Field emergency rescue medical big data management system

Abstract

The invention discloses a field emergency rescue medical big data management system, and relates to the technical field of medical data management. The system comprises a data acquisition module, a data transmission module, a data storage module, an analysis processing module and a rescue decision support module, wherein the acquisition module acquires physiological and environmental data by using a multi-modal sensor array and performs edge preprocessing; the transmission module constructs a multi-link communication architecture and encrypts transmission; the storage module adopts distributed encryption and block chain storage; the analysis processing module combines depth migration and federal learning to mine data; the decision support module provides a scheme based on a knowledge graph and reinforcement learning, and VR / AR auxiliary decision is fused. The system comprehensively and accurately collects data, stably transmits and stores data in various field environments, deeply mines data values to assist in accurate decision making, supports remote cooperation, realizes functions of intelligent early warning, equipment management, authority control and the like, and comprehensively improves the medical efficiency and accuracy of field emergency rescue.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data management, and particularly to a big data management system for field emergency rescue medicine. Background Art

[0002] Traditional means of collecting field medical rescue data are relatively single, mostly relying on manual recording of the physiological indicators and symptoms of the injured. This is not only inefficient but also prone to errors and omissions. The performance of existing sensor devices in complex field environments is also not satisfactory. For example, signals are easily interfered with, resulting in inaccurate or lost data. Moreover, different types of medical data are scattered and recorded, making it difficult to form comprehensive and systematic information and unable to provide strong support for rescue decision-making.

[0003] In terms of data transmission and storage, the field environment often has weak or even no network signal, which makes it difficult to transmit medical data at the rescue site to the rear medical team in real time, delaying the best diagnosis and treatment time. Even if some data can be transmitted, there are also problems such as slow transmission speed and poor stability. In terms of data storage, there is a lack of unified and standardized management, and data storage is scattered and easily lost, which is not conducive to subsequent medical analysis and experience summary.

[0004] In addition, the decision-making process for field emergency rescue lacks scientific and effective support. When facing complex conditions and limited resources, rescue personnel are difficult to make accurate decisions quickly. Due to the inability to comprehensively obtain on-site information, rear medical experts are also difficult to provide accurate remote guidance. Moreover, there is a lack of an efficient coordination mechanism among various rescue links, resulting in low rescue efficiency and posing a great threat to the lives of the injured. Generally speaking, the existing field emergency rescue medical system has obvious deficiencies in data management and decision support, and there is an urgent need for an innovative solution to improve the rescue level. Summary of the Invention

[0005] The big data management system for field emergency rescue medicine proposed by the present invention aims to solve the problems mentioned in the above prior art.

[0006] To achieve the above object, the present invention adopts the following technical solutions: A big data management system for field emergency rescue medicine, comprising: A data acquisition module: equipped with an integrated sensor array, integrating wearable flexible sensors, implantable microelectromechanical sensors, and environmental perception sensors, realizing the location positioning of the rescue site through wireless transmission technology combined with satellite positioning and inertial navigation technology, using edge computing technology to perform preliminary preprocessing on the collected data at the sensor end, and applying an algorithm combining wavelet transform and singular value decomposition to remove noise interference and extract features; Data Transmission Module: Build a link adaptive communication architecture, comprehensively utilize 5G, satellite communication, and ad-hoc network communication technologies. In the area covered by 5G signals, carrier aggregation and Massive MIMO technologies are used to achieve data transmission. For satellite communication, a low-earth orbit satellite constellation combined with beamforming technology is selected, and the beam direction is dynamically adjusted according to the location of the rescue site and the satellite signal strength to enhance the anti-interference ability of communication. In remote areas without network coverage, ad-hoc network communication is activated, and node devices automatically build a network through a dynamic routing algorithm. A data encryption mechanism combining quantum key distribution (QKD) and homomorphic encryption is introduced; Data Storage Module: Adopt a distributed encrypted storage system to store data fragments on servers in different geographical locations. Storage nodes are used to record the operation logs and traceability information of the data. The consortium chain architecture combined with the Byzantine Fault Tolerance (BFT) algorithm is used to improve the processing efficiency. Different compression algorithms are used for structured physiological data and unstructured medical image data during data storage. For structured data, a hybrid algorithm based on dictionary coding and run-length encoding is used, and for unstructured data, a deep learning-driven image compression algorithm is adopted; Data Analysis and Processing Module: Apply a technical framework combining deep transfer learning and federated learning. Through the construction of a fusion neural network, fusion analysis of physiological data, environmental data, and medical image data is realized. In disease prediction, a prediction model based on Long Short-Term Memory (LSTM) network and attention mechanism is adopted, combined with time series analysis and causal inference algorithms.

[0007] Furthermore, it also includes: Rescue Decision Support Module: Build a decision-making system based on knowledge graph and reinforcement learning. The knowledge graph integrates medical knowledge, rescue experience, geographical information, and medical resource data to form a structured knowledge network. The reinforcement learning algorithm optimizes the decision by interacting with the rescue environment, considering the rescue scenario during the decision-making process. The Bayesian optimization algorithm is introduced to optimize the parameters of the decision model; Data Fusion Module: Adopt a multi-data fusion algorithm based on deep learning. By constructing a fusion model combining a multi-layer perceptron (MLP) and a convolutional neural network (CNN), the correlation relationships between different data sources are automatically learned. For different modalities of data such as physiological data, environmental data, and medical image data, feature extraction is first performed through their respective feature extraction networks, and then the extracted features are input into the fusion model for fusion. During the fusion process, the attention mechanism is used to weight different features to highlight the influence of features on decision-making.

[0008] Furthermore, it also includes: Remote medical collaboration module: Realize remote real-time consultation through the transmission of 5G and satellite communication. Adopt 3D video reconstruction technology to provide the perspective of the rescue scene for experts in the rear. Experts conduct remote consultations through virtual reality devices; utilize tactile Internet technology for experts to perceive the operation feedback of on-site medical devices in real time, and at the same time remotely control the devices to perform medical operations through remote operation robot technology; Early warning module: Based on deep learning anomaly detection models and time series prediction models, conduct real-time monitoring and early warning on the physiological data of the injured and the environmental data of the rescue scene; the anomaly detection model adopts a deep learning algorithm based on autoencoders to automatically detect abnormal data points by learning the distribution patterns of normal data. The time series prediction model adopts a prediction algorithm based on Transformer, combined with the attention mechanism to model the time series of physiological data and environmental data, and predict the change trend in the future; at the same time, utilize data fusion technology to comprehensively consider physiological data, environmental data, and medical device data to improve the accuracy of early warning; Device management module: Utilize Internet of Things technology to manage data acquisition devices, communication devices, and medical devices. Real-time monitor the operating status, power, signal strength, and temperature parameters of the devices through the sensors and chips built into the devices, and conduct preliminary analysis and processing through edge computing; adopt a fault diagnosis model based on deep learning to learn and analyze the operating data of the devices to predict device failures in advance; User permission management module: Adopt attribute-based encryption ABE and user permission management mechanisms. Attribute-based encryption allows encrypting and decrypting data according to the attributes of users, enabling users with corresponding attributes to access data. The storage node is used to record the permission change records and access logs of users, and automatically execute permission verification and access control through contract technology; at the same time, introduce an identity authentication method that combines biometric technology and dynamic passwords to enhance the accuracy of user identity authentication.

[0009] Furthermore, it also includes: Data visualization module; Use virtual reality, augmented reality, and mixed reality technologies to realize the visual display of medical data. At the same time, adopt a data-driven visual layout optimization algorithm to automatically adjust the layout of visual elements according to the importance and relevance of the data, improving the efficiency of data viewing and analysis; Data traceability module: Utilize distributed ledger technology and zero-knowledge proof technology to achieve privacy protection of medical data. The storage node records the collection, transmission, processing, and storage information of the data. Zero-knowledge proof technology allows proving the authenticity of the data without disclosing the data; at the same time, enhance the accuracy of data traceability by introducing timestamp and digital signature technologies; Emergency Material Management Module: Construct an emergency material management system based on the Internet of Things and big data analysis. By installing Internet of Things tags on emergency materials, the location, quantity, and shelf life information of the materials can be monitored in real time. Using big data analysis technology, predict the material requirements for different regions and different types of rescue tasks based on historical rescue data, geographical information, and seasonal factors. Adopt an optimization model for material dispatching based on the genetic algorithm to optimize the material allocation plan according to factors such as material inventory, transportation distance, and urgency of rescue requirements. At the same time, establish an emergency material replenishment system to automatically trigger the replenishment process when the material inventory is lower than the safety threshold. System Self-Optimization Module: By collecting performance data, user feedback data, and environmental data during the system operation process, use reinforcement learning and adaptive control technologies to automatically optimize the system's parameter settings, algorithm selection, and resource allocation. At the same time, improve the system's performance and adaptability through online learning and model update mechanisms.

[0010] Compared with the existing technologies, the beneficial effects of the present invention are: In terms of data collection, the integrated multi-modal sensor array can comprehensively and accurately collect various types of data. Whether it is the weak physiological signals of the human body or complex environmental information, it can be accurately obtained, providing a rich and reliable data basis for subsequent medical analysis. In terms of data transmission and storage, the multi-link adaptive communication architecture ensures stable and fast data transmission in various field environments. The system combining distributed encrypted storage and blockchain not only guarantees the security of the data but also realizes the traceability and efficient management of the data.

[0011] The data analysis and processing module uses cutting-edge technologies to deeply explore the data value, improving the accuracy of disease prediction and diagnosis, and helping rescue personnel better grasp the condition of the injured. The rescue decision support module uses technologies such as knowledge graphs and reinforcement learning to provide scientific and reasonable rescue plans for rescue personnel. At the same time, assist in decision-making through technologies such as virtual reality to make rescue decisions more efficient and accurate.

[0012] The remote medical collaboration module breaks through the time and space limitations, enabling rear experts to participate in the rescue in real time and achieving remote precise guidance through a variety of advanced technologies, significantly improving the diagnostic accuracy of difficult diseases. The intelligent early warning module discovers potential risks in advance, winning precious time for the rescue; the equipment management module ensures the stable operation of various equipment; the user permission management module ensures data security; the data visualization module facilitates rescue personnel to view and analyze data; the data traceability module enhances data credibility; the emergency material management module ensures sufficient material supply; the system self-optimization module continuously improves the system performance. Brief Description of the Drawings

[0013] Figure 1 It is a schematic block diagram of the field emergency rescue medical big data management system proposed by the present invention; Figure 2 Chart information on the data transmission speed of a field emergency rescue medical big data management system proposed by the present invention under different communication environments; Figure 3 Chart information on the comparison of disease prediction accuracy rates of a field emergency rescue medical big data management system proposed by the present invention. Detailed implementation manners

[0014] 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.

[0015] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by terms such as "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation of the present invention.

[0016] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality" means two or more, unless otherwise specifically defined. In addition, the terms "installation", "connection", and "connection" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected, or indirectly connected through an intermediate medium, and it may be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. Next, the present invention will be further described in detail with reference to the accompanying drawings.

[0017] Refer to Figure 1 and Figure 3 for the detailed implementation manners of the field emergency rescue medical big data management system: The system of this application consists of five major modules: data acquisition, transmission, storage, analysis and processing, and rescue decision support. Through the close cooperation of each module, efficient management of field medical data and scientific rescue decisions are achieved. The following is a detailed implementation method.

[0018] I. Data Acquisition Module In the field rescue scenario, rescue personnel carry devices integrated with multi-modal sensor arrays. This array contains two types of sensors: physiological parameter sensors, such as heart rate, blood oxygen, and body temperature sensors; environmental parameter sensors, like temperature and humidity, air pressure, and GPS positioning sensors.

[0019] Taking the heart rate sensor as an example, it collects data per second. Assuming data is collected during the time period from to , the number of acquisitions n is equal to , and the time point of each acquisition is , thus obtaining the heart rate data sequence , where is the heart rate data value collected at the time point .

[0020] The collected data will first be preprocessed using edge computing on the device side. Here, the Kalman filter algorithm is used to remove noise interference. Briefly, this algorithm works in two steps: first, predict the data state at the next moment, and then update the prediction result by combining the actual observed data to obtain more accurate data.

[0021] To improve data acquisition efficiency, the system sets a trigger mechanism. For example, when the heart rate suddenly changes significantly, that is, when the heart rate change rate exceeds the pre-set threshold α, the sensor will automatically increase the acquisition frequency to ensure that key data is not missed. The following is the data representation of the system of this application:

[0022] II. Data Transmission Module The data transmission module adopts a multi-link adaptive communication architecture, integrating technologies such as 5G and satellite communication. The system assigns priorities to the collected data, with the vital sign data having a higher priority than the environmental data.

[0023] When in the 5G network coverage area, the 5G network is preferentially used to transmit data. The system packs the data into data packets conforming to the TCP / IP protocol. Assuming the data packet size is S bytes and the 5G network transmission rate is , then the transmission time . If in an area with poor 5G signal or no signal, the system will automatically switch to the satellite communication link. The satellite communication transmission time is equal to the data transmission time plus the satellite relay delay time , where is the transmission rate of satellite communication.

[0024] To ensure the security of data during transmission, the system adopts the AES-256 encryption algorithm. The sender encrypts the original data D with the key K into , and the receiver decrypts with the same key K to prevent data from being stolen or tampered with. The following are the characterizations of data transmission in different environments:

[0025] III. Data storage module The data storage module combines distributed encrypted storage and blockchain technology. The system first classifies the data, such as physiological data , environmental data , etc., and then compresses different types of data. Taking physiological data as an example, the LZ77 compression algorithm is adopted, and the compression ratio , is the size of the data before compression, is the size of the data after compression.

[0026] In terms of storage, the Ceph distributed storage system is used to disperse and store the data on multiple nodes, and each node stores different copies of the data. At the same time, using blockchain technology, a unique hash value is generated for each piece of data and recorded in the blockchain. When verifying the data integrity later, only need to recalculate the hash value and compare it with the record in the blockchain. If they are the same, it means the data has not been tampered with. The following are the characterizations of the data storage of the system in this application:

[0027] IV. Data analysis and processing module This module analyzes data using deep transfer learning, federated learning, and multi-modal fusion neural networks. First, a convolutional neural network (CNN) is used to extract features from physiological data (such as electrocardiogram signals), and a fully connected neural network is used to extract features from environmental data.

[0028] Then, the two types of features are combined into by means of weighted fusion, and the values of and are adjusted through training to find the optimal fusion method, is the weight coefficient of physiological data features; is the weight coefficient of environmental data features.

[0029] Deep transfer learning transfers the model parameters trained on a large amount of medical data to this system, reducing the training time and data requirements. Federated learning enables devices at different rescue sites to train a model together without sharing the original data, and updates the model parameters according to the formula Update the model parameters, represents the global model parameters at the th round of training; is the number of devices or nodes participating in federated learning; is used to index the devices participating in the training; The th device has the number of training samples; is the total number of samples of all devices participating in the training; The th device calculates the gradient of the loss function based on the local data and the current global model parameters ; realizing the in-depth mining of data value.

[0030] The following is the characterization of the disease prediction accuracy of the system of this application:

[0031] V. Rescue Decision Support Module The rescue decision support module formulates rescue plans based on the knowledge graph and reinforcement learning. First, construct a medical rescue knowledge graph, and store information such as disease symptoms, treatment methods, and rescue resources in the form of "(Entity 1, Relationship, Entity 2)", such as (heart disease, treatment method, cardiopulmonary resuscitation).

[0032] Then, use the reinforcement learning algorithm to continuously try in the state space S and the action space A through the reward function R to find the optimal rescue strategy.

[0033] In addition, with the help of VR / AR technology, the rescue site environment and patient information are presented in a three-dimensional visualization form. Through VR / AR devices, rescue personnel can intuitively view the changes in patients' physiological data and the operation steps of the rescue plan, realizing scientific and efficient rescue decisions.

[0034] Through the coordinated operation of each module, the system of this application can complete the full-process management of medical data from collection, transmission, storage to analysis and processing in a complex field environment, and provide strong support for rescue decisions.

[0035] The above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. A field emergency rescue medical big data management system, characterized in that, Including: Data acquisition module: Equipped with an integrated sensor array, it realizes the location positioning of the rescue site through wireless transmission technology combined with satellite positioning and inertial navigation technology. Using edge computing technology, it preliminarily preprocesses the collected data at the sensor end, and uses an algorithm combining wavelet transform and singular value decomposition to remove noise interference and extract features; Data transmission module: Builds a link adaptive communication architecture. In the 5G signal coverage area, it uses carrier aggregation and Massive MIMO technology to achieve data transmission. For satellite communication, it selects a low-earth orbit satellite constellation combined with beamforming technology, and dynamically adjusts the beam direction according to the rescue site location and satellite signal strength to improve the communication anti-interference ability. In remote areas without network, it starts ad-hoc network communication, and node devices automatically build a network through a dynamic routing algorithm; Data storage module: Adopts a distributed encrypted storage system to store data fragments on servers in different geographical locations. Storage nodes are used to record the operation logs and traceability information of the data. It uses a consortium chain architecture combined with the Byzantine fault tolerance algorithm BFT to improve processing efficiency. Different compression algorithms are used for structured physiological data and unstructured medical image data during data storage. For structured data, a hybrid algorithm based on dictionary coding and run-length coding is used, and for unstructured data, a deep learning-driven image compression algorithm is used; Data analysis and processing module: Utilizes a technical framework combining deep transfer learning and federated learning. By constructing a fusion neural network, it conducts fusion analysis on physiological data, environmental data, and medical image data. In disease prediction, it uses a prediction model based on long short-term memory network LSTM and attention mechanism, combined with time series analysis and causal inference algorithms.

2. The outdoor emergency rescue medical big data management system according to claim 1, characterized in that, It also includes: Rescue decision support module: Builds a decision-making system based on knowledge graph and reinforcement learning. The knowledge graph integrates medical knowledge, rescue experience, geographical information, and medical resource data to form a structured knowledge network. The reinforcement learning algorithm optimizes the decision by interacting with the rescue environment and takes into account the rescue scenario during the decision-making process. The Bayesian optimization algorithm is introduced to optimize the parameters of the decision model; Data fusion module: Adopts a multi-data fusion algorithm based on deep learning. By constructing a fusion model combining multi-layer perceptron MLP and convolutional neural network CNN, it automatically learns the correlation relationships between different data sources. For different modalities of data such as physiological data, environmental data, and medical image data, first, feature extraction is performed through their respective feature extraction networks, and then the extracted features are input into the fusion model for fusion. During the fusion process, the attention mechanism is used to weight different features to highlight the impact of features on decision-making.

3. The outdoor emergency rescue medical big data management system according to claim 1, characterized in that It also includes: Remote medical collaboration module: Achieves remote real-time consultation through the transmission of 5G and satellite communication. It uses 3D video reconstruction technology to provide the rear experts with the perspective of the rescue site. Experts conduct remote consultations through virtual reality devices; Utilizing tactile Internet technology, experts can real-time sense the operation feedback of on-site medical equipment, and at the same time remotely control the equipment to perform medical operations through remote operation robot technology.

4. The outdoor emergency rescue medical big data management system according to claim 1, characterized in that, It also includes: Early warning module: Based on deep learning anomaly detection models and time series prediction models, it monitors and gives early warnings to the physiological data of the injured and the environmental data at the rescue site in real time; The anomaly detection model uses a deep learning algorithm based on autoencoders to automatically detect abnormal data points by learning the distribution patterns of normal data. The time series prediction model uses a prediction algorithm based on Transformers, combines the attention mechanism to model the time series of physiological data and environmental data, and predicts the change trends in the future. At the same time, data fusion technology is used to comprehensively consider physiological data, environmental data, and medical device data to improve the accuracy of early warnings.

5. The outdoor emergency rescue medical big data management system according to claim 1, characterized in that, It also includes: Device management module: Using Internet of Things technology to manage data collection devices, communication devices, and medical devices, and real-time monitoring the operating status, power, signal strength, and temperature parameters of the devices through the sensors and chips built into the devices, and performing preliminary analysis and processing through edge computing; Using a fault diagnosis model based on deep learning to learn and analyze the operating data of the devices to predict device failures in advance.

6. The outdoor emergency rescue medical big data management system according to claim 1, characterized in that, It also includes: User permission management module: Adopting attribute-based encryption ABE and user permission management mechanisms, attribute-based encryption allows data to be encrypted and decrypted according to the attributes of users, enabling users with corresponding attributes to access data. The storage node is used to record the permission change records and access logs of users, and automatically performs permission verification and access control through contract technology; At the same time, an identity authentication method combining biometric technology and dynamic passwords is introduced to enhance the accuracy of user identity authentication.

7. The outdoor emergency rescue medical big data management system according to claim 1, characterized in that, It also includes: Data visualization module; Using virtual reality, augmented reality, and mixed reality technologies to achieve visual display of medical data. At the same time, a data-driven visual layout optimization algorithm is adopted to automatically adjust the layout of visual elements according to the importance and relevance of the data, improving the efficiency of data viewing and analysis.

8. The outdoor emergency rescue medical big data management system according to claim 1, characterized in that, It also includes: Data traceability module: Using distributed ledger technology and zero-knowledge proof technology to achieve privacy protection of medical data. The storage node records the collection, transmission, processing, and storage information of the data. Zero-knowledge proof technology allows the authenticity of the data to be proven without revealing the data; At the same time, the accuracy of data traceability is enhanced by introducing timestamp and digital signature technologies.

9. The field emergency rescue medical big data management system according to claim 1, characterized in that It also includes: Emergency supplies management module: Construct an emergency supplies management system based on the Internet of Things and big data analysis. By installing Internet of Things tags on emergency supplies, the location, quantity, and shelf life information of the supplies are monitored in real time; Using big data analysis technology, predict the supplies needs for different regions and different types of rescue tasks based on historical rescue data, geographical information, and seasonal factors; Adopt a supplies scheduling optimization model based on genetic algorithms to optimize the supplies allocation plan according to factors such as the inventory of supplies, transportation distance, and urgency of rescue needs. At the same time, establish an emergency supplies replenishment system to automatically trigger the replenishment process when the supplies inventory is lower than the safety threshold.

10. The outdoor emergency rescue medical big data management system according to claim 1, characterized in that It also includes: System self-optimization module: By collecting performance data, user feedback data, and environmental data during system operation, it automatically optimizes the system's parameter settings, algorithm selection, and resource allocation using reinforcement learning and adaptive control technologies; at the same time, it improves the system's performance and adaptability through an online learning and model update mechanism.

Images