Emergency rescue room equipment state intelligent monitoring and early warning system

Abstract

The invention discloses an emergency rescue room equipment state intelligent monitoring and early warning system, which comprises an embedded sensor embedded in medical rescue equipment, and the output end of the embedded sensor is bidirectionally and electrically connected with an Internet of Things gateway. According to the invention, a system architecture composed of an embedded sensor, an Internet of Things gateway, a server and a series of internal intelligent algorithm models (a message queue, a long and short term memory network, a label generation sub-module, a model training and selection module, an equipment life cycle and performance degradation model, a curve fitting and regression model and a degradation trajectory prediction model) is constructed; the method achieves the effect of prospective prediction and judgment of the equipment state, solves the limitation that an existing system can only carry out afterward comparison alarm, achieves the conversion from passive response to active prediction, can give out early warning before the equipment performance has a substantial fault, avoids sudden shutdown in the rescue process, and improves the rescue efficiency. And the continuity and the efficiency of rescue are ensured.

Description

Technical Field

[0001] This invention relates to the field of medical equipment monitoring technology, specifically to an intelligent monitoring and early warning system for the status of equipment in emergency resuscitation rooms. Background Technology

[0002] Medical equipment monitoring refers to the entire process of real-time or near-real-time data collection, transmission, analysis, and management of medical equipment's operating status, usage efficiency, geographical location, and safety performance, utilizing technologies such as the Internet of Things, big data, and cloud computing.

[0003] For example, the patent application number published on the China Patent Network is 202310125392.3, and the patent title is: "Method and Related Device for Monitoring the Status of Medical Equipment Based on the Internet of Things". The method includes: filtering and extracting features from a set of status information to obtain a set of status features; performing vector transformation on the set of status features to obtain a target status vector, and inputting the target status vector into a preset medical equipment status analysis model to perform medical equipment status analysis, thereby obtaining a target status analysis result; acquiring historical periodic status data of multiple medical devices, and generating a standard status analysis result based on the historical periodic status data and the target monitoring period; comparing the result features of the standard status analysis result and the target status analysis result to obtain a target feature comparison result; generating equipment control strategies for multiple medical devices based on the target feature comparison result, and adjusting the equipment parameters of multiple medical devices according to the equipment control strategies.

[0004] However, the existing system mainly relies on comparison with historical data for judgment. When the data shows a large error, it can only prove that there is a problem with the current operating status, and it is necessary to shut down the system immediately for testing. This affects the efficiency of emergency rescue and can easily cause harm.

[0005] Therefore, it is necessary to design and modify the intelligent monitoring and early warning system for the status of equipment in the emergency resuscitation room. Summary of the Invention

[0006] To address the problems mentioned in the background art, the present invention aims to provide an intelligent monitoring and early warning system for the status of equipment in emergency resuscitation rooms. This system has the advantage of predicting and judging the status, and solves the problem that existing systems mainly rely on comparing with historical data for judgment. When the data shows a large error, it can only prove that there is a problem with the current operating status, requiring immediate shutdown for testing, which affects the efficiency of emergency resuscitation and may easily cause harm.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an intelligent monitoring and early warning system for the status of equipment in an emergency resuscitation room, including embedded sensors embedded inside medical resuscitation equipment; The output of the embedded sensor is bidirectionally electrically connected to an IoT gateway. The output of the IoT gateway is bidirectionally electrically connected to a server. The server internally runs a message queue. The output of the message queue is bidirectionally electrically connected to a Long Short-Term Memory (LSTM) network. The output of the LSM network is bidirectionally electrically connected to a tag generation submodule. The output of the tag generation submodule is bidirectionally electrically connected to a model training and selection module. The output of the model training and selection module is bidirectionally electrically connected to a device lifecycle and performance degradation model. The output of the device lifecycle and performance degradation model is bidirectionally electrically connected to a curve fitting and regression model. The output of the curve fitting and regression model is bidirectionally electrically connected to a degradation trajectory prediction model.

[0008] In a preferred embodiment of the present invention, the output of the model training and selection module is bidirectionally electrically connected to a training set, a validation set, and a test set; the outputs of the training set, validation set, and test set are bidirectionally electrically connected to an online inference and real-time early warning module; the output of the online inference and real-time early warning module is bidirectionally electrically connected to a model evaluation and update module; and the input of the online inference and real-time early warning module is bidirectionally electrically connected to the output of the decay trajectory prediction model.

[0009] As a preferred embodiment of the present invention, the output end of the model training and selection module is bidirectionally electrically connected to an early warning rule engine, the output end of the early warning rule engine is bidirectionally electrically connected to an operation and maintenance management platform, and the input end of the operation and maintenance management platform is bidirectionally electrically connected to a mobile application and a real-time monitoring and visualization screen.

[0010] As a preferred embodiment of the present invention, the early warning rule engine consists of three levels of early warning alarms: Level 1 warning (blue / notification): a certain parameter of the equipment deviates slightly from the normal value, requiring attention, and the equipment department engineer is notified; Level 2 warning (yellow / warning): the risk of failure is significantly increased, requiring preparation for intervention, and the equipment department and head nurse are notified; Level 3 warning (red / emergency): the failure needs to be handled immediately, a pop-up window is displayed on the real-time monitoring and visualization screen, and a text message / App push is sent to relevant medical staff and the equipment department, while suggesting the location of backup equipment.

[0011] As a preferred embodiment of the present invention, the equipment lifecycle and performance degradation model consists of a usage intensity quantification submodule, an environmental data association submodule, and a maintenance history integration submodule.

[0012] As a preferred embodiment of the present invention, the embedded sensor consists of a temperature and humidity sensor, a three-phase inductive sensor, a pressure sensor, a flow sensor, a timer, and an RFID tag, and the input terminal of the IoT gateway is bidirectionally electrically connected to a camera.

[0013] As a preferred embodiment of the present invention, the output end of the IoT gateway is bidirectionally electrically connected to a wireless transmission module and a standard data interface, respectively. The output end of the standard data interface is connected to a standard data cable. Both the standard data cable and the wireless transmission module are bidirectionally electrically connected to the server. The wireless transmission module consists of a 5G transmission module and a Wi-Fi transmission module.

[0014] In a preferred embodiment of the present invention, the server is internally configured with a message queue. The output of the message queue is bidirectionally electrically connected to a real-time database and a data cleaning submodule. The output of the data cleaning submodule is bidirectionally electrically connected to a feature extraction submodule. The output of the real-time database is bidirectionally electrically connected to a model management and scheduling platform. The outputs of both the model management and scheduling platform and the feature extraction submodule are bidirectionally electrically connected to a time-series database. The output of the time-series database is bidirectionally electrically connected to the input of a long short-term memory network.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention achieves the effect of proactively predicting and judging the status of equipment by constructing a system architecture consisting of embedded sensors, IoT gateways, servers, and a series of internal intelligent algorithm models (message queue, long short-term memory network, tag generation submodule, model training and selection module, equipment life cycle and performance degradation model, curve fitting and regression model, and degradation trajectory prediction model). It solves the limitation of existing systems that can only perform post-event comparison alarms, realizes the transformation from passive response to proactive prediction, and can issue early warnings before the equipment performance suffers substantial failures, avoid sudden shutdowns during rescue, and ensure the continuity and efficiency of rescue.

[0016] 2. This invention achieves the effect of scientifically training and continuously optimizing the prediction model by explicitly setting training, validation, and test sets in the model training process and connecting them to the online inference and real-time early warning module and the model evaluation and update module. This ensures the accuracy and generalization ability of the selected model, realizes the real-time application of prediction capabilities through online inference, and forms a closed-loop optimization through the model evaluation and update module. This solves the problem that the model may fail over time or due to changes in equipment status, and enables the system to have the ability to learn and continuously improve itself.

[0017] 3. This invention connects the output of the model training and selection module to the early warning rule engine, and further connects it to the operation and maintenance management platform, mobile application, and real-time monitoring and visualization screen. This achieves the effect of effectively transforming intelligent analysis results into specific action instructions and visualized information, solving the problem of the disconnect between predictive information and actual operation and maintenance management. It ensures that early warning information can be accurately and efficiently conveyed to different responsible personnel and visualized in key locations, realizing closed-loop management of prediction, early warning, and operation and maintenance.

[0018] 4. This invention defines the early warning rule engine as consisting of three levels of warning alarms: blue, yellow, and red. It clarifies the triggering conditions, notification targets, and action suggestions for each level, achieving a hierarchical, precise, and efficient early warning effect. This solves the problem of over-warning or under-warning that may occur with traditional binary alarm methods. Through hierarchical early warning, it avoids wasting resources on responses to minor anomalies and ensures that the most efficient response mechanism can be activated in emergency situations, significantly improving the speed of emergency response and the scientific nature of decision-making.

[0019] 5. This invention concretizes the equipment lifecycle and performance degradation model into three sub-modules: quantification of usage intensity, correlation with environmental data, and integration of maintenance history. This achieves a multi-dimensional and refined modeling effect for equipment performance degradation. It overcomes the limitation of potentially inaccurate predictions relying solely on a single operating parameter. By comprehensively quantifying equipment wear and tear, environmental impact, and the "repair" effect of maintenance, the performance degradation model more closely reflects the actual physical degradation process of the equipment, thereby significantly improving the accuracy and reliability of predictions.

[0020] 6. This invention, by listing various types of embedded sensors including temperature and humidity sensors, three-phase inductors, pressure sensors, flow sensors, timers, and RFID sensors, and linking them with cameras, achieves comprehensive, multi-parameter data acquisition of equipment status. This solves the problem of single-dimensional monitoring data that cannot fully reflect the complex operating status of equipment. The rich sensor data provides a solid data foundation for subsequent intelligent analysis, enabling the model to comprehensively assess the health status of equipment from multiple perspectives such as electrical, mechanical, environmental, and usage duration. RFID and cameras provide auxiliary information such as equipment identity, location, and usage scenario.

[0021] 7. This invention, by specifying that the IoT gateway has two output methods—a wireless transmission module (5G and Wi-Fi) and a standard data interface—achieves flexible, reliable, and high-performance data transmission, solving the problems of data transmission adaptability and stability in different scenarios. Wireless transmission provides deployment flexibility and mobility, while the wired interface ensures the absolute stability of data transmission from fixed devices. The combination of the two ensures that massive amounts of monitoring data can be delivered to the server uninterruptedly and completely, providing a channel guarantee for real-time early warning.

[0022] 8. This invention achieves efficient data diversion, buffering, cleaning, and feature extraction of incoming data by setting up a message queue inside the server and constructing a dual-path data processing pipeline that leads to a real-time database and then to a time-series database via data cleaning and feature extraction submodules. This solves the problems of system bottlenecks that may be caused by concurrent processing of massive amounts of real-time data, as well as the problems of low quality raw data and poor performance when directly used for model training. The message queue smooths out peaks and valleys, and the dual-path processing takes into account both real-time display and in-depth analysis needs. Data cleaning and feature extraction provide high-quality, high-value data input for subsequent advanced algorithms such as LSTM, which is a key preprocessing step for the entire intelligent prediction model to work effectively. Attached Figure Description

[0023] Figure 1 This is a structural system diagram of the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] like Figure 1 As shown, the intelligent monitoring and early warning system for the status of emergency resuscitation room equipment provided by the present invention includes embedded sensors embedded inside the medical resuscitation equipment; The output of the embedded sensor is bidirectionally electrically connected to an IoT gateway. The output of the IoT gateway is bidirectionally electrically connected to a server. The server internally runs a message queue. The output of the message queue is bidirectionally electrically connected to a Long Short-Term Memory (LSTM) network. The output of the LSM network is bidirectionally electrically connected to a tag generation submodule. The output of the tag generation submodule is bidirectionally electrically connected to a model training and selection module. The output of the model training and selection module is bidirectionally electrically connected to a device lifecycle and performance degradation model. The output of the device lifecycle and performance degradation model is bidirectionally electrically connected to a curve fitting and regression model. The output of the curve fitting and regression model is bidirectionally electrically connected to a degradation trajectory prediction model.

[0026] refer to Figure 1 The output of the model training and selection module is bidirectionally electrically connected to the training set, validation set, and test set. The outputs of the training set, validation set, and test set are bidirectionally electrically connected to the online inference and real-time early warning module. The output of the online inference and real-time early warning module is bidirectionally electrically connected to the model evaluation and update module. The input of the online inference and real-time early warning module is bidirectionally electrically connected to the output of the decay trajectory prediction model.

[0027] As a technical optimization solution of the present invention, by explicitly setting training sets, validation sets, and test sets in the model training process and connecting them to the online inference and real-time early warning module and the model evaluation and update module, the effect of scientific training and continuous optimization of the prediction model is achieved, ensuring the accuracy and generalization ability of the selected model. The prediction capability is applied in real time through online inference, and a closed-loop optimization is formed through the model evaluation and update module, which solves the problem that the model may fail over time or due to changes in equipment status, enabling the system to have the ability to learn and continuously improve itself.

[0028] refer to Figure 1 The output of the model training and selection module is bidirectionally electrically connected to the early warning rule engine, the output of the early warning rule engine is bidirectionally electrically connected to the operation and maintenance management platform, and the input of the operation and maintenance management platform is bidirectionally electrically connected to the mobile application and the real-time monitoring and visualization screen.

[0029] As a technical optimization solution of the present invention, by connecting the output of the model training and selection module to the early warning rule engine, and further connecting it to the operation and maintenance management platform, mobile application and real-time monitoring and visualization screen, the intelligent analysis results are effectively transformed into specific action instructions and visualized information. This solves the problem of the disconnect between predictive information and actual operation and maintenance management, ensures that early warning information can be accurately and efficiently conveyed to different responsible personnel, and is visualized in key locations, thus realizing closed-loop management of prediction, early warning and operation and maintenance.

[0030] refer to Figure 1 The early warning rule engine consists of three levels of early warning alerts: Level 1 alert (blue / notification): A parameter of the equipment deviates slightly from the normal value and needs attention. The equipment department engineer is notified. Level 2 alert (yellow / warning): The risk of failure is significantly increased and intervention is required. The equipment department and head nurse are notified. Level 3 alert (red / emergency): The failure needs to be dealt with immediately. A pop-up window appears on the real-time monitoring and visualization screen, and a text message / app push is sent to relevant medical staff and the equipment department. At the same time, the location of backup equipment is suggested.

[0031] As a technical optimization of this invention, the early warning rule engine is specifically defined as consisting of three levels of early warning alarms: blue, yellow, and red. The triggering conditions, notification targets, and action suggestions for each level are clearly defined, achieving the effect of hierarchical, accurate, and efficient early warning. This solves the problem of over-warning or under-warning that may occur with traditional binary alarm methods. Through hierarchical early warning, it avoids the waste of resources in responding to minor anomalies and ensures that the most efficient response mechanism can be activated in emergency situations, significantly improving the speed of emergency response and the scientific nature of decision-making.

[0032] refer to Figure 1The equipment lifecycle and performance degradation model consists of a usage intensity quantification submodule, an environmental data association submodule, and a maintenance history integration submodule.

[0033] As a technical optimization of this invention, the equipment lifecycle and performance degradation model is concretized into three sub-modules: usage intensity quantification, environmental data correlation, and maintenance history integration. This achieves multi-dimensional and refined modeling of equipment performance degradation. This overcomes the limitation of potentially inaccurate predictions relying solely on a single operating parameter. By comprehensively quantifying equipment wear and tear, environmental impact, and the "repair" effect of maintenance, the performance degradation model more closely reflects the actual physical degradation process of the equipment, thereby significantly improving the accuracy and reliability of predictions.

[0034] refer to Figure 1 The embedded sensor consists of a temperature and humidity sensor, a three-phase inductive sensor, a pressure sensor, a flow sensor, a timer, and an RFID tag. The input terminal of the IoT gateway is bidirectionally electrically connected to a camera.

[0035] As a technical optimization of this invention, by listing various types of embedded sensors, including temperature and humidity sensors, three-phase inductors, pressure sensors, flow sensors, timers, and RFID sensors, and linking them with cameras, the invention achieves comprehensive, multi-parameter data acquisition of equipment status. This solves the problem of single-dimensional monitoring data that cannot fully reflect the complex operating status of equipment. The rich sensor data provides a solid data foundation for subsequent intelligent analysis, enabling the model to comprehensively assess the health status of equipment from multiple perspectives, including electrical, mechanical, environmental, and usage duration. RFID and cameras provide auxiliary information such as equipment identity, location, and usage scenario.

[0036] refer to Figure 1 The output end of the IoT gateway is bidirectionally electrically connected to a wireless transmission module and a standard data interface. The output end of the standard data interface is connected to a standard data cable. Both the standard data cable and the wireless transmission module are bidirectionally electrically connected to the server. The wireless transmission module consists of a 5G transmission module and a Wi-Fi transmission module.

[0037] As a technical optimization of this invention, by specifying that the IoT gateway has two output methods—a wireless transmission module (5G and Wi-Fi) and a standard data interface—it achieves flexible, reliable, and high-performance data transmission, solving the problems of data transmission adaptability and stability in different scenarios. Wireless transmission provides deployment flexibility and mobility, while the wired interface ensures the absolute stability of data transmission from fixed devices. The combination of the two ensures that massive amounts of monitoring data can be delivered to the server uninterruptedly and completely, providing a channel guarantee for real-time early warning.

[0038] refer to Figure 1The server has an internal message queue. The output of the message queue is bidirectionally connected to a real-time database and a data cleaning submodule. The output of the data cleaning submodule is bidirectionally connected to a feature extraction submodule. The output of the real-time database is bidirectionally connected to a model management and scheduling platform. The outputs of both the model management and scheduling platform and the feature extraction submodule are bidirectionally connected to a time-series database. The output of the time-series database is bidirectionally connected to the input of the Long Short-Term Memory network.

[0039] As a technical optimization of this invention, by setting up a message queue inside the server and constructing a dual-path data processing pipeline leading to a real-time database and then to a time-series database via data cleaning and feature extraction submodules, the goal of efficiently diverting, buffering, cleaning, and extracting features from the incoming data is achieved. This solves the system bottlenecks that may result from concurrent processing of massive amounts of real-time data, as well as the problems of low quality raw data and poor performance when directly used for model training. The message queue smooths out peaks and valleys, and the dual-path processing takes into account both real-time display and in-depth analysis needs. Data cleaning and feature extraction provide high-quality, high-value data input for subsequent advanced algorithms such as LSTM, which is a key preprocessing step for the entire intelligent prediction model to work effectively.

[0040] The working principle and usage process of this invention: An embedded sensor cluster is used within various medical emergency equipment. These sensors act as the nerve endings of the system, responsible for collecting the most raw and comprehensive equipment operation data. This cluster includes: a temperature and humidity sensor to monitor the operating environment of key components inside the equipment; excessively high temperatures may indicate poor heat dissipation or component aging; a three-phase inductive sensor primarily used to monitor electrical parameters such as current, voltage, and power in equipment with motors; abnormal fluctuations may indicate excessive motor load or impending motor failure; pressure and flow sensors are crucial for gas delivery equipment such as ventilators and anesthesia machines, monitoring gas pressure and flow in real time to ensure output accuracy meets set values; deviations may indicate pipeline leaks or valve malfunctions; and a timer. Accurately recording the cumulative operating time, single continuous operating time, and standby time of equipment is fundamental to assessing equipment usage intensity and estimating its lifespan. RFID tags assign a unique digital identity to each device, facilitating rapid identification and location, and linking it to subsequent maintenance and usage records. Furthermore, the IoT gateway's input end is bidirectionally connected to a camera. This camera is not used to collect patient privacy information but rather to assist in determining the equipment's usage scenario and provide contextual information for data analysis. The IoT gateway, acting as a local data hub, is responsible for aggregating data from all sensors and cameras and transmitting it stably and efficiently to a remote server via its wireless transmission module (5G / Wi-Fi) or standard data interface (such as Ethernet cable). 5G technology ensures continuous and low-latency data transmission during mobile ward rounds or equipment transfers. After data arrives at the server, it first enters a message queue. This queue acts as a buffer, effectively handling the large volume of instantaneous data and high concurrency characteristic of emergency room equipment, ensuring system stability and preventing crashes. Data from the message queue is divided into two paths: one enters a real-time database to store the latest status data requiring extremely rapid response, providing data support for the real-time monitoring dashboard; the other enters a data cleaning and feature extraction pipeline. The raw data typically contains noise, outliers, and missing values. The data cleaning submodule filters, smooths, and completes the data to ensure quality. Subsequently, the feature extraction submodule extracts features from the cleaned data. Features such as vibration spectrum characteristics, harmonic components of current, short-term standard deviation of operating parameters, and operating frequency of specific components are processed into high-quality data, which is then stored in a time-series database. This database is specifically optimized for storing massive amounts of data in chronological order, providing efficient data retrieval services for subsequent time-series analysis models. A Long Short-Term Memory (LSTM) network learns time-series patterns. The system inputs historical operating data of the equipment from the time-series database (such as vibration, temperature, and current sequences from the past few months) into an LSTM network. The LSTM network can learn the complex patterns of parameter changes throughout the entire process of the equipment's progression from a "new" state to "slight degradation" and then to "severe degradation." The label generation submodule utilizes historical, past equipment maintenance records and fault reports.Historical data is labeled, and the labeled data is divided into training, validation, and test sets by the model training and selection module for training and optimizing prediction models such as LSTM. The module tries various algorithms and evaluates their performance through the validation set, ultimately selecting the most accurate model as the final equipment lifecycle and performance degradation model. The equipment lifecycle and performance degradation model is not a single model, but is composed of three key sub-modules. The intensity quantization sub-module quantifies the running time, start-stop times, etc. recorded by the timer into a "usage intensity coefficient". The environmental data association sub-module associates environmental factors such as temperature and humidity with performance data to assess the impact of the environment on equipment degradation. The maintenance history integration sub-module integrates the records of previous maintenance and repairs into the model. A good maintenance may "reset" certain degradation indicators. Combining this multi-dimensional information, the system uses curve fitting and regression models to extrapolate future performance parameters, thereby forming a predictive degradation trajectory that can intuitively show the possible changes in the equipment's state in the next few days or weeks. The trained model is deployed on the model management and scheduling platform to perform online inference on the new data that comes in in real time, that is, to calculate the current "health score" of the equipment in real time and predict its future state. The inference results are fed into the early warning rule engine, which has a three-level early warning mechanism: Level 1 (Blue / Alert): Triggered when the model predicts that a certain parameter of the equipment begins to deviate slightly from the normal range, but there is no risk of failure in the short term. The system will record this trend and notify the equipment department engineer, suggesting that the equipment be monitored during planned inspections. Level 2 (Yellow / Warning): Triggered when the model judges that the risk of failure has increased significantly and the equipment may need maintenance in the next few days. The system will simultaneously notify the equipment department and the head nurse, reminding them to prepare intervention measures, such as allocating spare equipment in advance or arranging maintenance windows. Level 3 (Red / Emergency): Triggered when the model predicts or detects that a failure is about to occur or has already occurred. When immediate action is required, the system triggers an alarm pop-up on the real-time monitoring and visualization dashboard, simultaneously sending SMS messages and pushing notifications via the app to relevant medical staff and equipment department heads. Crucially, the system automatically suggests the location of backup equipment, significantly reducing equipment retrieval time during rescue efforts and directly safeguarding life-saving efforts. Alarm information and equipment status are centrally managed through the operations and maintenance management platform and distributed to mobile applications and the real-time monitoring and visualization dashboard. Furthermore, all alarm results and subsequent manual confirmation and maintenance feedback form a closed loop, flowing into the model evaluation and update module for continuous model performance evaluation and iterative optimization, making the system increasingly intelligent and accurate.

[0041] In summary, this intelligent monitoring and early warning system for emergency resuscitation room equipment status achieves the effect of proactively predicting and judging equipment status by constructing a system architecture consisting of embedded sensors, IoT gateways, servers, and a series of internal intelligent algorithm models (message queues, long short-term memory networks, tag generation submodules, model training and selection modules, equipment lifecycle and performance degradation models, curve fitting and regression models, and degradation trajectory prediction models). This overcomes the limitations of existing systems that can only perform post-event comparison alarms, realizing a shift from passive response to proactive prediction. As a result, it can issue early warnings before substantial equipment performance failures occur, avoiding sudden shutdowns during resuscitation and ensuring the continuity and efficiency of resuscitation.

[0042] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0043] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent monitoring and early warning system for the status of equipment in the emergency resuscitation room, including embedded sensors embedded inside the medical resuscitation equipment; Its features are: The output of the embedded sensor is bidirectionally electrically connected to an IoT gateway. The output of the IoT gateway is bidirectionally electrically connected to a server. The server internally runs a message queue. The output of the message queue is bidirectionally electrically connected to a Long Short-Term Memory (LSTM) network. The output of the LSM network is bidirectionally electrically connected to a tag generation submodule. The output of the tag generation submodule is bidirectionally electrically connected to a model training and selection module. The output of the model training and selection module is bidirectionally electrically connected to a device lifecycle and performance degradation model. The output of the device lifecycle and performance degradation model is bidirectionally electrically connected to a curve fitting and regression model. The output of the curve fitting and regression model is bidirectionally electrically connected to a degradation trajectory prediction model.

2. The intelligent monitoring and early warning system for the status of equipment in the emergency resuscitation room according to claim 1, characterized in that: The output of the model training and selection module is bidirectionally electrically connected to the training set, validation set, and test set. The outputs of the training set, validation set, and test set are bidirectionally electrically connected to the online inference and real-time early warning module. The output of the online inference and real-time early warning module is bidirectionally electrically connected to the model evaluation and update module. The input of the online inference and real-time early warning module is bidirectionally electrically connected to the output of the decay trajectory prediction model.

3. The intelligent monitoring and early warning system for the status of equipment in the emergency resuscitation room according to claim 2, characterized in that: The output of the model training and selection module is bidirectionally electrically connected to an early warning rule engine, the output of the early warning rule engine is bidirectionally electrically connected to an operation and maintenance management platform, and the input of the operation and maintenance management platform is bidirectionally electrically connected to a mobile application and a real-time monitoring and visualization screen.

4. The intelligent monitoring and early warning system for the status of equipment in the emergency resuscitation room according to claim 3, characterized in that: The early warning rule engine consists of three levels of early warning alerts: Level 1 alert (blue / notification): A parameter of the equipment deviates slightly from the normal value and needs attention; the equipment department engineer is notified. Level 2 alert (yellow / warning): The risk of failure is significantly increased and intervention is required; the equipment department and head nurse are notified. Level 3 alert (red / emergency): The failure needs to be handled immediately; a pop-up window appears on the real-time monitoring and visualization screen, and an SMS / App push is sent to relevant medical staff and the equipment department, while also suggesting the location of backup equipment.

5. The intelligent monitoring and early warning system for the status of equipment in the emergency resuscitation room according to claim 4, characterized in that: The equipment lifecycle and performance degradation model consists of a usage intensity quantification submodule, an environmental data association submodule, and a maintenance history integration submodule.

6. The intelligent monitoring and early warning system for the status of equipment in the emergency resuscitation room according to claim 5, characterized in that: The embedded sensor consists of a temperature and humidity sensor, a three-phase inductive sensor, a pressure sensor, a flow sensor, a timer, and an RFID tag. The input terminal of the IoT gateway is bidirectionally electrically connected to a camera.

7. The intelligent monitoring and early warning system for the status of equipment in the emergency resuscitation room according to claim 6, characterized in that: The output terminals of the IoT gateway are bidirectionally electrically connected to a wireless transmission module and a standard data interface. The output terminal of the standard data interface is connected to a standard data cable. Both the standard data cable and the wireless transmission module are bidirectionally electrically connected to the server. The wireless transmission module consists of a 5G transmission module and a Wi-Fi transmission module.

8. The intelligent monitoring and early warning system for the status of equipment in the emergency resuscitation room according to claim 7, characterized in that: The server is internally equipped with a message queue. The output of the message queue is bidirectionally electrically connected to a real-time database and a data cleaning submodule. The output of the data cleaning submodule is bidirectionally electrically connected to a feature extraction submodule. The output of the real-time database is bidirectionally electrically connected to a model management and scheduling platform. The outputs of the model management and scheduling platform and the feature extraction submodule are both bidirectionally electrically connected to a time-series database. The output of the time-series database is bidirectionally electrically connected to the input of a long short-term memory network.

Images