Wearable device-assisted remote rehabilitation monitoring and intervention methods

Abstract

The present invention belongs to the field of telemedicine rehabilitation, and in particular, relates to a remote rehabilitation monitoring and intervention method assisted by wearable devices. The method firstly automatically configures monitoring indicators and their collection frequencies according to the rehabilitation parts, collects and pre-processes the monitoring data, and then encrypts and uploads them to a cloud server. Secondly, a machine learning evaluation model is used to analyze the pre-processed data to obtain a comprehensive health score of the rehabilitated person and a risk score of the main monitoring indicators. Thirdly, the rehabilitation stage of the rehabilitated person is determined according to the rehabilitation goals and scores, and a corresponding rehabilitation guidance strategy is generated through a strategy generation model. Finally, the training process of the rehabilitated person is monitored in real time through the wearable device, and the data is fed back to the online physician for real-time strategy adjustment. The method realizes intelligent monitoring and intervention of remote rehabilitation, and improves the efficiency and accuracy of rehabilitation.

Description

Technical Field

[0001] The present invention belongs to the field of remote medical rehabilitation, and in particular to a remote rehabilitation monitoring and intervention method assisted by a wearable device. Background Art

[0002] With the rapid development of modern medical technology, remote rehabilitation monitoring and intervention has become an important research direction in the field of rehabilitation medicine. In particular, driven by technologies such as the Internet of Things, big data, and cloud computing, wearable devices play an increasingly important role in remote rehabilitation monitoring and intervention. These devices can monitor the patient's physiological data in real time, such as heart rate, blood pressure, blood sugar, etc., and transmit the data to the cloud or the doctor's terminal for analysis and processing through wireless communication technology. For example, the patent with publication number CN109545325A proposes a wearable upper limb intelligent rehabilitation device for hemiplegic patients. The device integrates components such as processors, sensors, and wireless communicators to achieve real-time monitoring of the patient's upper limb movements and simple rehabilitation guidance. At the same time, it supports information exchange with the central node, providing a hardware foundation for the formulation and implementation of personalized rehabilitation plans. On the other hand, the patent with publication number CN109256204A discloses a remote intervention rehabilitation training system, which drives rehabilitation training through a motor module, monitors the rehabilitation process through a sensor module, and cooperates with a remote monitoring module to achieve remote monitoring of rehabilitation training data and expert remote intervention guidance.

[0003] The above existing technologies have the following problems: 1) Existing wearable devices usually lack specificity when collecting data, and cannot set specific collection indicators according to specific rehabilitation parts, which makes the collected data not only redundant but also useless, and reduces the use time of wearable devices; 2) The application of wearable devices in remote rehabilitation monitoring and intervention is mostly still in the data collection and transmission stage, lacking intelligent analysis and third-party feedback guidance; In order to solve the above problems, the present invention provides a remote rehabilitation monitoring and intervention method assisted by wearable devices. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention proposes a remote rehabilitation monitoring and intervention method assisted by wearable devices. The method first automatically configures monitoring indicators and their collection frequencies according to the rehabilitation parts, collects and preprocesses the monitoring data, and then encrypts and uploads them to the cloud server. Secondly, the preprocessed data is analyzed using a machine learning evaluation model to obtain the comprehensive health score of the rehabilitated person and the risk score of the main monitoring indicators. Thirdly, the rehabilitation stage of the rehabilitated person is determined according to the rehabilitation goals and scores, and the corresponding rehabilitation guidance strategy is generated through the strategy generation model. Finally, the training process of the rehabilitated person is monitored in real time through wearable devices, and the data is fed back to the online physician for real-time strategy adjustment. The method realizes intelligent monitoring and intervention of remote rehabilitation and improves the efficiency and accuracy of rehabilitation.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] Wearable device-assisted remote rehabilitation monitoring and intervention methods include:

[0007] S1. According to the rehabilitation part, the wearable device automatically configures the main monitoring index and the auxiliary monitoring index and their corresponding collection frequency, obtains the monitoring index data of the rehabilitation part of the rehabilitation person, and pre-processes the obtained monitoring index data, encrypts it and uploads it to the cloud server for storage;

[0008] S2. Build and train an evaluation model through machine learning, input the preprocessed monitoring indicator data into the evaluation model, and obtain the comprehensive health score of the rehabilitated person and the risk score of the main monitoring indicator;

[0009] S3. According to the rehabilitation goal, configure the multi-level standard rehabilitation stage, obtain the corresponding rehabilitation stage of the rehabilitated person according to the obtained comprehensive health score and the configured multi-level standard rehabilitation stage, feed back the obtained corresponding rehabilitation stage of the rehabilitated person and the multi-indicator risk score to the strategy generation model configured by the cloud server, and generate the corresponding rehabilitation guidance strategy;

[0010] S4. Feedback the generated rehabilitation guidance strategy to the rehabilitation person, and monitor the monitoring indicator data of the rehabilitation person in the rehabilitation training process according to the rehabilitation guidance strategy in real time through wearable devices, obtain the change trend of the real-time comprehensive health score and multi-indicator risk score, and feed back the obtained change trend to the online rehabilitation physician to adjust the generated rehabilitation guidance strategy in real time;

[0011] The specific steps for automatically configuring the main monitoring indicators and auxiliary monitoring indicators and their corresponding collection frequencies in S1 include:

[0012] S101, automatically judging the rehabilitation type through the rehabilitation part and historical medical records data input by the rehabilitation patient and using the discrimination algorithm configured in the device to obtain the rehabilitation type corresponding to the rehabilitation part;

[0013] S102, input the acquired rehabilitation type and rehabilitation part into the built-in dynamic indicator collection matching library, obtain the main monitoring indicator and auxiliary monitoring indicator corresponding to the rehabilitation part and rehabilitation type and their corresponding collection frequency through the matching algorithm, and deploy the acquired main monitoring indicator and auxiliary monitoring indicator and their corresponding collection frequency to the monitoring sensor corresponding to the wearable device;

[0014] The steps of constructing S102 dynamic indicator collection and matching library include:

[0015] S1021. Obtain rehabilitation diagnosis and treatment medical record data corresponding to different levels of rehabilitation types for different rehabilitation parts, and simultaneously obtain monitoring indicator data collected during historical rehabilitation training corresponding to different levels of rehabilitation types and their corresponding collection frequencies, and perform cleaning and deduplication preprocessing on the collected data to obtain an indicator library data set;

[0016] S1022, constructing and pre-training an entity-relationship extraction model through a Chinese pre-trained Bert sub-model and a correlation sub-model, inputting the obtained indicator library data set into the Chinese pre-trained Bert sub-model in the entity-relationship extraction model, and obtaining rehabilitation parts, rehabilitation types and their corresponding monitoring indicator entity sets and relationship data;

[0017] S1023, input the acquired rehabilitation type and the corresponding monitoring indicator entity set and relationship data into the correlation sub-model in the entity-relationship extraction model, and calculate the correlation coefficient p between different monitoring indicators and rehabilitation types in the monitoring indicator entity set. ij , where p ij represents the correlation coefficient between the i-th monitoring indicator and the j-th rehabilitation type;

[0018] S1024. Set the index coefficient threshold and when Then the i-th monitoring indicator is the main monitoring indicator set of the j-th rehabilitation type. When , the i-th monitoring indicator is the auxiliary monitoring indicator of the j-th rehabilitation type. When , the i-th monitoring indicator is not a monitoring indicator of the j-th rehabilitation type;

[0019] S1025, constructing an indicator triplet set according to the obtained rehabilitation parts, rehabilitation types and their corresponding monitoring indicator entity sets and relationship data and the main monitoring indicator and auxiliary monitoring indicator relationships obtained in S1024;

[0020] S1026. Input the constructed indicator triple set into the graph database to obtain a dynamic indicator collection and matching library, and set a five-level indicator node in the dynamic indicator collection and matching library;

[0021] S1027. When the person who is recovering inputs the corresponding rehabilitation part and medical record data through the wearable device interactive interface, the corresponding main monitoring index and auxiliary monitoring index are automatically matched and configured in the monitoring sensor of the wearable device through the dynamic index collection and matching library;

[0022] S1028. When the main monitoring indicators and auxiliary monitoring indicators corresponding to the rehabilitation part are not in the dynamic indicator collection and matching library, the main monitoring indicators and auxiliary monitoring indicators with the greatest correlation with the rehabilitation type corresponding to the rehabilitation part are configured to the corresponding monitoring sensors through a matching algorithm.

[0023] Specifically, the specific steps of obtaining the comprehensive health score of the recovered person and the risk score of the main monitoring indicators in S2 include:

[0024] S201, construct and train a comprehensive evaluation model based on a comprehensive fuzzy algorithm, input the main monitoring indicators and auxiliary monitoring indicators of the rehabilitation type corresponding to the rehabilitation part of the rehabilitated person into the comprehensive evaluation model to calculate the comprehensive health score of the rehabilitation part of the rehabilitated person;

[0025] S202, obtaining a standard normal interval of a main monitoring indicator and historical main monitoring indicator data, and using the obtained standard normal interval and historical main monitoring indicator data to calculate an interval upper limit error or an interval lower limit error;

[0026] S203, using the obtained interval upper limit error or interval lower limit error to train the risk assessment model constructed by the support vector machine to obtain a trained risk assessment model;

[0027] S204. Input the main monitoring indicator data obtained in real time by the wearable device into the risk assessment model, calculate the real-time interval upper limit error or interval lower limit error, multiply the calculated interval upper limit error or interval lower limit error by 10, and obtain the corresponding main monitoring indicator risk score.

[0028] Specifically, the specific steps of calculating the upper limit error or the lower limit error of the interval include:

[0029] S2021. Let [ak, bk] be the standard normal interval of the kth main monitoring indicator among the main monitoring indicators, and ak and bk be the lower threshold and upper threshold of the standard normal interval respectively;

[0030] S2022. Set the obtained kth main monitoring indicator data to ck. When ck<ak, the lower limit error of the interval is ek1=ak-ck. When ck>bk, the upper limit error of the interval is ek2=bk-ck.

[0031] Specifically, the five-level indicator nodes set in S1026 are: rehabilitation part is the first-level indicator node, rehabilitation type is the second-level indicator node, different levels of rehabilitation types are the third-level indicator node, main monitoring indicators and auxiliary monitoring indicators are the fourth-level indicator nodes, and the monitoring collection frequency corresponding to the main monitoring indicator and the monitoring collection frequency corresponding to the auxiliary monitoring indicator are the fifth-level indicator nodes.

[0032] Specifically, the specific steps of S4 include:

[0033] S401, obtaining the real-time primary monitoring index data and auxiliary monitoring index data of the rehabilitated person during the rehabilitation training process according to the rehabilitation guidance strategy, and calculating the real-time comprehensive health score and the primary monitoring index risk score of the rehabilitated person through the processes S201-S204;

[0034] S402, comparing the obtained real-time comprehensive health score and main monitoring indicator risk score with the comprehensive health score and main monitoring indicator risk score at the previous moment, when the real-time comprehensive health score is greater than the comprehensive health score at the previous moment and the main monitoring indicator risk score is less than the main monitoring indicator risk score at the previous moment, it indicates that the current rehabilitation guidance strategy is effective, and the current rehabilitation guidance strategy is continued to be used for rehabilitation training until the rehabilitation goal is achieved;

[0035] S403. When the real-time comprehensive health score is less than or equal to the comprehensive health score at the previous moment or the main monitoring indicator risk score is greater than or equal to the main monitoring indicator risk score at the previous moment, it means that the current rehabilitation guidance strategy has defects. The real-time comprehensive health score, the main monitoring indicator risk score and the current rehabilitation guidance strategy are fed back to the online rehabilitation physician, and the current rehabilitation guidance strategy is adjusted until the rehabilitation goal is achieved.

[0036] A computer-readable storage medium stores computer instructions, which, when executed, execute a wearable device-assisted remote rehabilitation monitoring and intervention method.

[0037] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor implements a wearable device-assisted remote rehabilitation monitoring and intervention method when executing the computer program.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] 1. Aiming at the deficiencies of the prior art, the present invention automatically configures the main monitoring indicators and auxiliary monitoring indicators and their corresponding collection frequencies through a dynamic indicator collection matching library, making the wearable device more targeted in data collection, thereby effectively avoiding data redundancy and uselessness, and improving the efficiency and accuracy of data collection. This not only reduces the burden on the rehabilitated person, but also prolongs the use time of the wearable device and reduces maintenance costs.

[0040] 2. In view of the deficiencies of the prior art, the present invention constructs and trains an evaluation model through machine learning technology, which can intelligently analyze the comprehensive health score and main monitoring indicator risk score of the rehabilitated person, which not only improves the accuracy and objectivity of the evaluation, but also provides the rehabilitated person with more scientific and accurate rehabilitation guidance. In addition, the present invention also understands the rehabilitation progress and potential risks of the rehabilitated person by real-time monitoring the process of the rehabilitated person performing rehabilitation training according to the rehabilitation guidance strategy, and obtains the changing trend of the real-time comprehensive health score and multi-indicator risk score, so that the online rehabilitation physician can make more targeted real-time adjustments to the rehabilitation guidance strategy, which greatly improves the scientificity and effectiveness of the rehabilitation process and provides the rehabilitated person with more comprehensive and personalized rehabilitation services. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A flow chart of the wearable device-assisted remote rehabilitation monitoring and intervention method of the present invention;

[0042] Figure 2 It is a five-level indicator node diagram set in the dynamic indicator collection and matching library of the present invention. DETAILED DESCRIPTION

[0043] Example 1

[0044] See also Figure 1 , an embodiment of the present invention provides: a wearable device-assisted remote rehabilitation monitoring and intervention method, comprising the following steps:

[0045] S1. According to the rehabilitation part, the wearable device automatically configures the main monitoring index and the auxiliary monitoring index and their corresponding collection frequency, obtains the rehabilitation part monitoring index data of the rehabilitation person, and pre-processes the acquired monitoring index data and encrypts and uploads it to the cloud server for storage; in this embodiment, the acquired monitoring index data is pre-processed and encrypted and uploaded to the cloud server for transmission via Bluetooth, Wi-Fi or mobile network;

[0046] Furthermore, the specific steps of automatically configuring the primary monitoring indicator and the secondary monitoring indicator and their corresponding collection frequencies in S1 include:

[0047] S101. Through the rehabilitation site and historical medical records data input by the rehabilitation person, the rehabilitation type is automatically judged through the discrimination algorithm configured in the device, and the rehabilitation type of the corresponding rehabilitation site is obtained; further, the discrimination algorithm in this embodiment includes a decision tree algorithm, a random forest and a K nearest neighbor algorithm; the rehabilitation site in this embodiment is specifically set by the specific rehabilitation person, and the rehabilitation type is set by the rehabilitation site, for example, the rehabilitation site is a knee joint, a wrist or a spine, and the corresponding rehabilitation type is muscle injury, arthritis or fracture healing, etc.; the wearable device includes a smart bracelet, an electrocardiogram patch or a blood pressure monitor, and the specific wearable device is selected according to the different rehabilitation sites;

[0048] S102, input the acquired rehabilitation type and rehabilitation part into the built-in dynamic indicator collection matching library, obtain the main monitoring indicator and auxiliary monitoring indicator corresponding to the rehabilitation part and rehabilitation type and their corresponding collection frequency through the matching algorithm, and deploy the acquired main monitoring indicator and auxiliary monitoring indicator and their corresponding collection frequency to the monitoring sensor corresponding to the wearable device. Further, the matching algorithm includes a matching algorithm constructed based on the Chinese pre-trained Bert model, a rule engine algorithm and a fuzzy logic algorithm;

[0049] Furthermore, in this embodiment, the step of constructing a dynamic indicator collection matching library in S102 includes:

[0050] S1021. Obtain rehabilitation diagnosis and treatment medical record data corresponding to different levels of rehabilitation types for different rehabilitation parts, and at the same time obtain monitoring indicator data collected during historical rehabilitation training corresponding to different levels of rehabilitation types and their corresponding collection frequencies, and perform cleaning and deduplication preprocessing on the collected data to obtain an indicator library data set; further, the monitoring indicator data in this embodiment is obtained from professional physician diagnosis books and medical databases, and different levels of rehabilitation types in this embodiment are obtained through the diagnosis and treatment medical records issued by doctors in the rehabilitation diagnosis and treatment medical record data. In this embodiment, different levels include mild, moderate and severe, such as mild Achilles tendon strain, moderate Achilles tendon strain and severe Achilles tendon strain, i.e. Achilles tendon rupture; these different levels of rehabilitation types are all included in the rehabilitation diagnosis and treatment medical record data, and can be obtained through the electronic diagnosis and treatment medical records or paper diagnosis and treatment medical records of the rehabilitated person;

[0051] S1022, constructing and pre-training an entity-relationship extraction model through the Chinese pre-trained Bert sub-model and the correlation sub-model, inputting the obtained indicator library data set into the Chinese pre-trained Bert sub-model in the entity-relationship extraction model, and obtaining the rehabilitation site, rehabilitation type and its corresponding monitoring indicator entity set and relationship data; further, in this embodiment, the correlation sub-model adopts the cosine correlation formula to calculate the correlation coefficient between the rehabilitation type and the corresponding monitoring indicator;

[0052] S1023, input the acquired rehabilitation type and the corresponding monitoring indicator entity set and relationship data into the correlation sub-model in the entity-relationship extraction model, and calculate the correlation coefficient p between different monitoring indicators and rehabilitation types in the monitoring indicator entity set. ij , where p ij represents the correlation coefficient between the i-th monitoring indicator and the j-th rehabilitation type;

[0053] S1024. Set the index coefficient threshold and when Then the i-th monitoring indicator is the main monitoring indicator set of the j-th rehabilitation type. When , the i-th monitoring indicator is the auxiliary monitoring indicator of the j-th rehabilitation type. When , the i-th monitoring indicator is not a monitoring indicator of the j-th rehabilitation type; further, the indicator coefficient threshold in this embodiment and It is set by a rehabilitation physician in this field, for example

[0054] S1025. Construct an indicator triple set according to the obtained rehabilitation site, rehabilitation type and its corresponding monitoring indicator entity set and relationship data and the relationship between the main monitoring indicator and the auxiliary monitoring indicator obtained in S1024. Exemplarily, in this embodiment, the indicator triple set is taken as an example of the rehabilitation site of the Achilles tendon, specifically [Achilles tendon, rehabilitation type, Achilles tendon strain], [Achilles tendon strain, grade, severe], [Achilles tendon strain, grade, severe Achilles tendon strain], [severe Achilles tendon strain, main monitoring indicator, main monitoring indicator 1], [severe Achilles tendon strain, auxiliary monitoring indicator, auxiliary monitoring indicator 1], [main monitoring indicator 1, acquisition frequency, frequency value 1], [auxiliary monitoring indicator 1, auxiliary monitoring indicator, auxiliary monitoring indicator 2], where the main monitoring indicator 1 and auxiliary monitoring indicator 1 are only references to specific indicators for the Achilles tendon, and are replaced by specific indicators in actual applications;

[0055] S1026. Input the constructed indicator triple set into the graph database to obtain a dynamic indicator collection and matching library, and set a five-level indicator node in the dynamic indicator collection and matching library; refer to Figure 2 The five-level indicator nodes set in this embodiment are specifically: rehabilitation part is a first-level indicator node, rehabilitation type is a second-level indicator node, different levels of rehabilitation types are third-level indicator nodes, main monitoring indicators and auxiliary monitoring indicators are fourth-level indicator nodes, and the monitoring acquisition frequency corresponding to the main monitoring indicator and the monitoring acquisition frequency corresponding to the auxiliary monitoring indicator are fifth-level indicator nodes; Figure 2 As shown in the figure, rehabilitation part 1 represents the specific part that needs to be rehabilitated during the actual rehabilitation process of the rehabilitator, such as the Achilles tendon. Rehabilitation type 1 and rehabilitation type 2 are the actual rehabilitation types corresponding to rehabilitation part 1, such as muscle injury and fracture healing. The third-level indicator node is a specific division corresponding to different levels of rehabilitation type 1 and rehabilitation type 2, such as mild, moderate and moderate. The fourth-level indicator node and the fifth-level indicator node are also specific divisions of different main monitoring indicators and auxiliary monitoring indicators and their corresponding collection frequencies. We will not go into details here. For details, please refer to Figure 2 ;

[0056] In this embodiment, the collection frequency is set according to the acquired historical data, such as the monitoring collection frequency corresponding to the main monitoring indicator and the monitoring collection frequency corresponding to the auxiliary monitoring indicator;

[0057] S1027. When the person who is recovering inputs the corresponding rehabilitation part and medical record data through the wearable device interactive interface, the corresponding main monitoring index and auxiliary monitoring index are automatically matched and configured in the monitoring sensor of the wearable device through the dynamic index collection and matching library;

[0058] S1028. When the main monitoring indicators and auxiliary monitoring indicators corresponding to the rehabilitation part are not in the dynamic indicator collection and matching library, the main monitoring indicators and auxiliary monitoring indicators with the greatest correlation with the rehabilitation type corresponding to the rehabilitation part are configured to the corresponding monitoring sensors through a matching algorithm.

[0059] S2. Build and train an evaluation model through machine learning, input the preprocessed monitoring index data into the evaluation model, and obtain the comprehensive health score of the rehabilitated person and the risk score of the main monitoring index; further, the specific steps of obtaining the comprehensive health score of the rehabilitated person and the risk score of the main monitoring index in S2 of this embodiment include:

[0060] S201, construct and train a comprehensive evaluation model based on a comprehensive fuzzy algorithm, input the main monitoring indicators and auxiliary monitoring indicators of the rehabilitation type corresponding to the rehabilitation part of the rehabilitated person into the comprehensive evaluation model to calculate the comprehensive health score of the rehabilitation part of the rehabilitated person;

[0061] S202, obtaining a standard normal interval of the main monitoring indicator and historical main monitoring indicator data, and calculating an upper limit error or a lower limit error of the interval using the obtained standard normal interval and historical main monitoring indicator data; further, the specific steps of calculating the upper limit error or the lower limit error of the interval include:

[0062] S2021. Let [ak, bk] be the standard normal interval of the kth main monitoring indicator among the main monitoring indicators, and ak and bk be the lower threshold and upper threshold of the standard normal interval respectively; the standard normal interval of the main monitoring indicator in this embodiment is set by the value range of the corresponding monitoring indicator in the existing medical library;

[0063] S2022, setting the obtained k-th main monitoring indicator data to ck, when ck<ak, the interval lower limit error is ek1=ak-ck, when ck>bk, the interval upper limit error is ek2=bk-ck;

[0064] S203, using the obtained interval upper limit error or interval lower limit error to train the risk assessment model constructed by the support vector machine to obtain a trained risk assessment model;

[0065] S204. Input the main monitoring indicator data obtained in real time by the wearable device into the risk assessment model, calculate the real-time interval upper limit error or interval lower limit error, multiply the calculated interval upper limit error or interval lower limit error by 10, and obtain the corresponding main monitoring indicator risk score.

[0066] S3. According to the rehabilitation goals, multi-level standard rehabilitation stages are configured, and the corresponding rehabilitation stages of the rehabilitated persons are obtained according to the obtained comprehensive health scores and the configured multi-level standard rehabilitation stages. The obtained corresponding rehabilitation stages of the rehabilitated persons and the multi-indicator risk scores are fed back to the strategy generation model configured by the cloud server to generate corresponding rehabilitation guidance strategies. Furthermore, the multi-level standard rehabilitation stages in this embodiment are set by professional physicians in this field based on the specific medical diagnosis data and rehabilitation goals of different rehabilitation parts of historical rehabilitated persons. Furthermore, the strategy generation model in this embodiment is implemented through the existing GPT model, the Chinese Bert model based on pre-training, or the Wenxin Yiyanda model. The corresponding rehabilitation guidance strategies include providing patients with customized rehabilitation exercise plans and dietary recommendations, etc.

[0067] S4, feeding back the generated rehabilitation guidance strategy to the rehabilitation person, and monitoring the monitoring index data of the rehabilitation person in the rehabilitation training process according to the rehabilitation guidance strategy in real time through wearable devices, obtaining the change trend of the real-time comprehensive health score and multi-index risk score, and feeding back the obtained change trend to the online rehabilitation physician to adjust the generated rehabilitation guidance strategy in real time. Further, the specific steps of S4 include:

[0068] S401, obtaining the real-time primary monitoring index data and auxiliary monitoring index data of the rehabilitated person during the rehabilitation training process according to the rehabilitation guidance strategy, and calculating the real-time comprehensive health score and the primary monitoring index risk score of the rehabilitated person through the processes S201-S204;

[0069] S402, comparing the obtained real-time comprehensive health score and main monitoring indicator risk score with the comprehensive health score and main monitoring indicator risk score at the previous moment, when the real-time comprehensive health score is greater than the comprehensive health score at the previous moment and the main monitoring indicator risk score is less than the main monitoring indicator risk score at the previous moment, it indicates that the current rehabilitation guidance strategy is effective, and the current rehabilitation guidance strategy is continued to be used for rehabilitation training until the rehabilitation goal is achieved;

[0070] S403: When the real-time comprehensive health score is less than or equal to the comprehensive health score at the previous moment or the main monitoring indicator risk score is greater than or equal to the main monitoring indicator risk score at the previous moment, it indicates that the current rehabilitation guidance strategy has defects, and the real-time comprehensive health score, the main monitoring indicator risk score and the current rehabilitation guidance strategy are fed back to the online rehabilitation physician to adjust the current rehabilitation guidance strategy until the rehabilitation goal is achieved. In this embodiment, adjusting the current rehabilitation guidance strategy includes adjusting the difficulty and intensity of the rehabilitation guidance strategy and the content and measurement of the dietary advice according to the real-time comprehensive health score and the main monitoring indicator risk score of the rehabilitator.

[0071] This process collects data in real time through wearable devices, calculates the real-time health score and risk score of the rehabilitated person, and compares it with the data at the previous moment, which can quickly determine the effectiveness of the rehabilitation guidance strategy. Once the strategy is found to be ineffective, the system can immediately feedback to the online rehabilitation physician for adjustment to ensure that the rehabilitated person can carry out rehabilitation training according to the strategy that best suits his or her individual situation, thereby accelerating the rehabilitation process and improving the quality of rehabilitation. This intelligent and personalized rehabilitation management method has greatly improved the efficiency and effectiveness of rehabilitation treatment.

[0072] Example 2

[0073] A computer-readable storage medium stores computer instructions, which, when executed, execute a wearable device-assisted remote rehabilitation monitoring and intervention method.

[0074] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor implements a wearable device-assisted remote rehabilitation monitoring and intervention method when executing the computer program.

[0075] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the scope of protection of the purpose of the present invention and the claims, and all of these are within the protection of the present invention.

[0076] If the disclosed technical solution involves personal information, the product using the disclosed technical solution has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the disclosed technical solution involves sensitive personal information, the product using the disclosed technical solution has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, when the personal information processing rules are notified by obvious signs / information, the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

Claims

1. A wearable device-assisted remote rehabilitation monitoring and intervention method, characterized in that the steps include: S1. According to the rehabilitation part, the wearable device automatically configures the main monitoring index and the auxiliary monitoring index and their corresponding collection frequency, obtains the monitoring index data of the rehabilitation part of the rehabilitation person, and pre-processes the obtained monitoring index data, encrypts it and uploads it to the cloud server for storage; S2. Build and train an evaluation model through machine learning, input the preprocessed monitoring indicator data into the evaluation model, and obtain the comprehensive health score of the rehabilitated person and the risk score of the main monitoring indicator; S3. According to the rehabilitation goal, configure the multi-level standard rehabilitation stage, obtain the corresponding rehabilitation stage of the rehabilitated person according to the obtained comprehensive health score and the configured multi-level standard rehabilitation stage, feed back the obtained corresponding rehabilitation stage of the rehabilitated person and the multi-indicator risk score to the strategy generation model configured by the cloud server, and generate the corresponding rehabilitation guidance strategy; S4. Feedback the generated rehabilitation guidance strategy to the rehabilitation person, and monitor the monitoring indicator data of the rehabilitation person in the rehabilitation training process according to the rehabilitation guidance strategy in real time through wearable devices, obtain the change trend of the real-time comprehensive health score and multi-indicator risk score, and feed back the obtained change trend to the online rehabilitation physician to adjust the generated rehabilitation guidance strategy in real time; The specific steps of automatically configuring the main monitoring indicators and auxiliary monitoring indicators and their corresponding collection frequencies in S1 include: S101, automatically judging the rehabilitation type through the rehabilitation part and historical medical records data input by the rehabilitation patient and using the discrimination algorithm configured in the device to obtain the rehabilitation type corresponding to the rehabilitation part; S102, input the acquired rehabilitation type and rehabilitation part into the built-in dynamic indicator collection matching library, obtain the main monitoring indicator and auxiliary monitoring indicator corresponding to the rehabilitation part and rehabilitation type and their corresponding collection frequency through the matching algorithm, and deploy the acquired main monitoring indicator and auxiliary monitoring indicator and their corresponding collection frequency to the monitoring sensor corresponding to the wearable device; The step of constructing the dynamic indicator collection and matching library in S102 includes: S1021. Obtain rehabilitation diagnosis and treatment medical record data corresponding to different levels of rehabilitation types for different rehabilitation parts, and simultaneously obtain monitoring indicator data collected during historical rehabilitation training corresponding to different levels of rehabilitation types and their corresponding collection frequencies, and perform cleaning and deduplication preprocessing on the collected data to obtain an indicator library data set; S1022, constructing and pre-training an entity-relationship extraction model through a Chinese pre-trained Bert sub-model and a correlation sub-model, inputting the obtained indicator library data set into the Chinese pre-trained Bert sub-model in the entity-relationship extraction model, and obtaining rehabilitation parts, rehabilitation types and their corresponding monitoring indicator entity sets and relationship data; S1023, input the acquired rehabilitation type and the corresponding monitoring indicator entity set and relationship data into the correlation sub-model in the entity-relationship extraction model, and calculate the correlation coefficient p between different monitoring indicators and rehabilitation types in the monitoring indicator entity set. ij , where p ij represents the correlation coefficient between the i-th monitoring indicator and the j-th rehabilitation type; S1024. Set the index coefficient threshold and when Then the i-th monitoring indicator is the main monitoring indicator set of the j-th rehabilitation type. When , the i-th monitoring indicator is the auxiliary monitoring indicator of the j-th rehabilitation type. When , the i-th monitoring indicator is not a monitoring indicator of the j-th rehabilitation type; S1025, constructing an indicator triplet set according to the obtained rehabilitation parts, rehabilitation types and their corresponding monitoring indicator entity sets and relationship data and the main monitoring indicator and auxiliary monitoring indicator relationships obtained in S1024; S1026. Input the constructed indicator triple set into the graph database to obtain a dynamic indicator collection and matching library, and set a five-level indicator node in the dynamic indicator collection and matching library; S1027. When the person who is recovering inputs the corresponding rehabilitation part and medical record data through the wearable device interactive interface, the corresponding main monitoring index and auxiliary monitoring index are automatically matched and configured in the monitoring sensor of the wearable device through the dynamic index collection and matching library; S1028. When the main monitoring indicators and auxiliary monitoring indicators corresponding to the rehabilitation part are not in the dynamic indicator collection and matching library, the main monitoring indicators and auxiliary monitoring indicators with the greatest correlation with the rehabilitation type corresponding to the rehabilitation part are configured to the corresponding monitoring sensors through a matching algorithm.

2. The wearable device-assisted remote rehabilitation monitoring and intervention method according to claim 1, characterized in that: The specific steps of obtaining the comprehensive health score of the recovered person and the risk score of the main monitoring indicators in S2 include: S201, construct and train a comprehensive evaluation model based on a comprehensive fuzzy algorithm, input the main monitoring indicators and auxiliary monitoring indicators of the rehabilitation type corresponding to the rehabilitation part of the rehabilitated person into the comprehensive evaluation model to calculate the comprehensive health score of the rehabilitation part of the rehabilitated person; S202, obtaining a standard normal interval of a main monitoring indicator and historical main monitoring indicator data, and using the obtained standard normal interval and historical main monitoring indicator data to calculate an interval upper limit error or an interval lower limit error; S203, using the obtained interval upper limit error or interval lower limit error to train the risk assessment model constructed by the support vector machine to obtain a trained risk assessment model; S204. Input the main monitoring indicator data obtained in real time by the wearable device into the risk assessment model, calculate the real-time interval upper limit error or interval lower limit error, multiply the calculated interval upper limit error or interval lower limit error by 10, and obtain the corresponding main monitoring indicator risk score.

3. The wearable device-assisted remote rehabilitation monitoring and intervention method according to claim 2, characterized in that: The specific steps of calculating the upper limit error or the lower limit error of the interval include: S2021. Let [ak, bk] be the standard normal interval of the kth main monitoring indicator among the main monitoring indicators, and ak and bk be the lower threshold and upper threshold of the standard normal interval respectively; S2022. Set the obtained kth main monitoring indicator data to ck. When ck<ak, the lower limit error of the interval is ek1=ak-ck. When ck>bk, the upper limit error of the interval is ek2=bk-ck.

4. The wearable device-assisted remote rehabilitation monitoring and intervention method according to claim 3, characterized in that: The five-level indicator nodes set in S1026 are specifically: rehabilitation part is a first-level indicator node, rehabilitation type is a second-level indicator node, different levels of rehabilitation types are third-level indicator nodes, main monitoring indicators and auxiliary monitoring indicators are fourth-level indicator nodes, and the monitoring collection frequency corresponding to the main monitoring indicator and the monitoring collection frequency corresponding to the auxiliary monitoring indicator are fifth-level indicator nodes.

5. The wearable device-assisted remote rehabilitation monitoring and intervention method according to claim 4, characterized in that: The specific steps of S4 include: S401, obtaining the real-time primary monitoring index data and auxiliary monitoring index data of the rehabilitated person during the rehabilitation training process according to the rehabilitation guidance strategy, and calculating the real-time comprehensive health score and the primary monitoring index risk score of the rehabilitated person through the processes S201-S204; S402, comparing the obtained real-time comprehensive health score and main monitoring indicator risk score with the comprehensive health score and main monitoring indicator risk score at the previous moment, when the real-time comprehensive health score is greater than the comprehensive health score at the previous moment and the main monitoring indicator risk score is less than the main monitoring indicator risk score at the previous moment, it indicates that the current rehabilitation guidance strategy is effective, and the current rehabilitation guidance strategy is continued to be used for rehabilitation training until the rehabilitation goal is achieved; S403. When the real-time comprehensive health score is less than or equal to the comprehensive health score at the previous moment or the main monitoring indicator risk score is greater than or equal to the main monitoring indicator risk score at the previous moment, it means that the current rehabilitation guidance strategy has defects. The real-time comprehensive health score, the main monitoring indicator risk score and the current rehabilitation guidance strategy are fed back to the online rehabilitation physician, and the current rehabilitation guidance strategy is adjusted until the rehabilitation goal is achieved.

6. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed, the wearable device-assisted remote rehabilitation monitoring and intervention method described in any one of claims 1-5 is executed.

7. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the wearable device-assisted remote rehabilitation monitoring and intervention method described in any one of claims 1 to 5 is implemented.

Images