CN-121583545-B - Real-time data-driven oxygen therapy exercise training and rehabilitation method and device

CN121583545BCN 121583545 BCN121583545 BCN 121583545BCN-121583545-B

Abstract

The application discloses a real-time data-driven oxygen therapy exercise training and rehabilitation method and device, belongs to the technical field of health management and oxygen enrichment rehabilitation training, and aims to solve the problems that a traditional oxygen therapy exercise scheme lacks individual suitability, is difficult to dynamically match with the physical state of a user and is delayed in abnormal monitoring response. The device integrates an intelligent rehabilitation monitoring and evaluating platform, a motion carrier, an oxygen inhaler, a physiological monitoring component and a brain-computer interface component, acquires basic physiological and health data of a user through cooperation of the components, matches a standard behavior baseline and corrects the basic physiological and health data into a personalized behavior baseline, combines the motion carrier parameters to generate a personalized oxygen-enriched rehabilitation exercise scheme, acquires data in real time in execution, discriminates abnormality through AI abnormality detection and generates a wake-up prompt, and corrects the baseline according to the abnormality type. The application realizes the dynamic adaptation of the oxygen therapy exercise scheme to the individual, and improves the training safety and the rehabilitation accuracy.

Inventors

BIAN XUELIAN
MAO DANDAN
WANG XIAO

Assignees

上海利康精准医疗技术有限公司

Dates

Publication Date: 20260512
Application Date: 20260123

Claims (8)

1. A real-time data-driven, oxygen therapy exercise training and rehabilitation method, comprising: Acquiring basic physiological information and health detection indexes of a user, and performing gradual intersection operation matching with a standard behavior baseline library to acquire a standard behavior baseline of the user; The standard behavior base line is a multidimensional parameter set which is formed by adopting a density clustering algorithm and is formed by carrying out hierarchical clustering according to age segmentation, basic disease types and health detection index intervals and reflects rehabilitation exercise capacity of corresponding people, and the multidimensional parameter set comprises exercise tolerance interval parameters, physiological response threshold interval parameters and exercise adaptation interval parameters; The gradual intersection operation is a hierarchical intersection operation mode of screening matched standard behavior baselines from a standard behavior baseline library step by step according to the sequence of age segmentation, basic disease types and health detection index intervals based on basic physiological information and health detection indexes of a user; Based on the user standard behavior base line obtained by matching, constructing a gradient test rehabilitation exercise scheme so as to correct the user standard behavior base line through response difference characteristics of the user in a non-oxygen inhalation environment and an oxygen inhalation environment and acquire a personalized behavior base line of the user; Setting physiological state targets and motion execution targets of each training stage by taking personalized behavior baselines as the basis and combining motion mode parameters of target motion carriers selected by a user, and generating a personalized oxygen-enriched rehabilitation motion scheme by adopting a multi-factor weighted decision method; when a user triggers the personalized oxygen-enriched rehabilitation exercise scheme to execute, acquiring real-time data of the user through a physiological monitoring assembly, a brain-computer interface assembly, a motion carrier and an oxygen inhalation machine, carrying out AI anomaly detection through deviation degree standardization processing and anomaly judgment critical value correction, identifying deviation data points, judging whether the user has an execution anomaly or not, quantifying the anomaly degree and dividing anomaly types, and further generating an AI awakening prompt; and when the user is judged to have the execution abnormality, classifying and correcting the personalized behavior base line of the user according to the abnormality type.
2. The method for real-time data-driven oxygen therapy athletic training and rehabilitation of claim 1, wherein the step of obtaining a standard behavioral baseline of the user comprises: Acquiring basic physiological information and health detection indexes of a user, wherein the basic physiological information comprises age, static heart rate and basic disease type, and the health detection indexes comprise susceptibility gene related data and telomere detection results; Performing format unification and numerical normalization processing on the basic physiological information and the health detection index which pass through the validity verification to form a structured data feature set; Based on age information in the structured data feature set, extracting all standard behavior baselines of the corresponding age segments from the standard behavior baseline library by combining with dynamic age segmentation rules of the standard behavior baseline library to form a first baseline subset; Extracting standard behavior baselines consistent with the basic disease types from the first baseline subset based on basic disease type information in the structured data feature set, and adjusting matching tolerance according to the distribution variance of the baseline samples corresponding to the basic disease types to form a second baseline subset; And extracting standard behavior baselines with the mapping coincidence degree of the health detection index interval and the health detection index of the user larger than a preset coincidence degree threshold value from the second baseline subset by adopting an interval mapping matching method based on the health detection index information in the structured data feature set, so as to form a third baseline subset.
3. The real-time data driven oxygen therapy athletic training and rehabilitation method of claim 2, wherein the specific step of obtaining a standard behavioral baseline of the user further comprises: performing multidimensional similarity fusion calculation on each standard behavior baseline of the third baseline subset, and distributing mapping weights according to mapping positions of the user health detection indexes in the corresponding behavior baseline index intervals; Calculating the matching degree of the user health detection index and the corresponding interval of each standard behavior baseline based on the mapping weight, and calculating the matching degree of the static heart rate, the basic disease associated exercise physiological parameter and the corresponding parameter of each standard behavior baseline; Fusing the fitness detection index fitness, the static heart rate matching degree and the basic disease associated exercise physiological parameter matching degree to obtain the comprehensive similarity of each standard behavior baseline of the third baseline subset; if the third baseline subset has standard behavior baselines with comprehensive similarity larger than the preset high-adaptation standard, determining the standard behavior baselines as standard behavior baselines matched with the user; if the standard behavior base line is not present, eliminating abnormal standard behavior base lines with the comprehensive similarity lower than the comprehensive similarity mean value of the third base line subset minus the standard deviation of the comprehensive similarity, recalculating the comprehensive similarity of the rest standard behavior base lines, selecting the standard behavior base lines with the comprehensive similarity greater than the preset adaptation threshold after recalculation, and determining the standard behavior base lines as the standard behavior base lines matched with the user.
4. The real-time data driven oxidative therapeutic exercise training and rehabilitation method according to claim 2, wherein the step of constructing a graded test rehabilitation exercise program comprises: Extracting core parameters in a user standard behavior baseline, wherein the core parameters comprise a motion tolerance interval parameter, a physiological response threshold interval parameter and a motion adaptation interval parameter; dividing a plurality of continuous gradient intervals according to the sequence from low to high based on the motion tolerance interval parameters of the standard behavior base line, dividing the gradient intervals according to the distribution density of the motion tolerance in the standard behavior base line, determining the total duration of single test motion by combining the motion adaptation interval parameters of the standard behavior base line, distributing the duration proportion according to the importance of each gradient interval, setting the change rule of the action frequency according to the physiological response threshold interval parameters of the standard behavior base line, and forming a graded test rehabilitation motion scheme.
5. The real-time data driven oxygen therapy athletic training and rehabilitation method of claim 1, wherein the step of deviant normalization and abnormality determination threshold correction comprises: When a user triggers the execution of a personalized oxygen-enriched rehabilitation exercise scheme, synchronously acquiring real-time data of the user, wherein the real-time data comprises exercise physiological dynamic data, brain-computer interface data, exercise execution data and oxygen supply real-time parameters; Adding a unified time stamp to the real-time data, dividing the data segments according to the training phases, and constructing a real-time data sequence corresponding to each training phase; Based on the physiological state target and the motion execution target of each training stage, combining the current oxygen supply real-time parameters, constructing a stage target dynamic threshold system to set a standard threshold interval of motion physiological dynamic data and motion execution data of each training stage; Performing deviation degree standardization processing on the real-time data sequence of the motion physiological dynamic data and the motion execution data, calculating absolute difference values of the motion physiological dynamic data and the motion execution data and the standard threshold value interval of the corresponding training stage, and converting the absolute difference values into corrected deviation degree values according to the proportion of the absolute difference values to the total span of the standard threshold value interval to form a deviation degree sequence of the motion physiological dynamic data and the motion execution data; Extracting the dynamic change trend of the alpha wave energy duty ratio and the beta wave energy duty ratio aiming at the real-time data sequence of brain-computer interface data, and quantifying to obtain the real-time fatigue value and concentration value of a user; based on the fatigue value and the concentration value, executing two-dimensional fault tolerance adjustment, dynamically correcting the preset fault tolerance of the user to generate the real-time fault tolerance of the user, calculating the adaptation deviation between the oxygen supply real-time parameter and the preset oxygen supply standard of the current stage, and generating an oxygen supply adaptation value; Setting an abnormality judgment critical value of each training stage, and correcting the abnormality judgment critical value of each training stage according to the oxygen supply adaptation value and the real-time fault tolerance of a user.
6. The method of real-time data driven oxygen therapy exercise training and rehabilitation according to claim 5, wherein the steps of AI anomaly detection, deviation data point identification, anomaly degree quantification and anomaly type classification, and further generating AI wake-up cues include: Comparing the deviation degree sequences of the motion physiological dynamic data and the motion execution data with the abnormal judgment critical values corrected in the corresponding training phase one by one, marking the deviation data points larger than the abnormal judgment critical values, counting the duty ratio of the deviation data points in a preset detection period, and judging that the current user has the execution abnormality if the duty ratio is larger than a preset duty ratio threshold value; Dividing the degree of abnormality according to the degree of deviation of each deviation data point from the abnormality judgment critical value and the duty ratio of the deviation data point in the detection period; The method comprises the steps of adopting a decision tree and gradient lifting tree fusion AI algorithm to extract abnormal characteristics of real-time data deviating from a time window where a data point is located and dividing abnormal types, wherein the abnormal characteristics comprise deviation degree, real-time fault tolerance and oxygen supply adaptation degree value of motion physiological dynamic data and motion execution data; and establishing a dynamic mapping library of the abnormality type, the abnormality degree, the user state and the wake-up strategy, and generating an AI wake-up prompt through the multi-mode interaction component according to the current abnormality type, the abnormality degree, the fatigue value and the concentration value of the user and the wake-up strategy.
7. The real-time data driven oxygen therapy athletic training and rehabilitation method of claim 1, wherein the step of classifying and correcting the personalized behavioral baseline of the user according to the type of anomaly comprises: Triggering personalized behavior baseline correction operation of a user when the user is judged to have abnormal execution, and constructing an abnormal data set when the current personalized oxygen-enriched rehabilitation exercise scheme is executed; positioning an unadapted target correction dimension parameter in a personalized behavior baseline based on a preset association mapping rule of an abnormal type and a baseline parameter; Extracting features of various time sequence data in the abnormal data set, screening feature indexes of target correction dimension parameters by combining functional attributes of the target correction dimension parameters, and constructing a correction basis feature set; according to the characteristic index of the target correction dimension parameter, quantifying the adaptive deviation degree of the target correction dimension parameter through multidimensional characteristic weighted deviation calculation, and determining the adjustment direction and the adjustment amplitude of the target correction dimension parameter to directionally adjust the target correction dimension parameter; and carrying out safety boundary verification on the corrected target correction dimension parameters, carrying out safety boundary correction, and integrating the target correction dimension parameters with the rest parameters of the personalized behavior base line to form the corrected personalized behavior base line.
8. The real-time data-driven oxygen therapy exercise training and rehabilitation device for realizing the real-time data-driven oxygen therapy exercise training and rehabilitation method according to any one of claims 1 to 7, which is characterized by comprising a rehabilitation intelligent monitoring and evaluating platform, a motion carrier, an oxygen inhalation machine, a physiological monitoring assembly and a brain-computer interface assembly; The intelligent rehabilitation monitoring and evaluating platform is used for constructing a personalized behavior baseline of a user, generating a personalized oxygen-enriched rehabilitation exercise scheme, performing abnormal diagnosis to feed back and correct the personalized behavior baseline, the exercise carrier is used for providing an exercise execution environment adapting to the personalized oxygen-enriched rehabilitation exercise scheme, the oxygen inhaler is used for dynamically outputting adapted oxygen supply parameters according to the personalized oxygen-enriched rehabilitation exercise scheme and real-time monitoring data, the physiological monitoring component is used for acquiring exercise physiological dynamic data of the user, and the brain-computer interface component is used for acquiring electroencephalogram characteristic data related to fatigue and concentration.

Description

Real-time data-driven oxygen therapy exercise training and rehabilitation method and device Technical Field The invention relates to the technical field of health management and oxygen-enriched rehabilitation training, in particular to a real-time data-driven oxygen therapy exercise training and rehabilitation method and device. Background The oxygen therapy exercise training is widely applied to the field of health intervention, is particularly suitable for rehabilitation old people such as senile dementia and postoperative care of tumors, but has obvious bottlenecks in the prior art, on one hand, the exercise scheme is formulated based on general crowd data, the individual differences such as age and basic health state of a user are not fully considered, the adaptability is insufficient, the characteristics of a target exercise carrier are not combined, the situation that the scheme is not matched with the carrier capability is easy to occur, the smooth exercise is directly affected, the active exercise function which is not required to be regulated by the user is not needed, the exercise requirements such as target gas concentration, exercise speed, strength and time are not required to be assisted by other people according to the health indexes, the use convenience is poor, on the other hand, the exercise abnormality monitoring is dependent on a fixed threshold value, the dimension is single, the physiological index abnormality and exercise execution deviation are difficult to accurately identify in time, the dynamic adjustment mechanism is lacking, once the exercise parameters are fixed, the exercise parameters cannot be corrected according to abnormal conditions, the change of the physical state of the user easily occurs after long-term use, the exercise safety and effect are reduced, in addition, the existing equipment is single in function, the actual equipment cannot be used, the integrated with the physical state change due to the physical state of the user, the physical state is not optimal, the physiological parameter is difficult to be well-controlled, and the physiological parameter is not suitable for the accurate and has high, and has the health parameter, and has poor safety and has high performance, and has low performance, and is difficult to be well. The Chinese patent application with publication number of CN117497131A discloses a COPD intelligent oxygen therapy cloud platform based on the Internet of things and a management method, wherein the system comprises an output end, an intelligent application layer and a perception Internet of things, and the intelligent application layer establishes a next oxygen therapy scheme by analyzing input end data and platform existing data, so that standardized management and on-line and off-line coordination of home oxygen therapy of COPD patients are realized. But still fails to solve the problems of single monitoring dimension, static scheme, lack of nerve feedback support and weak remote synergy. Even though research indicates the potential value of nerve feedback and oxygen therapy combination on rehabilitation, and oxygen therapy exercise training of special people such as COPD and the like is in urgent need of multi-dimensional monitoring and dynamic adaptation, the prior patent technology still does not form a systematic solution. Accordingly, to overcome these limitations, the present invention proposes a real-time data-driven method and apparatus for oxygen therapy exercise training and rehabilitation. Disclosure of Invention Aiming at the defects existing in the prior art, the invention aims to provide the real-time data-driven oxygen therapy exercise training and rehabilitation method and device, which solve the problems of how to accurately acquire the rehabilitation exercise related parameters of the individual of the adaptive user for the oxygen therapy exercise training, generate the rehabilitation exercise scheme of the individual of the adaptive user and the target exercise carrier, timely identify the abnormal execution in the training process, and dynamically correct the related parameters based on the abnormal type to ensure the suitability and the safety of the training. In order to achieve the above purpose, the present invention provides the following technical solutions: A real-time data-driven oxygen therapy exercise training and rehabilitation method comprising: Acquiring basic physiological information and health detection indexes of a user, and performing gradual intersection operation matching with a standard behavior baseline library to acquire a standard behavior baseline of the user; Based on the user standard behavior base line obtained by matching, constructing a gradient test rehabilitation exercise scheme so as to correct the user standard behavior base line through response difference characteristics of the user in a non-oxygen inhalation environment and an oxygen inhalation environment and acquire a personalized