ICU multi-device aerosol disinfection and killing cooperative control method based on dynamic risk field

Abstract

The application provides an ICU multi-device aerosol disinfection collaborative control method based on a dynamic risk field, comprising the following steps: acquiring and fusing biological aerosol detection data and positioning data of a mobile disinfection robot in real time, and constructing a dynamic three-dimensional infection risk heat map; monitoring the dynamic three-dimensional infection risk heat map, triggering an event-driven decision immediately when the risk value of any spatial point is detected to exceed a preset emergency threshold or the risk value change rate exceeds a preset sharp increase threshold, and generating corresponding high-priority abstract instructions; triggering a global optimization decision at a preset period, acquiring the state information of all disinfection devices and the current risk heat map at the current period, running a multi-objective optimization model under the condition of meeting a preset static responsibility area matrix constraint condition, and solving and generating a normal cruise abstract instruction set of each disinfection device in the next period; and delivering the high-priority abstract instructions or the normal cruise abstract instruction set to corresponding devices for execution.

Description

Technical Field

[0001] This invention belongs to the field of air disinfection, and in particular relates to a collaborative control method for aerosol disinfection of multiple ICU devices based on a dynamic risk field. Background Technology

[0002] The ICU is a high-risk area for hospital-acquired infections, as patients often carry highly pathogenic bacteria such as multidrug-resistant bacteria. Aerosols in the ICU environment primarily originate from patient respiratory activities and airway-related medical procedures, including coughing, mechanical ventilation, endotracheal intubation, and suctioning. Simultaneously, the operation of medical equipment and the activities of healthcare workers lead to the diffusion and resuspension of aerosols, ensuring the continuous presence and dynamic distribution of pathogenic particles within the ward space. Therefore, ICU aerosols are characterized by multi-source generation, continuous release, and dynamic spatial changes, placing high demands on the real-time performance, coverage, and coordination of disinfection systems.

[0003] Existing ICU air purification technologies mainly include:

[0004] Static disinfection equipment, such as fixed ultraviolet sterilizers and air purifiers, is typically installed in fixed locations in patient rooms and operates continuously. Its disadvantages include: fixed locations have coverage blind spots, making it difficult to respond to sudden localized contamination events, and high energy consumption during long-term operation.

[0005] Mobile disinfection robots: These robots can autonomously navigate to target areas for disinfection. In existing technologies, mobile robots typically operate according to preset paths or are remotely controlled by humans. They lack the ability to adaptively adjust to the real-time aerosol distribution, resulting in delayed responses and difficulty in timely intervention during aerosol outbreaks (such as during suctioning procedures).

[0006] Single control strategy: Existing disinfection equipment mostly adopts timed start-stop or simple concentration threshold control, lacking coordinated scheduling among multiple devices, and cannot achieve three-dimensional prevention and control by permanent coverage of fixed equipment and dynamic supplementation of mobile equipment.

[0007] The invention with application number CN202411938158A discloses a control system for synergistic disinfection using ultraviolet light and low-temperature plasma. This system monitors environmental parameters in real time using sensors and compares them with preset values, adjusting the output of the disinfection equipment to achieve closed-loop control. However, this solution primarily targets the disinfection of products on conveyor belts and does not address the collaborative scheduling of multiple devices in the complex dynamic environment of an ICU, nor does it incorporate spatial positioning and dynamic risk distribution for decision-making.

[0008] The invention with application number CN119839853A discloses a virtual-physical collaborative control method for a bricklaying robot based on digital twins. This method updates the physical entity's state in real time through a virtual entity model and optimizes the control strategy within the virtual environment. While this solution involves collaborative control, it is applied to the field of industrial robots and does not address aerosol disinfection or multi-device scheduling.

[0009] In summary, there is currently a lack of a method that can perceive the dynamic risk distribution in the ICU space in real time and coordinate the scheduling of fixed and mobile disinfection equipment accordingly, so as to achieve rapid response to sudden risks and optimization of overall energy efficiency. Summary of the Invention

[0010] To address the above technical problems, this invention provides a collaborative control method for aerosol disinfection of multiple devices in an ICU based on a dynamic risk field. The specific technical solution is as follows:

[0011] A collaborative control method for aerosol disinfection of multiple devices in an ICU based on a dynamic risk field includes the following steps:

[0012] Step S1: Real-time acquisition and fusion of bioaerosol detection data and mobile disinfection robot positioning data to construct a dynamic three-dimensional infection risk heat map covering the ICU space;

[0013] Step S2: Monitor the dynamic three-dimensional infection risk heat map. When the risk value of any spatial point exceeds the preset emergency threshold, or the risk value change rate exceeds the preset sudden increase threshold, immediately trigger event-driven decision-making and generate the corresponding high-priority abstract instruction.

[0014] Step S3: Trigger global optimization decisions at preset intervals. At the current time of the cycle, obtain the status information of all disinfection equipment and the current risk heat map. Under the condition of satisfying the preset static responsibility area matrix constraints, run the multi-objective optimization model to solve and generate the normal patrol abstract instruction set of each disinfection equipment in the next cycle.

[0015] Step S4: Translate the generated high-priority abstract instructions or normal patrol abstract instruction set into specific task packages that can be executed by each disinfection device, and send them to the corresponding devices for execution.

[0016] Compared with the prior art, the present invention has the following beneficial effects:

[0017] (1) Dynamic risk perception and quantification: By integrating the data from the real-time tracing system of three-dimensional reconstruction of bioaerosols and the SLAM positioning data of the mobile disinfection robot, a dynamic three-dimensional infection risk heat map covering the entire ICU space is constructed using a spatiotemporal interpolation algorithm. This transforms the abstract aerosol distribution into a quantifiable and calculable spatial risk field, providing a precise basis for subsequent optimization decisions.

[0018] (2) Hybrid Triggering Collaborative Decision Making: The hybrid decision making mechanism of “event triggering + periodic triggering” is adopted, which can not only achieve rapid command response (emergency elimination / diffusion blocking) to sudden local pollution (such as the moment of suctioning operation), but also achieve optimal system energy efficiency and equipment load balance through periodic global optimization, thus solving the problem of response blind spots and low energy efficiency in single mode.

[0019] (3) Multi-device collaborative three-dimensional prevention and control: By combining the offline planned static responsibility area matrix with online dynamic optimization, a three-dimensional prevention and control system is realized, with fixed equipment permanently covering the core area and mobile equipment dynamically filling the mobile area, which significantly improves the disinfection coverage and response speed in the complex environment of the ICU.

[0020] (4) Multi-objective optimization and energy efficiency improvement: With the optimization objectives of maximizing risk elimination, minimizing total energy consumption, and balancing the load, the system effectively reduces long-term operating energy consumption and extends equipment lifespan while ensuring the disinfection effect.

[0021] (5) Three-dimensional prevention and control: Fixed equipment is stationed to cover the core area, and mobile equipment is dynamically deployed to fill the mobile area, with a coverage rate of ≥98%.

[0022] (6) Modular architecture: easy to expand and maintain, adaptable to ICU wards of different sizes. Attached Figure Description

[0023] Figure 1 This is a flowchart of the collaborative control method according to an embodiment of the present invention;

[0024] Figure 2 This is a schematic diagram illustrating the construction of a dynamic risk field according to an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram illustrating the event triggering decision-making process according to an embodiment of the present invention;

[0026] Figure 4 This is a schematic diagram of periodic global optimization in an embodiment of the present invention;

[0027] Figure 5 This is a system module structure diagram of an embodiment of the present invention;

[0028] Figure 6 This is a schematic diagram of the fixed disinfection device of the present invention;

[0029] Figure 7 This is a schematic diagram of the mobile disinfection robot structure of the present invention;

[0030] Figure 8 This is a schematic diagram of intelligent disinfection in a double room at the PLA General Hospital. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other. To achieve the above objectives, this invention adopts the following technical solution.

[0032] This invention provides a collaborative control method for aerosol disinfection of multiple devices in an ICU based on a dynamic risk field, such as... Figure 1 , Figure 5 As shown, it includes the following steps:

[0033] Step S1: Real-time acquisition and fusion of bioaerosol detection data and mobile disinfection robot positioning data to construct a dynamic three-dimensional infection risk heat map covering the ICU space;

[0034] Step S2: Monitor the dynamic three-dimensional infection risk heat map. When the risk value of any spatial point exceeds the preset emergency threshold, or the risk value change rate exceeds the preset sudden increase threshold, immediately trigger event-driven decision-making and generate the corresponding high-priority abstract instruction.

[0035] Step S3: Trigger global optimization decisions at preset intervals. At the current time of the cycle, obtain the status information of all disinfection equipment and the current risk heat map. Under the condition of satisfying the preset static responsibility area matrix constraints, run the multi-objective optimization model to solve and generate the normal patrol abstract instruction set of each disinfection equipment in the next cycle.

[0036] Step S4: Translate the generated high-priority abstract instructions or normal patrol abstract instruction set into specific task packages that can be executed by each disinfection device, and send them to the corresponding devices for execution.

[0037] Furthermore, such as Figure 2 As shown, the specific steps in step S1 for constructing a dynamic three-dimensional infection risk heat map include:

[0038] Receive discrete concentration data from a real-time bioaerosol 3D reconstruction and tracing system deployed in the ICU. or risk probability data ;in t represents the spatial grid index, and t represents time.

[0039] Receive pose data, including position, from the mobile disinfection device in real time via SLAM algorithm. and attitude Q(t); These are the x, y, and z axis coordinates.

[0040] The discrete concentration data and the pose data are spatiotemporally aligned, and a three-dimensional meshed risk probability field covering the entire ICU space is reconstructed using a spatiotemporal interpolation algorithm. This serves as the dynamic three-dimensional infection risk heat map.

[0041] Furthermore, such as Figure 3 As shown, the high-priority abstract instructions in step S2 include:

[0042] When the detected risk value exceeds the preset emergency threshold When the time comes, an "emergency annihilation" command is generated. This command takes the current risk peak point as the initial target. As the mobile device moves toward the target, it continuously adjusts the target position according to the real-time updated risk heat map until it reaches the high concentration area and performs comprehensive disinfection.

[0043] When a continuous diffusion trajectory of bioaerosols is detected, and the rate of change of the risk value at any point on the trajectory exceeds a preset surge threshold, the risk is considered to be increasing. When the spread path is blocked, a "spread blocking" command is generated. This command uses the end point of the current spread path as the initial interception point. As the mobile device moves toward the interception point, it continuously tracks the changes in the end point of the path based on the real-time updated risk heat map and dynamically adjusts the target position until the spread path is blocked.

[0044] Furthermore, such as Figure 4 As shown, the objective function of the multi-objective optimization model in step S3 includes: maximizing the amount of risk elimination within the period, minimizing the total energy consumption of the system within the period, and minimizing the load imbalance between each disinfection device;

[0045] The static responsibility area matrix is ​​predetermined by an offline multi-objective optimization algorithm. This matrix defines the main coverage area of ​​different fixed disinfection devices and the coverage radius and mobile standby area of ​​mobile disinfection robots.

[0046] Furthermore, step S4, translating abstract instructions into concrete task packages, includes:

[0047] Based on the type of abstract instruction and the target area, the preset parameter mapping table is queried to determine the target operating parameters of each disinfection device, including the target coordinates, disinfection power level, and running time.

[0048] The target operating parameters are encapsulated into executable instructions according to the device communication protocol and then sent via a wireless network.

[0049] Furthermore, the disinfection equipment includes at least: a fixed disinfection device and a mobile disinfection robot. For example... Figure 6As shown, in a fixed disinfection device, air enters through the air inlet, passes through the disinfection module and the barrier module, and exits through the air outlet. An air sensor is installed to detect air quality, and a control module is also included. Figure 7 As shown, the mobile disinfection robot includes a fixed disinfection device and a mobile disinfection robot base, which is movable. The disinfection module uses cold plasma disinfection technology.

[0050] Example 1: System Deployment and Initialization

[0051] like Figure 8 As shown, this embodiment uses a double room in the ICU of the Fifth Medical Center of the People's Liberation Army as an example. The room is approximately 25 square meters in size and has two beds, equipped with medical equipment such as ventilators and monitors. The room is pre-configured as follows:

[0052] Two fixed disinfection devices were installed next to the beds in the wards.

[0053] One mobile disinfection robot is on standby near the nurses' station;

[0054] A real-time bioaerosol 3D reconstruction tracking system is distributed in the four corners of the ward and around the beds, outputting real-time bioaerosol concentration field data of the ICU space at a refresh rate of no less than 10Hz. ;

[0055] During system initialization, offline layout planning is performed: based on the 3D model of the ward, basic airflow field data obtained from CFD simulation, and equipment performance parameters (effective radius of action, power consumption, etc.), the NSGA-II multi-objective optimization algorithm is used to solve for the optimal deployment positions of fixed devices and the optimal standby points of mobile robots, with the objectives of maximizing risk coverage, minimizing global response time, and maximizing system energy efficiency, generating a static responsibility area matrix. This matrix clarifies that: fixed device 1 is mainly responsible for the area near bed A, fixed device 2 is mainly responsible for the area near bed B, and the mobile robot is responsible for the blind spots covered by the fixed devices and serves as a global mobile response unit.

[0056] Example 2: Construction of Dynamic Risk Field (corresponding to step S1)

[0057] During system operation, the following operations are performed at 100ms intervals:

[0058] (1) Data acquisition: Concentration data of the real-time tracer system for three-dimensional reconstruction of bioaerosols is read in real time via RS485 bus. Simultaneously, it receives pose data in real time from the mobile disinfection robot via wireless network, output through the SLAM algorithm. .

[0059] (2) Spatiotemporal alignment: Since there may be a slight deviation between the detector sampling time and the robot positioning time, linear interpolation is used to align the two to a unified timestamp. .

[0060] (3) Spatiotemporal interpolation reconstruction: the current time Discrete concentration data and their corresponding spatial coordinates are used as input. A continuous three-dimensional gridded risk field covering the entire ward space is reconstructed using the spatio-temporal kriging interpolation algorithm. The spatio-temporal kriging model is as follows:

[0061] ;

[0062] in, The point to be estimated at time [time] The risk value, For the risk value of the known observation point, The weighting coefficients are obtained by solving the spatiotemporal variogram model.

[0063] (4) Grid-based output: The ward space is divided into a 10cm×10cm×10cm grid, and each grid point is assigned a risk value. The value ranges from 0 to 1, representing the relative infection risk probability at that location at the current moment, forming a dynamic three-dimensional infection risk heat map.

[0064] Example 3: Event-triggered decision (corresponding to step S2)

[0065] The system monitors the dynamic risk field at 1-second intervals and determines in real time whether event-driven commands are triggered.

[0066] (1) Emergency threshold determination: Set an emergency threshold When a risk value R ≥ 0.8 is detected at any grid point, an "Emergency Annihilation" command is immediately generated. This command points to the coordinates of the risk point and assigns it the highest priority.

[0067] (2) Sudden increase threshold determination: Set a sudden increase threshold. This means the risk value increases by more than 0.5 per second. When the rate of change of risk value at any grid point is detected... When the system detects a risk point, it triggers a "diffusion blocking" command. The system uses the currently detected risk point as the initial target location for the mobile disinfection robot. As the robot moves towards the target, it continuously receives high-refresh-rate (≥10Hz) updated risk heatmap data and dynamically adjusts its target location based on the latest risk distribution until it reaches the actual risk source. This "walking and updating" strategy eliminates the need for airflow field prediction, avoids model errors, and enables precise tracking of sudden bioaerosol events.

[0068] Once the event is triggered, the generated abstract command (such as "Emergency Annihilation: Target Coordinates (2.5, 3.0, 1.2)") immediately proceeds to step S4 for execution, without waiting for the next cycle. Figure 3 The flowchart illustrating the event triggering logic is shown.

[0069] Example 4: Periodic global optimization (corresponding to step S3)

[0070] The system triggers a global optimization decision every T=10 minutes. At the beginning of each cycle... Perform the following operations:

[0071] (1) Status Acquisition: Obtain the current status of all devices, including:

[0072] Fixed equipment 1: Current power P1 = 150W, cumulative running time t1 = 2.3h;

[0073] Fixed equipment 2: Current power P2 = 150W, cumulative running time t2 = 2.3h;

[0074] Mobile robot: Current location (1.2,0.5,0), current battery level 85%, current task status "idle".

[0075] (2) Risk field integral: Predicting the risk change in the next period T based on the current risk field. Based on historical trends, estimate the cumulative risk at each grid point within the period if no intervention is taken.

[0076] (3) Construct a multi-objective optimization model:

[0077] Decision variables:

[0078] Power settings of fixed devices 1 and 2 , (0 indicates standby, 1 indicates low power, and 2 indicates high power).

[0079] The task sequence of the mobile robot includes the coordinates of several task points and their corresponding power levels, as well as path planning.

[0080] Objective function:

[0081] ;

[0082] in, This is the estimated amount of risk to be eliminated within the period. This refers to the disinfection efficiency coefficient corresponding to the equipment power. For spatial points At time t, the power level for eliminating [something] is [level]. For spatial points The relative probability of infection at time t;

[0083] Total energy consumption within the cycle. The power level of the device at time t. For the equipment in power mode Instantaneous power;

[0084] For equipment load imbalance, Cumulative running time for a single device Average uptime for all devices;

[0085] α, β, γ are weighting coefficients. In this embodiment, α = 0.5, β = 0.3, and γ = 0.2.

[0086] Constraints:

[0087] Power constraints: The power of the mobile robot is dynamically adjusted according to the task, but the peak value is ≤400W.

[0088] Power constraint: Total energy consumption of the mobile robot task sequence ≤ current battery level × 15%;

[0089] Responsibility area constraints: Fixed equipment mainly covers its static responsibility area, while mobile robots are responsible for the mobile area;

[0090] Time constraint: Total time of mobile robot task sequence ≤ T.

[0091] (4) Solving the optimization model: The above multi-objective optimization problem is solved using a genetic algorithm to obtain the Pareto optimal solution set. Based on the project preference (prioritizing the elimination effect), the solution with the largest risk elimination is selected to generate the "normal patrol" instruction set for the next cycle.

[0092] The solution obtained in this embodiment is:

[0093] Fixed disinfection device 1: High-power continuous operation (covering area A of the hospital bed);

[0094] Fixed disinfection device 2: Low power operation (covering area B of the ward, as this area has a lower risk).

[0095] Mobile disinfection robot: The task sequence is [(1.5,2.0,0) low-power disinfection for 60s → (3.0,1.5,0) low-power disinfection for 30s], and the path planning adopts A. The algorithm avoids areas with hospital beds and medical equipment.

[0096] Example 5: Instruction Translation and Execution (corresponding to step S4)

[0097] (1) Instruction translation: The system maintains a parameter mapping table to convert abstract instructions into concrete parameters:

[0098]

[0099] (2) Instruction encapsulation: The main controller (STM32MP1) calls the corresponding private communication protocol according to the target device ID to encapsulate the specific parameters into a binary instruction frame. For example, the target coordinate instruction format sent to the mobile robot is: [Frame header 0xAA][Device ID 0x01][Instruction code 0x10][X coordinate 4 bytes][Y coordinate 4 bytes][Z coordinate 4 bytes][Checksum 2 bytes][Frame tail 0x55].

[0100] (3) Command issuance: The encapsulated command is sent to the corresponding device via wireless local area network (Wi-Fi). After the device executes the command, it returns a status message (including pose, battery level, power, task progress, etc.). The main controller parses the message and reports it to the system platform, forming a closed-loop control.

[0101] Example 6: Synergistic Disinfection Effect

[0102] In a simulation test conducted in a double room of the ICU at the Fifth Medical Center of the People's Liberation Army, the collaborative control method of this embodiment achieved the following results:

[0103] Response time for sudden aerosol incidents (simulated suctioning operation): ≤5 seconds (from the outbreak of the risk to the start of the mobile robot).

[0104] Energy consumption during periodic normal operation: 32% lower than that of stationary equipment operating at high power all time;

[0105] Spatial coverage: Fixed equipment covers the core area (around the beds), and mobile robots cover the mobile area (corridors, gaps between equipment), with an overall coverage rate of ≥98%;

[0106] Equipment load balancing: The difference in cumulative running time between the two fixed devices and the mobile robot is ≤15%.

[0107] Example 7: System Module Structure

[0108] like Figure 5 As shown, the system for implementing the above method includes:

[0109] Risk heat map construction module: used to acquire and integrate aerosol detection data and mobile disinfection equipment location data in real time to construct a dynamic three-dimensional infection risk heat map;

[0110] Hybrid trigger decision module: used to monitor the risk heatmap, and when the risk value exceeds the emergency threshold or the risk value change rate exceeds the sudden increase threshold, it triggers event-driven decision generation to generate high-priority abstract instructions, and triggers global optimization decision according to a preset cycle;

[0111] Global optimization solution module: used to obtain the status of all disinfection equipment and the current risk heat map at the current cycle time, run a multi-objective optimization model under the constraints of the static responsibility area matrix, and generate a normal patrol abstract instruction set;

[0112] Command translation and distribution module: used to translate abstract commands into specific task packages that can be executed by various disinfection devices and then distribute them;

[0113] Multiple disinfection devices: used to receive and execute the specific task package.

[0114] In this invention, alternatives to spatiotemporal interpolation algorithms include spatiotemporal kriging interpolation and other data interpolation methods that can reconstruct continuous fields from discrete points.

[0115] In this invention, alternatives to multi-objective optimization algorithms include genetic algorithms and other multi-objective optimization solution algorithms.

[0116] In this invention, alternatives to trajectory recognition and endpoint tracking can be used: in addition to real-time recognition based on heatmaps, other methods can also be employed to track the diffusion front.

[0117] In this invention, the static responsibility region matrix can be replaced by other methods besides offline multi-objective optimization to pre-divide the equipment responsibility region.

[0118] In this invention, alternative disinfection technologies, besides cold plasma, can also be used.

[0119] In this invention, alternatives to communication protocols and instruction encapsulation can be used, in addition to the RS485 bus, to employ other industrial buses or communication protocols.

[0120] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A collaborative control method for aerosol disinfection of multiple devices in an ICU based on a dynamic risk field, characterized in that, Includes the following steps: Step S1: Real-time acquisition and fusion of bioaerosol detection data and mobile disinfection robot positioning data to construct a dynamic three-dimensional infection risk heat map covering the ICU space; Step S2: Monitor the dynamic three-dimensional infection risk heat map. When the risk value of any spatial point exceeds the preset emergency threshold, or the risk value change rate exceeds the preset sudden increase threshold, immediately trigger event-driven decision-making and generate the corresponding high-priority abstract instruction. Step S3: Trigger global optimization decisions at preset intervals. At the current time of the cycle, obtain the status information of all disinfection equipment and the current risk heat map. Under the condition of satisfying the preset static responsibility area matrix constraints, run the multi-objective optimization model to solve and generate the normal patrol abstract instruction set of each disinfection equipment in the next cycle. Step S4: Translate the generated high-priority abstract instructions or normal patrol abstract instruction set into specific task packages that can be executed by each disinfection device, and send them to the corresponding devices for execution; The high-priority abstract instructions mentioned in step S2 include: When the detected risk value exceeds the preset emergency threshold When the time comes, an "emergency annihilation" command is generated. This command takes the current risk peak point as the initial target. As the mobile device moves toward the target, it continuously adjusts the target position according to the real-time updated risk heat map until it reaches the high concentration area and performs comprehensive disinfection. When a continuous diffusion trajectory of bioaerosols is detected, and the rate of change of the risk value at any point on the trajectory exceeds a preset surge threshold, the risk is considered to be increasing. When the spread path is blocked, a "spread blocking" command is generated. This command uses the end point of the current spread path as the initial interception point. As the mobile device moves toward the interception point, it continuously tracks the changes in the end point of the path according to the real-time updated risk heat map and dynamically adjusts the target position until the spread path is blocked. The disinfection equipment includes at least: a fixed disinfection device and a mobile disinfection robot.

2. The method for coordinated control of aerosol disinfection of multiple ICU devices based on a dynamic risk field according to claim 1, characterized in that, Step S1, which involves constructing a dynamic three-dimensional infection risk heatmap, specifically includes: Receive discrete concentration data from a real-time bioaerosol 3D reconstruction and tracing system deployed in the ICU. or risk probability data ;in This is a spatial grid index, where t is time; Receive real-time pose data from mobile disinfection equipment, including location. and attitude Q(t); These are the x, y, and z axis coordinates. The discrete concentration data and the pose data are spatiotemporally aligned, and a three-dimensional meshed risk probability field covering the entire ICU space is reconstructed using a spatiotemporal interpolation algorithm. This serves as the dynamic three-dimensional infection risk heat map.

3. The method for coordinated control of aerosol disinfection of multiple ICU devices based on a dynamic risk field according to claim 1, characterized in that, The objective function of the multi-objective optimization model in step S3 includes: maximizing the amount of risk elimination within the period, minimizing the total energy consumption of the system within the period, and minimizing the load imbalance between the various disinfection devices; The static responsibility area matrix is ​​predetermined by an offline multi-objective optimization algorithm. This matrix defines the main coverage area of ​​different fixed disinfection devices and the coverage radius and mobile standby area of ​​mobile disinfection robots.

4. The method for coordinated control of aerosol disinfection of multiple ICU devices based on a dynamic risk field according to claim 1, characterized in that, Step S4, which translates abstract instructions into concrete task packages, includes: Based on the type of abstract instruction and the target area, the preset parameter mapping table is queried to determine the target operating parameters of each disinfection device, including the target coordinates, disinfection power level, and running time. The target operating parameters are encapsulated into executable instructions according to the device communication protocol and then sent via a wireless network.

5. The method for coordinated control of aerosol disinfection of multiple ICU devices based on a dynamic risk field according to claim 1, characterized in that, Linear interpolation was used to interpolate the concentration data. pose data Align to a uniform timestamp .

6. The method for coordinated control of aerosol disinfection of multiple ICU devices based on a dynamic risk field according to claim 1, characterized in that, Spatiotemporal interpolation algorithm reconstruction includes: converting the current time... Using discrete concentration data and corresponding spatial coordinates as input, a spatiotemporal kriging interpolation algorithm is employed to reconstruct a continuous three-dimensional gridded risk field covering the entire ward space. The spatiotemporal kriging model is as follows: ； in, The point to be estimated at time [time] The risk value, The risk value for the known observation point. The weighting coefficients are obtained by solving the spatiotemporal variogram model.

7. The method for coordinated control of aerosol disinfection of multiple ICU devices based on a dynamic risk field according to claim 1, characterized in that, Set emergency threshold Set a sudden increase threshold That is, the risk value increases by more than 0.5 per second.

8. The method for coordinated control of aerosol disinfection of multiple ICU devices based on a dynamic risk field according to claim 1, characterized in that, In step S3, Objective function: ； in, This is the estimated amount of risk to be eliminated within the period. The disinfection efficiency coefficient corresponding to the equipment power; For spatial points At time t, the power level for eliminating [something] is [level]. For spatial points The relative probability of infection at time t; Total energy consumption within the cycle. The power level of the device at time t. For the equipment in power mode Instantaneous power; For equipment load imbalance, Cumulative running time for a single device Average uptime for all devices; α, β, γ are weighting coefficients; Constraints: Power constraints: , The power settings are for fixed disinfection devices; the power of mobile disinfection robots is dynamically adjusted according to the task, but the peak value is ≤400W. Power constraint: Total energy consumption of the mobile robot task sequence ≤ current battery level × 15%; Responsibility area constraints: Fixed equipment mainly covers its static responsibility area, while mobile robots are responsible for the mobile area; Time constraint: The total time of the mobile robot task sequence is ≤ T, where T is the triggering period.

Images