Infectious disease prevention and control whole process decision simulation evaluation integrated platform based on digital twinning

Abstract

The present application relates to the technical field of personnel evacuation, in particular to an infectious disease prevention and control whole-process decision simulation evaluation integrated platform based on digital twinning. The platform comprises: a data acquisition module for acquiring people flow state data of a target place; a risk evaluation module for determining risk quantitative values of each sub-region based on the people flow state data; a strategy generation module for obtaining a shunt adaptation degree of each sub-region to each exit according to a risk diffusion trend factor, a risk quantitative value of each sub-region and a distance between the sub-region and the exit; screening an optimal adaptation exit of each sub-region by using the shunt adaptation degree, merging sub-regions adjacent to and matching the same optimal adaptation exit into a target region, and generating a corresponding people flow guiding strategy; and an effect evaluation module for evaluating global total risks before and after implementation of the people flow guiding strategy respectively, and evaluating the guiding strategy effect based on the global total risks before and after implementation. The present application improves the accuracy of evaluation results of the people flow guiding strategy effect.

Description

Technical Field

[0001] This invention relates to the field of personnel evacuation technology, specifically to an integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins. Background Technology

[0002] Digital twin technology can accurately replicate the spatiotemporal distribution, flow, and dynamic evolution of disease transmission within a given location. It can simulate and extrapolate the entire process of prevention and control decisions, such as crowd control, verifying the potential effects of these decisions and providing quantifiable scientific evidence for infectious disease prevention and control decisions. The risk of infectious disease transmission is closely related to the flow of people within a location and is constantly changing. Relying solely on static crowd flow data at a single moment for risk assessment cannot reflect the real-time changes in disease transmission, leading to inaccurate evaluations of crowd control strategies. Summary of the Invention

[0003] To address the inaccurate evaluation results of existing methods when assessing the effectiveness of crowd control strategies, the present invention aims to provide an integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins. The specific technical solution adopted is as follows: This invention provides an integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins. The platform includes: The data acquisition module is used to collect pedestrian flow data in different sub-areas of the target location; The risk assessment module is used to evaluate the risk diffusion trend factor of each sub-region based on the pedestrian flow status data; and to determine the risk quantification value of each sub-region based on the risk diffusion trend factor. The strategy generation module is used to obtain the diversion adaptation degree from each sub-region to each exit based on the risk diffusion trend factor of each sub-region, the risk quantification value, and the spatial distance between each sub-region and each exit; use the diversion adaptation degree to filter the optimal adaptation exit of each sub-region, and merge adjacent sub-regions that match the same optimal adaptation exit into a target region to generate a corresponding pedestrian flow guidance strategy. The effectiveness evaluation module is used to evaluate the overall risk before and after the implementation of the crowd guidance strategy, and evaluate the effectiveness of the guidance strategy based on the overall risk before and after implementation.

[0004] Preferably, the step of evaluating the risk diffusion trend factor of each sub-region based on the pedestrian flow status data includes: The transmission risk value of the candidate sub-region is determined based on the pedestrian flow status data of the candidate sub-region; Calculate the spatial rate of change of the propagation risk value along the horizontal axis and the spatial rate of change along the vertical axis, and calculate the temporal rate of change of the propagation risk value along the time axis. Based on the spatial rate of change along the horizontal axis, the spatial rate of change along the vertical axis, and the time rate of change, a risk diffusion trend factor for candidate sub-regions is obtained by combining a neural network model. The candidate sub-region can be any sub-region.

[0005] Preferably, obtaining the propagation risk value of the candidate sub-region includes: The product of the crowd density and the degree of disorder in pedestrian movement in the candidate sub-region is used as the propagation risk value of the candidate sub-region; wherein, the degree of disorder in pedestrian movement is determined based on the spatial variation rate of the pedestrian flow components along the lateral and longitudinal directions and the cross variation rate between the lateral and longitudinal flow components.

[0006] Preferably, the step of obtaining the risk diffusion trend factor of the candidate sub-region based on the spatial change rate along the horizontal axis, the spatial change rate along the vertical axis, and the time change rate, combined with a neural network model, includes: The propagation risk value, the spatial change rate along the horizontal axis, the spatial change rate along the vertical axis, and the population density of the candidate sub-region are input into a convolutional neural network to obtain a spatial feature vector that characterizes the correlation strength between the candidate sub-region and different directions. The propagation risk values ​​at multiple collection times within the current period, the time change rate, and the population density of the candidate sub-region are input into the long short-term memory network to obtain temporal feature values ​​that characterize the risk evolution rate of the candidate sub-region. By integrating the spatial feature vector, the temporal feature value, the population density of the candidate sub-region, and the time change rate, the risk diffusion trend factor of the candidate sub-region is obtained.

[0007] Preferably, determining the risk quantification value of each sub-region based on the risk diffusion trend factor includes: The risk amplification factor is obtained based on the rate of change of the propagation risk value of the candidate sub-region along the time axis and the risk diffusion trend factor. By combining the temporal feature values ​​and the risk amplification factor, the risk quantification value of the candidate sub-region is obtained.

[0008] Preferably, obtaining the diversion suitability from each sub-region to each exit based on the risk diffusion trend factor of each sub-region, the risk quantification value, and the spatial distance between each sub-region and each exit includes: Based on the risk diffusion trend factor and the risk quantification value of the candidate sub-region, a risk fit value is obtained; The negative correlation mapping result between the spatial distance of the candidate sub-region and the exit is used as the distance adaptation value; By combining the risk adaptation value and the distance adaptation value, the diversion adaptation degree from the candidate sub-region to the exit is obtained.

[0009] Preferably, the step of using the traffic adaptation degree to filter the optimal adaptation exit of each sub-region includes: taking the exit corresponding to the maximum value of the traffic adaptation degree from the candidate sub-region to all exits as the optimal adaptation exit of the candidate sub-region.

[0010] Preferably, the evaluation of the overall risk before and after the implementation of the crowd guidance strategy includes: The total global risk before the implementation of the crowd flow guidance strategy is obtained by averaging the risk quantification values ​​of all sub-regions at all times in the current period before the implementation of the guidance strategy. The total global risk after the implementation of the crowd flow guidance strategy is obtained by averaging the risk quantification values ​​of all sub-areas within a preset time period after the current time period after the implementation of the guidance strategy.

[0011] Preferably, obtaining the population density of the candidate sub-region includes: The ratio of the number of people in a candidate sub-region to the area of ​​the candidate sub-region is used as the population density of the candidate sub-region.

[0012] Preferably, the effect of the strategy guided by the overall risk assessment before and after implementation includes: The difference between the total global risk before implementation and the total global risk after implementation is taken as the risk reduction amount, and the relative risk change rate is obtained based on the risk reduction amount and the total global risk before implementation. The effectiveness of the guidance strategy is determined by comparing the relative rate of change of risk with a preset rate of change threshold.

[0013] The present invention has at least the following beneficial effects: The digital twin-based integrated platform for decision-making simulation and evaluation of infectious disease prevention and control provides this invention. Through a data acquisition module, it collects real-time pedestrian flow data from different sub-areas of a target location, ensuring the dynamism and realism of the foundational data for subsequent risk assessment and strategy generation. The risk assessment module evaluates the risk diffusion trend factors of each sub-area based on the pedestrian flow data, thereby determining the risk quantification value of each sub-area. This accurately reflects the real-time spread and spatial distribution characteristics of infectious diseases in complex pedestrian environments. The strategy generation module combines the risk diffusion trend factors, risk quantification values, and spatial distance between each sub-area and its exit to calculate the diversion suitability from each sub-area to its respective exit, and selects the optimal suitability based on this diversion suitability. The system merges adjacent sub-regions that match the same optimal exit into a target region, thereby generating targeted crowd guidance strategies and achieving intelligent, dynamic, and precise control of crowd flow. The effect evaluation module further evaluates the overall global risk before and after the implementation of the guidance strategy, and quantifies the effect of the guidance strategy by comparing the overall global risk before and after implementation. This forms a closed-loop decision simulation process of data collection, evaluation, decision-making, and verification. The digital twin-based infectious disease prevention and control full-process decision simulation evaluation integration platform provided by this invention overcomes the problem of inaccurate evaluation results caused by relying on static crowd flow data for risk assessment in existing technologies, and improves the scientific nature of infectious disease prevention and control decisions and the reliability of crowd guidance strategy effect evaluation. Attached Figure Description

[0014] To more clearly illustrate the technical solutions and advantages 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.

[0015] Figure 1 This is a structural block diagram of the integrated decision-making simulation and evaluation platform for the entire process of infectious disease prevention and control based on digital twins, provided in an embodiment of the present invention. Detailed Implementation

[0016] The following description, in conjunction with the accompanying drawings, details the specific scheme of the integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins provided by this invention.

[0017] Example of an integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins: The specific scenario addressed in this embodiment is: shopping malls and other areas with high traffic volume are prone to localized crowd gatherings, which can easily lead to the spread of infectious diseases. This embodiment will guide people to appropriate entrances and exits based on the distribution of people, avoiding excessive crowding and providing scientific support for the hygiene and safety management of public places such as shopping malls.

[0018] This embodiment proposes an integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins, such as... Figure 1 As shown, the digital twin-based infectious disease prevention and control decision simulation and evaluation integration platform of this embodiment includes a data acquisition module, a risk assessment module, a strategy generation module, and an effect evaluation module.

[0019] The data acquisition module is used to collect pedestrian flow data in different sub-areas of the target location. The risk assessment module is used to evaluate the risk diffusion trend factor of each sub-region based on the pedestrian flow status data; and to determine the risk quantification value of each sub-region based on the risk diffusion trend factor. The strategy generation module is used to obtain the diversion adaptation degree from each sub-region to each exit based on the risk diffusion trend factor of each sub-region, the risk quantification value, and the spatial distance between each sub-region and each exit; use the diversion adaptation degree to filter the optimal adaptation exit of each sub-region, and merge adjacent sub-regions that match the same optimal adaptation exit into a target region to generate a corresponding pedestrian flow guidance strategy. The effectiveness evaluation module is used to evaluate the overall risk before and after the implementation of the crowd guidance strategy, and evaluate the effectiveness of the guidance strategy based on the overall risk before and after implementation.

[0020] I. Data Acquisition Module, used to collect pedestrian flow data in different sub-areas of the target location.

[0021] First, the target location is divided into sub-regions. The target location is a public area where people gather, such as a shopping mall. As a specific implementation method, the target location can be divided into multiple square sub-regions with a side length of 4 meters. In specific applications, the implementer can set the size of the sub-regions according to the size of the target location.

[0022] Then, for each sub-area, the data acquisition module collects real-time pedestrian flow data for each sub-area of ​​the target location. This pedestrian flow data includes information such as the number of people and their direction of movement. Simultaneously, it obtains the location of the center point of each sub-area and uses this location as the position of that sub-area. As a specific implementation, the data acquisition frequency is set to once every 0.1 seconds, meaning pedestrian flow data is collected every 0.1 seconds, with a time interval of 0.1 seconds between two adjacent acquisition times. In specific applications, the implementer can adjust this setting according to the specific circumstances.

[0023] II. Risk assessment module, used to evaluate the risk diffusion trend factor of each sub-region based on the pedestrian flow status data; and to determine the risk quantification value of each sub-region based on the risk diffusion trend factor.

[0024] To achieve full-process decision simulation and evaluation of infectious disease prevention and control in target locations, we first construct risk judgment criteria that reflect the relationship between personnel activities and risk spread, based on the laws of infectious disease risk transmission in actual locations. Secondly, based on the characteristics of global risk distribution, and taking into account prevention and control safety and traffic flow efficiency, we generate personnel diversion and guidance strategies that have risk avoidance and reasonable passage. Finally, we compare the overall risk changes before and after the implementation of the strategies to complete the quantitative evaluation of prevention and control measures.

[0025] Considering that the spread of infectious diseases within enclosed shopping malls primarily relies on close, frequent, and disorderly contact between people, the more concentrated the crowds, the more chaotic the movement, and the more instances of intermingling and crowding, the higher the probability of close contact between individuals and the longer the duration of such contact. This also increases the likelihood of virus transmission during this process. Therefore, based on this phenomenon, a correlation between crowd flow and transmission risk can be established using the overall characteristics of real-time crowd flow.

[0026] Different areas such as shopping mall lobbies and corridors may experience situations where people stop and avoid each other due to differences in the direction of movement and frequency of movement. Such spatial interactions between people are key factors that can lead to close contact and increase the risk of transmission. Therefore, the risk assessment module combines crowd density with the degree of interaction between people in the space to quantify the situation.

[0027] This example uses one sub-region as an example; the method provided in this embodiment can be used to process other sub-regions as well.

[0028] Any sub-region is designated as a candidate sub-region. The propagation risk value of the candidate sub-region is determined based on the pedestrian flow data of the candidate sub-region. Specifically, the ratio of the number of people in the candidate sub-region to the area of ​​the candidate sub-region is taken as the population density of the candidate sub-region. The product of the population density of the candidate sub-region and the degree of disorder in pedestrian movement is taken as the propagation risk value of the candidate sub-region. The degree of disorder in pedestrian movement is determined based on the spatial variation rate of the pedestrian flow components along the horizontal and vertical directions and the cross variation rate between the horizontal and vertical flow components.

[0029] As a specific implementation method, the degree of disorder in personnel movement can be determined as follows: Extract the average directional deviation angle of all personnel in the candidate sub-region (based on the positive x-axis of space), and combine it with the number of personnel in the candidate sub-region. Then, use trigonometric functions to decompose the flow components of personnel in the candidate sub-region along the horizontal direction and the flow components along the vertical direction. The specific calculation formula is as follows: =Φ×cos(θ), =Φ×sin(θ), where, This represents the lateral flow component of personnel in the candidate sub-region. The vertical flow component of personnel in the candidate sub-region is represented by Φ, the number of personnel in the candidate sub-region is represented by θ, the average directional deviation angle of all personnel in the candidate sub-region is represented by cosine, and sin is represented by sine. The horizontal and vertical implementers are set according to specific circumstances.

[0030] The ratio of the lateral flow component of people in a candidate sub-region to the length (side length) of the candidate sub-region along the horizontal axis is taken as the spatial variation rate of the lateral flow component of people in the candidate sub-region; the ratio of the longitudinal flow component of people in a candidate sub-region to the length (side length) of the candidate sub-region along the vertical axis is taken as the spatial variation rate of the longitudinal flow component of people in the candidate sub-region. The ratio of the lateral flow component of people in a candidate sub-region to the length (side length) of the candidate sub-region along the vertical axis is denoted as the first ratio, and the ratio of the longitudinal flow component of people in a candidate sub-region to the length (side length) of the candidate sub-region along the horizontal axis is denoted as the second ratio. Combining the spatial variation rates of the lateral and longitudinal flow components of people in a candidate sub-region, the first ratio, and the second ratio, the degree of disorder in the movement of people in the candidate sub-region is obtained; the degree of disorder in the movement of people in the candidate sub-region can be expressed as:

[0031] in, This indicates the degree of disorder in the movement of people in the candidate sub-region. This represents the lateral flow component of personnel in the candidate sub-region. This represents the longitudinal flow component of personnel in the candidate sub-region. This represents the length of the candidate sub-region along the horizontal axis. This represents the length of the candidate sub-region along the vertical axis. Indicates the first ratio. This represents the second ratio. This represents the rate of change between the horizontal and vertical flow components of the candidate sub-region.

[0032] The risk of transmission within a shopping mall does not exist in isolation within a single sub-area or at a single moment. An increase in risk in one sub-area can propagate to adjacent grids through pedestrian flow, and the risk dynamically changes with pedestrian flow and time. Therefore, the spatial transmission trend of risk can be represented by quantifying the three-dimensional spatiotemporal gradient of risk.

[0033] Specifically, the propagation risk value of a candidate sub-region is calculated by subtracting its propagation risk value from the propagation risk value of the next adjacent sub-region along the horizontal axis. The ratio of this difference to the length (side length) of the candidate sub-region along the horizontal axis is used as the spatial rate of change of the propagation risk value of the candidate sub-region along the horizontal axis. Similarly, the propagation risk value of a candidate sub-region is calculated by subtracting its propagation risk value from the propagation risk value of the next adjacent sub-region along the vertical axis. This ratio is used as the spatial rate of change of the propagation risk value of the candidate sub-region along the vertical axis. Finally, the propagation risk value of a candidate sub-region at the current acquisition time is calculated by subtracting its propagation risk value at the previous acquisition time. The ratio of this difference to the time interval between two adjacent acquisition times is used as the temporal rate of change of the propagation risk value of the candidate sub-region along the time axis.

[0034] Considering that risk diffusion within shopping malls is the result of both spatial transmission and temporal dynamic evolution, the risk diffusion trend factor of candidate sub-regions is obtained by combining a neural network model based on the spatial rate of change of the propagation risk value of candidate sub-regions along the horizontal axis, the spatial rate of change of the propagation risk value of candidate sub-regions along the vertical axis, and the temporal rate of change of the propagation risk value of candidate sub-regions along the time axis.

[0035] Specifically, the propagation risk value of the current candidate sub-region, the spatial rate of change of the propagation risk value of the candidate sub-region along the horizontal axis, the spatial rate of change of the propagation risk value of the candidate sub-region along the vertical axis, and the population density of the candidate sub-region are input into the convolutional neural network (CNN) to obtain a spatial feature vector representing the correlation strength between the candidate sub-region and the orientation. ,For example: , representing horizontal and vertical spatial correlation strengths of 0.5 and 0.6, respectively. The network structure and training process of convolutional neural networks are existing technologies and will not be elaborated further here.

[0036] This embodiment uses a time period with a preset duration, where the current time is the last time point, as the current time period. The preset duration can be 3 minutes. Using the method provided in this embodiment, the propagation risk value for each acquisition time point within the current time period of the candidate sub-region can be obtained.

[0037] The propagation risk values ​​of all data collected in the current time period for the candidate sub-region, the rate of change of the propagation risk values ​​of the candidate sub-region along the time axis, and the population density of the candidate sub-region are arranged in chronological order to form a time-series data sequence. This time-series data sequence is input into a Long Short-Term Memory (LSTM) network. After layered processing through LSTM hidden layers, Dropout layers, and linear mapping layers, a time-series feature value is output to characterize the rate of risk evolution of the candidate sub-region. For example, the time-series feature value for peak periods is 0.9, and the time-series feature value for off-peak periods is 0.5. The network structure and training process of the LSTM network are existing technologies and will not be elaborated further here.

[0038] In the offline training phase of the neural network, to guide the network in learning and representing the spatial correlation strength and risk evolution rate characteristics, supervision labels are extracted and constructed in advance based on the simulation data retained from the historical twin operation of the target location. Specifically, the proportion of the relative number of triggers caused by the actual movement and gathering of people around the sub-region at each time point is statistically analyzed to calculate the real spatial transmission weight, which serves as the training supervision label for the correlation strength of that sub-region. The tangent curvature gradient value of the historical risk fluctuation recorded at that point is measured and used as the training supervision label to represent the risk evolution rate. Then, the mean square error between the forward calculation result and the constructed actual supervision label is used to calculate the loss, and the backpropagation correction of the error gradient is applied to constrain the two types of models to converge in the direction with the specified representation ability.

[0039] Since the spread of risk in shopping malls is a process involving both spatial transmission and temporal dynamic evolution, it is necessary to fuse spatial feature vectors and temporal feature values. In public places like shopping malls, spatial characteristics determine whether risk can spread from one area to another; without spatial connections, risk cannot be transmitted between different locations. Temporal characteristics, on the other hand, are responsible for changing the speed of spread. For example, during peak hours when crowds are concentrated and changes are rapid, risk spread is accelerated. Therefore, spatial feature vectors and temporal feature values ​​should be fused with spatial characteristics as the primary driver and temporal changes as a corrective factor.

[0040] Specifically, the magnitude of the spatial feature vector is calculated, and the density constraint factor is determined based on the population density of the candidate sub-region. The risk diffusion trend factor of the candidate sub-region is obtained based on the magnitude of the spatial feature vector, the temporal feature value, the rate of change of the propagation risk value of the candidate sub-region along the time axis, and the density constraint factor.

[0041] As a specific implementation method, a specific formula for calculating the risk diffusion trend factor is given. The risk diffusion trend factor of the candidate sub-region can be expressed as:

[0042] in, The risk diffusion trend factor represents the candidate sub-region. This represents the magnitude of the spatial feature vector of the candidate sub-region. Represents the temporal feature values ​​of candidate sub-regions. This represents the rate of change of the propagation risk value of the candidate sub-region along the time axis over time. This represents the normalized result of the population density in the candidate sub-region. This represents the density constraint factor.

[0043] In this embodiment, the data is normalized using the maximum and minimum value normalization method. The maximum and minimum values ​​used are the maximum and minimum population densities of all sub-regions in the historical statistics, respectively.

[0044] Because the characteristics of pedestrian flow vary in different areas of the shopping mall, local risk indicators within a sub-area cannot fully depict the relationship between pedestrian flow and risk in the grid. Therefore, the overall risk within a sub-area is quantified by combining the above-mentioned multi-dimensional characteristics.

[0045] Specifically, a risk amplification factor is obtained based on the rate of change of the propagation risk value of the candidate sub-region along the time axis. That is, the sum of the rate of change of the propagation risk value of the candidate sub-region along the time axis, the risk diffusion trend factor of the candidate sub-region, and a constant 1 is used as the risk amplification factor. Combining the time-series characteristic value and the risk amplification factor, the risk quantification value of the candidate sub-region is obtained. As a specific implementation, the product of the time-series characteristic value and the risk amplification factor is used as the risk quantification value of the candidate sub-region.

[0046] Using the above methods, we can obtain the risk quantification value for each sub-area. The risk quantification value integrates multi-dimensional characteristics such as static risk, diffusion trend, and pedestrian density, and realizes the quantification of the risk level of each sub-area of ​​the target site. It can objectively reflect the risk distribution of each sub-area and provide a unified quantitative basis for the subsequent formulation of shopping mall pedestrian diversion plan.

[0047] III. Strategy generation module, used to obtain the diversion adaptation degree from each sub-region to each exit based on the risk diffusion trend factor of each sub-region, the risk quantification value, and the spatial distance between each sub-region and each exit; using the diversion adaptation degree to filter the optimal adaptation exit of each sub-region, and merging adjacent sub-regions that match the same optimal adaptation exit into the target region, generating the corresponding pedestrian flow guidance strategy.

[0048] When matching sub-areas with entrances and exits, from an epidemic prevention and safety perspective, the first considerations are the current overall risk level at the exit and the future risk spread trend. A higher overall risk level at the exit indicates greater population density in that area, making it less suitable as a guidance target. A stronger risk spread trend at the exit indicates that congestion and risk are more likely to spread to surrounding areas, further expanding the epidemic prevention danger zone. Simultaneously, to ensure the guidance plan is reasonable and feasible, the distance between the grid and the exit must also be considered. The greater the distance, the longer it takes for people to walk, and the higher the probability of contact, which could actually increase the risk of transmission.

[0049] The following explanation will still use the candidate sub-region as an example. Other sub-regions can be processed using the method provided in this embodiment.

[0050] Specifically, risk adaptation values ​​are obtained based on the risk diffusion trend factor and risk quantification value of the candidate sub-regions; spatial coordinates of all exits of the target site are extracted; the negative correlation mapping result of the spatial distance between the candidate sub-regions and each exit is used as the distance adaptation value; and the diversion adaptation degree from the candidate sub-regions to each exit is obtained by combining the risk adaptation value and the distance adaptation value.

[0051] As a specific implementation method, a specific formula for calculating the flow adaptation degree is given. The flow adaptation degree from the candidate sub-region to the i-th exit can be expressed as:

[0052] in, This represents the flow adaptation degree from the candidate sub-region to the i-th exit. This represents the risk quantification value of the candidate sub-region. The risk diffusion trend factor represents the candidate sub-region. This represents the spatial distance between the candidate sub-region and the i-th exit. This represents an exponential function with the natural constant as its base.

[0053] When the risk quantification value of the candidate sub-region is larger, the risk diffusion trend factor is larger, and the spatial distance between the candidate sub-region and the i-th exit is greater, the people in the candidate sub-region are less suitable for evacuation from the i-th exit, and the diversion fit between the candidate sub-region and the i-th exit is smaller.

[0054] Using the above method, the flow adaptation degree from candidate sub-regions to each exit can be obtained. The exit corresponding to the maximum flow adaptation degree from the candidate sub-region to all exits is taken as the optimal adaptation exit of the candidate sub-region; sub-regions that are adjacent in location and match the same optimal adaptation exit are merged into a target region.

[0055] Furthermore, the optimal exit for each target area is translated into specific on-site guidance instructions, such as "Personnel currently near Gate 1, please proceed to Gate 2 (Main Gate)" or "Personnel located around a certain shop, please move to Gate 5 (Side Gate)." These instructions are then pushed to the corresponding area's public electronic information display screens, broadcasting system, or via Bluetooth beacons within the mall to mobile devices entering that area. The specific on-site guidance instructions are pre-set manually.

[0056] The strategy generation module constructs a pedestrian flow guidance strategy based on risk diffusion trend factors, risk quantification values, and spatial distance. It transforms abstract risk values ​​into executable guidance plans. By matching the optimal exit for each target area, it achieves balanced distribution and orderly guidance of pedestrian flow, providing data support for shopping malls to carry out routine safe guidance while ensuring normal operation.

[0057] IV. Effect Evaluation Module: This module is used to evaluate the overall risk before and after the implementation of the crowd guidance strategy, and to evaluate the effectiveness of the guidance strategy based on the overall risk before and after implementation.

[0058] The effectiveness evaluation module will assess the risks before and after the implementation of the crowd guidance strategy, and evaluate the effectiveness of the guidance strategy based on the changes in risks before and after the implementation.

[0059] Specifically, the risk quantification values ​​of all sub-areas before the implementation of the guidance strategy are averaged at all times within the current time period to obtain the total global risk before the implementation of the crowd flow guidance strategy; the risk quantification values ​​of all sub-areas after the implementation of the guidance strategy are averaged at all times within a preset time period after the current time period to obtain the total global risk after the implementation of the crowd flow guidance strategy, wherein the preset time period can be the same as the duration of the current time period.

[0060] After obtaining the total global risk before and after the implementation of the crowd guidance strategy, firstly, the difference between the total global risk before implementation and the total global risk after implementation is taken as the risk reduction amount. Then, based on the risk reduction amount and the total global risk before implementation, the relative risk change rate is obtained. Specifically, the ratio of the risk reduction amount to the total global risk before implementation is taken as the relative risk change rate. In particular, if the total global risk before implementation is 0, the relative risk change rate is directly set to 0.

[0061] Furthermore, the effectiveness of the guidance strategy is judged by comparing the relative rate of change of risk with a preset rate of change threshold. Specifically, if the relative rate of change of risk is greater than the preset rate of change threshold, it indicates that the current crowd guidance strategy is effective in reducing the risk of transmission throughout the area, meaning the strategy is effective. If the relative rate of change of risk is less than or equal to the preset rate of change threshold, it indicates that the strategy is ineffective. In this embodiment, the preset rate of change threshold is 0.6. In specific applications, implementers can set this threshold according to specific circumstances.

[0062] Finally, all data generated by the above calculation and evaluation process is stored in the platform database, and the final generated crowd guidance strategy is synchronized to the digital twin visualization interface to achieve data traceability and intuitive command execution.

[0063] The digital twin-based infectious disease prevention and control decision simulation and evaluation integrated platform provided in this embodiment collects real-time pedestrian flow data from different sub-areas of the target location through a data acquisition module, ensuring the dynamism and authenticity of the foundational data for subsequent risk assessment and strategy generation. The risk assessment module evaluates the risk diffusion trend factor of each sub-area based on the pedestrian flow data, thereby determining the risk quantification value of each sub-area, accurately reflecting the real-time spread and spatial distribution characteristics of infectious diseases in complex pedestrian environments. The strategy generation module combines the risk diffusion trend factor, risk quantification value, and spatial distance between each sub-area and its exit to calculate the diversion suitability from each sub-area to its exit, and selects the optimal suitability based on the diversion suitability. The system merges adjacent sub-regions that match the same optimal exit into a target region, thereby generating targeted crowd guidance strategies and achieving intelligent, dynamic, and precise control of crowd flow. The effect evaluation module further evaluates the overall global risk before and after the implementation of the guidance strategy, and quantifies the effect of the guidance strategy by comparing the overall global risk before and after implementation. This forms a closed-loop decision simulation process of data collection, evaluation, decision-making, and verification. The digital twin-based infectious disease prevention and control full-process decision simulation evaluation integration platform provided in this embodiment overcomes the problem of inaccurate evaluation results caused by relying on static crowd flow data for risk assessment in the prior art, and improves the scientific nature of infectious disease prevention and control decisions and the reliability of crowd guidance strategy effect evaluation.

[0064] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A digital twin-based integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control, characterized in that, The platform includes: The data acquisition module is used to collect pedestrian flow data in different sub-areas of the target location; The risk assessment module is used to evaluate the risk diffusion trend factor of each sub-region based on the pedestrian flow status data; and to determine the risk quantification value of each sub-region based on the risk diffusion trend factor. The strategy generation module is used to obtain the diversion adaptation degree from each sub-region to each exit based on the risk diffusion trend factor of each sub-region, the risk quantification value, and the spatial distance between each sub-region and each exit; use the diversion adaptation degree to filter the optimal adaptation exit of each sub-region, and merge adjacent sub-regions that match the same optimal adaptation exit into a target region to generate a corresponding pedestrian flow guidance strategy. The effectiveness evaluation module is used to evaluate the overall risk before and after the implementation of the crowd guidance strategy, and evaluate the effectiveness of the guidance strategy based on the overall risk before and after implementation.

2. The integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins as described in claim 1, characterized in that, The evaluation of risk diffusion trend factors for each sub-region based on the pedestrian flow status data includes: The transmission risk value of the candidate sub-region is determined based on the pedestrian flow status data of the candidate sub-region; Calculate the spatial rate of change of the propagation risk value along the horizontal axis and the spatial rate of change along the vertical axis, and calculate the temporal rate of change of the propagation risk value along the time axis. Based on the spatial rate of change along the horizontal axis, the spatial rate of change along the vertical axis, and the time rate of change, a risk diffusion trend factor for candidate sub-regions is obtained by combining a neural network model. The candidate sub-region can be any sub-region.

3. The integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins as described in claim 2, characterized in that, The acquisition of the propagation risk value of the candidate sub-region includes: The product of the crowd density and the degree of disorder in pedestrian movement in the candidate sub-region is used as the propagation risk value of the candidate sub-region; wherein, the degree of disorder in pedestrian movement is determined based on the spatial variation rate of the pedestrian flow components along the lateral and longitudinal directions and the cross variation rate between the lateral and longitudinal flow components.

4. The integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins as described in claim 2, characterized in that, The method of obtaining risk diffusion trend factors for candidate sub-regions based on the spatial rate of change along the horizontal axis, the spatial rate of change along the vertical axis, and the time rate of change, combined with a neural network model, includes: The propagation risk value, the spatial change rate along the horizontal axis, the spatial change rate along the vertical axis, and the population density of the candidate sub-region are input into a convolutional neural network to obtain a spatial feature vector that characterizes the correlation strength between the candidate sub-region and different directions. The propagation risk values ​​at multiple collection times within the current period, the time change rate, and the population density of the candidate sub-region are input into the long short-term memory network to obtain temporal feature values ​​that characterize the risk evolution rate of the candidate sub-region. By integrating the spatial feature vector, the temporal feature value, the population density of the candidate sub-region, and the time change rate, the risk diffusion trend factor of the candidate sub-region is obtained.

5. The integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins as described in claim 4, characterized in that, The step of determining the risk quantification value of each sub-region based on the risk diffusion trend factor includes: The risk amplification factor is obtained based on the rate of change of the propagation risk value of the candidate sub-region along the time axis and the risk diffusion trend factor. By combining the temporal feature values ​​and the risk amplification factor, the risk quantification value of the candidate sub-region is obtained.

6. The integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins as described in claim 2, characterized in that, The step of obtaining the diversion suitability from each sub-region to each exit based on the risk diffusion trend factor of each sub-region, the risk quantification value, and the spatial distance between each sub-region and each exit includes: Based on the risk diffusion trend factor and the risk quantification value of the candidate sub-region, a risk fit value is obtained; The negative correlation mapping result between the spatial distance of the candidate sub-region and the exit is used as the distance adaptation value; By combining the risk adaptation value and the distance adaptation value, the diversion adaptation degree from the candidate sub-region to the exit is obtained.

7. The integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins as described in claim 2, characterized in that, The step of using the traffic adaptation degree to filter the optimal adaptation exit of each sub-region includes: taking the exit corresponding to the maximum value of the traffic adaptation degree from the candidate sub-region to all exits as the optimal adaptation exit of the candidate sub-region.

8. The integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins as described in claim 2, characterized in that, The evaluation of the overall risk before and after the implementation of the crowd guidance strategy includes: The total global risk before the implementation of the crowd flow guidance strategy is obtained by averaging the risk quantification values ​​of all sub-regions at all times in the current period before the implementation of the guidance strategy. The total global risk after the implementation of the crowd flow guidance strategy is obtained by averaging the risk quantification values ​​of all sub-areas within a preset time period after the current time period after the implementation of the guidance strategy.

9. The integrated platform for decision simulation and evaluation of infectious disease prevention and control based on digital twins as described in claim 3, characterized in that, Obtaining the population density of the candidate sub-region includes: The ratio of the number of people in a candidate sub-region to the area of ​​the candidate sub-region is used as the population density of the candidate sub-region.

10. The integrated platform for decision simulation and evaluation of the entire process of infectious disease prevention and control based on digital twins as described in claim 1, characterized in that, The effectiveness of the strategy guidance based on the overall risk assessment before and after implementation includes: The difference between the total global risk before implementation and the total global risk after implementation is taken as the risk reduction amount, and the relative risk change rate is obtained based on the risk reduction amount and the total global risk before implementation. The effectiveness of the guidance strategy is determined by comparing the relative rate of change of risk with a preset rate of change threshold.

Images