Search

CN-121999973-A - Orthopedics rehabilitation intelligent training optimizing system

CN121999973ACN 121999973 ACN121999973 ACN 121999973ACN-121999973-A

Abstract

The invention belongs to the technical field of artificial intelligence and medical rehabilitation, in particular relates to an intelligent training optimization system for orthopedic rehabilitation, and aims to solve the problems of insufficient individuation, feedback lag, subjective evaluation and the like in traditional rehabilitation training. The system collects motion and physiological data through the wearable sensing unit, combines cloud individual modeling, dynamic risk assessment and reinforcement learning driven self-adaptive optimization algorithm to generate and adjust a training scheme, realizes action real-time correction through multi-mode feedback, supports doctor-patient coordination and group intelligent analysis through the clinical management platform, and improves rehabilitation safety, effectiveness and accessibility.

Inventors

  • LI YUANFENG
  • JIANG XIAOGUANG
  • KANG JIE

Assignees

  • 黑龙江中医药大学

Dates

Publication Date
20260508
Application Date
20260119

Claims (10)

  1. 1. An orthopedics rehabilitation intelligent training optimizing system, which is characterized by comprising: the wearable sensing unit is used for being deployed at a critical anatomical position of a limb of a patient and continuously collecting kinematic parameters and physiological signals of the patient; the edge computing terminal is used for receiving an original signal stream of the wearable sensing unit, performing preliminary filtering, feature extraction and time sequence alignment operation, and uploading the processed structured data packet to the cloud intelligent engine; The cloud intelligent engine is used as a core decision center of the system, receives the structured data packet and generates an optimized training instruction based on an individual rehabilitation model, a real-time risk assessment result and a historical training response; The interactive feedback module is used for receiving the training instruction issued by the cloud intelligent engine and converting the training instruction into a multi-mode feedback signal so as to correct the real-time behavior of the patient; The clinical management platform is used for providing a patient data visual interface and an intervention interface for medical professionals so as to realize remote collaborative decision-making; The cloud intelligent engine comprises an individuation rehabilitation modeling module, a dynamic risk assessment module, a self-adaptive training optimization module and a multi-mode feedback generation module, wherein the individuation rehabilitation modeling module is used for constructing and maintaining a skeleton-muscle dynamics digital twin model of a patient, the dynamic risk assessment module is used for carrying out multi-dimensional anomaly detection and triggering hierarchical early warning on real-time motion data, the self-adaptive training optimization module is used for updating a training strategy by using a reinforcement learning algorithm, and the multi-mode feedback generation module is used for converting an optimized training instruction into a guide signal suitable for different sensing channels.
  2. 2. The orthopedic rehabilitation intelligent training optimizing system according to claim 1, wherein the wearable sensing unit comprises an inertia measuring assembly, a surface myoelectricity sensing node and a pressure sensing array, wherein the inertia measuring assembly is used for collecting joint angular speed, linear acceleration and space orientation information, the surface myoelectricity sensing node is used for collecting myoelectricity signals of target muscle groups, and the pressure sensing array is used for collecting pressure distribution data of soles or supporting surfaces.
  3. 3. The orthopedic rehabilitation intelligent training optimizing system according to claim 2, wherein the edge computing terminal is further configured to locally run a lightweight inference framework to complete gesture resolution and preliminary anomaly screening, and to buffer data to support breakpoint continuous transmission when the network is interrupted.
  4. 4. The orthopedic rehabilitation intelligent training optimizing system according to claim 3, wherein the personalized rehabilitation modeling module is used for initializing the digital twin model based on preoperative imaging data, operation records, basic physical performance test results and initial gait analysis reports of patients, introducing a migration learning strategy, screening similar cases from a similar case database to map movement pattern parameters, and shortening a cold start period of the model.
  5. 5. The orthopedic rehabilitation intelligent training optimizing system according to claim 4, wherein the dynamic risk assessment module is internally provided with a multi-level judgment logic and comprises a physiological tolerance boundary checking unit, a fatigue accumulation effect analysis unit and an action continuity evaluation unit, wherein the physiological tolerance boundary checking unit is used for presetting a safety envelope curve of a joint movement range according to individual characteristics of a patient and performing overrun detection, the fatigue accumulation effect analysis unit is used for analyzing a neural muscle system function decline trend based on electromyographic signals, and the action continuity evaluation unit is used for comparing the similarity of a current action sequence and a standard template to judge technical failure.
  6. 6. The system of claim 5, wherein the adaptive training optimization module is configured to construct a reinforcement learning framework with a historical training completion rate, pain score, risk event frequency, and function improvement slope as a state space, a training difficulty coefficient adjustment vector as an action space, and continuously update a training strategy based on a reward function that comprehensively considers task achievement degree, risk index, and patient compliance.
  7. 7. The orthopedic rehabilitation intelligent training optimizing system according to claim 6, wherein the multi-modal feedback generation module has context awareness capability for dynamically switching a dominant feedback channel in voice prompts, haptic vibrations or graphical guidance based on ambient light intensity, background noise level and user focus of attention.
  8. 8. The orthopedic rehabilitation intelligent training optimizing system according to claim 7, wherein the interactive feedback module comprises a wearable touch feedback belt for simulating a correct force direction through a local vibration sequence, a wireless earphone for playing voice prompts or beat tones, and augmented reality glasses for projecting virtual track lines and target areas to guide space positioning actions.
  9. 9. The orthopedic rehabilitation intelligence training optimizing system according to claim 8, wherein the clinical management platform is provided with a group intelligence analysis component for performing cluster analysis on rehabilitation data of registered patient groups, identifying typical recovery trajectory patterns, and pushing evidence-based medical advice packages to relevant responsible physicians when a general rehabilitation platform period is detected for a specific subgroup.
  10. 10. The orthopedic rehabilitation intelligent training optimizing system according to claim 9, wherein the clinical management platform is further used for recording audit logs generated by interaction operations of all doctors and patients and storing the audit logs into blockchain nodes so as to ensure traceability and non-falsification of diagnosis and treatment behaviors, the cloud intelligent engine implements a federal learning framework in the model training process, and allows a plurality of participants to jointly optimize a core algorithm on the premise of not sharing original data.

Description

Orthopedics rehabilitation intelligent training optimizing system Technical Field The invention belongs to the technical field of artificial intelligence and medical rehabilitation, and particularly relates to an intelligent training optimization system for orthopedics rehabilitation. Background Orthopedics rehabilitation is an important branch in modern medical systems and aims at promoting the recovery and reconstruction of the motor function of a patient through scientific and systematic intervention means. Along with the aging of population and the high incidence of traumatic diseases, the demand for efficient and accurate rehabilitation training schemes is growing. The traditional rehabilitation mode mainly depends on experience judgment and manual operation of therapists, the standardization degree of the training process is low, the individuation adaptation capability is insufficient, and a real-time dynamic monitoring and feedback mechanism for the physiological state and the athletic performance of a patient is lacking. In recent years, the development of technologies such as intelligent perception, data driving modeling, man-machine cooperative control and the like injects new technical kinetic energy into rehabilitation engineering, and promotes rehabilitation equipment to evolve towards an automation and intelligent direction. The intelligent training system based on the sensor network and biomechanics analysis becomes one of the key technical paths in the orthopedics rehabilitation field. The system aims at constructing a personalized training feedback model by collecting kinematic parameters (such as joint angles and gait characteristics) and physiological signals (such as myoelectric signals and pressure distribution) of a patient and adjusting training intensity and auxiliary force by combining an adaptive control algorithm, so that rehabilitation efficiency and safety are improved. In an ideal state, the system should be able to dynamically identify the functional recovery phase of the patient, autonomously optimize the training strategy, and maximize the activation effect of the neuromuscular system while ensuring the normative actions. In the prior art, basic motion monitoring and auxiliary execution functions are realized, but the method has the obvious defects of delayed response to state change of a patient, difficulty in realizing fine-granularity functional evaluation and accurate intervention, dependence on preset rules or off-line analysis for training strategy updating, lack of on-line learning and self-adaptive adjustment capability, loose information interaction among all modules of a system, high perception-decision-execution link delay and poor coordination, and meanwhile, neglects integration of subjective feeling and psychological excitation factors of the patient, and influences training compliance and long-term curative effect. The problems are particularly prominent in complex rehabilitation scenes, and the clinical applicability and popularization value of the intelligent rehabilitation system are severely restricted. Therefore, there is a need for an orthopedic rehabilitation intelligent training optimization system with depth perception, autonomous optimization and closed-loop regulation and control capabilities to break through the existing bottleneck. Disclosure of Invention The invention aims to provide an intelligent training optimization system for orthopedics rehabilitation, which aims to solve the key problems of insufficient individuation, loss of real-time feedback, low doctor-patient coordination efficiency, subjectivity in rehabilitation process evaluation and the like of the training scheme in the current orthopedics postoperative and sports injury rehabilitation process. In the prior art, a static training plan is generally formulated by depending on experience of doctors, and patients lack of effective supervision and dynamic adjustment mechanisms when executing in a family or community environment, so that training actions are not standard, load control is inaccurate, pain risk is increased, rehabilitation effect is finally affected, and secondary injury is possibly caused. In addition, the collection of rehabilitation data is mostly discrete and discontinuous records, a complete biomechanics and physiological response closed loop is difficult to form, and the accurate medical idea is restricted to be applied to the field of rehabilitation. The technical scheme of the invention is that an orthopedics rehabilitation intelligent training optimizing system integrating biological signal perception, movement gesture analysis, personalized model driving and remote collaborative decision is constructed. The system is composed of five core parts of a wearable sensing unit, an edge computing terminal, a cloud intelligent engine, an interactive feedback module and a clinical management platform, wherein the data interconnection and instruction synchronization are