AI empowerment-based intelligent operating room air quality prediction and supervision system

Abstract

The application discloses an intelligent operating room air quality prediction and supervision system based on AI empowerment, which comprises a multidimensional data perception layer, an AI analysis engine and an intelligent supervision platform. The multidimensional data perception layer comprises a particle monitoring network and a parameter acquisition module. The particle monitoring network is arranged in layers around the operating bed to collect multi-particle size particle concentration data through sensors. The parameter acquisition module acquires environmental parameters. The AI analysis engine comprises a time series database, a multi-factor correlation analysis model and a prediction and early warning module. The time series database constructs a four-dimensional data storage structure. The multi-factor correlation analysis model establishes the correlation between the air quality index and the environmental parameters. The prediction and early warning module predicts the particle concentration trend based on the four-dimensional data storage structure and triggers a hierarchical early warning. The intelligent supervision platform generates control instructions according to the hierarchical early warning and collects data to optimize the model. The application realizes accurate prediction, active control and closed-loop optimization of the air quality in the operating room, and improves the control efficiency and medical safety.

Description

Technical Field

[0001] This invention relates to the field of intelligent air environment monitoring technology, specifically an AI-enabled intelligent operating room air quality prediction and monitoring system. Background Technology

[0002] With the continuous development of medical technology, the air quality in operating rooms, as the core treatment area of ​​hospitals, is directly related to the infection rate at surgical sites and is crucial for postoperative recovery and medical safety. Currently, air quality management in operating rooms mainly relies on fixed-point sampling or periodic manual testing, depending on traditional testing equipment and human experience. However, this traditional method is difficult to capture in real time the spatial distribution differences of air quality in operating rooms, and it cannot correlate particulate matter concentration with dynamic factors such as personnel activity and surgical type. This results in an inability to predict and proactively intervene in air quality changes, leading to a lag in control and failing to meet the needs of modern operating rooms for refined air quality management. Summary of the Invention

[0003] Based on this, it is necessary to provide an AI-enabled intelligent operating room air quality prediction and monitoring method, system, computer equipment, and storage medium that can capture and acquire elbow joint motion data and extract multi-dimensional motion control features to generate a comprehensive quantitative rehabilitation score.

[0004] This invention discloses an AI-enabled intelligent operating room air quality prediction and monitoring system, comprising: The multi-dimensional data sensing layer includes a particle monitoring network and a parameter acquisition module. The particle monitoring network is deployed with multiple particle sensors layered around the operating table to collect multi-size particle concentration data. The parameter acquisition module is used to collect environmental parameters, including at least personnel activity data, surgical type data, cleaning operation data, and material delivery data. The AI ​​analysis engine, connected to the multi-dimensional data perception layer, includes a time-series database, a multi-factor correlation analysis model, and a prediction and early warning module. The time-series database stores data collected by the multi-dimensional data perception layer and constructs a four-dimensional data storage structure encompassing time, space, particle size, and concentration. The multi-factor correlation analysis model establishes a correlation between the air quality index and environmental parameters based on the four-dimensional data storage structure. The prediction and early warning module predicts the trend of particulate matter concentration changes at different spatial locations within a preset future time period based on the four-dimensional data storage structure and triggers tiered early warnings based on the prediction results. The intelligent monitoring platform, connected to the AI ​​analysis engine, is used to generate control instructions based on the graded early warning and send them to the execution equipment for execution. It collects air quality data before and after the control and feeds back the control effect verification data to the AI ​​analysis engine to optimize the multi-factor correlation analysis model.

[0005] In one embodiment, the particle monitoring network is deployed in a layered manner around the operating table, in the core operating area, the auxiliary instrument area, and the personnel activity area.

[0006] In one embodiment, the expression of the multi-factor association analysis model is: , in, Let be the air quality index at time t. for The concentration matrix of multi-size particles at time t. for The intensity index of human activity at any given time. for The risk factor of the surgical type at any given time. for Factors affecting cleaning operations at any given time for The degree of disruption to logistics activities at any given time. This is the error term.

[0007] In one embodiment, the personnel activity intensity index is obtained by quantifying the personnel activity data, which includes the frequency of medical staff entering and leaving, their movement trajectory, and the duration of their stay in a preset area.

[0008] In one embodiment, the surgical type risk coefficient is obtained by associating the surgical type data, which includes the surgical code and its corresponding infection risk level.

[0009] In one embodiment, the cleaning operation impact factor is obtained by quantifying the cleaning operation data, which includes cleaning time, cleaning method, and disinfectant type; the logistics activity interference degree is obtained by quantifying the material delivery data, which includes the instrument pack opening time, auxiliary material delivery route, and logistics vehicle travel time.

[0010] In one embodiment, the prediction and early warning module includes a data preparation unit, a trend prediction unit, and an early warning triggering unit; The data preparation unit is connected to the time series database and is used to extract particle concentration data for historical periods and the current period from the four-dimensional data storage structure, and to perform noise reduction, completion and normalization preprocessing on the extracted data. The trend prediction unit is connected to the data preparation unit and is used to input the processed data into the LSTM or Transformer time series prediction model and output the predicted values ​​of particulate matter concentration at different spatial locations within the future preset time period. The early warning triggering unit is connected to the trend prediction unit and is used to compare the predicted particulate matter concentration with a preset air quality risk threshold. Based on the comparison result, it generates a prompt-level warning, a warning-level warning, or a severe warning and pushes the generated warning information to the intelligent monitoring platform.

[0011] In one embodiment, the intelligent monitoring platform includes a control strategy knowledge base, which stores recommended control measures corresponding to different warning levels, different surgical types, and different pollution scenarios; the intelligent monitoring platform matches and generates control instructions from the control strategy knowledge base according to the graded warning.

[0012] In one embodiment, the recommended control measures include at least one of the following: Start the laminar flow purification system and adjust the purification airflow and purification time; Adjust the surgical schedule; Adjust the execution time or method of the cleaning and disinfection process; or Adjust the delivery route or delivery time of the goods.

[0013] In one embodiment, the AI ​​analysis engine further includes a graph generation module, which is connected to the time-series database and is used to generate a heat map of the spatial distribution of particle concentration based on the four-dimensional data storage structure, and to overlay and display the trajectory information corresponding to the personnel activity data on the heat map of the spatial distribution of particle concentration.

[0014] The aforementioned AI-powered intelligent operating room air quality prediction and monitoring system collects multi-size particle concentration data and multi-dimensional environmental parameters through a multi-dimensional data perception layer, providing a comprehensive data foundation for air quality analysis. The AI ​​analysis engine utilizes a time-series database to construct a four-dimensional data storage structure and establishes a correlation between the air quality index and environmental parameters based on a multi-factor correlation analysis model. The prediction and early warning module then predicts the trend of particulate matter concentration changes at different spatial locations in the future and triggers tiered early warnings, achieving accurate prediction and early warning of operating room air quality. The intelligent monitoring platform generates and executes control instructions based on tiered early warnings, while simultaneously collecting data before and after control and feeding it back to the AI ​​analysis engine to optimize the correlation model, forming a closed-loop management system. This improves the timeliness and accuracy of operating room air quality control, reduces the risk of surgical site infection, and ensures medical safety. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a structural block diagram of an AI-enabled intelligent operating room air quality prediction and monitoring system in one embodiment. Detailed Implementation

[0017] To facilitate understanding of the technical solutions provided in the embodiments of this application, the background technology involved in the embodiments of this application will be described below.

[0018] With the continuous development of medical technology, the air quality in operating rooms, as the core treatment area of ​​hospitals, is directly related to the infection rate at surgical sites and is crucial for patients' postoperative recovery and medical safety. Currently, air quality management in operating rooms mainly adopts fixed-point sampling tests or periodic manual tests, relying on traditional testing equipment and human experience for regulation.

[0019] However, this traditional method has significant technical limitations in practical applications. First, fixed-point sampling makes it difficult to accurately reflect spatial distribution differences within the operating room, particularly the particulate matter concentration gradient between the critical area centered on the operating table and surrounding auxiliary areas. This leads to discrepancies between the monitoring data and the actual air quality in the core operating area. Second, traditional methods focus only on particulate matter concentration, failing to incorporate dynamic environmental factors such as personnel activity, surgical procedures, cleaning operations, and material delivery. This makes it impossible to establish a correlation between monitoring data and pollution sources, hindering the rapid identification of causes and the implementation of targeted measures when air quality anomalies occur. More critically, existing technologies can only provide historical data statistics and retrospective analysis, lacking the ability to predict future air quality trends. This means that control measures can only be implemented passively after pollution occurs, failing to achieve proactive intervention.

[0020] The root cause of the aforementioned technical deficiencies lies in the fact that traditional operating room air quality management lacks the ability to integrate and perceive multi-source data, comprehensively analyze complex factors, and predict future trends. This leaves air quality control in a reactive state, making it difficult to meet the demands of modern operating rooms for refined and intelligent air quality management. Especially during high-risk surgeries, fluctuations in air quality can directly impact the success or failure of the operation, highlighting the urgent need to address the technical issues of lagging control.

[0021] Based on this, this embodiment provides an AI-enabled intelligent operating room air quality prediction and monitoring system that can achieve multi-source data fusion, accurate prediction, and proactive control.

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present application.

[0023] Furthermore, in this embodiment, the collection, storage, processing, and use of all data comply with relevant laws and regulations and are carried out with the authorization of patients, medical staff, and related personnel to ensure personal information security and privacy protection.

[0024] In summary, existing operating room air quality management methods suffer from a lack of multi-source data fusion and perception capabilities, as well as the ability to predict future trends, resulting in lagging regulation and difficulty in meeting the needs of refined management. To address this issue, this embodiment provides an AI-enabled intelligent operating room air quality prediction and monitoring system, such as... Figure 1 As shown, it includes a multi-dimensional data perception layer, an AI analysis engine, and an intelligent supervision platform.

[0025] The multidimensional data sensing layer includes a particle monitoring network and a parameter acquisition module. The particle monitoring network is deployed in layers around the operating table to collect multi-size particle concentration data. The parameter acquisition module is used to collect environmental parameters, including at least personnel activity data, surgical type data, cleaning operation data, and material delivery data.

[0026] In this embodiment, the multidimensional data sensing layer is the foundation for data acquisition. The particle sensor can be a miniature intelligent particle sensor based on the principle of laser scattering, such as the PMS5003 sensor, which has a diameter of less than 5 cm and features small size, high accuracy, and low power consumption. Other similar sensors of the same type can also be selected in practical applications; this embodiment does not limit the choice. Each sensor collects the particle number concentration in real time across multiple particle size ranges, including but not limited to PM0.3, PM0.5, PM1.0, PM2.5, PM5.0, and PM10, with the concentration unit denoted as # / m³. 3Personnel activity data can be obtained through UWB positioning systems or infrared counting devices. For example, medical staff wear UWB tags, and UWB readers deployed in the operating room record their entry and exit frequency, movement trajectory, and dwell time in real time. Surgical type data is obtained by interfaceing with the hospital's HIS system to obtain surgical codes and associating them with surgical infection risk levels according to preset rules. Cleaning operation data is collected through IoT cleaning equipment or manual input, including cleaning time, cleaning method (wet / dry), and disinfectant type. Material delivery data is recorded through a material delivery tracking unit, recording information such as instrument pack opening time, auxiliary material delivery route, and logistics vehicle travel time.

[0027] The AI ​​analysis engine connects to a multi-dimensional data perception layer, including a time-series database, a multi-factor correlation analysis model, and a prediction and early warning module. The time-series database stores data collected by the multi-dimensional data perception layer and constructs a four-dimensional data storage structure encompassing time, space, particle size, and concentration. The multi-factor correlation analysis model establishes the correlation between the air quality index and environmental parameters based on the four-dimensional data storage structure. The prediction and early warning module predicts the trend of particulate matter concentration changes at different spatial locations within a preset future time period based on the four-dimensional data storage structure and triggers tiered early warnings based on the prediction results.

[0028] In this embodiment, the AI ​​analysis engine is responsible for data processing and analysis. A high-performance time-series database (such as InfluxDB) is used to store all collected data and construct a four-dimensional data storage structure. This structure organizes data according to four dimensions: time, space (sensor location), particle size, and concentration, facilitating rapid subsequent retrieval and analysis. A multi-factor correlation analysis model establishes the correlation between the Air Quality Index (AQI) and environmental parameters based on historical data in the four-dimensional data storage structure. This model can be implemented using methods such as multiple regression and neural networks. The prediction and early warning module uses time-series prediction algorithms (such as LSTM) based on real-time and historical data in the four-dimensional data storage structure to predict the trend of particulate matter concentration changes at different spatial locations over a future period. The prediction period can be set according to the type of surgery and actual needs, such as 30 minutes or 60 minutes. After comparing the prediction results with a preset air quality risk threshold, an early warning of the corresponding level is triggered, such as a prompt, warning, or severe level.

[0029] Among them, the intelligent monitoring platform is connected to the AI ​​analysis engine. It is used to generate control instructions based on the graded early warning and send them to the execution equipment for execution. It collects air quality data before and after the control and feeds back the control effect verification data to the AI ​​analysis engine to optimize the multi-factor correlation analysis model.

[0030] In this embodiment, the intelligent monitoring platform receives tiered early warning information and generates control instructions based on the warning level and the current surgical type from its built-in control strategy knowledge base. These control instructions are then sent to the executing devices, including laminar flow purification systems, surgical scheduling systems, and cleaning equipment control systems. For example, the instructions can control the laminar flow purification system to adjust its fan speed or start-up time, or notify management personnel to adjust surgical schedules. Simultaneously, the platform collects air quality data before and after the control measures are implemented and feeds the verification data back to the AI ​​analysis engine to optimize the parameters of the multi-factor correlation analysis model, forming a closed-loop management system.

[0031] Based on the above, a multi-dimensional data perception layer was used to collect comprehensive data on operating room air quality and related environmental factors. The AI ​​analysis engine used a four-dimensional data storage structure and a multi-factor correlation model to achieve data fusion and trend prediction. The intelligent monitoring platform then actively adjusted and provided feedback optimization based on the prediction results. This solved the problems of traditional methods being unable to predict in advance or actively intervene, and improved the timeliness and accuracy of operating room air quality control.

[0032] To further improve the accuracy of particle monitoring, in one embodiment, the particle monitoring network is deployed in a layered manner around the operating table, covering the core operating area, the auxiliary instrument area, and the personnel activity area.

[0033] In this embodiment, the spatial deployment of the particle monitoring network is centered on the operating table and divided into three concentric circular areas. The core operating area, with a radius of 1.5 meters, primarily covers the space directly involved in surgical procedures. The largest number of miniature intelligent particle sensors are deployed in this area, for example, four, to ensure data accuracy in critical areas. The auxiliary instrument area, a ring-shaped area with a radius of 1.5 to 3 meters, covers the area where instruments are placed and transferred. Three sensors are deployed in this area to monitor particulate contamination that may be generated during instrument operation. The personnel activity area, a ring-shaped area with a radius of 3 to 5 meters, covers the area where medical staff frequently move around. Two sensors are deployed in this area to monitor particulate matter diffusion caused by personnel activity. The sensor deployment density in each area can be adjusted according to the actual area and layout of the operating room; for example, the number of sensors can be appropriately increased for large operating rooms. This layered deployment effectively captures the spatial gradient changes in particulate matter diffusion from the core area to the periphery, providing more refined spatial dimension data for subsequent analysis.

[0034] Based on the above, by deploying miniature intelligent particle sensors in a spatial grid, the system can accurately identify pollution sources and diffusion paths. For example, when the particulate matter concentration in the core operating area increases, the source of pollution can be determined by analyzing data from the auxiliary equipment area and personnel activity area, thereby providing precise positioning for control measures and improving the pertinence and effectiveness of control.

[0035] To quantify the relationship between air quality and multiple factors, in one embodiment, the expression of the multifactor correlation analysis model is as follows: , in, Let be the air quality index at time t. for The concentration matrix of multi-size particles at time t. for The intensity index of human activity at any given time. for The risk factor of the surgical type at any given time. for Factors affecting cleaning operations at any given time for The degree of disruption to logistics activities at any given time. This is the error term.

[0036] In this embodiment, the multi-factor association analysis model maps multiple input variables to the air quality index through a function f. . It is a comprehensive indicator that can be calculated based on the weighted sum of the concentrations of particles of different sizes, for example, using the formula... ,in For the first Weight of each particle size segment for The concentration of that particle size range at any given time. The weights can be preset according to the type of surgery and hospital standards. It is a multidimensional matrix where rows represent different spatial locations and columns represent different particle size ranges, storing the particle concentration at each location and particle size at time t. It is an index derived from the quantification of personnel activity data. It is a risk coefficient associated with the type of surgery. and These are the influencing factors for cleaning operations and logistics activities, respectively. (Function) It can take many forms, such as a linear regression model: ,in These are regression coefficients, obtained through training on historical data; alternatively, nonlinear models such as neural networks can be used. Error term. The model is used to correct for stochastic fluctuations that cannot be explained by the model, and is determined through residual analysis. This model reveals the combined impact of various factors on air quality, providing a theoretical basis for subsequent forecasting and regulation.

[0037] Based on the above, through a multi-factor correlation analysis model, the system can identify the main factors affecting air quality within a specific time period. For example, when the AQI rises abnormally at a certain moment, and the particle concentration in the core area of ​​the multi-size particle concentration matrix increases, while the human activity intensity index shows frequent human activity, it can be determined that the pollution mainly comes from human activity, thereby strengthening human behavior management or adjusting the operating parameters of the purification equipment in a targeted manner.

[0038] To clarify the specific quantification method of the personnel activity intensity index, based on the aforementioned multi-factor correlation analysis model, personnel activity data includes the frequency of entry and exit of medical staff, their movement trajectory, and the duration of their stay in the preset area.

[0039] In detail, the calculation of the personnel activity intensity index H(t) comprehensively considers the frequency of entry and exit, movement trajectory, and duration of stay. Specifically, the location information of each medical staff member can be obtained in real time through an Ultra Wide Band (UWB) positioning system, and statistics can be compiled. Total number of people in the operating room at any given time (e.g., in a 5-minute time window) And the speed of movement of each person. and the length of stay in each region One method of quantification is: ,in and These are weighting coefficients, and different values ​​can be set according to the personnel's roles (such as surgeon, nurse, anesthesiologist). A simpler model can also be used: ,in For entry and exit frequency, The total distance traveled, The average length of stay, These are empirical coefficients. In practical applications, these coefficients can also be learned from historical data using machine learning methods. This index quantifies the degree of air disturbance caused by human activity, providing an important input for multi-factor models.

[0040] Based on the above, this quantitative method enables the system to transform human dynamics into calculable indicators, allowing the model to objectively reflect the impact of human activities on air quality and providing data support for subsequent prediction and regulation.

[0041] To reflect the different air quality requirements of different surgical procedures, based on the aforementioned multi-factor association analysis model, the surgical procedure data includes the surgical code and its corresponding infection risk level.

[0042] In detail, the surgical type risk coefficient is obtained by connecting to the Hospital Information System (HIS) to retrieve the surgical code for the current procedure and based on a pre-defined infection risk level mapping table. For example, for high-risk surgeries such as joint replacement surgery and organ transplantation, the risk coefficient is set to 1.0; for medium-risk surgeries such as open-chest surgery, it is set to 0.6; and for low-risk surgeries such as surface surgeries, it is set to 0.2. This mapping relationship can be established by the hospital's infection control department according to guidelines and can be dynamically adjusted. During the surgical procedure... The system remains constant from the start to the end of the surgery, but if the type of surgery changes during the procedure (such as the addition of high-risk procedures), the system can update in real time. This coefficient is used to adjust the thresholds for air quality forecasting and warnings; for example, high-risk surgeries require stricter air quality standards.

[0043] Based on the above, by introducing a surgical type risk coefficient, the system can dynamically adjust the control strategy according to the infection risk of different surgeries, ensuring that the air quality of high-requirement surgeries is given priority, which reflects the concept of differentiated management.

[0044] To quantify the impact of cleaning operations and logistics activities, based on the aforementioned multi-factor correlation analysis model, cleaning operation data includes cleaning time, cleaning method, and disinfectant type; the interference of logistics activities is obtained by quantifying material distribution data, which includes the opening time of medical kits, the delivery route of auxiliary materials, and the travel time of logistics vehicles.

[0045] In detail, factors affecting cleaning operations Used to evaluate the effectiveness of cleaning operations in improving air quality. Cleaning operation data is obtained through IoT cleaning devices or manual input. One quantification method is to monitor the rate of decrease in particle concentration in each area for a period of time after the cleaning operation (e.g., 30 minutes) and compare it with the historical average rate of decrease to obtain the effectiveness coefficient of this cleaning. Base values ​​can also be set directly based on the cleaning method: wet cleaning is more effective than dry cleaning, and disinfectants containing hydrogen peroxide are more effective than alcohol. It can be defined as: ,in For cleaning time, This is the cleaning method coefficient. This is the disinfectant type coefficient. As weighted. Logistics activity interference degree Used to assess the air disturbance caused by material distribution. By recording the opening time of equipment packs and the time periods of logistics vehicle passage, the number of logistics events occurring at time t and the duration of each event can be statistically analyzed. For example, ,in For the number of events, This refers to the total time the logistics vehicle spends in the operating room. These quantified factors, when input into a multi-factor model, can more accurately predict changes in air quality.

[0046] Based on the above, by quantifying cleaning and logistics factors and incorporating them into the model, the system can comprehensively consider various interfering factors, improve the accuracy of predictions, and provide a basis for optimizing cleaning plans and logistics routes.

[0047] To specifically implement the prediction and early warning function, in one embodiment, the prediction and early warning module includes a data preparation unit, a trend prediction unit, and an early warning triggering unit.

[0048] Specifically, the data preparation unit connects to the time series database to extract particle concentration data for historical and current periods from the four-dimensional data storage structure, and performs denoising, completion, and normalization preprocessing on the extracted data.

[0049] In this embodiment, the data preparation unit first reads particle concentration data for the most recent period (e.g., the past 6 hours) and the current moment from the four-dimensional data storage structure of the time-series database. This data is organized by time, spatial location, and particle size. Then, preprocessing is performed: noise reduction uses median filtering or wavelet transform to remove abnormal peaks; completion uses linear interpolation or the mean of adjacent sensors to fill missing values; normalization maps the data to the [0,1] interval to eliminate the influence of dimensions.

[0050] Specifically, the trend prediction unit is connected to the data preparation unit, which is used to input the processed data into the LSTM or Transformer time series prediction model and output the predicted values ​​of particulate matter concentration at different spatial locations within a preset future time period.

[0051] In this embodiment, the trend prediction unit inputs the preprocessed data into the time series prediction model. If an LSTM model is used, its input dimension is (time step, number of spatial locations × number of particle sizes), and the output is the predicted concentration sequence for each spatial location and each particle size within a preset future time period (e.g., 30 minutes). The LSTM model is trained using historical data, and the loss function is the mean squared error. The Transformer model, on the other hand, utilizes a self-attention mechanism to capture long-term temporal dependencies.

[0052] Specifically, the early warning triggering unit is connected to the trend prediction unit, which compares the predicted particulate matter concentration with the preset air quality risk threshold, generates a prompt-level warning, a warning-level warning, or a severe warning based on the comparison results, and pushes the generated warning information to the intelligent monitoring platform.

[0053] In this embodiment, the early warning triggering unit compares the predicted value with a preset threshold. The threshold can be dynamically adjusted according to the risk coefficient of the surgical type: the threshold is more stringent for high-risk surgeries. For example, the comparison results are divided into three levels: when the predicted value exceeds the threshold but does not exceed 1.2 times, a prompt-level early warning is triggered, indicating a slight exceedance that requires continuous monitoring; when it exceeds 1.2 times but does not exceed 1.5 times, a warning-level early warning is triggered, suggesting intervention; when it exceeds 1.5 times, a severe warning is triggered, requiring immediate action. The early warning information is pushed to the intelligent monitoring platform and relevant personnel through sound and light, SMS, system pop-ups, etc.

[0054] Based on the above, this prediction and early warning mechanism enables the system to detect pollution before it occurs, providing a window of opportunity for proactive control and effectively preventing pollution from interfering with the surgical procedure.

[0055] To achieve intelligent control, in one embodiment, the intelligent monitoring platform includes a control strategy knowledge base, which stores recommended control measures corresponding to different warning levels, different surgical types, and different pollution scenarios; the intelligent monitoring platform matches and generates control instructions from the control strategy knowledge base based on the graded warning.

[0056] In this embodiment, the control strategy knowledge base is a structured database that stores recommended measures under multi-dimensional conditions. For example, for the scenario of "warning-level alert + joint replacement surgery + core area contamination," the recommended measure is "activate the laminar flow purification system in high-speed mode, suspend unnecessary personnel entry and exit, and strengthen local suction." The rules in the knowledge base are formulated by hospital infection control experts based on experience and can be continuously updated according to actual results. When the intelligent monitoring platform receives a graded alert, it retrieves matching control measures from the knowledge base based on conditions such as the alert level, the current surgery type, and the contamination area (determined by the spatial location in the four-dimensional data storage structure), and then generates specific control instructions. The instructions can be control signals automatically sent to the execution equipment or operational suggestions sent to management personnel. For example, the instruction "laminar flow purification system_wind speed_80%" directly controls the equipment, while "suggest adjusting the surgery schedule" is sent to the operating room head nurse via SMS. Through matching in the knowledge base, the standardization and intelligence of control measures are achieved, avoiding the arbitrariness of manual decision-making.

[0057] Based on the above, the scientific nature and consistency of the control measures were ensured, and the response efficiency was improved.

[0058] To clarify the specific content of the recommended regulatory measures, based on the aforementioned regulatory strategy knowledge base, the recommended regulatory measures include at least one of the following: Start the laminar flow purification system and adjust the purification airflow and purification time; Adjust the surgical schedule; Adjust the execution time or method of the cleaning and disinfection process; or Adjust the delivery route or delivery time of the goods.

[0059] In this embodiment, the recommended control measures cover multiple aspects, including equipment control, process optimization, and management adjustments. Activating the laminar flow purification system and adjusting the airflow speed and duration is the most direct intervention method, such as increasing the airflow speed from 40% to 80% of normal mode and extending the operating time to 30 minutes. Adjusting the surgical schedule refers to scheduling high-risk surgeries during periods of predicted lower contamination or postponing non-emergency surgeries to avoid contamination peaks. Adjusting the cleaning and disinfection process includes cleaning in advance, using more efficient disinfectants, or increasing the frequency of cleaning. Adjusting the material delivery route or time period refers to guiding logistics vehicles to avoid core operating areas or postponing deliveries to the surgical interval. These measures can be used individually or in combination, depending on the specific scenario. After matching with the knowledge base, the system can issue multiple instructions simultaneously to achieve comprehensive control. For example, during a severe warning, the system may simultaneously activate the laminar flow purification system, notify the nurse station to suspend personnel entry and exit, and prompt the logistics center to postpone deliveries.

[0060] Based on the above, the diverse control measures enable the system to flexibly respond to different situations and ensure that air quality quickly returns to a safe level.

[0061] To visually demonstrate the correlation between air quality and human activity, in one embodiment, the AI ​​analysis engine also includes a graph generation module. The graph generation module is connected to a time-series database and is used to generate a heat map of the spatial distribution of particle concentration based on a four-dimensional data storage structure. The trajectory information corresponding to the human activity data is then overlaid on the heat map of the spatial distribution of particle concentration.

[0062] In this embodiment, the map generation module utilizes data visualization technology to read particle concentration data and personnel activity trajectory data from the four-dimensional data storage structure of the time-series database, generating intuitive graphics. The spatial distribution heatmap uses an operating room floor plan as a background, with different colors representing the current particle concentration levels in each area: red for high concentration and blue for low concentration. Simultaneously, the historical movement trajectories of medical personnel are overlaid on the heatmap, for example, lines representing paths taken over the past 5 minutes and dots representing current locations. This display allows managers to clearly see the overlap between contaminated hotspots and personnel activity paths, enabling rapid determination of whether contamination is caused by personnel activity. Furthermore, the module can generate time-series line graphs, showing the changes in concentration in different areas over time, and displaying them synchronously with the personnel activity intensity curve. These maps are presented through the monitoring platform's display terminal, supporting interactive operations such as zooming and time rewinding.

[0063] Based on the above, the system uses visualization techniques to transform complex data into easily understandable graphics, assisting managers in making quick decisions and improving the system's usability and user experience.

[0064] Those skilled in the art will understand that all or part of the functions of the system and its modules described in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed by a processor, the computer program can implement the functions of the system described in the above embodiments. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0065] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0066] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. An AI-enabled intelligent operating room air quality prediction and monitoring system, characterized in that, include: The multi-dimensional data sensing layer includes a particle monitoring network and a parameter acquisition module. The particle monitoring network is deployed with multiple particle sensors layered around the operating table to collect multi-size particle concentration data. The parameter acquisition module is used to collect environmental parameters, including at least personnel activity data, surgical type data, cleaning operation data, and material delivery data. The AI ​​analysis engine, connected to the multi-dimensional data perception layer, includes a time-series database, a multi-factor correlation analysis model, and a prediction and early warning module. The time-series database stores data collected by the multi-dimensional data perception layer and constructs a four-dimensional data storage structure encompassing time, space, particle size, and concentration. The multi-factor correlation analysis model establishes a correlation between the air quality index and environmental parameters based on the four-dimensional data storage structure. The prediction and early warning module predicts the trend of particulate matter concentration changes at different spatial locations within a preset future time period based on the four-dimensional data storage structure and triggers tiered early warnings based on the prediction results. The intelligent monitoring platform, connected to the AI ​​analysis engine, is used to generate control instructions based on the graded early warning and send them to the execution equipment for execution. It collects air quality data before and after the control and feeds back the control effect verification data to the AI ​​analysis engine to optimize the multi-factor correlation analysis model.

2. The system according to claim 1, characterized in that, The particle monitoring network is deployed in a layered manner around the operating table, covering the core operating area, auxiliary instrument area, and personnel activity area.

3. The system according to claim 1, characterized in that, The expression for the multi-factor association analysis model is: ， in, Let be the air quality index at time t. for The concentration matrix of multi-size particles at time t. for The intensity index of human activity at any given time. for The risk factor of the surgical type at any given time. for Factors affecting cleaning operations at any given time. for The degree of disruption to logistics activities at any given time. This is the error term.

4. The system according to claim 3, characterized in that, The personnel activity intensity index is obtained by quantifying the personnel activity data, which includes the frequency of entry and exit of medical staff, their movement trajectory, and the duration of their stay in the preset area.

5. The system according to claim 3, characterized in that, The surgical type risk coefficient is obtained by associating the surgical type data, which includes the surgical code and its corresponding infection risk level.

6. The system according to claim 3, characterized in that, The cleaning operation influencing factors are obtained by quantifying the cleaning operation data, which includes cleaning time, cleaning method, and disinfectant type; the logistics activity interference is obtained by quantifying the material distribution data, which includes the equipment pack opening time, auxiliary material delivery route, and logistics vehicle travel time.

7. The system according to claim 1, characterized in that, The prediction and early warning module includes a data preparation unit, a trend prediction unit, and an early warning triggering unit; The data preparation unit is connected to the time series database and is used to extract particle concentration data for historical periods and the current period from the four-dimensional data storage structure, and to perform noise reduction, completion and normalization preprocessing on the extracted data. The trend prediction unit is connected to the data preparation unit and is used to input the processed data into the LSTM or Transformer time series prediction model and output the predicted values ​​of particulate matter concentration at different spatial locations within the future preset time period. The early warning triggering unit is connected to the trend prediction unit and is used to compare the predicted particulate matter concentration with a preset air quality risk threshold. Based on the comparison result, it generates a prompt-level warning, a warning-level warning, or a severe warning and pushes the generated warning information to the intelligent monitoring platform.

8. The system according to claim 1, characterized in that, The intelligent monitoring platform includes a control strategy knowledge base, which stores recommended control measures corresponding to different warning levels, different surgical types, and different pollution scenarios; the intelligent monitoring platform matches and generates control instructions from the control strategy knowledge base based on the graded warning.

9. The system according to claim 8, characterized in that, The recommended control measures include at least one of the following: Start the laminar flow purification system and adjust the purification airflow and purification time; Adjust the surgical schedule; Adjust the execution time or method of the cleaning and disinfection process; or Adjust the delivery route or delivery time of the goods.

10. The system according to claim 1, characterized in that, The AI ​​analysis engine also includes a graph generation module, which is connected to the time-series database and is used to generate a heat map of the spatial distribution of particle concentration based on the four-dimensional data storage structure, and to overlay and display the trajectory information corresponding to the personnel activity data on the heat map of the spatial distribution of particle concentration.

Images