Remote medical care monitoring and early warning method and system
By performing lightweight early warning model processing and critical data judgment on edge devices, the data delay and leakage problems of telemedicine monitoring system when network instability is solved, rapid response and security early warning are achieved, and system flexibility and security are improved.
Patent Information
- Application Number
- CN202510263378.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-06
- Publication Date
- 2025-07-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
When the network condition of the telemedicine nursing monitoring and early warning system is unstable, the delay in data transmission affects the timeliness of early warning and there is a risk of data leakage.
Lightweight early warning model processing is carried out on edge devices, and patient body metric information is analyzed through fuzzy rules and machine learning models. Only critical data is sent to the cloud, and non-critical data is stored locally. Edge computing is used to reduce network bandwidth pressure and reduce the risk of data leakage.
It realizes rapid warning response when network fluctuations, reduces data delays and leakage risks, and improves the timeliness of warnings and system flexibility and security.
Smart Images
Figure CN120356593A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of medical monitoring and early warning, and more specifically, the present invention relates to a multi-remote medical care monitoring and early warning method and system. Background Art
[0002] With the continuous progress of the Internet and information technology, telemedicine has become an important medical service model. It greatly meets the needs of patients for medical services through remote monitoring, consultation and treatment. In recent years, with the rapid development of the Internet, mobile communication and sensing technologies, telemedicine has been widely applied.
[0003] However, for a remote medical care monitoring and early warning system, real-time data transmission may be affected by the network condition, resulting in data delay and affecting the timeliness of early warning. The system highly depends on a stable Internet connection, and network interruption will affect the monitoring and early warning functions. In addition, during the data transmission and storage processes, there is a risk of being hacked and data leakage, which may endanger the privacy of patients.
[0004] In view of the above problems, the present invention proposes a solution. Summary of the Invention
[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a remote medical care monitoring and early warning method and system to solve the problems raised in the above background art.
[0006] To achieve the above object, the present invention provides the following technical solutions:
[0007] A remote medical care monitoring and early warning method includes the following steps:
[0008] Step S1, analyze and determine the network state, normalize the network state data, after processing, use the preference ranking organization method for enrichment evaluation (PROMETHEE) to assign weights, and use the weighted sum of the weight assignment to obtain a comprehensive score to analyze and determine the network state;
[0009] Step S2, when the network state is good, normally send the monitoring data to the cloud; when the network state is poor, obtain the patient's physical index information and corresponding cases; comprehensively analyze the physical index information and the corresponding cases to obtain the importance degree of the physical index information; a lightweight early warning model is installed on the edge device, and the early warning detection result is obtained through the processing of the lightweight early warning model;
[0010] Step S3, comprehensively evaluate the importance degree of the patient's physical index information and the early warning detection result by using fuzzy rules to obtain the key information to be sent to the cloud; for non-critical information, it is stored locally and sent to the cloud after the network state returns to good.
[0011] In a preferred embodiment, in step S1, the network status is determined by comprehensively analyzing network status data. The network status data with different dimensions is normalized, and then the weight is assigned by using the preference ranking organization method for enrichment evaluation (PROMETHEE). The network scores are obtained by adding the weights of different network status data. The network status is obtained according to the magnitude of the network scores.
[0012] In a preferred embodiment, in step S2, the importance degree of the patient's physical index information is obtained by comprehensively analyzing the patient's physical index information and the corresponding case analysis. The random forest algorithm is selected for this machine learning model. The importance degree of various physical index information of the patient for the corresponding case patient is output through the random forest algorithm, and the evaluation of the importance degree of the index information is obtained. And the early warning detection result is obtained after being processed by the lightweight early warning model.
[0013] In a preferred embodiment, in step S3, after receiving the patient's physical index information and the processing result of the lightweight early warning model, the patient's physical index information and the early warning detection result are defined as input variables, and they are respectively divided into different fuzzy sets. The key information judgment result is defined as the output variable and is divided into a fuzzy set. A set of fuzzy rules is formulated to describe the influence of different input variables on the output variable. Fuzzy reasoning is performed according to the fuzzy rules to determine the key information judgment result.
[0014] A remote medical care monitoring and early warning system for implementing the above-mentioned remote medical care monitoring and early warning method, including a data acquisition module, a data analysis module, and a data storage module.
[0015] The data acquisition module is used to obtain the patient's physical index information and the corresponding case, and send them to the data processing module to ensure the operation of the subsequent data processing module.
[0016] The data processing module is used to determine the importance degree of the patient's physical index information and the discrimination of sensitive information according to the data collected by the data acquisition module, and comprehensively analyze and discriminate the key data according to the importance degree of the patient's physical index information and the early warning detection result.
[0017] The data storage module is used to store all the data generated during the processing of the remote medical care detection and early warning system.
[0018] A remote medical care monitoring and early warning method and system of the present invention perform preliminary analysis and calculation on edge devices: when the data transmission network is normal, intelligently capture and analyze the patient's physical index information, realize local processing of sensitive data, thereby reducing the opportunity of transmission to the cloud and reducing the risk of data leakage. When the data transmission network is abnormal, the edge device plays a role in completing the preliminary analysis of the patient's physical index information. Through the evaluation of the patient's physical index information by the lightweight early warning model, medical staff can quickly and conveniently discover the patient's danger. By discriminating the key information in the patient's index information, only the processed key information is sent to the cloud, and other data is stored by the local edge device and sent to the cloud when the network is normal to reduce the pressure on the network bandwidth. At the same time, the complex model in the cloud will be regularly synchronized to the edge device to ensure that the lightweight early warning model of the edge device is kept updated.
[0019] A remote medical care monitoring and early warning method and system, the real-time data transmission may be affected by the network condition, resulting in data delay and affecting the timeliness of early warning. The system highly depends on a stable Internet connection, and network interruption will affect the monitoring and early warning functions. Detailed implementation mode
[0020] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.
[0021] Embodiment 1, the present invention includes a remote medical care monitoring and early warning method, as Figure 1 shown, including the following steps:
[0022] In step S1, tools such as iperf can be used to test the network throughput to see if there are fluctuations in the throughput. The instability of the throughput is often one of the manifestations of network fluctuations. If there are large fluctuations in the throughput (such as suddenly dropping from 10 Mbps to 1 Mbps), it indicates that the network is unstable.
[0023] Packet loss is one of the important indicators of network fluctuations. Use Ping to monitor the packet loss rate. If the packet loss rate exceeds 1%-2%, it may affect the normal application experience. You can also use tools such as mtr (Linux / macOS) or WinMTR (Windows) for packet loss monitoring. In the case of high latency and packet loss, the network quality will significantly decline, resulting in performance problems of the application program.
[0024] For applications that require high availability (such as IoT devices, real-time communication applications, etc.), heartbeat packets (keep-alive signals) can be sent regularly. If there are abnormal fluctuations in the response time of the heartbeat packets or the packet loss rate increases, the degree of network fluctuations can be judged.
[0025] Classify according to the network requirements in specific scenarios. Vital sign monitoring: such as real-time monitoring of vital signs such as heart rate, blood pressure, and blood oxygen, requires extremely low latency and high stability, especially the impact of the heartbeat packet loss rate and latency is the most significant. Medical data transmission: such as the transmission of big data such as medical records and images (X-ray films, CT scans, etc.), requires high throughput and a low data packet loss rate.
[0026] The process of comprehensively analyzing and determining the network status is as follows:
[0027] Receive the real-time data volume of monitored network throughput, packet loss rate, and Ping latency, and evaluate the important indicators of network performance:
[0028] Use linear normalization to map the real-time network throughput to the interval [0,1]. The formula is: In the formula, Normalized_Score is the value after normalization.
[0029] Use linear normalization to map the real-time latency and real-time packet loss rate to the interval [0,1]. The formula is:
[0030] In the formula, Normalized_Score is the value after normalization. Here we use the form of 1 - to ensure that the lower the latency and packet loss rate, the larger the normalized value, because lower latency and packet loss rate usually mean better network performance.
[0031] Specifically, considering the importance scores of latency, throughput, and packet loss rate in network performance, the weight assignment is carried out respectively according to the preference ranking method as shown in Table 1 below:
[0032]
[0033] Table 1
[0034] It should be noted that the setting of the preference ranking method can be adjusted according to the actual situation. For example, in the scenario of real-time vital sign monitoring, Ping latency and packet loss rate need to be considered first. Excessive latency or packet loss rate will lead to inaccurate real-time data and affect the doctor's decision-making. The transmission of medical images (such as CT, X-ray, MRI, etc.) data requires a relatively high bandwidth. Data packet loss may cause a decrease in image quality or loss of some image information, affecting the diagnostic accuracy.
[0035] After the importance scores of the three indicators in network performance are determined, the comprehensive score can be calculated: the comprehensive score of network status S = ∑(weighted scores of each normalized network performance indicator);
[0036] According to the comprehensive score of network status, set the evaluation method for network fluctuation status: when the comprehensive score is lower than 80%, the network status is poor; when the comprehensive score is higher than 80%, the network status is good. The weighted scoring method is an effective comprehensive analysis method. By assigning weights to multiple network performance indicators and combining the normalized values to calculate the weighted scores, the comprehensive score is finally obtained to evaluate the network condition. By reasonably setting the weights, scoring criteria, and calculation process, reliable evaluation results can be obtained.
[0037] Step S2, when the network status is poor, obtain the patient's physical index information and corresponding cases. Select a machine learning model, and through the machine learning model, output and judge the importance of various physical index information of the patient for the corresponding case patient, obtain the evaluation of the importance level, and obtain the warning detection result after processing by the lightweight warning model on the edge device. When the lightweight warning model detects and triggers the warning and reminder mechanism, it will issue sound prompts, vibration reminders, etc. to remind medical staff and patients.
[0038] It should be noted that the lightweight warning model is built on the edge device and responds to the changes in the patient's health in real time. The warning model can process data and issue alarms in an extremely short time. Through the collaboration of edge computing and the cloud, the lightweight warning model can not only ensure the fast response and low power consumption of the device side, but also rely on the powerful computing power of the cloud to continuously optimize and update the model. Technical means such as real-time update mechanism, dynamic threshold adjustment, and intelligent decision feedback enable the medical monitoring system to adapt to the changes of patients and the needs of the environment at any time, providing more accurate and personalized support for medical warnings. This combination provides the medical monitoring system with a high degree of flexibility, scalability, and security, while reducing the dependence on hardware resources. Such models usually face limitations in processing power, computing resources, data transmission bandwidth, etc. Therefore, when designing the lightweight warning model, while ensuring the accuracy and real-time nature of the warning, the consumption of computing and storage resources should be minimized as much as possible. The warning detection results of the lightweight warning model are of two types: abnormal and normal. The effective application of these results can significantly improve the patient's health management level and quality of life.
[0039] The warning results of the lightweight warning model are divided into two types: abnormal and normal. When an abnormality is detected, the edge device issues an alarm.
[0040] In step S2, an appropriate machine learning model can be selected, such as a support vector machine (SVM), a decision tree, a neural network, etc., to comprehensively analyze the body index information and the corresponding cases to determine the importance degree of the body index information.
[0041] Taking the random forest algorithm as an example for the machine learning model in this embodiment, in step S3, the following steps are further included:
[0042] Step S2.1: Let D be a data set containing m samples, and each sample j has k features cj1, cj2, cj3... cjk; and let the target variable be Tj, where Tj is the discrimination of important indicators.
[0043] Step S2.2: Preprocess the data by normalization to ensure the accuracy of the model. Randomly divide the data set D into a training set Dtrain and a test set Dtest, usually using a ratio of 70%-30%. The training set is used to build the model, and the test set is used to verify the model. The training set contains p samples, and the test set contains q samples, where p + q = m.
[0044] Step S2.3: Construct W decision trees, and each tree Uw is independently constructed. For each tree, a sample set of size p is drawn from the training set Dtrain by bootstrap sampling. At each splitting node, instead of using all k features, randomly select the square root of k features from the k features to find the best split.
[0045] Step S2.4: Each tree Uw generates a model Kw based on its training data. For a given test sample c, each tree Uw will output a prediction result Tw(c).
[0046] Step S2.5: Output all tree prediction outputs T(c); T(c) = mode T1(c), T2(c),…, Tw(c); and take the average value of all tree prediction outputs as the prediction result.
[0047] Tj is a binary label (0 or 1), indicating the non-important or important state.
[0048] It should be noted that the specific process of normalizing the data is similar to the previous text and will not be elaborated here.
[0049] Step S3: Comprehensively evaluate the important indicators and early warning detection results of the patient using fuzzy rules. The evaluation process is as follows:
[0050] Step S3.1, after receiving the patient's physical index information and processing it with the lightweight warning model, define the patient's physical index information and the warning detection result as input variables, and divide them into different fuzzy sets respectively.
[0051] For example, "Moderate", "Important" for the patient's physical index information; "Normal", "Abnormal" for the warning detection result.
[0052] Step S3.2, define the key information judgment result as the output variable, and divide it into a fuzzy set. For example, "Local stored", "Critical".
[0053] Step S3.3, formulate a set of fuzzy rules to describe the influence of different input variables on the output variable. The definition of the rules can be based on professional knowledge or obtained through data analysis and experiments. For example:
[0054] Mark the patient's physical index information as A, the warning detection result to be marked as B, and the key information judgment to be marked as C, then it can be defined as
[0055] Rule 1: IF (A is Moderate) AND (B is Normal) THEN (C is Local stored)
[0056] Rule 2: IF (A is Important) AND (B is Abnormal) THEN (C is Critical)
[0057] Rule 3: IF (A is Moderate) AND (B is Abnormal) THEN (C is Local stored) ...
[0059] Step S3.4, perform fuzzy reasoning according to the fuzzy rules to determine the key information judgment result.
[0060] Furthermore, for the judgment of different situations of the patient's physical index information and the warning detection result, it can be judged according to the actual situation. For example, when the body temperature data in the patient's physical index information exceeds 38.1°C, it is abnormal. For patients with fever symptoms, the importance of body temperature is relatively high. At this time, the patient's physical index information output is Important; if the warning detection result also outputs abnormal, Abnormal, then it is judged that this data is "Critical", that is, important data, and needs to be sent to the cloud, etc., which will not be elaborated here.
[0061] It should be noted that the division of the fuzzy sets can be adjusted according to the actual situation. For example, although three fuzzy sets are taken as an example in this embodiment, in fact, the patient's physical index information, early warning detection results, and key information judgment results can be divided into more than three sets to facilitate better discrimination of key information according to different situations.
[0062] Embodiment 2. The present invention also provides a remote medical care detection and early warning system, as Figure 2 shown, which includes a data acquisition module, a data processing module, and a data storage module;
[0063] The data acquisition module is used to obtain the patient's physical index information and corresponding cases, and send them to the data processing module to ensure the operation of the subsequent data processing module;
[0064] The data processing module is used to determine the discrimination of the patient's important indicators and sensitive information according to the data collected by the data acquisition module, and comprehensively analyze and discriminate the key data according to the patient's important indicators and early warning detection results;
[0065] The data storage module is used to store all the data generated during the processing of the remote medical care detection and early warning system.
[0066] The above formulas are all dimensionless and take their numerical values for calculation. The formula is a formula obtained by collecting a large amount of data for software simulation to approximate the real situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.
[0067] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.
[0068] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application of the technical solution and the inventive constraints. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.
[0069] In addition, in each embodiment of the present application, the various functional modules can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.
[0070] As described above, it is only the specific implementation manner of the present application. However, the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims described above.
[0071] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.
Claims
1. A remote medical care monitoring and warning method, characterized in that It includes the following steps: Step S1: Analyze and determine the network status, normalize the network status data, assign weights using the preference ranking organization method (PROMETHEE) after processing, and use weighted summation of the weight assignment to obtain a comprehensive score, and analyze and determine the network status. Step S2: When the network status is good, normally send the monitoring data to the cloud. When the network status is poor, obtain the patient's physical index information and corresponding cases; comprehensively analyze the physical index information and the corresponding cases on the edge device to obtain the importance degree of the physical index information; a lightweight warning model is installed on the edge device, and the warning detection result is obtained through the processing of the lightweight warning model. Step S3: Comprehensively evaluate the importance degree of the patient's physical index information and the warning detection result using fuzzy rules to obtain the key information to be sent to the cloud; for non-critical information, it is stored locally and sent to the cloud after the network status returns to good.
2. A remote medical care monitoring and warning method according to claim 1, wherein: In step S1, comprehensively analyze the network status data to determine the network status, normalize the network status data with different dimensions, and then use the preference ranking organization method (PROMETHEE) to assign weights. The network status data with different weights are added to obtain a network score; the network status is obtained according to the size of the network score.
3. A remote medical care monitoring and warning method according to claim 1, wherein: In step S2, comprehensively analyze the patient's physical index information and corresponding cases to obtain the importance degree of the patient's physical index information; in this case, the machine learning model selects the random forest algorithm, and the importance degree of various physical index information of the patient for the corresponding case patient is output through the random forest algorithm to obtain the evaluation of the importance degree of the index information; and the warning detection result is obtained after being processed by the lightweight warning model.
4. A remote medical care monitoring and warning method according to claim 1, wherein: In step S3, after receiving the patient's physical index information and the processing result of the lightweight warning model, define the patient's physical index information and the warning detection result as input variables, and divide them into different fuzzy sets respectively; define the key information judgment result as the output variable and divide it into a fuzzy set; formulate a set of fuzzy rules to describe the influence of different input variables on the output variable; perform fuzzy reasoning according to the fuzzy rules to determine the key information judgment result.
5. A remote medical care monitoring and warning system, based on the remote medical care monitoring and warning method according to any one of claims 1-4, characterized in that, It includes a data acquisition module, a data analysis module, and a data storage module; The data acquisition module is used to obtain the patient's physical index information and corresponding cases, and send them to the data processing module to ensure the operation of the subsequent data processing module; The data processing module is used to determine the discrimination of the patient's important indicators and sensitive information according to the data collected by the data acquisition module, and comprehensively analyze and discriminate the key data according to the patient's important indicators and the warning detection result; The data storage module is used to store all the data generated during the processing of the remote medical care detection and warning system.