system based on medical and health big data sharing, according to an embodiment of this application. Figure 2 This is a schematic diagram of a limb state sensing module according to an embodiment of this application; Figure 3 This is a schematic diagram of a medical big data analysis center according to an embodiment of this application; Figure 4 This is a schematic diagram of an intelligent guidance decision engine provided according to an embodiment of this application; Figure 5 This is a schematic diagram of a limb posture dynamic adjustment unit provided according to an embodiment of this application; Figure 6 This is a schematic diagram of a medical and health data sharing module according to an embodiment of this application; Figure 7 This is a schematic diagram of a wheelchair user limb alignment guidance system based on medical and health big data sharing, according to an embodiment of this application. Figure 8 This is a flowchart illustrating a method for guiding wheelchair users to limb alignment based on the sharing of medical and health big data, according to an embodiment of this application. Figure 9 This is a schematic diagram of a method for guiding wheelchair users to limb alignment based on the sharing of medical and health big data, according to an embodiment of this application. Figure 10 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] The following describes an embodiment of the wheelchair user limb alignment guidance system based on medical and health big data sharing, with reference to the accompanying drawings. Addressing the poor adaptability issue mentioned in the background art, this application provides a wheelchair user limb alignment guidance system based on medical and health big data sharing. In this system, a limb state perception module accurately collects multi-dimensional data on limb joint angles, pressure distribution, and trunk posture, providing complete basic data for subsequent analysis and judgment. The medical big data analysis center, using a weighted fusion algorithm and machine learning model, performs deep fusion processing on multi-source data, accurately identifying potential health risks such as pressure sores and worsening scoliosis, locating abnormal force points and posture deviation sources, significantly improving the accuracy of risk prediction and abnormal location. The intelligent guidance decision engine deeply integrates preset rehabilitation rules, user historical data, and multi-center similar case data, dynamically generating data including angle adjustment, ... Personalized alignment plans and adaptive rehabilitation suggestions, including details such as timing, upgrade wheelchair users from generalized correction to personalized, precise guidance. The dynamic limb posture adjustment unit, through multi-actuator collaborative drive and multi-modal guidance feedback including voice, vision, and touch, ensures precise implementation of alignment movements and active user cooperation, further enhancing adjustment accuracy with a real-time calibration mechanism. The medical and health data sharing module, through encrypted transmission protocols and role-based access control models, enables secure data sharing and hierarchical access management across hospitals and rehabilitation institutions, synchronously updating user rehabilitation progress and risk changes in real time. This effectively protects data privacy and security while improving cross-institutional collaboration efficiency, enhancing the standardization of wheelchair users' limb positioning, reducing the risk of complications, accelerating the rehabilitation process, and optimizing the quality of collaborative allocation of medical resources. Thus, it solves the problems of poor adaptability and sharing difficulties in existing technologies.

[0020] Figure 1 This is a schematic diagram of a limb alignment guidance system for wheelchair users based on the sharing of medical and health big data, provided as an embodiment of this application.

[0021] This application provides a wheelchair user limb alignment guidance system based on the sharing of medical and health big data. The system 10 includes: The system includes a limb status perception module 100, a medical big data analysis center 200, an intelligent guidance and decision-making engine 300, a limb posture dynamic adjustment unit 400, and a medical and health data sharing module 500.

[0022] The system comprises the following components: a limb status perception module 100, which collects data on the limb joint angles, limb pressure distribution, and trunk posture of wheelchair users; a medical big data analysis center 200, which integrates these data to identify potential health risks such as pressure sores, worsening scoliosis, and muscle spasms caused by improper sitting posture, and locates abnormal force points and postural deviation sources; an intelligent guidance decision engine 300, which dynamically generates personalized limb alignment guidance plans and rehabilitation assistance suggestions based on potential health risks, abnormal force points, and postural deviation sources, combined with preset rehabilitation rules and user historical data; a limb posture dynamic adjustment unit 400, which drives the wheelchair adjustment mechanism based on the personalized limb alignment guidance plans and rehabilitation assistance suggestions, and provides simultaneous alignment guidance feedback; and a medical health data sharing module 500, which shares personalized limb alignment guidance plans and rehabilitation assistance suggestions, patient medical records, and potential health risks, abnormal force points, and postural deviation sources across institutions, simultaneously updates user status information, and manages access permissions.

[0023] It is understood that in this embodiment, the limb state perception module accurately collects multi-dimensional data on limb joint angles, pressure distribution, and trunk posture, providing complete basic data for subsequent analysis and judgment. The medical big data analysis center, using a weighted fusion algorithm and machine learning model, performs deep fusion processing on multi-source data, accurately identifying potential health risks such as pressure sores and worsening scoliosis, locating abnormal force points and postural deviation sources, and significantly improving the accuracy of risk prediction and abnormal location. The intelligent guidance decision engine deeply integrates preset rehabilitation rules, user historical data, and multi-center similar case data to dynamically generate personalized alignment plans and adaptive rehabilitation suggestions, including details such as adjustment angles and timing. The system has now been upgraded from generalized correction to personalized, precise guidance. The dynamic limb posture adjustment unit, through multi-actuator collaborative drive and multi-modal guidance feedback including voice, vision, and touch, ensures accurate implementation of alignment movements and active user cooperation. Combined with a real-time calibration mechanism, it further enhances adjustment accuracy. The medical and health data sharing module, through encrypted transmission protocols and role-based access control models, enables secure data sharing and hierarchical access management across hospitals and rehabilitation institutions. It synchronously updates user rehabilitation progress and risk changes in real time, effectively protecting data privacy and security while improving cross-institutional collaboration efficiency. This improves the standardization of wheelchair users' limb positioning, reduces the risk of complications, accelerates the rehabilitation process, and optimizes the quality of collaborative allocation of medical resources. Thus, it solves the problems of poor adaptability and sharing difficulties in existing technologies.

[0024] In this embodiment of the application, the limb state sensing module 100 includes: Figure 2 As shown, the components include a joint angle sensor group, a pressure sensor array, and a torso posture acquisition unit.

[0025] Among them, the joint angle sensor group is used to collect rotation angle data of shoulder, elbow, hip and knee joints in real time; the pressure sensor array is laid in the wheelchair seat and backrest area to collect pressure distribution data between the limbs and the contact surface; the trunk posture acquisition unit is used to capture trunk pitch angle, tilt angle and rotation angle data.

[0026] It is understood that the embodiments of this application use a joint angle sensor array to accurately capture the real-time rotation angles of the four major joints of the shoulder, elbow, hip, and knee, providing precise quantitative basis for limb movement trajectory analysis and functional assessment; the pressure sensor array uses distributed acquisition in the seat and backrest areas to dynamically monitor the distribution of limb contact pressure, providing data support for posture correction, pressure injury warning, and comfort optimization; the trunk posture acquisition unit captures the trunk pitch, tilt, and rotation angles in real time, accurately judging the body's balance state and posture abnormalities, improving the comprehensiveness, accuracy, and real-time nature of the perception of the human limb state.

[0027] For example, in intelligent rehabilitation training for stroke patients, the trunk posture acquisition unit can capture real-time three-dimensional posture data of the patient's trunk, including pitch, tilt, and rotation angles. When a patient exhibits excessive trunk pitch when standing due to insufficient core muscle strength, the acquisition unit can accurately capture the angle deviation value and fluctuation frequency. Compared to traditional manual observation, which can only qualitatively determine postural abnormalities, this unit can more quantitatively reflect balance control ability. If a patient exhibits compensatory trunk tilt or asymmetrical rotation during walking training, the acquisition unit can predict the risk of solidifying abnormal movements through changes in posture parameters. This dynamic posture data provides precise input for the motion simulation of the dynamic twin model of rehabilitation training, helping the intelligent rehabilitation analysis center to identify early signs of postural abnormalities, thereby adjusting the training plan in a timely manner and triggering posture correction reminders. This reduces secondary injuries caused by the solidification of incorrect postures and significantly improves the scientific rigor and safety of rehabilitation training.

[0028] In this embodiment of the application, the medical big data analysis center 200 includes: Figure 3 As shown, there are a data fusion processing unit, a health risk identification unit, and an anomaly location unit.

[0029] The data fusion processing unit uses a weighted fusion algorithm to perform spatiotemporal alignment and redundancy removal on limb joint angle, limb pressure distribution, and trunk posture data to generate a unified limb state dataset. The health risk identification unit identifies pressure ulcer risk levels, scoliosis progression trends, and the probability of muscle spasms based on preset thresholds and machine learning models. The abnormality localization unit analyzes pressure distribution hotspots and posture deviation data and uses correlation analysis algorithms to locate the specific locations of abnormal force points on the limbs and the sources of posture deviations in the trunk and limbs.

[0030] It is understood that the data fusion processing unit in this application embodiment performs spatiotemporal alignment and redundancy removal on limb joint angle, pressure distribution, and trunk posture data through a weighted fusion algorithm to generate a standardized limb state dataset. Compared with traditional data processing methods that can only analyze data in a single dimension, this unit can more comprehensively integrate the correlation of multi-source information. The health risk identification unit accurately identifies the risk level of pressure ulcers, the progression trend of scoliosis, and the probability of muscle spasm attacks based on preset thresholds and machine learning models, and performs quantitative assessment and trend prediction of potential health problems. The abnormality localization unit analyzes pressure distribution hotspots and posture deviation data, and uses correlation analysis algorithms to accurately locate abnormal force points on the limbs and the sources of posture deviations in the trunk and limbs, providing clear targets for targeted interventions, helping medical personnel to efficiently formulate intervention strategies, reduce the probability of complications, and improve the scientificity and accuracy of rehabilitation management.

[0031] It should be noted that when the health risk identification unit operates, it utilizes a collaborative mechanism based on preset thresholds and a machine learning model (where the preset thresholds are critical value ranges corresponding to various risks determined according to clinical diagnosis and treatment standards and health physiological indicators, and the machine learning model is an intelligent analysis model trained on historical health data and risk cases). Upon receiving limb state data (including pressure distribution, trunk posture, joint angles, etc.) from the data fusion processing unit, it performs data matching, analysis, and judgment for three types of health risks: For pressure ulcer risk level identification, it initially filters high-pressure-risk data by comparing limb pressure distribution data with preset "pressure ulcer risk thresholds" (e.g., single-point pressure > 32 mmHg and duration > 2 hours); then, combined with a machine learning classification model (trained on clinical pressure ulcer cases and pressure-injury correlation data), it integrates features such as pressure intensity and distribution evenness, mapping the results to low, medium, high, and extremely high pressure ulcer risk levels; simultaneously, leveraging the pressure-tissue damage correlation patterns in the model, it identifies high-risk pairs. The system identifies target points for stress intervention by analyzing corresponding body contact areas. For scoliosis progression trend identification, it compares continuous trunk posture data with a preset "normal trunk tilt angle fluctuation threshold" (e.g., static tilt angle > 5° is abnormal) to initially identify abnormal posture data. Then, it combines a machine learning regression model (trained with follow-up posture data of scoliosis patients) to fit a tilt angle change curve and calculate the progression rate. Simultaneously, it compares the rate with a preset "scoliosis progression critical threshold" (e.g., slope > 0.5° / month is rapid progression) to determine the scoliosis progression status, providing a basis for adjusting the intervention plan. For muscle spasm probability identification, it compares joint angle and posture dynamic data with a preset "abnormal fluctuation threshold" (e.g., joint fluctuation > 15° within 1 second is abnormal) to initially screen high-risk fluctuation data. Then, it combines a machine learning probability model (trained with limb data during spasm attacks) to output an attack probability value. Simultaneously, when the probability exceeds a preset "high attack risk threshold" (e.g., > 70%), it triggers an alert, providing a time window for preventative intervention. By establishing a baseline reference for risk assessment through preset thresholds, and leveraging machine learning models to achieve intelligent data analysis, abstract limb data is transformed into precise health risk information that can guide clinical intervention. When the abnormal location unit operates, based on the physiological correlation logic between pressure distribution hotspots and limb posture deviations (i.e., abnormal local force on a limb will trigger corresponding compensatory posture deviations of the trunk / limbs, and the two have a clear correspondence in spatial location and numerical change trend), after receiving limb pressure distribution data and trunk and limb posture data output by the data fusion processing unit, it will combine these data with a correlation analysis algorithm for matching and location: through the numerical characteristics of pressure distribution data (such as pressure amplitude and regional distribution pattern), it will identify "pressure hotspot areas" that exceed a preset threshold, and at the same time extract "posture deviation characteristics" that deviate from the normal physiological range from the posture data. The system identifies "symptoms" (such as excessive trunk tilt angles and abnormal joint angle fluctuations). Then, using a correlation analysis algorithm, it matches pressure hotspot areas with postural deviation features to pinpoint the specific location of the "abnormal force point" corresponding to the pressure hotspot in the limb's physical space. Based on the location information of this force point, and combined with the mechanical transmission laws of postural deviations (such as the trunk compensatory tilt logic caused by uneven local force), it traces the "postural deviation source" of the trunk and limbs that caused the deviation. Simultaneously, using the confidence results of the correlation analysis, it marks the correlation strength between the deviation source and the abnormal force point, providing precise locational basis for subsequent targeted posture correction and force adjustment.

[0032] Weighted fusion algorithm formula:

[0033] in, For a unified limb state dataset; These are the weighting coefficients for the joint angle data; This refers to joint angle data; Weighting coefficients for limb pressure distribution data; Data on limb pressure distribution; These are the weighting coefficients for the torso posture data; This is trunk posture data.

[0034] Specific machine learning model formulas:

[0035]

[0036]

[0037] in, For input features The probability of muscle spasm attacks; It is a feature vector composed of features such as joint angle fluctuation amplitude and trunk posture stability; The feature weight vector; For bias terms; It is a natural constant; This is the transpose operation of the feature weight vector; for Predicted trunk tilt angle at any given moment; The slope of scoliosis progression; For time; This is the initial roll angle; A comprehensive risk score for pressure ulcers; For the first The weight of each indicator; For the first The scores for each indicator; This refers to the index number.

[0038] Association analysis algorithm formula:

[0039] in, The confidence level of the association between pressure distribution anomaly A and attitude deviation B; This represents the probability of hotspot areas appearing in the pressure distribution. This represents the joint probability of both abnormal pressure and attitude deviation occurring simultaneously. This indicates an abnormal pressure distribution; This represents a postural deviation of the torso or limbs.

[0040] In this embodiment of the application, the intelligent guidance decision engine 300 includes: Figure 4 As shown, there are three units: guidance plan generation unit, rehabilitation suggestion adaptation unit, and medical data matching unit.

[0041] The guidance plan generation unit generates personalized limb alignment guidance plans based on potential health risk levels, abnormal stress points, and sources of postural deviations, combined with preset rehabilitation rules and user historical data. These plans include limb adjustment angles, timing, and duration of hold. The rehabilitation suggestion adaptation unit generates targeted muscle relaxation training and joint mobility assistance suggestions using a user feature adaptation algorithm, combined with the user's age, medical history, and rehabilitation stage. The medical data matching unit accesses a multi-center medical and health database and uses a similarity calculation algorithm to match intervention plans for similar cases, optimizing the current guidance plan and rehabilitation suggestions.

[0042] It is understood that the guidance scheme generation unit in this application accurately identifies the user's potential health risk level, abnormal force points, and postural deviation sources. Combining preset rehabilitation rules with the user's historical rehabilitation data, it generates a personalized limb alignment guidance scheme containing quantitative parameters such as limb adjustment angle, timing, and duration of hold. This provides directly executable operational guidelines for real-time posture correction, avoiding secondary injuries caused by blind adjustments. The rehabilitation suggestion adaptation unit, through a user feature adaptation algorithm, deeply integrates individual differences such as the user's age, medical history, and rehabilitation stage to generate targeted muscle relaxation training and joint mobility assistance suggestions. This achieves precise matching between the rehabilitation scheme and the user's physical condition, improving the safety and suitability of rehabilitation training. The medical data matching unit accesses a multi-center medical and health database and matches clinical intervention schemes for similar cases using a similarity calculation algorithm. This integrates clinically validated effective experience into the current guidance scheme and rehabilitation suggestions, breaking the limitations of single user data, improving the scientific rigor and effectiveness of the scheme, and providing data-driven auxiliary decision-making support for medical personnel, comprehensively improving the efficiency and quality of rehabilitation intervention.

[0043] It should be noted that the guidance plan generation unit first receives the potential health risk level (such as high risk of pressure ulcers, risk of scoliosis progression) output by the health risk identification unit, the abnormal force points on the limbs (such as pressure points on the right hip of the chair) and the sources of postural deviation (such as rightward trunk tilt caused by insufficient core muscle strength) determined by the abnormal positioning unit. At the same time, it retrieves the system's built-in preset rehabilitation rules (covering medical evidence such as clinical posture correction standards and pressure injury prevention guidelines) and the user's historical rehabilitation data (past posture adjustment records, feedback on the effectiveness of the plan adaptation, etc.). Then, it prioritizes the risk level and the abnormal location and deviation source (such as force points corresponding to high-risk pressure ulcers being included in the adjustment scope first). The system combines preset rules to determine the core objectives of the adjustment, and then uses verified adjustment parameters from user history data (such as the user's adaptability to a specific angle and the effect of maintaining the adjusted posture) to generate a personalized limb alignment guidance plan that is both targeted and feasible. The plan clearly includes precise limb adjustment angles (such as 4° hip abduction and 3° trunk tilt correction), scientific adjustment timing (such as active adjustment every 25 minutes and immediate triggering when the pressure value exceeds the threshold), and reasonable holding time (such as maintaining the adjusted posture for 12 minutes). This avoids the problem of insufficient adaptability of general plans and ensures that users can accurately execute the plan through quantitative indicators, effectively reducing the probability of health risk deterioration.

[0044] User feature adaptation algorithm formula:

[0045]

[0046] in, These are the standardized user feature values; Original values ​​for user characteristics; This is the minimum reference value for this feature; This is the maximum reference value for this feature; Rate the fit between the user and the rehabilitation recommendations; The weighting coefficients for age characteristics; The standardized user age; The weighting coefficients for medical history characteristics; Standardized medical history scoring; The weighting coefficients for the rehabilitation phase; These are the standardized quantitative values ​​for the rehabilitation stage.

[0047] Similarity calculation algorithm formula:

[0048]

[0049] in, The weighted Euclidean distance between the current user and the database cases; Indexed by feature dimensions; This represents the total number of feature dimensions. For the first The weight coefficients of each feature; For the current user Standardized values ​​of each feature; For the database case number Standardized values ​​of each feature; This represents the cosine similarity between the current user and the case.

[0050] When the medical data matching unit operates, it uses cross-institutional case association logic based on a multi-center medical and health database and current user data (i.e., the multi-center database integrates massive rehabilitation cases from medical institutions of different regions and levels; each case includes user characteristics, health risk type, intervention plan, and clinical effect feedback, providing practically validated rehabilitation experience references). Upon receiving the initial guidance plan and rehabilitation suggestions for the current user from the intelligent guidance decision engine, as well as data such as the user's limb status and health risks generated by the previous data fusion processing unit, it combines this data with a similarity calculation algorithm for case matching and plan optimization: First, the core characteristics of the current user (age, medical history, rehabilitation stage, potential health risk level, abnormal stress points, etc.) are standardized according to preset standards (eliminating dimensional differences between different characteristics through feature standardization formulas); then, a similarity calculation algorithm (quantifying the numerical differences in features between the current user and the case using weighted Euclidean distance, and matching their features using cosine similarity) is used. To ensure consistency in trends, the system traverses case data from a multi-center database, calculates the overall similarity between the current user and each case, and filters out "highly matched cases" with similarity exceeding a preset threshold (e.g., overall similarity ≥ 0.8). It then extracts the intervention plans corresponding to these highly matched cases and, combined with case feedback data (e.g., rehabilitation effectiveness, user tolerance, complication rate), analyzes the core strategies in the plans that are suitable for the current user (e.g., effective limb adjustment frequency in similar high-risk pressure ulcer cases, optimized joint activity intensity in similar progressive scoliosis cases). These clinically validated strategies are then integrated into the current personalized limb alignment guidance plan and rehabilitation recommendations. For example, the duration of limb adjustments is adjusted to the optimal range in similar cases, or muscle relaxation training movements that received positive user feedback in similar cases are added. Simultaneously, precautions in similar case plans are noted (e.g., training intensities that users with a history of certain chronic diseases should avoid), further improving the safety and suitability of the plan.

[0051] For example, in the scenario of stroke patient rehabilitation management, taking a 62-year-old patient with a 5-year history of hypertension and in the mid-stage of 3 months of post-stroke rehabilitation as an example, the rehabilitation suggestion adaptation unit operates accurately and is tailored to individual characteristics: The unit first receives core feature information such as the patient's age, history of hypertension, and rehabilitation stage. Using the feature standardization formula in the user feature adaptation algorithm, it standardizes 62 years old (reference age range 18-80 years) to 0.71, hypertension history to 0.6, and mid-stage of rehabilitation to 0.6; then, it calculates the appropriate recommendations based on preset weights (medical history 0.4, age 0.3, rehabilitation stage 0.3). The matching score yielded a comprehensive fit score of 0.62. Based on this score, the unit was matched with the corresponding rehabilitation suggestion library. Considering the patient's history of hypertension and the need to avoid strenuous movements, a muscle relaxation program was generated: "Twice daily, 10 minutes each time, gentle neck and shoulder stretching exercises (range of motion controlled within 30°)." Simultaneously, taking into account the joint tolerance corresponding to age 62 and the functional recovery needs in the mid-stage of rehabilitation, a joint mobility assistance suggestion was generated: "Three sets of slow-paced wrist flexion and extension exercises daily, eight repetitions per set, plus 5 minutes of local heat application after training," with the added consideration of "blood pressure must be controlled below 140 / 90 mmHg during training." This set of suggestions not only suits the patient's physical condition but also avoids the risks associated with hypertension, making rehabilitation training safer and more suited to the patient's recovery pace.

[0052] In this embodiment, the limb posture dynamic adjustment unit 400 includes, as follows: Figure 5 As shown, there is an adjustment mechanism drive unit, a guidance feedback providing unit, and an attitude calibration unit.

[0053] The adjustment mechanism drive unit receives personalized alignment guidance and drives the wheelchair seat lifting, backrest angle adjustment, and tilt compensation actuator through trajectory planning algorithms. The guidance feedback unit guides the user to actively cooperate with the alignment action through voice prompts, flashing visual indicator lights, and tactile vibration feedback. The posture calibration unit collects the adjusted limb state data in real time, compares it with the target posture parameters, and calibrates the adjustment accuracy.

[0054] It is understood that, in this embodiment, the adjustment mechanism drive unit receives personalized alignment guidance plans and uses a trajectory planning algorithm to transform abstract parameters such as limb adjustment angles and positions in the plan into precise mechanical execution trajectories. This drives the seat lifting, backrest angle adjustment, and tilt compensation mechanisms to complete the adaptation actions, replacing the crude operation of manual adjustment and achieving quantitative precision in posture adjustment, avoiding errors from manual operation. The guidance feedback unit guides the user to actively cooperate with the alignment actions through multimodal methods such as voice prompts, flashing visual indicator lights, and tactile vibration feedback, linking the passive adjustment of the device with the active cooperation of the user, improving the smoothness of posture adjustment and the user's sense of participation. The posture calibration unit collects limb state data after adjustment in real time and compares it with the target posture parameters in the plan. Once a deviation is detected, the actuator is fine-tuned to continuously calibrate the adjustment accuracy, ensuring that the posture always conforms to the rehabilitation needs. This not only improves the accuracy and efficiency of posture adjustment but also enhances the user's cooperation experience, while reducing the operational burden on caregivers and effectively assisting in the correction and maintenance of limb posture.

[0055] It should be noted that the trajectory planning algorithm formula is as follows:

[0056]

[0057]

[0058] in, The target position / angle of the actuator at time t; Time planning for the trajectory; The coefficient of the zeroth term of the polynomial; The coefficient of the linear term in the polynomial; The coefficients of the quadratic term in the polynomial; The coefficients of the cubic terms in the polynomial; The coefficients of the fourth-degree term in the polynomial; The coefficients of the fifth-degree term in the polynomial; The position / angle of the actuator at the initial moment (t=0); The initial position / angle of the actuator; The velocity of the actuator at the initial moment (t=0); The initial velocity of the actuator; The acceleration of the actuator at the initial moment (t=0); The initial acceleration of the actuator; The position / angle of the actuator at the target time (t=T); T is the total duration of a single adjustment. The target position / angle of the executing agency; The speed of the actuator at the target time (t=T); The target speed for the implementing agency; The acceleration of the actuator at the target time (t=T); To accelerate the objectives of the implementing agency.

[0059] The posture calibration unit first performs multi-dimensional comparison and error quantification analysis on the three types of information: the adjusted limb state data (such as trunk pitch / tilt angles acquired by the inertial measurement unit (IMU), seat / backrest contact pressure distribution acquired by the flexible pressure sensor, and limb center of gravity coordinates and joint postures captured by the visual depth camera), the preset target posture parameters of the personalized alignment guidance plan (such as the trunk neutral angle for scoliosis patients, the center of gravity balance coordinates for lower limb paralysis patients, and the limb contact posture threshold for postoperative rehabilitation patients), and the execution feedback data of the adjustment mechanism (such as seat lifting displacement, backrest adjustment angle, and tilt compensation amount). If the deviation between the actual limb posture and the target posture (such as angle error ≤ 0.5°, center of gravity offset ≤ 5mm) is within the preset accuracy threshold, it can be determined that the adjustment is qualified and no additional calibration is required; if the deviation exceeds the threshold, further analysis is performed based on the execution feedback data. According to the source of the deviation, if the deviation originates from the displacement error of the actuator caused by the mechanical transmission clearance, it may point to the deviation of the motor step angle or the wear of the transmission components. If the deviation is accompanied by abnormal fluctuations in pressure distribution, it may point to the user's autonomous micro-movement of the limbs or the deviation in posture perception. By distinguishing the type and degree of deviation, for systematic deviations caused by mechanical transmission errors, precise calibration instructions are generated to fine-tune the trajectory planning algorithm parameters (such as correcting the target position of the fifth-order polynomial interpolation and optimizing the PID feedback compensation coefficient), driving the corresponding actuator to perform secondary fine-tuning. For dynamic deviations caused by the user's micro-movement of the limbs, the linkage guidance feedback unit generates voice prompts, tactile vibrations and other collaborative guidance instructions to guide the user to adjust the limb posture, effectively offsetting the influence of various interference factors on the accuracy of posture adjustment, ensuring the accurate implementation of personalized posture guidance programs, and improving the standardization and comfort of posture in rehabilitation training or daily use.

[0060] For example, in a scenario where a 58-year-old scoliosis patient uses a smart rehabilitation wheelchair for posture correction, the guidance feedback unit precisely guides the patient to actively adjust their posture through multimodal collaborative feedback: when the posture calibration unit detects that the patient's torso is tilted to the left by more than 5° from the target angle, the unit immediately triggers a synchronous feedback mechanism. The voice module plays a gentle prompt in real time, "Please slowly adjust your posture to the right, keeping your back against the backrest," while the green indicator light on the right armrest of the wheelchair flashes at a frequency of 1 time per second, visually indicating the direction of adjustment; the tactile vibration module built into the left seat cushion vibrates continuously at a moderate intensity, precisely corresponding to the left side area where the patient experiences abnormal force, enhancing the adjustment guidance through tactile perception. Based on the synchronous prompts of voice, vision, and touch, the patient actively and slowly adjusts their torso to the right. When the posture calibration unit detects that the torso tilt angle has returned to the target range, the unit automatically stops vibrating and the indicator light flashing, and the voice module immediately provides feedback, "Posture has been calibrated, please maintain the current posture," completing one closed-loop guidance cycle. Throughout the process, multimodal feedback not only avoids the perceptual blind spots that may exist in single feedback methods, but also enables patients to quickly understand adjustment needs and actively participate in posture correction through precise directional prompts and tactile positioning. This not only improves adjustment efficiency, but also enhances patients' initiative and cooperation in the rehabilitation process.

[0061] In this embodiment of the application, the medical and health data sharing module 500 includes, as follows: Figure 6 As shown, there are three units: data sharing unit, access control unit, and status synchronization unit.

[0062] The data sharing unit is used to share personalized limb alignment guidance plans, rehabilitation assistance suggestions, core patient medical records and risk assessment results across hospitals and rehabilitation institutions through encrypted transmission protocols, ensuring privacy and security during transmission; the access control unit is used to set different data access permissions for medical staff, rehabilitation therapists and patients' families based on a role-based access control model; and the data synchronization and update unit is used to synchronize user posture adjustment effect data, health risk change data and rehabilitation progress data in real time, ensuring data consistency during cross-institutional collaboration.

[0063] Understandably, the data sharing unit in this application embodiment securely shares personalized limb alignment guidance plans, rehabilitation assistance suggestions, core medical record data, and risk assessment results across hospitals and rehabilitation institutions based on an encrypted transmission protocol. This breaks down data barriers between institutions while strengthening patient privacy protection at the transmission level. The access control unit, based on a role-based access control model, precisely assigns differentiated data access permissions to medical staff, rehabilitation therapists, and patients' families. This ensures that medical staff can obtain complete diagnostic and treatment data to develop scientific plans while preventing unauthorized access to sensitive data, achieving a balance between compliance and security in data use. The status synchronization unit synchronizes dynamic data such as the effect of user posture adjustment, changes in health risks, and rehabilitation progress in real time. This ensures that the data obtained by each entity remains consistent during cross-institutional collaboration, avoiding conflicts in diagnostic and treatment suggestions due to data lag or deviation. This improves the efficiency and accuracy of multidisciplinary team collaboration, allowing patients to receive consistent personalized services when transferred or continuing rehabilitation between different institutions.

[0064] It should be noted that the role-based access control model is as follows:

[0065]

[0066]

[0067]

[0068]

[0069] in, Assign functions for role-permission roles; A collection of roles in a medical setting, including (Medical staff) (Rehabilitation Therapist) (Patient's family member); This is a set of medical data operation permissions, including (Read) (Revise), (shared), (View); 1 indicates that permission assignment is in effect, 0 indicates that no assignment has been made; For permission-resource association sets; A collection of medical data resources, including (Personalized guidance plan for body alignment) (Rehabilitation support suggestions) (Core data in patient medical records) (Risk assessment results); Indicates permissions Resources Perform the corresponding operation; This is an access permission verification function; For specific operational actions (such as reading medical records or modifying rehabilitation plans); Functions that allow child roles to inherit permissions from their parent roles; For a sub-role (such as a resident physician); The role of a father (e.g., chief physician); This is a permission constraint function; The number of permissions a character has.

[0070] For example, taking the scenario of stroke patient Ms. Zhang being transferred from an acute phase treatment at a tertiary hospital to a community rehabilitation center, the data synchronization and update unit achieves real-time unified connection of rehabilitation data across institutions: core data recorded at the hospital, such as "80% compliance rate for shoulder and neck adjustment angles," "pressure ulcer risk reduced from intermediate to low," and "60% completion of acute phase rehabilitation," are synchronized to the community rehabilitation center's management platform after the unit completes format standardization conversion and hash integrity verification; when the rehabilitation center monitors a temporary fluctuation in Ms. Zhang's posture adjustment angle (slightly conflicting with the hospital's historical records), the unit uses a timestamp priority arbitration mechanism to prioritize... The latest clinically confirmed data from the hospital was used as the benchmark for calibration. Subsequently, the "posture adjustment effect data (synchronized at the second level)" and "rehabilitation progress data (synchronized at the minute level)" were simultaneously pushed to the community rehabilitation therapist's workstation, the hospital neurologist's remote diagnosis and treatment system, and the mobile terminal of Ms. Zhang's family. The community rehabilitation therapist directly adjusted the training intensity based on the hospital's historical data, the hospital physician remotely viewed the real-time progress without repeated assessments, and the family kept track of the rehabilitation progress. This not only avoided the program conflicts caused by information gaps between institutions, but also enabled seamless connection of Ms. Zhang's rehabilitation services between the hospital and the community, ensuring the continuity and accuracy of the rehabilitation process.

[0071] This application proposes a limb alignment guidance system for wheelchair users based on shared medical and health big data. The system uses a limb state perception module to accurately collect multi-dimensional data on limb joint angles, pressure distribution, and trunk posture, providing comprehensive foundational data for subsequent analysis and judgment. A medical big data analysis center, employing a weighted fusion algorithm and machine learning model, deeply integrates multi-source data to accurately identify potential health risks such as pressure sores and worsening scoliosis, locating abnormal stress points and postural deviation sources, significantly improving the accuracy of risk prediction and abnormality localization. An intelligent guidance decision engine deeply integrates preset rehabilitation rules, user historical data, and multi-center similar case data to dynamically generate personalized alignment systems that include details such as adjustment angles and timing. The proposed solutions and adaptive rehabilitation recommendations represent an upgrade from generalized correction to personalized, precise guidance. The dynamic limb posture adjustment unit, through multi-actuator collaborative drive and multi-modal guidance feedback including voice, vision, and touch, ensures precise implementation of alignment movements and active user cooperation, further enhancing adjustment accuracy with a real-time calibration mechanism. The medical and health data sharing module, through encrypted transmission protocols and role-based access control models, enables secure data sharing and tiered access management across hospitals and rehabilitation institutions. It synchronously updates user rehabilitation progress and risk changes in real time, effectively protecting data privacy and improving cross-institutional collaboration efficiency. This enhances the standardization of limb positioning for wheelchair users, reduces the risk of complications, accelerates the rehabilitation process, and optimizes the quality of collaborative allocation of medical resources. Thus, it solves the problems of poor adaptability and sharing difficulties in existing technologies.

[0072] The following will illustrate a specific embodiment of a wheelchair user limb alignment guidance system based on the sharing of medical and health big data. Figure 7 As shown, it includes: In the rehabilitation and nursing ward of a public nursing home, Mr. Li, a 68-year-old patient with sequelae of stroke, has been using a regular wheelchair for daily activities. Due to improper sitting posture, he has concentrated pressure on his left shoulder and mild scoliosis. Traditional manual care methods are difficult to monitor his posture changes in real time, and rehabilitation plans lack personalized dynamic adjustments, which can easily lead to complications such as pressure sores and muscle spasms. Therefore, the nursing home introduced a wheelchair user limb alignment guidance system based on the sharing of medical and health big data to achieve intelligent rehabilitation management throughout the entire process.

[0073] The limb status perception module serves as the system's data entry point. It consists of a joint angle sensor group, a pressure sensor array, and a trunk posture acquisition unit, forming a multi-dimensional limb data acquisition network. The joint angle sensor array uses miniature MEMS sensors, worn on Grandpa Li's shoulder, elbow, hip, and knee joints respectively. It collects rotation angle data of each joint in real time via Bluetooth 5.0 protocol (measurement range 0°~180°, accuracy ±0.5°), with a sampling frequency of 10Hz, which can accurately capture subtle changes in limb movement. The pressure sensor array uses flexible thin-film pressure sensors, laid in a 20×20 matrix on the wheelchair seat and backrest area. The single sensor range is 0~500kPa, with a resolution of 1kPa, which can collect pressure distribution data of limbs and contact surfaces in real time and identify areas of concentrated pressure (pressure value ≥300kPa is judged as a high-risk area). The trunk posture acquisition unit integrates a three-axis accelerometer and a gyroscope, worn on Grandpa Li's back belt, with a sampling frequency of 50Hz, which can accurately capture trunk pitch angle (measurement range -30°~30°), tilt angle (measurement range -20°~20°), and rotation angle (measurement range -45°~45°) data, with an accuracy of ±0.1°. This data is transmitted to edge computing nodes via the nursing home's internal 5G private network. After preprocessing (filtering, noise reduction, and format conversion), it is uploaded to the cloud-based medical and health database and simultaneously synchronized to the nursing home's care terminal and the mobile devices of Mr. Li's children, ensuring the real-time nature and integrity of the data and providing high-quality data support for subsequent analysis and decision-making.

[0074] The medical big data analysis center, through the collaborative work of a data fusion processing unit, a health risk identification unit, and an anomaly localization unit, achieves in-depth analysis of limb status data and accurate identification of health risks. The data fusion processing unit employs a weighted fusion algorithm to perform spatiotemporal alignment and redundancy removal on three heterogeneous data types: joint angles, pressure distribution, and trunk posture. Using the timestamp of trunk posture data as a benchmark, a linear interpolation algorithm synchronizes joint angle and pressure data at different sampling frequencies (time synchronization error ≤10ms). Weights for each data dimension are dynamically allocated based on the entropy weight method (pressure distribution weight 0.4, trunk posture weight 0.35, joint angle weight 0.25), generating a unified limb status dataset. The health risk identification unit, based on preset thresholds and an improved random forest classification model, quantifies three types of potential health risks: pressure ulcer risk is calculated based on the duration and peak value of pressure distribution data (pressure ≥300kPa lasting more than 20 minutes is considered high risk), spinal... The severity of scoliosis was assessed by the cumulative deviation of trunk tilt angle (a mean tilt angle ≥5° over 1 hour was considered progressive). The probability of muscle spasm attacks was predicted by the frequency of joint angle mutations (≥5 angle mutations within 10 minutes were considered high probability). After training with 50,000 sets of clinical rehabilitation data, the model achieved a risk identification accuracy of 92%. The abnormal localization unit analyzed pressure distribution hotspots and posture deviation data, and used the Apriori association analysis algorithm to mine the association rules between pressure concentration areas and posture deviations (confidence ≥0.85), locating the specific location of abnormal force points in the limbs (localization error ≤2cm) and the sources of posture deviations in the trunk and limbs. For example, it identified a strong correlation between the pressure concentration in Grandpa Li's left shoulder and the leftward tilt of the trunk, accurately locating the source of deviation as uneven force caused by trunk tilt.

[0075] The intelligent guidance decision engine consists of a guidance plan generation unit, a rehabilitation suggestion adaptation unit, and a medical data matching unit, enabling the precise generation and optimization of personalized rehabilitation guidance plans. The guidance plan generation unit, based on the potential health risk level output by the health risk identification unit, the abnormal force points and postural deviation sources determined by the abnormal location unit, and combined with preset rehabilitation rules and Mr. Li's historical rehabilitation data, uses a multi-objective optimization algorithm to generate a personalized limb alignment guidance plan that includes limb adjustment angles, timing, and holding duration. For example, addressing Mr. Li's left shoulder pressure concentration and trunk tilting issues, it generates a specific plan: "Adjust sitting posture every 30 minutes, tilt the trunk 3° to the right, relax the left shoulder, and hold for 5 minutes," with a plan generation time of ≤1 second. The rehabilitation suggestion adaptation unit, through a user feature adaptation algorithm, deeply integrates Mr. Li's age (68 years old), medical history (post-stroke sequelae), rehabilitation stage (recovery period), and other individual differences to generate targeted muscle relaxation training and joint mobility assistance. Suggestions include, for example, "Perform a pendulum exercise on the left shoulder twice a day for 10 minutes each time (swing amplitude 30°)" and "Use hot compresses (temperature 40℃~45℃) to promote local blood circulation." The system also calculates the suitability score of each suggestion based on the analytic hierarchy process (suggestions with a suitability score ≥0.8 are given priority). The medical data matching unit accesses a national multi-center medical and health database for stroke rehabilitation (containing 100,000+ similar cases). Using a weighted Euclidean distance similarity calculation algorithm, it matches cases with a similarity score ≥0.8 to Mr. Li's age, medical history, and rehabilitation stage, extracting effective intervention programs (such as "sitting posture adjustment combined with shoulder stretching training can reduce the risk of pressure ulcers by 65%)." These clinically validated experiences are integrated into the current guidance plan, optimizing the adjustment frequency and duration, significantly improving the scientific validity and effectiveness of the plan.

[0076] The limb posture dynamic adjustment unit, through the collaboration of the adjustment mechanism drive unit, the guidance feedback unit, and the posture calibration unit, achieves precise execution and real-time calibration of personalized guidance plans. After receiving the guidance plan, the adjustment mechanism drive unit uses a cubic polynomial trajectory planning algorithm to convert parameters such as adjustment angles and holding times into a smooth motion trajectory for the wheelchair actuators: driving the seat lifting motor (lifting range 0~10cm, speed 0.5cm / s) to adjust the seat height, driving the backrest angle adjustment motor (adjustment range 90°~120°, speed 1° / s) to adjust the backrest tilt, and driving the tilt compensation actuator (compensation range -5°~5°, speed 0.5° / s) to correct the sitting posture, ensuring a smooth and impact-free movement (acceleration ≤0.2m / s²). The guidance feedback unit simultaneously guides Grandpa Li to actively cooperate with the correction movements through multimodal methods such as voice prompts, flashing visual indicator lights, and tactile vibration feedback. The voice module plays a gentle prompt, "Please slowly adjust your sitting posture to the right." The green indicator light on the right armrest of the wheelchair flashes at a frequency of 1 time per second to indicate the adjustment direction. The vibration module built into the left seat cushion vibrates continuously at a medium intensity (vibration frequency 50Hz), accurately corresponding to abnormal force areas, with a feedback response time of ≤0.5 seconds. The posture calibration unit collects the adjusted limb state data in real time and compares it with the target posture parameters. It uses a PID control algorithm to calculate the posture deviation and generate a compensation amount (compensation accuracy ±0.1°). When the trunk tilt angle deviation is detected to exceed 1°, the tilt compensation actuator is automatically fine-tuned to ensure that the posture always conforms to the target parameters, with an adjustment accuracy error of ≤0.5°, realizing a closed-loop adjustment of "precise drive - multimodal guidance - real-time calibration".

[0077] The medical and health data sharing module achieves secure sharing and real-time synchronization of rehabilitation data across institutions through the linkage of a data sharing unit, an access control unit, and a status synchronization unit. The data sharing unit uses a transmission protocol based on the SM4 encryption algorithm to securely share Mr. Li's personalized limb alignment guidance plan, rehabilitation assistance suggestions, core medical record data, and risk assessment results to the neurology departments of his contracted tertiary hospitals, community rehabilitation centers, and nursing home care systems. The encrypted transmission rate is ≥10Mbps, reducing the risk of privacy leakage during data transmission to zero. The access control unit, based on the RBAC role-based access control model, sets differentiated data access permissions for different collaborating entities: the hospital's neurologist has full access to read and modify the data; the community rehabilitation therapist has access to view the plan and provide adjustment suggestions; and Mr. Li's family members only have access to view the rehabilitation progress. The access verification response time is ≤0.1 seconds, ensuring the compliance and security of data use. Status synchronization... The unit employs a timestamp priority arbitration mechanism to synchronize Mr. Li's posture adjustment effect data, health risk change data, and rehabilitation progress data in real time. High-frequency dynamic data such as posture adjustment effect is synchronized at the second level, while phased summary data such as rehabilitation progress is synchronized at the minute level. When data conflicts occur (such as the pressure ulcer risk level recorded by the nursing home being inconsistent with the hospital's real-time monitoring data), the system automatically calibrates based on the latest clinically validated data to ensure complete consistency of data between the hospital, rehabilitation center, nursing home, and family's mobile devices, with a synchronization error of ≤5 seconds. This completely breaks down data silos in cross-institutional collaboration, allowing Mr. Li to receive a coherent and unified full-process rehabilitation service when transferred between nursing homes, community rehabilitation centers, and hospitals, ensuring the continuity and safety of the rehabilitation process.

[0078] In summary, this application's embodiments collect limb data through a limb state perception module, providing precise input. The medical big data analysis center performs data parsing, risk identification, and anomaly localization, solving the problem of traditional manual care's difficulty in real-time posture monitoring. This significantly reduces the risk of complications such as pressure sores and muscle spasms caused by improper sitting posture. The intelligent guidance decision engine generates personalized posture correction plans and integrates experience from similar cases, avoiding the drawbacks of rehabilitation plans lacking dynamic adjustment and improving the scientific rigor and adaptability of the plans. The limb posture dynamic adjustment unit improves the accuracy of sitting posture adjustment and user cooperation through precise driving, multimodal guidance, and real-time calibration, enhancing the actual effect of posture correction. The medical and health data sharing module enables secure sharing and real-time synchronization across institutions, breaking down data silos, ensuring consistent and unified rehabilitation services across institutions, guaranteeing the continuity and safety of the rehabilitation process, reducing the burden on nursing staff, and improving rehabilitation efficiency and service quality.

[0079] Next, referring to the accompanying drawings, a method for guiding wheelchair users to correct limb alignment based on the sharing of medical and health big data is described according to an embodiment of this application.

[0080] like Figure 8As shown, this method for guiding wheelchair users in limb alignment based on the sharing of medical and health big data includes the following steps: In step S101, data on the limb joint angles, limb pressure distribution, and trunk posture of the wheelchair user are acquired.

[0081] It is understood that the embodiments of this application acquire data on the joint angles, pressure distribution, and trunk posture of wheelchair users to monitor their limb activity, contact force, and trunk posture deviations in real time. This allows for the accurate capture of abnormal force points, sources of posture deviation, and potential health risk signs caused by improper sitting posture. This provides high-quality data for subsequent medical big data fusion processing, health risk identification and abnormal location, generation and dynamic adjustment of personalized limb alignment guidance plans. Real-time perception and early intervention of the user's limb posture can reduce complications such as pressure sores and worsening of scoliosis caused by improper sitting posture. This lays a solid data foundation for the intelligent rehabilitation management of wheelchair users throughout the entire process and improves the accuracy and foresight of rehabilitation guidance.

[0082] In step S102, the data on the limb joint angles, limb pressure distribution, and trunk posture of the wheelchair user are fused and processed. Based on preset thresholds and machine learning models, the risk of pressure sores, abnormal spinal load, and potential health risks of muscle fatigue caused by improper sitting posture are assessed. The abnormal force points of the limbs and the sources of posture deviations are located using correlation analysis algorithms.

[0083] Among them, association analysis algorithm is a data mining algorithm that can accurately identify the potential association between abnormal force points of limbs and postural deviation sources in medical rehabilitation scenarios by mining the association rules (such as frequent itemsets and causal relationships) between different variables in the dataset. This provides data support for health risk positioning and personalized rehabilitation plan generation, and improves the accuracy and scientific nature of rehabilitation intervention.

[0084] It is understood that the embodiments of this application, by employing correlation analysis algorithms, can deeply mine the potential correlation rules between multi-dimensional data such as limb joint angles, pressure distribution, and trunk posture of wheelchair users, accurately locate the causal relationship between abnormal force points of limbs and sources of posture deviation, and provide clear data support for tracing the source of health risks; at the same time, it avoids the limitations of viewing single data in isolation in traditional analysis, making health risk assessment more targeted, providing a scientific basis for the accurate generation of subsequent personalized limb alignment guidance plans, effectively improving the targeting and effectiveness of rehabilitation intervention, and reducing the risk of complications such as pressure sores and abnormal spinal load caused by improper sitting posture.

[0085] For example, taking the wheelchair posture monitoring scenario of Mr. Li, a 68-year-old patient with sequelae of stroke, as an example, the application of the association analysis algorithm accurately realized data association mining and deviation source localization: The system first collected real-time data of Mr. Li's trunk tilt angle, left shoulder pressure distribution, shoulder joint angle, and other dimensions. By mining the frequent itemsets and association rules of these data, the association analysis algorithm found that when the trunk tilted to the left by ≥4°, the confidence level of the left shoulder pressure ≥300kPa reached 0.92, and the support level of the two occurring simultaneously reached 0.88. At the same time, it also identified a strong correlation between the limited shoulder joint range of motion and the pressure concentration under this posture. Based on these association results, the system accurately located the core deviation source of Mr. Li's left shoulder pressure concentration as trunk tilt, rather than simply improper limb positioning. This provided clear data support for the subsequent generation of a personalized alignment plan combining "trunk tilt correction + shoulder relaxation," avoiding the misjudgment of deviation sources caused by viewing single data in isolation in traditional analysis, and improving the accuracy of health risk positioning and the pertinence of rehabilitation plans.

[0086] In step S103, based on potential health risks, abnormal force points on limbs, and sources of postural deviations, and combined with a multi-center medical and health database, a personalized limb alignment guidance plan and rehabilitation assistance suggestions are dynamically generated. Under user confirmation and safety monitoring, the wheelchair adjustment mechanism is driven to make adjustments and provide alignment guidance feedback according to the personalized limb alignment guidance plan and rehabilitation assistance suggestions. At the same time, the personalized limb alignment guidance plan and rehabilitation assistance suggestions, patient medical record data, and potential health risks, abnormal force points on limbs, and sources of postural deviations are shared securely across institutions, and user status information is updated synchronously.

[0087] Among them, the multi-center medical and health database is a distributed database that integrates diverse medical and health data such as case data, rehabilitation plans, and risk assessment results from medical institutions and rehabilitation institutions in different regions and at different levels.

[0088] It is understood that the embodiments of this application utilize a multi-center medical and health database, which can integrate massive amounts of case data, rehabilitation intervention plans, and risk assessment results of stroke, scoliosis, etc., from medical institutions and rehabilitation institutions at different levels and in different regions across the country. This provides clinically validated data support for the generation of personalized limb alignment guidance plans and rehabilitation assistance suggestions. By matching cases that are highly similar to the current user's age, medical history, and rehabilitation stage, effective intervention strategies are extracted and integrated into the plan, significantly improving the scientificity, effectiveness, and clinical suitability of the plan, and avoiding plan deviations caused by the limited experience of a single institution. At the same time, relying on the distributed architecture of the database, cross-institutional data secure sharing is achieved, breaking down data silos between hospitals, rehabilitation centers, and nursing homes. This ensures that when users are transferred or continue rehabilitation, all collaborating entities can obtain unified rehabilitation data, guaranteeing the continuity and consistency of rehabilitation services, reducing duplicate assessments and plan conflicts, improving the efficiency of cross-institutional rehabilitation collaboration, and providing wheelchair users with full-process, high-quality, personalized rehabilitation services.

[0089] It should be noted that personalized limb alignment guidance plans and rehabilitation assistance suggestions, patient medical records and potential health risks, abnormal limb stress points and postural deviation sources are securely shared across institutions, and user status information is updated synchronously. Personalized limb alignment guidance plans, rehabilitation assistance suggestions (such as adjusting angles and training movements), core patient medical record data (medical history, rehabilitation stage) and potential health risks (pressure ulcer level, scoliosis progression), abnormal limb stress points (such as the pressure concentration area of ​​the left shoulder), and postural deviation sources (trunk tilt) are encrypted using the SM4 encryption transmission protocol to ensure the privacy and security of data during transmission across hospitals, rehabilitation institutions, and nursing homes, and to prevent the leakage of sensitive information. At the same time, based on the RBAC role-based access control model, differentiated access permissions are assigned to different collaborating entities. Hospital physicians can view and modify complete data to optimize treatment plans, rehabilitation therapists can only access plans and risk data to adjust training plans, and patient family members can only view rehabilitation progress, achieving compliant control over data use. In the synchronization update phase, a timestamp priority arbitration mechanism is adopted to perform layered synchronization of dynamic data such as user posture adjustment effects (second-level synchronization), changes in health risks (minute-level synchronization), and rehabilitation progress (hour-level synchronization). When cross-institutional data conflicts occur (such as inconsistencies between pressure ulcer risk recorded by nursing homes and real-time monitoring data from hospitals), the latest clinically validated data is automatically used as the benchmark for calibration, ensuring complete consistency of data across hospitals, rehabilitation centers, nursing homes, and family mobile devices. In addition, synchronized data is pushed to the information systems and smart rehabilitation equipment of each institution in real time, enabling cross-institutional collaborators to formulate and adjust rehabilitation plans based on the same set of real-time and accurate data. This completely breaks down information silos in the traditional model, eliminating the need for repeated assessments when patients are transferred or continue rehabilitation, achieving seamless integration of rehabilitation services. This not only ensures the continuity and safety of the rehabilitation process but also improves the efficiency and accuracy of cross-institutional collaboration.

[0090] According to the embodiments of this application, a method for guiding wheelchair users to limb alignment based on medical and health big data sharing is proposed. This method uses a limb state perception module to accurately collect multi-dimensional data on limb joint angles, pressure distribution, and trunk posture, providing complete basic data for subsequent analysis and judgment. The medical big data analysis center uses a weighted fusion algorithm and machine learning model to deeply fuse multi-source data, accurately identifying potential health risks such as pressure sores and worsening scoliosis, locating abnormal force points and sources of posture deviation, and significantly improving the accuracy of risk prediction and abnormal location. The intelligent guidance decision engine deeply integrates preset rehabilitation rules, user historical data, and multi-center similar case data to dynamically generate personalized alignment guidance that includes details such as adjustment angles and timing. The proposed solutions and adaptive rehabilitation recommendations upgrade from generalized correction to personalized, precise guidance. The dynamic limb posture adjustment unit, through multi-actuator collaborative drive and multi-modal guidance feedback including voice, vision, and touch, ensures accurate implementation of alignment movements and active user cooperation, further enhancing adjustment precision with a real-time calibration mechanism. The medical and health data sharing module, through encrypted transmission protocols and role-based access control models, enables secure data sharing and hierarchical access management across hospitals and rehabilitation institutions. It synchronously updates user rehabilitation progress and risk changes in real time, effectively protecting data privacy and improving cross-institutional collaboration efficiency. This enhances the standardization of limb positioning for wheelchair users, reduces the risk of complications, accelerates the rehabilitation process, and optimizes the quality of collaborative allocation of medical resources. Thus, it solves the problems of poor adaptability and sharing difficulties in existing technologies.

[0091] The following will illustrate a specific embodiment of a method for guiding wheelchair users on limb alignment based on the sharing of medical and health big data. Figure 9 As shown, it includes: We selected Grandma Wang, a 65-year-old patient in the stroke rehabilitation ward of a rehabilitation hospital, as our service recipient. Due to right-sided hemiplegia, she has been using a smart adjustable wheelchair for a long time, and has problems such as concentrated pressure on her right hip and mild left lateral scoliosis caused by improper sitting posture. Traditional rehabilitation guidance relies on manual observation, which lacks personalization and cross-institutional data sharing. Based on this, we deployed a wheelchair user limb alignment guidance system based on the sharing of medical and health big data.

[0092] The data acquisition process employs a multi-sensor collaborative deployment to achieve precise, multi-dimensional capture of limb status. For limb joint angle acquisition, miniature MEMS posture sensors (model MPU6050) are used, fixed to Grandma Wang's left shoulder, elbow, hip, and knee joints respectively. They communicate with the data acquisition terminal via I2C bus, with a sampling frequency of 10Hz, a measurement range of 0°~180°, and an accuracy of ±0.5°. Real-time acquisition of rotation angle data for each joint is conducted, with a focus on monitoring changes in passive movement angles of the hemiplegic limb. For limb pressure distribution acquisition, a flexible thin-film pressure sensor array (model FSR402) is used, laid in a 20×20 matrix on the wheelchair seat and backrest corresponding to the areas of contact with the human body. The system uses a single sensor with a range of 0-500 kPa, a resolution of 1 kPa, and a sampling frequency of 5 Hz. Data is transmitted via the SPI protocol, enabling accurate identification of high-risk areas with pressure ≥300 kPa. Trunk posture acquisition utilizes a wearable module (model BMI160) integrating a triaxial accelerometer and gyroscope, fixed at the T6-T8 intervertebral space on Grandma Wang's back. The sampling frequency is 50 Hz, and the measurement range covers pitch angles of -30° to 30°, tilt angles of -20° to 20°, and rotation angles of -45° to 45°, with an accuracy of ±0.1° for all angles, capturing trunk posture shift data in real time. All sensor data is aggregated by the wheelchair's built-in edge computing terminal and uploaded to the local medical data server via the hospital's 5G private network (transmission latency ≤20ms). High-frequency sensor data is stored in an InfluxDB time-series database (retained for 7 days), and structured data such as patient medical records is stored in a MySQL database. A data verification mechanism (CRC32 checksum) ensures data integrity during transmission, constructing a complete data acquisition and storage system.

[0093] The data processing stage utilizes a multi-algorithm collaboration to achieve data fusion, risk assessment, and deviation localization. The data fusion process employs a weighted fusion algorithm, using the timestamps of trunk posture data as a baseline. A linear interpolation algorithm is used to spatiotemporally align joint angle (10Hz) and pressure distribution (5Hz) data, with time synchronization errors controlled within 10ms. Entropy weighting is used to dynamically allocate weights to each data dimension, with pressure distribution data having a weight of 0.4, trunk posture data a weight of 0.35, and joint angle data a weight of 0.25. A unified limb state dataset is generated through weighted summation. Simultaneously, the 3σ criterion is used to remove outlier data (such as instantaneous false alarms from pressure sensors), and missing data (missing rate <3%) is filled using a weighted average of the nearest five points. The health risk assessment is based on preset thresholds and an improved random forest model. Pressure ulcer risk is determined by a combination of pressure value and duration in the pressure concentration area (pressure ≥300 kPa and lasting for more than 20 minutes is considered high risk). Abnormal spinal load is assessed by the cumulative deviation of trunk tilt angle (average tilt angle ≥5° for 1 consecutive hour is considered progressive stage). Muscle fatigue risk is predicted by the frequency of joint angle mutations (≥5 angle mutations within 10 minutes is considered high probability). The model was trained on rehabilitation data from 50,000 stroke patients. The input is a fused 18-dimensional feature vector (6 joint angle features + 8 pressure distribution features + 4 trunk posture features), and the output is a quantitative level of three risk categories (0-10 points). The accuracy on the test set reached 92%. Anomaly localization was performed using the Apriori association analysis algorithm, with a minimum support of 0.8 and a minimum confidence of 0.85. Association rules between pressure distribution hotspots and posture deviation data were mined. Through frequent itemset mining, a strong correlation was found between the concentrated pressure on Grandma Wang's right hip (support 0.88) and the leftward tilt of the torso ≥4° (support 0.91), with a confidence of 0.93. Ultimately, the abnormal stress point was accurately located in the ischial tuberosity region of the right hip, and the source of posture deviation was the uneven stress caused by the leftward tilt of the torso.

[0094] The solution generation and execution phases rely on multi-center data support to achieve personalized guidance and precise control. Personalized solution generation is based on risk assessment results and anomaly location information, combined with a multi-center medical and health database (integrating data from over 100,000 stroke rehabilitation cases from 12 top-tier hospitals and 20 rehabilitation institutions nationwide), and uses a multi-objective optimization algorithm to dynamically generate solutions. First, a weighted Euclidean distance algorithm was used to match similar cases, and 200 cases with a similarity of ≥0.8 to Grandma Wang in age (±5 years), medical history (right hemiplegia), and rehabilitation stage (recovery period) were selected. Effective sitting posture adjustment strategies were extracted from these cases (such as the "trunk right tilt correction + hip decompression combination" scheme, which had an effectiveness rate of 78%). Combining the preset rehabilitation rules (such as the sitting posture adjustment angle for hemiplegic patients not exceeding 5°) with Grandma Wang's historical rehabilitation data (posture adjustment records for the past month), a specific plan was generated: "Adjust sitting posture every 25 minutes, tilt the trunk to the right by 3°, place a soft pillow under the right hip to assist in decompression, and maintain this position for 5 minutes." At the same time, rehabilitation assistance suggestions were generated: "Twice a day, 10 minutes each time, right hip pendulum exercise (swing amplitude 25°), combined with 42℃~45℃ hot compress to promote blood circulation." The plan generation time was ≤1 second. Before implementation, the plan details were displayed on the wheelchair touchscreen. After confirmation by Grandma Wang and her caregivers, the plan was initiated. The adjustment mechanism drive unit employed a cubic polynomial trajectory planning algorithm to convert the plan parameters into the motion trajectory of the wheelchair actuators. This drove the seat lifting motor (model 28BYJ-48, lifting range 0~10cm, speed 0.5cm / s), the backrest angle adjustment motor (model 39BYJ-40, adjustment range 90°~120°, speed 1° / s), and the tilt compensation actuator (compensation range -5°~5°, speed 0.5° / s) to work in tandem, with the acceleration controlled within... With a speed of ≤0.2m / s², the adjustment process is smooth and shock-free. The guidance feedback unit synchronously activates multimodal guidance, and the voice module plays a gentle prompt, "Please slowly adjust your sitting posture to the right." The green indicator light on the right armrest of the wheelchair flashes at a frequency of 1 time per second, and the vibration module built into the right seat cushion (frequency 50Hz, medium intensity) vibrates continuously, with a feedback response time of ≤0.5 seconds. The posture calibration unit collects limb status data after adjustment in real time, uses a PID control algorithm to calculate the deviation, and automatically fine-tunes the actuator when the torso tilt angle deviation exceeds 1°, with an adjustment accuracy error of ≤0.5°, achieving closed-loop control.

[0095] The cross-institutional data sharing and status synchronization process achieves unified data across the entire collaborative chain through secure encryption and access control. The data sharing unit employs a transmission protocol based on the SM4 encryption algorithm to securely share Grandma Wang's personalized limb alignment guidance plan, rehabilitation assistance suggestions, core medical record data (medical history, examination reports), risk assessment results, and abnormal location information to her contracted community health service center, family doctor workstation, and her children's mobile devices. The encrypted transmission rate is ≥10Mbps. During data transmission, identity authentication (UKey + password dual verification) and data anonymization (hiding sensitive information such as patient ID number and home address) are used to ensure privacy and security. The access control unit, based on the RBAC role-based access control model, assigns differentiated permissions to different collaborative entities: rehabilitation hospital physicians have full access to read and modify data; community rehabilitation therapists have access to view the plan and submit adjustment suggestions; family doctors have access to view risk data and follow-up records; and the patient's children only have access to view rehabilitation progress. Access verification response time is ≤0.1 seconds to prevent unauthorized access to sensitive data. The status synchronization unit employs a timestamp priority arbitration mechanism to perform layered synchronization of posture adjustment effect data (second-level synchronization), health risk change data (minute-level synchronization), and rehabilitation progress data (hour-level synchronization). When data conflicts occur (such as inconsistencies between rehabilitation progress recorded by the community health service center and real-time monitoring data from the hospital), the system automatically calibrates using the latest clinically validated data as the benchmark, with a synchronization error of ≤5 seconds. Synchronized data is pushed to the terminal systems of each collaborating entity in real time. Rehabilitation hospital physicians can remotely view the community rehabilitation effect and optimize the plan, community rehabilitation therapists can obtain precise guidance from the hospital, and children can monitor the rehabilitation progress in real time. This completely breaks down data silos across institutions, ensuring that Grandma Wang receives consistent and unified rehabilitation services throughout her hospital rehabilitation, community follow-up care, and home recuperation, avoiding duplicate assessments and plan conflicts, and improving the efficiency of cross-institutional collaboration and the quality of rehabilitation services.

[0096] In summary, this application's embodiments utilize multi-sensor collaborative data collection of limbs and multi-algorithm collaborative processing to accurately assess health risks and locate abnormal stress points and postural deviation sources, effectively reducing the risk of complications such as pressure sores and abnormal spinal load caused by improper sitting posture. Personalized guidance plans are generated based on a multi-center medical and health database, combined with closed-loop control and multimodal guidance to improve the accuracy and adaptability of sitting posture adjustments, addressing the shortcomings of traditional rehabilitation plans lacking personalization. A cross-institutional secure sharing and hierarchical synchronization mechanism breaks down data silos, avoids redundant assessments and plan conflicts, ensures the continuity and consistency of rehabilitation services across hospitals, communities, and homes, and comprehensively improves the efficiency of cross-institutional collaboration and the quality of rehabilitation services for wheelchair users.

[0097] Figure 10 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: The memory 1001, the processor 1002, and the computer program stored on the memory 1001 and capable of running on the processor 1002.

[0098] When the processor 1002 executes the program, it implements the method for guiding wheelchair users to straighten their limbs based on the sharing of medical and health big data provided in the above embodiments.

[0099] Furthermore, electronic devices also include: Communication interface 1003 is used for communication between memory 1001 and processor 1002.

[0100] The memory 1001 is used to store computer programs that can run on the processor 1002.

[0101] The memory 1001 may include high-speed RAM (Random Access Memory) and may also include non-volatile memory, such as at least one disk storage.

[0102] If the memory 1001, processor 1002, and communication interface 1003 are implemented independently, then the communication interface 1003, memory 1001, and processor 1002 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 10 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0103] Optionally, in a specific implementation, if the memory 1001, processor 1002, and communication interface 1003 are integrated on a single chip, then the memory 1001, processor 1002, and communication interface 1003 can communicate with each other through an internal interface.

[0104] The processor 1002 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of this application.

[0105] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for guiding wheelchair users to limb alignment based on the sharing of medical and health big data.

[0106] In addition, this application also provides a computer program product, including a computer program or instructions, which, when executed, implement the above-mentioned method for guiding wheelchair users to limb alignment based on the sharing of medical and health big data.

[0107] In the description of this specification, the references to "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0108] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0109] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0110] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any of the following techniques known in the art, or a combination thereof: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0111] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0112] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A limb alignment guidance system for wheelchair users based on the sharing of medical and health big data, characterized in that, include: The system includes a limb status perception module, a medical big data analysis center, an intelligent guidance and decision-making engine, a limb posture dynamic adjustment unit, and a medical and health data sharing module; among these components... The limb status sensing module is used to collect data on the limb joint angles, limb pressure distribution, and trunk posture of wheelchair users. The medical big data analysis center is used to integrate and process the data of the wheelchair user's limb joint angles, limb pressure distribution, and trunk posture to identify the risk of pressure sores, aggravation of scoliosis, and potential health risks of muscle spasms caused by improper sitting posture, and to locate abnormal force points and sources of posture deviation. The intelligent guidance decision engine is used to dynamically generate personalized limb alignment guidance plans and rehabilitation assistance suggestions based on the potential health risks, abnormal force points of limbs and sources of postural deviations, combined with preset rehabilitation rules and user historical data. The limb posture dynamic adjustment unit is used to drive the wheelchair adjustment mechanism and provide alignment guidance feedback simultaneously, based on the personalized limb alignment guidance plan and rehabilitation assistance suggestions. The medical and health data sharing module is used to share the personalized limb alignment guidance plan and rehabilitation assistance suggestions, patient medical record data, potential health risks, abnormal limb force points and postural deviation sources across institutions, synchronously update user status information and manage access permissions.

2. The wheelchair user limb alignment guidance system based on medical and health big data sharing according to claim 1, characterized in that, The limb state perception module includes a joint angle sensor group, a pressure sensor array, and a trunk posture acquisition unit. The joint angle sensor group is used to collect rotation angle data of the shoulder joint, elbow joint, hip joint, and knee joint in real time. The pressure sensor array is installed in the wheelchair seat and backrest area to collect pressure distribution data between the limb and the contact surface. The trunk posture acquisition unit is used to capture trunk pitch angle, tilt angle, and rotation angle data.

3. The wheelchair user limb alignment guidance system based on medical and health big data sharing according to claim 1, characterized in that, The medical big data analysis center includes a data fusion processing unit, a health risk identification unit, and an anomaly localization unit. The data fusion processing unit uses a weighted fusion algorithm to perform spatiotemporal alignment and redundancy removal on limb joint angle, limb pressure distribution, and trunk posture data to generate a unified limb state dataset. The health risk identification unit identifies pressure ulcer risk levels, scoliosis progression trends, and the probability of muscle spasm attacks based on preset thresholds and machine learning models. The anomaly localization unit analyzes pressure distribution hotspots and posture deviation data, and uses correlation analysis algorithms to locate the specific locations of abnormal force points on the limbs and the sources of posture deviations in the trunk and limbs.

4. The wheelchair user limb alignment guidance system based on medical and health big data sharing according to claim 1, characterized in that, The intelligent guidance decision engine includes a guidance plan generation unit, a rehabilitation suggestion adaptation unit, and a medical data matching unit. The guidance plan generation unit generates a personalized limb alignment guidance plan, including limb adjustment angles, timing, and duration, based on potential health risk levels, abnormal stress points, and sources of postural deviations, combined with preset rehabilitation rules and user historical data. The rehabilitation suggestion adaptation unit generates targeted muscle relaxation training and joint mobility assistance suggestions using a user feature adaptation algorithm, combined with the user's age, medical history, and rehabilitation stage. The medical data matching unit accesses a multi-center medical and health database and uses a similarity calculation algorithm to match intervention plans for similar cases, optimizing the current guidance plan and rehabilitation suggestions.

5. A wheelchair user limb alignment guidance system based on medical and health big data sharing as described in claim 1, characterized in that, The limb posture dynamic adjustment unit includes an adjustment mechanism drive unit, a guidance feedback providing unit, and a posture calibration unit. The adjustment mechanism drive unit receives personalized alignment guidance plans and drives the wheelchair seat lifting, backrest angle adjustment, and tilt compensation actuators through a trajectory planning algorithm. The guidance feedback providing unit guides the user to actively cooperate with the alignment movements through voice prompts, flashing visual indicator lights, and tactile vibration feedback. The posture calibration unit collects limb state data after adjustment in real time, compares it with target posture parameters, and calibrates the adjustment accuracy.

6. The wheelchair user limb alignment guidance system based on medical and health big data sharing according to claim 1, characterized in that, The medical and health data sharing module includes a data sharing unit, an access control unit, and a status synchronization unit. The data sharing unit is used to share personalized limb alignment guidance plans, rehabilitation assistance suggestions, core patient medical record data, and risk assessment results across hospitals and rehabilitation institutions via an encrypted transmission protocol, ensuring privacy and security during transmission. The access control unit is used to set different data access permissions for medical staff, rehabilitation therapists, and patients' families based on a role-based access control model. The data synchronization unit is used to synchronize user posture adjustment effect data, health risk change data, and rehabilitation progress data in real time, ensuring data consistency during cross-institutional collaboration.

7. A method for applying a wheelchair user limb alignment guidance system based on medical and health big data sharing as described in any one of claims 1-6, characterized in that, The method includes: Acquire data on the joint angles of wheelchair users, limb pressure distribution, and trunk posture. The data on the limb joint angles, limb pressure distribution, and trunk posture of the wheelchair user are fused and processed. Based on preset thresholds and machine learning models, the risk of pressure sores, abnormal spinal load, and potential health risks of muscle fatigue caused by improper sitting posture are assessed. The correlation analysis algorithm is used to locate abnormal force points and sources of posture deviation. Based on the potential health risks, abnormal limb stress points, and sources of postural deviations, and combined with a multi-center medical and health database, a personalized limb alignment guidance plan and rehabilitation assistance suggestions are dynamically generated. Under user confirmation and safety supervision, the wheelchair adjustment mechanism is driven to adjust according to the personalized limb alignment guidance plan and rehabilitation assistance suggestions, and alignment guidance feedback is provided. At the same time, the personalized limb alignment guidance plan and rehabilitation assistance suggestions, patient medical record data, and the potential health risks, abnormal limb stress points, and sources of postural deviations are shared securely across institutions, and user status information is updated synchronously.

8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the program to implement the method for guiding wheelchair users to limb alignment based on the sharing of medical and health big data as claimed in claim 7.

9. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When a computer program or instruction is executed, it implements the method for guiding wheelchair users to limb alignment based on the sharing of medical and health big data, as claimed in claim 7.

10. A computer program product, comprising a computer program or instructions, characterized in that, When a computer program or instruction is executed, it implements the method for guiding wheelchair users to limb alignment based on the sharing of medical and health big data, as claimed in claim 7.

Images