Deep learning-based venue flow control methods, systems, and readable media

Abstract

This invention discloses a deep learning-based method, system, and readable medium for venue traffic control. By acquiring personnel data, venue data, and office data, it determines corresponding spatial, mobility, and health factors based on personnel data, and constructs a radar chart based on these factors to determine personnel health status scores. Using historical venue traffic data, a trained LSTNet model is used to predict future traffic data for corresponding moments, and this prediction is combined with actual traffic data to determine the venue's load status. Based on personnel health status scores, venue load status, and office data, the method determines the personnel's access rights to the venue. Finally, based on personnel health status scores and venue load status, a queuing theory model is used to calculate the average waiting time for personnel to enter the venue. This method offers finer granularity, provides information on waiting times, avoids overcrowding, and facilitates the management of crowd flow within venues.

Description

Technical Field

[0001] This invention relates to the field of deep learning, and specifically to a deep learning-based method, system, and readable medium for venue traffic control. Background Technology

[0002] As the virus spreads globally, its transmission methods are becoming increasingly diverse and rapid, with many regions evolving from point-to-point transmission to cluster infections. Extensive resources have been invested in large-scale contact tracing, which often reveals that many current infection events originate from overly dense gatherings within specific locations, causing the virus to spread from isolated points to wider outbreaks. Therefore, it is crucial to record visitor information when people enter a location to monitor foot traffic, ensuring the rational allocation of limited epidemic prevention resources. Furthermore, tracing an infected individual's movements allows for identification of their visited locations and potential close contacts, enabling timely intervention to cut off possible sources of infection. Thus, achieving real-time monitoring of personnel flow within locations, rational allocation of spatial resources, and efficient contact tracing have become critical challenges.

[0003] However, the currently widely used methods still have some shortcomings. For example, in many places, visitors sign in, which can easily lead to people staying in one place, sharing pens, posing significant security risks, and hindering the monitoring of venue traffic. Furthermore, since each venue maintains its own sign-in book, subsequent contact tracing consumes a large amount of manpower and resources, is time-consuming, and prone to errors. In some places, visitors need to download a dedicated app to obtain a health code, and only by displaying a green health code can they gain access to the venue. This method also has drawbacks, namely, visitor entry events cannot be recorded, making venue traffic monitoring impossible, lacking a reference for the allocation of epidemic prevention resources, and making it impossible to trace people's travel trajectories. In other places, visitors record their sign-in information by swiping cards or using facial recognition on access control systems. These sign-in records can be used to monitor venue traffic and trace travel. However, this method requires venues to install access control equipment in advance, which is costly. Therefore, a system that can provide efficient traffic monitoring, rational allocation of spatial resources, and risk tracing functions at low cost is needed. Summary of the Invention

[0004] In view of the aforementioned technical problems, the purpose of the embodiments of this application is to propose a deep learning-based venue traffic control method, system, and readable medium to solve the technical problems mentioned in the background section.

[0005] In a first aspect, embodiments of this application provide a venue traffic control method based on deep learning, comprising the following steps:

[0006] S1: Acquire personnel data, location data, and office data; determine corresponding spatial factors, mobility factors, and health factors based on personnel data; construct a radar chart based on spatial factors, mobility factors, and health factors; and determine the health status score of personnel based on the radar chart.

[0007] S2, based on the historical traffic data of the venue, the trained LSTNet model is used to predict the traffic data for the corresponding future time, and the actual traffic data is combined to determine the load status of the venue.

[0008] S3 determines the person's access rights to the premises based on their health status score, the site's load status, and office data.

[0009] S4 calculates the average waiting time for a person to enter the venue using a queuing theory model, based on the person's health status score and the venue's load status.

[0010] Preferably, step S1 further includes:

[0011] The Neo4j graph database is used to store personnel data, location data, and office data, and a relationship graph between personnel and locations is constructed.

[0012] Preferably, the indicators in the radar chart include spatial factors, movement factors, and health factors corresponding to the person. Each indicator in the radar chart has a corresponding score. The score for spatial factors includes the score corresponding to the risk level of the geographical space where the person lives, works, or studies. The score for movement factors includes the score corresponding to the risk level of the person's recent movement trajectory. The score for health factors includes the score corresponding to the person's physical health status and contact status.

[0013] As a preferred option, the formula for calculating the health status score is as follows:

[0014]

[0015] Where n is the number of indicators, S i％n and S represents the score of the i-th indicator and the maximum score of that indicator for person u, respectively. (i+1)％n and These represent the score of the (i+1)th indicator and the maximum score of that indicator, respectively, with % representing the remainder symbol.

[0016] As a preferred option, the formula for calculating the load condition is as follows:

[0017]

[0018] Where T and T' represent the current actual traffic data of location v and the traffic data predicted by the trained LSTNet model, respectively.

[0019] Preferably, step S3 specifically includes:

[0020] The health status score of the person is compared with the first threshold. Based on the comparison result, it is determined whether the person can enter the place. If the health status score is less than the first threshold, the person cannot enter the place. If the health status score is greater than the first threshold, the person can enter the place.

[0021] The load condition of the location is compared with the second threshold. Based on the comparison result, it is determined whether the location is overloaded. If the load condition is greater than the second threshold, the location is overloaded and personnel cannot enter the location. If the load condition is less than the second threshold, the location is not overloaded and personnel can enter the location.

[0022] In response to determining that the person's health status score is less than a first threshold and the venue's load status is less than a second threshold, the system determines whether the person has permission to enter the venue based on the permission relationship between the person and the venue in the office data. The permission relationship includes venue whitelist management and venue reservation application management.

[0023] Preferably, step S4 specifically includes:

[0024] Prioritization is determined based on the individual's health status score, using the following formula:

[0025]

[0026] The available capacity of a location in the queuing theory model is determined based on the location's load conditions, using the following formula:

[0027] s=ω(v)·N c ;

[0028] Where Nc represents the maximum capacity of the venue;

[0029] Based on the available capacity of the venue, the average waiting time in the queue for people with priority k (k = 1, 2, ..., m) is calculated as W. k :

[0030]

[0031] in,

[0032]

[0033] λ i =(n i / n)λ;

[0034] μ = 1 / T;

[0035] Where λ represents the arrival rate of all priority personnel, λ i Let n represent the arrival rate of people with priority i following a Poisson distribution, where n is the number of people arriving at the same location. i Let i represent the number of people with priority i, n represent the total number of people, μ represent the service rate of the venue per hour, and T represent the average time each person spends in the venue, which follows an exponential distribution.

[0036] Secondly, embodiments of this application provide a deep learning-based venue traffic control system, comprising:

[0037] The health status assessment module is configured to acquire personnel data, location data, and office data; determine corresponding spatial factors, mobility factors, and health factors based on personnel data; construct a radar chart based on spatial factors, mobility factors, and health factors; and determine the personnel's health status score based on the radar chart.

[0038] The load assessment module is configured to use a trained LSTNet model to predict the traffic data for the corresponding future time based on the historical traffic data of the venue, and combine it with the actual traffic data to determine the load status of the venue.

[0039] The access control module is configured to determine a person's access rights to the premises based on their health status score, the venue's load status, and office data.

[0040] The queuing time calculation module is configured to calculate the average waiting time for a person to enter the venue based on the person's health status score and the venue's load status using a queuing theory model.

[0041] Thirdly, embodiments of this application provide an electronic device including one or more processors; and a storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the method described in any implementation of the first aspect.

[0042] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method as described in any of the implementations of the first aspect.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] (1) The deep learning-based venue flow control method of the present invention can evaluate the health status of personnel in multiple dimensions, has finer granularity, and monitor the load status of the venue in real time. It determines the final permissions by comprehensively considering the health status of personnel, the load status of the venue, and the permission relationship between personnel and the venue, provides waiting time for personnel, avoids crowding, protects the venue from potential harm of the virus, and effectively improves the efficiency of epidemic prevention and control.

[0045] (2) The deep learning-based venue traffic control method of the present invention provides strong support for the management of venue traffic and personnel in a highly efficient and low-cost manner.

[0046] (3) The deep learning-based venue traffic control method of the present invention can provide efficient venue traffic monitoring, reasonable scheduling of spatial resources and risk tracing functions at low cost. When the user scans the QR code generated by the system for the venue, the health status of the personnel and the load status of the venue can be comprehensively evaluated, providing technical means for dynamic regulation of spatial resources in the venue and fine-grained tracing management of close contact flow. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0048] Figure 1 This is an exemplary device architecture diagram in which an embodiment of this application can be applied;

[0049] Figure 2 This is a flowchart illustrating a deep learning-based venue traffic control method according to an embodiment of the present invention.

[0050] Figure 3 This is a system block diagram of a deep learning-based venue traffic control method according to an embodiment of the present invention.

[0051] Figure 4 This is a schematic diagram illustrating the multi-source information and user profile of a deep learning-based venue traffic control method according to an embodiment of the present invention.

[0052] Figure 5 This is a schematic diagram of the LSTNet model of the deep learning-based venue traffic control method according to an embodiment of the present invention.

[0053] Figure 6 This is a schematic diagram of abnormal traffic detection in a deep learning-based venue traffic control method according to an embodiment of the present invention.

[0054] Figure 7 This is a schematic diagram of the access control rule method of the venue flow control method based on deep learning, as an embodiment of the present invention.

[0055] Figure 8 This is a schematic diagram of a risk tracing visualization management interface for a deep learning-based venue traffic control method according to an embodiment of the present invention.

[0056] Figure 9 This is a schematic diagram of a deep learning-based venue traffic control system according to an embodiment of the present invention;

[0057] Figure 10 This is a schematic diagram of the structure of a computer device suitable for implementing the electronic devices of the present application. Detailed Implementation

[0058] 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. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0059] Figure 1 An exemplary device architecture 100 is shown in which the deep learning-based venue traffic control method or the deep learning-based venue traffic control system of the present application embodiments can be applied.

[0060] like Figure 1 As shown, the device architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the terminal devices 101, 102, and 103 and the server 105. The network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0061] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various applications, such as data processing applications and file processing applications, can be installed on terminal devices 101, 102, and 103.

[0062] Terminal devices 101, 102, and 103 can be either hardware or software. When terminal devices 101, 102, and 103 are hardware, they can be various electronic devices, including but not limited to smartphones, tablets, laptops, and desktop computers. When terminal devices 101, 102, and 103 are software, they can be installed in the electronic devices listed above. They can be implemented as multiple software programs or software modules (e.g., software programs or software modules used to provide distributed services) or as a single software program or software module. No specific limitations are imposed here.

[0063] Server 105 can be a server that provides various services, such as a background data processing server that processes files or data uploaded by terminal devices 101, 102, and 103. The background data processing server can process the acquired files or data and generate processing results.

[0064] It should be noted that the deep learning-based venue traffic control method provided in this application embodiment can be executed by server 105 or by terminal devices 101, 102, and 103. Correspondingly, the deep learning-based venue traffic control system can be set in server 105 or in terminal devices 101, 102, and 103.

[0065] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Any number of terminal devices, networks, and servers can be included depending on implementation needs. If the data being processed does not need to be retrieved remotely, the above architecture may not include a network, requiring only servers or terminal devices.

[0066] Figure 2 An embodiment of this application illustrates a deep learning-based venue traffic control method, comprising the following steps:

[0067] S1: Acquire personnel data, location data, and office data; determine corresponding spatial factors, mobility factors, and health factors based on personnel data; construct a radar chart based on spatial factors, mobility factors, and health factors; and determine the health status score of personnel based on the radar chart.

[0068] In a specific embodiment, step S1 further includes:

[0069] The Neo4j graph database is used to store personnel data, location data, and office data, and a relationship graph between personnel and locations is constructed.

[0070] For details, please refer to Figure 3The data layer, responsible for providing various databases and basic data, consists of a personnel database, a location database, and an office database. It utilizes the Neo4j graph database for storage, comprising the personnel database, location database, and office database. Neo4j is a database based on graph data structures for data storage and retrieval. Its basic storage units are nodes and edges, and attributes can be added to nodes and edges to store business data. Compared to traditional relational databases, it has a natural performance advantage in relational queries of complex data models. The personnel database stores personnel data, mainly including basic information such as name and employee ID; the location database stores location data, consisting of location name and geographical location; and the office database stores office data, providing fine-grained management for location whitelists and location reservation applications.

[0071] refer to Figure 3 At the logical layer, the system evaluates personnel health status, estimates venue load, and determines permission relationships to provide a reference for spatial resource allocation. It also constructs a relationship graph between personnel and venues based on the Neo4j graph database. This part is responsible for analyzing and processing data from the data layer and consists of three parts: personnel health status evaluation methods, venue load estimation methods, and access permission rule management methods. To evaluate personnel health status, the system integrates and analyzes multi-source personnel information, including visitor identity, address, and trajectory information, to comprehensively evaluate their health status. To estimate venue load, an LSTNet model is constructed based on multi-source heterogeneous mobile data to predict venue flow patterns, assess venue load pressure, and achieve real-time monitoring and early warning of abnormal flow. Based on the novel Neo4j graph database technology, combined with a venue whitelist and venue reservation application mechanism, fine-grained traceability management of close contacts moving between venues is achieved. Personnel on the whitelist have the right to enter venues, while those not on the whitelist can obtain entry permissions by submitting a reservation application.

[0072] In a specific embodiment, the indicators in the radar chart include spatial factors, movement factors, and health factors corresponding to the person. Each indicator in the radar chart has a corresponding score. The score for spatial factors includes the score corresponding to the risk level of the geographical space where the person lives, works, or studies. The score for movement factors includes the score corresponding to the risk level of the person's recent movement trajectory. The score for health factors includes the score corresponding to the person's physical health status and contact status.

[0073] For details, please refer to Figure 4The system acquires and evaluates an individual's health status through access restrictions and other means. The data is categorized into three factors: spatial factors, mobility factors, and health factors. Spatial factors refer to the geographical location of the individual's residence, work, or study, including whether they live in a high-risk community or work / study in a high-risk area. Mobility factors refer to the individual's recent travel history, including whether they have recently traveled to high-risk countries or regions and the modes of transportation they have used. Health factors refer to the individual's health condition, including whether they have any suspected symptoms related to the virus and whether they have been in close contact with confirmed cases. A dynamically adjustable scoring system assigns different scores to each indicator to evaluate the individual's health status. For example, in the early stages of the epidemic, due to the strictest safety controls, a person with a cough would receive a relatively low score. However, as the epidemic eases, if there are no new cases in the area for an extended period, the corresponding score will relatively increase, as coughs caused by colds or other reasons are common. Finally, radar charts are used to fuse and analyze these features, forming a user health profile to evaluate the individual's health status. The six vertices of the radar chart represent spatial factors, movement factors, and health factors. The value of each indicator is determined by its corresponding score. The higher the score of each indicator, the larger the area they form on the radar chart. For an individual, the ratio of the area formed by each indicator on the radar chart to the maximum area of ​​the radar chart is used as their health status score. A higher health status score indicates a worse health status for the individual.

[0074] In a specific embodiment, the formula for calculating the health status score is as follows:

[0075]

[0076] Where n is the number of indicators, S i％n and S represents the score of the i-th indicator and the maximum score of that indicator for person u, respectively. (i+1)％n and Let represent the score of the (i+1)th indicator and the maximum score of that indicator, respectively, with % representing the modulo operator. The health status score in the above formula is obtained by the ratio of the area formed by the radar chart to the maximum area of ​​the radar chart. The area of ​​the radar chart can be seen as the sum of many triangles, and the area of ​​each triangle is _____.

[0077] S2, based on the historical traffic data of the venue, uses a trained LSTNet model to predict the traffic data for the corresponding future time, and combines it with the actual traffic data to determine the load status of the venue.

[0078] Specifically, the LSTNet model is trained on historical traffic data for a given location to obtain a trained LSTNet model suitable for traffic prediction. This model monitors the location's load in real time and promptly detects abnormal traffic. The input to the LSTNet model is historical traffic data for a corresponding time period, and the output is traffic data for the corresponding future time period. Figure 5 As shown, the LSTNet model is a deep learning neural network used to learn and predict time series data that mixes long and short-term patterns. It consists of a Convolutional Neural Network (CNN), a Recurrent Neural Network (RNN), and an Autoregressive Neural Network (AR). The RNN part includes recurrent component components and recurrent jump component components. The CNN part aims to extract dependency patterns from short-term, local time series data, and its output is fed into both the recurrent component components and the recurrent jump component components. The recurrent jump component components are used to capture dependency patterns from complex long-term time series data. For example, since daily traffic data for a venue shows clear patterns, to predict Wednesday's traffic, the traffic on Wednesdays of each week in the historical period is used to predict the traffic on a future Wednesday. The recurrent jump component components can capture long-term dependency patterns by adding jump links between hidden units in the same stage of adjacent periods and the current hidden unit. Then, the outputs of the recurrent component components and the recurrent jump component components are combined into the output of the recurrent neural network part. Furthermore, the autoregressive neural network part is used as a linear component to address the scaling problem of local time series data. Finally, by combining the outputs of the recurrent neural network and the autoregressive neural network, the prediction results of the LSTNet model are obtained. The ratio of the actual current traffic flow at the location to the traffic flow data predicted by the LSTNet model is used as the load status of the location.

[0079] In a specific embodiment, the formula for calculating the load condition is as follows:

[0080]

[0081] Here, T and T' represent the current actual traffic data of venue v and the traffic data predicted by the trained LSTNet model, respectively. A second threshold is set: when the venue's load exceeds the second threshold, it indicates that the venue is under high load, and the epidemic prevention resources within the venue may be strained, suggesting that people should not enter. When the venue's load is below the second threshold, it indicates that the venue is under low load, meaning that the epidemic prevention resources within the venue are sufficient, and people can be allowed to enter.

[0082] like Figure 6As shown, when the second threshold is set to 1.3, it means that if the current actual traffic exceeds 30% of the model's predicted traffic, it will be considered an abnormal traffic point. The detection of 9:00 AM on June 18th suggests a possible abnormal traffic point, as the actual traffic and the model's predicted traffic at that time were 195 and 140 respectively, with a load factor of [missing information]. It exceeded the second threshold.

[0083] S3 determines a person's access rights to the premises based on their health status score, the site's load status, and office data.

[0084] In a specific embodiment, step S3 specifically includes:

[0085] The health status score of the person is compared with the first threshold. Based on the comparison result, it is determined whether the person can enter the place. If the health status score is less than the first threshold, the person cannot enter the place. If the health status score is greater than the first threshold, the person can enter the place.

[0086] The load condition of the location is compared with the second threshold. Based on the comparison result, it is determined whether the location is overloaded. If the load condition is greater than the second threshold, the location is overloaded and personnel cannot enter the location. If the load condition is less than the second threshold, the location is not overloaded and personnel can enter the location.

[0087] In response to determining that the person's health status score is less than a first threshold and the venue's load status is less than a second threshold, the system determines whether the person has permission to enter the venue based on the permission relationship between the person and the venue in the office data. The permission relationship includes venue whitelist management and venue reservation application management.

[0088] Specifically, the access control rule management method provides a fine-grained control scheme for the permission relationship between personnel and venues. For example... Figure 7 As shown, after a person scans the venue code, their health status is first assessed based on a health status score. If the score is below a first threshold, the person is considered to be at risk and entry is prohibited. Next, the venue's load is assessed based on actual traffic volume and traffic data predicted by a trained LSTNet model. If the venue's load exceeds a second threshold, it is considered overloaded, and entry is discouraged. When a person is healthy and the venue's load is normal, a final assessment is made based on the permission relationship between the person and the venue. If the person has the necessary permissions, entry is permitted. This system consists of two parts: venue whitelist management and venue reservation application management. Administrators can dynamically manage the venue whitelist by uploading files containing personnel information. Personnel on the whitelist have the right to enter. For personnel not on the whitelist, temporary access can be obtained by submitting a reservation application, which includes the requested time period and reason.

[0089] S4 calculates the average waiting time for a person to enter the venue using a queuing theory model, based on the person's health status score and the venue's load status.

[0090] For details, please refer to Figure 3 The application layer provides services such as personnel check-in, venue traffic monitoring, space resource allocation, and risk tracing. This part is responsible for providing application services to users, consisting of client-side personnel check-in services and management-side traffic monitoring, reasonable allocation of space resources, and risk tracing functions. To protect users' personal privacy and security, the system integrates a central identity verification and authorization service (CAS) and develops a check-in application based on a mini-program. When personnel scan the venue's QR code, they refer to... Figure 7 The system sequentially employs methods for evaluating personnel health status, estimating venue load, and managing access control rules to provide a basis for deciding whether to recommend personnel to enter the venue. A red or yellow card indicates that entry is not recommended, while a green card indicates that entry is safe. Real-time monitoring of venue traffic is achieved based on check-in information. A dynamic allocation scheme for venue space resources such as seats and windows is constructed based on queuing theory models to achieve low-risk social distancing control. Visualization technology enables visualized management of risk tracing, providing administrators with intuitive decision support. By tracing the movements of at-risk individuals, managers can mark venues visited by these individuals as potentially high-risk venues. Furthermore, individuals who repeatedly check in at the same venues as at-risk individuals within a short period will be marked as potential close contacts.

[0091] In a specific embodiment, step S4 specifically includes:

[0092] Prioritization is determined based on the individual's health status score, with lower k values ​​corresponding to higher priorities, as shown in the following formula:

[0093]

[0094] The available capacity of a location in the queuing theory model is determined based on the location's load conditions, using the following formula:

[0095] s=ω(v)·N c ;

[0096] Where Nc represents the maximum capacity of the venue;

[0097] Based on the available capacity of the venue, the average waiting time in the queue for people with priority k (k = 1, 2, ..., m) is calculated as W. k :

[0098]

[0099] in,

[0100]

[0101] λ i =(n i / n)λ;

[0102] μ = 1 / T;

[0103] Where λ represents the arrival rate of all priority personnel, λ i Let n represent the arrival rate of people with priority i following a Poisson distribution, where n is the number of people arriving at the same location. i Let represent the number of people with priority i, n represent the total number of people, and the service rate μ refers to the number of people the venue can serve per hour, which can be determined by the venue's resources, such as area and number of seats, and can be calculated as 1 / T. T represents the average stay time of each person in the venue, which follows an exponential distribution. In practice, it is assumed that people arriving at the venue form a queue, and the dynamically changing average waiting time W k By allocating personnel in the queue, an appropriate time is provided for each person to enter the venue, thereby keeping the number of people in the venue at a low level and reducing the risk of infection, thus achieving a reasonable allocation of space resources.

[0104] like Figure 8 As shown, the visual management interface for risk tracing consists of two parts. The first part is the query bar at the top, where administrators can specify query criteria to search for locations visited by at-risk individuals within a specified time period, as well as other individuals who checked in at the same locations as the at-risk individuals within the specified time interval. The second part is the search results section. On the right is a visual canvas for risk tracing based on Neo4jd3 technology, a powerful JavaScript library for graphical visualization based on D3.js. The canvas contains red nodes representing at-risk individuals, blue nodes representing locations, and green nodes representing other individuals who checked in. These nodes are connected by directed edges to form a relationship graph. When the mouse hovers over a node or edge, the top left corner of the canvas displays the detailed attributes of that node or edge. The left side contains a list of potential high-risk locations and close contacts, sorted in descending order by the frequency of locations visited by at-risk individuals and by the number of times potential close contacts appeared.

[0105] Further reference Figure 9 As an implementation of the methods shown in the above figures, this application provides an embodiment of a deep learning-based venue traffic control system, which is similar to... Figure 2 Corresponding to the method embodiments shown, the system can be specifically applied to various electronic devices.

[0106] This application provides a deep learning-based venue traffic control system, including:

[0107] The health status assessment module 1 is configured to acquire personnel data, location data, and office data; determine corresponding spatial factors, mobility factors, and health factors based on personnel data; construct a radar chart based on spatial factors, mobility factors, and health factors; and determine the personnel's health status score based on the radar chart.

[0108] Load status assessment module 2 is configured to predict future traffic data based on historical traffic data of the venue using a trained LSTNet model, and combine this with actual traffic data to determine the load status of the venue.

[0109] The permission determination module 3 is configured to determine a person's access permission to the premises based on their health status score, the premises' load status, and office data.

[0110] The queuing time calculation module 4 is configured to calculate the average waiting time for a person to enter the venue based on the person's health status score and the venue's load status using a queuing theory model.

[0111] The following is for reference. Figure 10 It illustrates an electronic device suitable for implementing embodiments of this application (e.g., Figure 1 A schematic diagram of the structure of a computer device 1000 (shown as a server or terminal device). Figure 10 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0112] like Figure 10 As shown, the computer device 1000 includes a central processing unit (CPU) 1001 and a graphics processing unit (GPU) 1002, which can perform various appropriate actions and processes according to programs stored in read-only memory (ROM) 1003 or programs loaded from storage section 1009 into random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the device 1000. The CPU 1001, GPU 1002, ROM 1003, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus 1005.

[0113] The following components are connected to I / O interface 1006: an input section 1007 including a keyboard, mouse, etc.; an output section 1008 including an LCD, speakers, etc.; a storage section 1009 including a hard disk, etc.; and a communication section 1010 including a network interface card, such as a LAN card, modem, etc. The communication section 1010 performs communication processing via a network such as the Internet. A drive 1011 may also be connected to I / O interface 1006 as needed. A removable medium 1012, such as a hard disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 1011 as needed so that computer programs read from it can be installed into storage section 1009 as needed.

[0114] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 1010, and / or installed from removable medium 1012. When the computer program is executed by central processing unit (CPU) 1001 and graphics processing unit (GPU) 1002, the functions defined in the methods of this application are performed.

[0115] It should be noted that the computer-readable medium described in this application can be a computer-readable signal medium, a computer-readable medium, or any combination thereof. A computer-readable medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor device, or any combination thereof. More specific examples of a computer-readable medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution device, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than a computer-readable medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution device, apparatus, or apparatus. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0116] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0117] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using dedicated hardware-based means to perform the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0118] The modules described in the embodiments of this application can be implemented in software or hardware. These modules can also be located within a processor.

[0119] In another aspect, this application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: acquire personnel data, location data, and office data; determine corresponding spatial factors, mobility factors, and health factors based on the personnel data; construct a radar chart based on the spatial factors, mobility factors, and health factors; determine the personnel's health status score based on the radar chart; predict future traffic data at corresponding future times using a trained LSTNet model based on historical traffic data from the location data; and determine the location's load status by combining this with actual traffic data; determine the personnel's access rights to the location based on their health status score, the location's load status, and office data; and calculate the average queuing time for the personnel to enter the location using a queuing theory model based on their health status score and the location's load status.

[0120] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A deep learning-based place traffic control method, characterized by, The method comprises the following steps: S1, obtaining personnel data, site data and office data, determining corresponding space factors, movement factors and health factors according to the personnel data, and constructing a radar chart based on the space factors, movement factors and health factors, determining the health status score of the personnel according to the radar chart, and the calculation formula of the health status score is as follows: ； wherein, , is the number of indicators, and denote the score of the u-th indicator of person u and the maximum score of this indicator, respectively, and denote the score of the u-th indicator of person u and the maximum score of this indicator, respectively, and denote the score of the u-th indicator of person u and the maximum score of this indicator, respectively, is the modulo symbol; S2, predicting the traffic data of the corresponding time in the future by using the trained LSTNet model according to the historical traffic data of the site data, and determining the load status of the site combined with the actual traffic data, and the calculation formula of the load status is as follows: ； wherein T and T' represent the current actual traffic data and the traffic data predicted by the trained LSTNet model of the place, respectively ; S3, determining the permission of the personnel to enter the site according to the health status score of the personnel, the load status of the site and the office data; S4, calculating the queuing average waiting time of the personnel to enter the site according to the health status score of the personnel and the load status of the site through a queuing theory model, which specifically comprises: determining the priority of the personnel according to the health status score of the personnel, and the formula is as follows: ； determining the available capacity of the site in the queuing theory model according to the load status of the site, and the formula is as follows: ； Wherein, Nc represents the maximum capacity of the site; The average waiting time for a person with priority level k in the queue is calculated as : ； wherein ; λ 𝑖3 = (n' 𝑖3 / n') λ; μ=1 / T1; where k = 1, 2, …, m, λ represents the arrival rate of all priority personnel, λ 𝑖3 represents the arrival rate of personnel with priority i3, where n 𝑖3 represents the number of personnel with priority i3, n represents the total number of personnel, the service rate μ refers to the number of personnel that can be served by the venue per hour, and Ti represents the average time of stay in the venue for each person, which follows an exponential distribution. 2.The deep learning-based place traffic control method of claim 1, wherein, The step S1 further comprises: storing the personnel data, site data and office data by using a neo4j graph database, and constructing a relationship graph between the personnel and the site. 3.The deep learning-based place traffic control method of claim 1, wherein, The indexes in the radar chart include the space factors, movement factors and health factors corresponding to the personnel, each index in the radar chart has a corresponding score, wherein the score of the space factors includes the score corresponding to the risk level of the geographical space where the personnel lives, works or studies, the score of the movement factors includes the score corresponding to the risk level of the recent movement trajectory of the personnel, and the score of the health factors includes the score corresponding to the physical health status and contact situation of the personnel. 4.The deep learning-based place traffic control method of claim 1, wherein, The step S3 specifically comprises: comparing the health status score of the personnel with a first threshold value, and determining whether the personnel can enter the site according to the comparison result, if the health status score is less than the first threshold value, the personnel cannot enter the site, if the health status score is greater than the first threshold value, the personnel can enter the site; comparing the load status of the site with a second threshold value, and determining whether the site is overloaded according to the comparison result, if the load status is greater than the second threshold value, the site is in an overloaded state, and the personnel cannot enter the site, if the load status is less than the second threshold value, the site is in an unoverloaded state, and the personnel can enter the site; in response to determining that the health status score of the personnel is greater than the first threshold value and the load status of the site is less than the second threshold value, determining whether the personnel has the permission to enter the site according to the permission relationship between the personnel and the site in the office data, and the permission relationship includes site whitelist management and site reservation application management.

5. A deep learning based place traffic control system, characterized by, It comprises: The health status evaluation module is configured to obtain personnel data, site data and office data, determine corresponding space factors, movement factors and health factors according to the personnel data, construct a radar chart based on the space factors, movement factors and health factors, determine a health status score of the personnel according to the radar chart, and the calculation formula of the health status score is as follows: ； in, , For the number of indicators, and Representing the first and second persons of person u respectively The scores of each indicator and the maximum score of that indicator. and They represent the first +1 score of indicator and the maximum score of that indicator, % is the modulo operator; The load condition evaluation module is configured to predict future traffic data at corresponding time according to historical traffic data of the site data by using a trained LSTNet model, and determine the load condition of the site in combination with actual traffic data, and the calculation formula of the load condition is as follows: ； wherein T and T' represent the current actual traffic data and the traffic data predicted by the trained LSTNet model of the site, respectively ; The permission determination module is configured to determine the permission of the personnel to enter the site according to the health status score of the personnel, the load condition of the site and the office data. The queuing time calculation module is configured to calculate the average queuing waiting time of the personnel to enter the site by using a queuing theory model according to the health status score of the personnel and the load condition of the site, and specifically includes: determining the priority of the personnel according to the health status score of the personnel, and the formula is as follows: ； determining the available capacity of the site in the queuing theory model according to the load condition of the site, and the formula is as follows: ； wherein, Nc represents the maximum capacity of the site; The average waiting time for a person with priority level k in the queue is calculated as : ； wherein ; λ 𝑖3 = (n' 𝑖3 / n') λ; μ = 1 / T1; where k = 1, 2, …, m, λ represents the arrival rate of all priority personnel, λ 𝑖3 represents the arrival rate of personnel with priority i3, where n 𝑖3 represents the number of personnel with priority i3, n represents the total number of personnel, the service rate μ refers to the number of personnel that can be served by the venue per hour, and Ti represents the average time of stay in the venue for each person, which follows an exponential distribution.

6. An electronic device, comprising: one or more processors; a storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1-4.

7. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the method of any one of claims 1-4.

Images