Electronic building identification based active community facility guide orientation method and system

Abstract

The application provides an activity community facility guidance method and system based on electronic building identification, and belongs to the technical field of intelligent community operation management. The method comprises the following steps: inputting initial real-time flow state and community physical space layout data into a facility heat evaluation model, calculating the heat level and space heat distribution of the facility; based on the heat level, space heat distribution and preset target, generating an electronic identification information collection strategy under the constraints of community physical space and network topology; collecting data based on the electronic identification information collection strategy to obtain initial facility data assets; obtaining the community historical real-time flow state corresponding to the initial facility data asset collection time, performing data cleaning and verification on the initial facility data assets through an anomaly detection algorithm to obtain effective facility data assets; and inputting the effective facility data assets into a path planning decision model to generate a facility guidance scheme within the boundary of community operation constraints.

Description

Technical Field

[0001] This application relates to the field of smart community operation and management technology, and in particular to a method and system for guiding and directing community facilities based on electronic building identification. Background Technology

[0002] As the construction of smart communities continues to deepen, the scale of urban communities and supporting facilities are constantly expanding, and residents' demand for facilities exhibits tidal characteristics and sudden fluctuations. While community infrastructure query and navigation services are now widely available, precise and efficient facility guidance matching real-time on-site conditions, improved facility utilization, and refined management of operation and maintenance costs have become core high-level needs that urgently need to be met in the field of refined smart community operation. The current industry standard solution involves deploying electronic building identification terminals at fixed locations within the community, with basic facility information pre-stored in the background. When a resident initiates a query, the system generates and displays a directional path based on the shortest spatial path. Property management manually inspects and monitors the facilities at fixed locations, updating the background information manually at fixed intervals to provide residents with basic guidance services. However, this solution suffers from significant discrepancies between the guidance and the actual on-site situation when there is a large influx of people or sudden changes in facility status, failing to meet high-level experience requirements.

[0003] Therefore, there is an urgent need for a method and system for guiding community facilities based on electronic building signage, in order to improve the real-time matching degree between community facility guidance schemes and actual site conditions. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides a method and system for guiding and directing community facilities based on electronic building signage.

[0005] A first aspect of this application provides a method for guiding and directing activity community facilities based on electronic building identification, comprising: The initial real-time pedestrian flow status within the community is obtained through a preset basic collection cycle; Input the initial real-time pedestrian flow status and community physical space layout data into the facility heat assessment model to calculate the facility heat level and spatial heat distribution; Based on heat level and spatial heat distribution, with the goal of maximizing data collection coverage and minimizing collection latency and construction costs, an electronic identification information collection strategy is generated under the constraints of community physical space and network topology. Data is collected based on electronic identification information collection strategies to obtain initial facility data assets; The community's historical real-time pedestrian flow status corresponding to the initial facility data asset collection time is obtained. The initial facility data assets are cleaned and verified through anomaly detection algorithms to obtain valid facility data assets. By inputting effective facility data assets into the path planning decision model, and solving a multi-objective optimization problem within the boundaries of community operation constraints, a facility guidance and orientation scheme is generated. Based on preset data asset utility evaluation indicators, the guidance effect of the facility guidance scheme is evaluated, and attribution analysis is used to identify the data feature dimensions that contribute the most to the generation of the facility guidance scheme. Based on the results of the guidance effect evaluation and the dimensions of data characteristics, the configuration parameters of the electronic identification information collection strategy and the internal parameters of the path planning decision model are adjusted using a pre-set feedback learning mechanism.

[0006] A second aspect of this application provides a guidance system for active community facilities based on electronic building identification, comprising: The data acquisition module is used to obtain the initial real-time pedestrian flow status within the community through a preset basic collection cycle; The heat assessment module is used to input initial real-time pedestrian flow status and community physical space layout data into the facility heat assessment model to calculate the facility's heat level and spatial heat distribution. The strategy generation module is used to generate electronic identification information collection strategies based on heat level and spatial heat distribution, with the goal of maximizing data collection coverage and minimizing collection latency and construction costs, under the constraints of community physical space and network topology. The asset construction module is used to collect data based on the electronic identification information collection strategy to obtain initial facility data assets. The cleaning and verification module is used to obtain the real-time historical population flow status of the community corresponding to the time of initial facility data asset collection, and to clean and verify the initial facility data assets through anomaly detection algorithms to obtain valid facility data assets. The route planning module is used to input effective facility data assets into the route planning decision model, solve multi-objective optimization problems within the boundaries of community operation constraints, and generate facility guidance and orientation schemes. The evaluation and analysis module is used to evaluate the guidance effect of the facility guidance scheme based on the preset data asset utility evaluation indicators, and to use attribution analysis to identify the data feature dimensions that contribute the most to the generation of the facility guidance scheme. The feedback adjustment module is used to adjust the configuration parameters of the electronic identification information collection strategy and the internal parameters of the path planning decision model based on the results and data feature dimensions of the guidance effect evaluation, using a preset feedback learning mechanism.

[0007] A third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the above-described activity community facility guidance method based on electronic building identification.

[0008] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method for guiding and directing activity community facilities based on electronic building identification.

[0009] The beneficial effects of the community facility guidance method and system based on electronic building identification provided in this application are as follows: This application can improve the real-time matching degree between the community facility guidance scheme and the actual on-site status, while strictly controlling the system's construction and operation costs, in scenarios where community pedestrian flow and facility usage status fluctuate dynamically. On the one hand, through real-time pedestrian flow collection, facility popularity assessment, and abnormal data verification, the timeliness and accuracy of the data on which the guidance scheme is based are ensured, effectively avoiding the problem of deviation between the guidance path and the on-site pedestrian flow and facility status, significantly improving residents' facility usage experience and passage efficiency, and simultaneously increasing the overall utilization rate of community activity facilities. On the other hand, through a multi-objective optimized dynamic collection strategy and a closed-loop feedback learning mechanism, dynamic adaptation and adjustment of collection and model parameters are achieved, avoiding resource idleness and ineffective cost expenditures in fixed operation modes. While ensuring the quality of guidance services, construction and operation costs are controlled, ensuring that the scheme adapts to the dynamic operation scenarios of the community in the long term. Attached Figure Description

[0010] Figure 1 A flowchart illustrating a method for guiding and directing activity community facilities based on electronic building signage, provided as an embodiment of this application; Figure 2 A structural block diagram of an activity community facility guidance and wayfinding system based on electronic building signage provided in an embodiment of this application; Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0011] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0012] To make the purpose, technical solution, and advantages of this application clearer, the following will be described in conjunction with the appendix. Figures 1-3 The following is an explanation using specific examples.

[0013] Please refer to Figure 1 , Figure 1A flowchart illustrating a method for guiding and directing activity community facilities based on electronic building signage, as provided in an embodiment of this application, includes: S101: Obtain the initial real-time pedestrian flow status within the community through a preset basic collection cycle.

[0014] In this embodiment, electronic building identification refers to an intelligent building identification device that integrates a pedestrian flow sensing unit, a data processing unit, and a communication unit. Its function is to sense the dynamic flow of people within the community in real time, process the collected data, and transmit the data to the backend system via the communication unit; it is the basic hardware for data collection.

[0015] This embodiment can employ infrared sensors, pressure sensors, or simple counters deployed near community entrances, main passageways, or facilities to count pedestrian flow at fixed time intervals. These sensors are configured to detect only the presence or passage of people and record the corresponding count data. Alternatively, periodic observation and recording can be achieved through manual patrols or fixed surveillance cameras, with staff manually inputting the pedestrian flow data. Both methods can provide information on the number of people in a specific area of ​​the community at a specific time.

[0016] S102: Input the initial real-time pedestrian flow status and community physical space layout data into the facility heat assessment model to calculate the facility heat level and spatial heat distribution.

[0017] In this embodiment, the facility popularity assessment model refers to a set of mathematical models or algorithms used to comprehensively analyze the pedestrian flow and physical spatial layout of facilities within a community, thereby calculating the popularity level and spatial popularity distribution of each facility. This model aims to quantify the popularity or usage intensity of facilities and characterize their spatial distribution features.

[0018] Initial real-time pedestrian flow data can include information such as pedestrian traffic and crowd density in each area. Community physical space layout data can include the location, area, and connecting paths of facilities. The facility popularity assessment model can be a rule-based system; for example, by performing a simple weighted average of pedestrian traffic data from different areas and normalizing it by combining it with the facility's area, an initial popularity value for each facility can be obtained. Subsequently, these initial popularity values ​​are matched with preset popularity level ranges to determine the facility's popularity level. Simultaneously, by combining the geographical coordinates of each facility, a simple two-dimensional heat map can be generated to represent the spatial popularity distribution.

[0019] S103: Based on heat level and spatial heat distribution, with the goal of maximizing data collection coverage and minimizing collection latency and construction costs, an electronic identification information collection strategy is generated under the constraints of community physical space and network topology.

[0020] In this embodiment, the electronic identification information collection strategy refers to a series of rules and parameters formulated to guide the electronic identification device in data collection, including but not limited to collection frequency, data upload cycle, deployment location, and quantity. This strategy aims to optimize the efficiency and quality of data collection to meet the needs of subsequent analysis and decision-making. This embodiment can employ a greedy algorithm to determine the deployment location and quantity of electronic identification points. The algorithm first selects deployment points in areas with high heat levels, then gradually expands to areas with lower heat levels, while considering the coverage area of ​​each deployment point until a preset coverage target is reached. When selecting deployment points, community physical space constraints are considered, such as whether the power supply is sufficient and whether the installation location is permissible. Network topology constraints are also considered, such as the communication distance and bandwidth limitations between the electronic identification points and the data aggregation point, to ensure timely data transmission. The collection strategy may include setting a fixed data collection frequency and upload cycle for each electronic identification point.

[0021] S104: Data is collected based on the electronic identification information collection strategy to obtain initial facility data assets.

[0022] In this embodiment, initial facility data assets refer to the raw data set directly collected by the electronic identification device according to the collection strategy. This data may contain noise or anomalies and needs to be cleaned and verified before it can be used for subsequent decision-making.

[0023] In this embodiment, based on the electronic tag location, quantity, collection frequency, and upload cycle determined in the electronic tag information collection strategy, each electronic tag device is activated and begins operation. The pedestrian flow sensing unit continuously monitors pedestrian flow, the data processing unit performs preliminary processing on the raw sensing data, and the communication unit uploads the processed data to the central server according to a preset cycle. This uploaded data, such as pedestrian flow and crowd density recorded by each electronic tag point within a specific time period, constitutes the initial facility data assets.

[0024] S105: Obtain the real-time historical pedestrian flow status of the community corresponding to the initial facility data asset collection time, and perform data cleaning and verification on the initial facility data asset through anomaly detection algorithm to obtain valid facility data assets.

[0025] In this embodiment, effective facility data assets refer to a set of data that, after being cleaned and verified by an anomaly detection algorithm, is deemed accurate, reliable, and usable for route planning decisions. This embodiment compares the currently collected initial facility data assets with historical pedestrian flow data from the same time period. The anomaly detection algorithm can employ a simple threshold verification method, for example, setting an upper and lower limit for pedestrian flow; any data point exceeding this range is marked as anomaly. Alternatively, a statistical method can be used, such as calculating the Z-score of data points and identifying data points with a Z-score exceeding a certain preset value as anomalies. Data points marked as anomaly can be deleted, replaced with historical averages, or corrected using interpolation methods to obtain more reliable effective facility data assets.

[0026] S106: Input effective facility data assets into the path planning decision model, solve the multi-objective optimization problem within the boundaries of community operation constraints, and generate facility guidance and orientation schemes.

[0027] In this embodiment, the route planning decision model refers to an algorithm or system used to receive effective facility data assets and solve a multi-objective optimization problem within the boundaries of community operational constraints to generate the optimal facility guidance and orientation scheme. This model aims to provide residents with efficient, balanced, and realistic guidance routes. For example, the route planning decision model can employ the classic Dijkstra's algorithm or A* algorithm, with the shortest path as the primary objective for route planning. Effective facility data assets can provide real-time congestion levels or real-time availability status of facilities for each route segment. Community operational constraints may include the maximum capacity of each route, one-way traffic restrictions, and restricted areas. Under these constraints, the model calculates the optimal route from the starting point to the target facility. The multi-objective optimization problem can be simplified to satisfying the shortest path requirement while avoiding currently congested areas as much as possible.

[0028] A facility guidance scheme refers to specific guidance suggestions or route planning results output by a route planning decision model. This scheme aims to guide residents to their target facilities while optimizing pedestrian flow and facility utilization within the community.

[0029] S107: Based on the preset data asset utility evaluation indicators, evaluate the guidance effect of the facility guidance scheme, and use attribution analysis to identify the data feature dimensions that contribute the most to the generation of the facility guidance scheme.

[0030] In this embodiment, data asset utility assessment metrics refer to a series of quantitative standards used to measure the quality and value of data assets. These metrics can assess the completeness, accuracy, timeliness of the data, and its contribution to the generation of the final guidance and direction scheme.

[0031] Attribution analysis is a statistical or machine learning method used to identify and quantify the contribution of different data feature dimensions to a given outcome (such as the generation of facility guidance schemes). Through attribution analysis, it's possible to discover which data are the key factors influencing the effectiveness of the scheme. Data asset utility evaluation indicators can include the proportion of actual pedestrian traffic following the guidance path after the scheme is generated, and whether the guidance path effectively alleviates congestion. Attribution analysis can employ simple correlation analysis methods, such as calculating the correlation coefficient between different types of data (e.g., pedestrian flow data, facility status data) and the final guidance effect; feature dimensions with higher correlation coefficients are considered to have made greater contributions.

[0032] S108: Based on the results and data feature dimensions of the guidance effect evaluation, and using a preset feedback learning mechanism, adjust the configuration parameters of the electronic identification information collection strategy and the internal parameters of the path planning decision model.

[0033] In this embodiment, the feedback learning mechanism refers to an adaptive system that can automatically adjust the configuration parameters of the electronic identification information collection strategy and the internal parameters of the path planning decision model based on the results of the guidance effect evaluation and the dimensions of data characteristics. This feedback learning mechanism aims to continuously improve the performance of the guidance method through continuous learning and optimization.

[0034] As can be seen from the above, this application achieves real-time matching between the guidance and orientation scheme and the actual on-site conditions by real-time sensing of community pedestrian flow, dynamic assessment of facility popularity, optimization of data collection strategies, refined data cleaning, intelligent path planning, and closed-loop feedback learning. This effectively solves the problem of large guidance deviations in traditional schemes during pedestrian flow fluctuations and sudden changes, improves the utilization efficiency of community facilities, and provides strong support for the refined operation of smart communities.

[0035] In one embodiment of this application, initial real-time pedestrian flow data and community physical space layout data are input into a facility popularity assessment model to calculate the facility's popularity level and spatial popularity distribution, including: Collect basic data on the real-time operation status of the community, including pedestrian flow, crowd density, and individual dwell time. The basic data is input into the facility popularity assessment model, which calculates the initial popularity value of the facility based on the weighted aggregation of the basic data. The initial heat value is matched to a preset heat level system to obtain the facility heat level; By combining spatial location information from the basic data with the physical spatial layout of the community, a spatial heat distribution is generated.

[0036] In this embodiment, when collecting basic data on the real-time operational status of the community, the basic data includes pedestrian flow, crowd density, and individual dwell time. Pedestrian flow refers to the number of people passing through a certain area or facility per unit time; crowd density refers to the number of people per unit area; and individual dwell time refers to the time an individual spends in a certain area or facility. This basic data can be monitored and statistically analyzed in real time using infrared sensors, millimeter-wave radar, and video analytics systems (combined with computer vision technology) deployed in various areas of the community to obtain pedestrian flow and crowd density data. Simultaneously, individual dwell time can be obtained through Wi-Fi / Bluetooth probe technology, RFID tag tracking, or video analytics-based personnel trajectory tracking algorithms. Furthermore, multiple data sources, such as the community's intelligent access control system and public area sensor networks, can be integrated for data collection to ensure the comprehensiveness and accuracy of the data.

[0037] Weighted aggregation is a method of comprehensively calculating data of different types or importance by assigning different weights to various basic data points to characterize their varying contributions to facility popularity. For example, a linear weighted summation approach can be used: initial popularity value = W1 × pedestrian flow + W2 × crowd density + W3 × individual dwell time, where W1, W2, and W3 are preset weight coefficients, which can be determined through expert experience, historical data analysis, or machine learning methods (such as regression analysis). Another approach is to use nonlinear functions (such as the Sigmoid function or ReLU function) to transform the basic data points before weighted aggregation to better characterize the nonlinear impact of different data on popularity. Facility popularity assessment models can be machine learning-based models, such as support vector machines or decision trees, that predict popularity values ​​by learning from historical data.

[0038] The popularity rating system aims to transform continuous initial popularity values ​​into discrete, easily understood, and operable ratings. For example, piecewise functions or threshold methods can be used to divide initial popularity values ​​into low, medium, high, and extremely high ratings, each corresponding to a popularity value range. The preset popularity rating system can be defined based on community operation experience, historical data distribution characteristics, or specific application scenario requirements. For instance, it can be defined as 0-20 for low popularity, 21-50 for medium popularity, 51-80 for high popularity, and 81-100 for extremely high popularity. This rating system can also be dynamically adjusted, adaptively adjusting according to overall community traffic trends or seasonal changes.

[0039] In one embodiment of this application, the popularity rating system is adaptively adjusted based on the overall community pedestrian flow trend or seasonal changes to ensure the timeliness and accuracy of popularity assessment, specifically including: Establish a popularity ranking system adjustment cycle (e.g., weekly or monthly). At the end of each adjustment cycle, obtain the overall historical pedestrian flow dataset for the community within that cycle. Analyze this overall historical pedestrian flow dataset to extract features that characterize the macro trend of pedestrian flow in the community. These features include, but are not limited to, average pedestrian flow within the cycle, peak pedestrian flow distribution, holiday effects, and seasonal fluctuation index.

[0040] Based on the extracted macro-trend characteristics, the heat value range corresponding to the preset heat level system is dynamically adjusted. Specific adjustment methods include: If a significant increase or decrease in the overall population flow trend in the community is detected compared to the previous period (such as entering the peak tourist season or the off-season in winter), the boundaries of the heat value intervals for all heat levels will be shifted upward or downward. The magnitude of the shift is proportional to the rate of change in the overall population flow trend.

[0041] If increased fluctuations in pedestrian traffic are detected (such as fluctuating pedestrian traffic during holidays), the range of heat values ​​for each heat level should be appropriately stretched to accommodate greater data fluctuations; if pedestrian traffic is detected to be stabilizing, the range should be appropriately compressed to improve the sensitivity of heat level classification.

[0042] Instead of using fixed absolute value ranges, the system dynamically defines grading boundaries based on the statistical distribution of initial popularity values ​​for all facilities within an adjustment period. For example, all popularity values ​​within a period are sorted from low to high, with the top 20% of facilities defined as extremely high popularity, the next 30% as high popularity, and so on. This method allows the popularity grading system to automatically adapt to long-term changes in the overall population flow in the community, maintaining its distinctiveness.

[0043] Furthermore, a time decay factor can be introduced to assign different weights to historical data, making recent data have a greater impact on the adjustment of the current ranking system, thus more sensitively capturing recent changes in population flow trends. Through this adaptive adjustment mechanism, it is ensured that the popularity ranking system always matches the dynamic operating context of the community.

[0044] Spatial heat distribution visually illustrates the geographical distribution characteristics of facility heat within the community's physical space. This can be achieved through a Geographic Information System (GIS), which associates the heat level of a facility with its actual geographic coordinates and visualizes it on the community's electronic map using color depth, heat maps, and other methods. Spatial location information can come from GPS data of sensor deployment locations, preset coordinates of facilities, etc.; community physical space layout data can come from CAD drawings, 3D models, etc. These data are fused to generate a spatial data layer that includes facility heat information. To generate a smoother, more continuous spatial heat distribution, a spatial interpolation algorithm (Kriging interpolation) can be used to process discrete facility heat values, generating a heat distribution surface across the entire community.

[0045] Through the above technical solution, this application collects basic data on the real-time operational status of the community, including pedestrian flow, crowd density, and individual dwell time. By covering multi-dimensional real-time indicators, the comprehensiveness and representativeness of the basic information are ensured, avoiding evaluation biases that may be caused by a single data source. The basic data is input into a facility popularity assessment model. The model calculates the initial popularity value of the facility based on a weighted aggregation of the basic data. The weighted aggregation method considers the differences in the contribution of different data types to popularity, enabling the initial popularity value to more objectively and comprehensively represent the actual usage intensity of the facility. The initial popularity value is matched to a preset popularity level system to obtain the facility popularity level. Through the standardization of the preset system, it is easy to uniformly apply and compare subsequent strategies, enhancing the operability of the popularity level. Combining the spatial location information in the basic data with the community's physical spatial layout, a spatial popularity distribution is generated. By integrating spatial location and physical layout data, the distribution characteristics of popularity in the community are intuitively displayed, providing spatial-dimensional decision support for the deployment of electronic identification points. This further improves the accuracy and adaptability of subsequent electronic identification information collection strategies, enabling the collection strategies to more accurately respond to the real-time dynamics of the community, thereby optimizing the efficiency and user experience of the entire guidance and orientation method.

[0046] In one embodiment of this application, the facility popularity assessment model is a graph neural network model that integrates community physical space constraints. Basic data is input into the facility popularity assessment model, and the model calculates the initial popularity value of the facility based on a weighted aggregation of the basic data, including: The various facilities and their connecting paths within the community are constructed as a spatial topology graph, where nodes represent facilities and edges represent the physical connections between facilities. The basic data is embedded into the spatial topology graph as the initial features of each node; By utilizing the message passing mechanism of graph neural networks, feature information of neighboring nodes is aggregated among nodes, and edge weights are introduced during the aggregation process; the edge weights are determined based on physical space constraints such as path length and ease of passage. After message passing and feature updating through a multi-layer graph neural network, the initial heat value of each facility node is output.

[0047] In this embodiment, the facility popularity assessment model aims to accurately characterize the actual activity level of each facility within the community. Its core lies in incorporating the community's physical spatial structure and the connectivity between facilities. Graph neural network models can effectively process non-Euclidean spatial data, naturally integrating physical constraints such as the spatial location and connectivity of facilities through node and edge representations. For example, this model can employ architectures such as message passing neural networks (MPNNs) to learn the deep spatial characteristics of facilities through multi-layer nonlinear transformations.

[0048] Constructing a spatial topology graph is a crucial step in graph neural network models for processing spatial data. Nodes represent specific facilities within the community, such as gyms, libraries, activity rooms, and restaurants. Edges represent physical connections or reachable paths between these facilities, such as corridors, passageways, stairs, and elevators. This graph structure can be stored in the form of an adjacency matrix or an adjacency list. For example, based on the community's architectural floor plan or 3D model, the center point of each facility can be abstracted as a node, and the traversable paths between facilities can be abstracted as edges; alternatively, edges can be established between strongly related facilities based on their functional zoning and pedestrian flow design.

[0049] Basic data serves as the raw input for assessing facility popularity, including pedestrian traffic, crowd density, and individual dwell time. Embedding this data as the initial features of nodes means that each facility node will have a feature vector at the input layer of the graph neural network. This feature vector includes information about the facility's current or recent pedestrian activity. For example, pedestrian traffic, crowd density, and individual dwell time can be directly used as the node's feature vector input; alternatively, this basic data can be preprocessed, such as normalized, standardized, or mapped using a small fully connected network, to generate more expressive initial feature vectors.

[0050] By leveraging the message-passing mechanism of graph neural networks, feature information from neighboring nodes is aggregated among nodes. This mechanism is the core operation of graph neural networks, allowing nodes to exchange and integrate information, thereby capturing local and global dependencies. In each message-passing iteration, each node receives information from its directly connected neighbors and updates it by combining this information with its own. For example, mean aggregation can be used, where each node averages the feature vectors of its neighbors and then combines them with its own features.

[0051] Introducing edge weights during the aggregation process is intended to characterize the impact of the physical connectivity between facilities on information transmission. Edge weights are determined based on physical spatial constraints such as path length and ease of access, quantifying the cost or efficiency of paths between facilities. For example, longer paths can have smaller weights, indicating a decrease in information transmission. Ease of access, such as whether stairs are required, whether the passage is narrow, or whether there are obstacles, can also be converted into weighting factors through empirical values ​​or preset rules. For instance, paths with higher accessibility have lower weights, indicating a smaller contribution to the spread of popularity.

[0052] Through message passing and feature updates within a multi-layered graph neural network, each node can aggregate information from more distant neighbors, thereby capturing a wider range of spatial dependencies. At each layer, node features are updated based on neighbor information and processed by a non-linear activation function. Finally, after multiple layers of learning, the model outputs a final feature representation for each facility node. This representation comprehensively characterizes its own activity state and its interactions within the community's physical space, thus outputting the initial popularity value for each facility node. For example, the last layer could be a fully connected layer that maps the learned node features to a scalar value, i.e., the facility's initial popularity value.

[0053] Through the aforementioned technical solution, this application constructs a spatial topology map of all facilities and their connecting paths within the community. Using basic data as the initial features of each node, the model leverages the message passing mechanism of a graph neural network to introduce edge weights determined by physical spatial constraints such as path length and ease of access when aggregating neighbor node feature information. This mechanism ensures that the mutual influence between facilities and the heat propagation process fully consider the actual physical accessibility and passage efficiency of the community, rather than merely abstract connections. After multiple layers of message passing and feature updates, the model can perform deep learning and output initial heat values ​​that accurately represent the actual usage of facilities and their interactions in physical space. This effectively solves the problem of inaccurate assessments caused by traditional heat assessment models neglecting physical spatial constraints, providing a more reliable and accurate data foundation for subsequent facility guidance and orientation schemes, improving the matching degree between the orientation scheme and the actual situation on site, and thus optimizing residents' community activity experience.

[0054] In one embodiment of this application, a method for guiding and directing activity community facilities based on electronic building identification further includes: The weights of edges are adjusted based on real-time pedestrian traffic data. When the pedestrian traffic on a certain path is detected to be greater than the preset traffic threshold, the weights of the corresponding edges on the path are reduced based on the first step length.

[0055] In this embodiment, adjusting edge weights based on real-time pedestrian flow data aims to dynamically represent actual traffic conditions by enabling the weights of edges in the graph neural network model that indicate connections between facilities. Specifically, pedestrian flow sensors, such as infrared beam counters or image recognition-based passenger flow statistics devices, can be deployed in the community's physical space to collect real-time pedestrian flow data along each path. This real-time data is periodically transmitted to a central processing unit to update the weights of corresponding edges in the spatial topology graph. Alternatively, historical pedestrian flow pattern data and local sampling data can be combined to estimate the real-time pedestrian flow along each path using a prediction algorithm, and edge weight adjustments can be triggered based on the estimation results.

[0056] When the pedestrian flow on a certain path is detected to be greater than the preset flow threshold, this condition serves as a mechanism to trigger weight adjustment, ensuring that weight modification is only performed when the path is congested or under high load.

[0057] The preset traffic threshold can be set based on the physical carrying capacity of the path, historical congestion data, or community management strategies. For example, the traffic threshold for a 2-meter-wide pedestrian path can be set to 50 people per minute. Furthermore, this traffic threshold can also be dynamically adjusted, for example, by flexibly setting it based on the average pedestrian flow at different times of the day or the expected pedestrian flow for specific events (such as large community events) to adapt to different operational scenarios.

[0058] Reducing the weight of edges corresponding to a path based on the first step length means decreasing the weight of the edge corresponding to that path in the spatial topology graph by a predetermined step size when the detected pedestrian flow exceeds a preset flow threshold. In the message passing mechanism of graph neural networks, edge weights typically represent the ease of information transmission or the attractiveness of a path. Therefore, reducing the edge weight means that the path's difficulty of passage increases or its attractiveness decreases, thereby weakening its impact on neighboring nodes during feature aggregation. The first step length can be a fixed value, such as decreasing by 0.1 or 0.05 each time, and this fixed value can be determined through experience or experimentation. Alternatively, the first step length can be dynamically determined based on the degree to which the pedestrian flow exceeds the preset flow threshold; that is, the greater the exceedance, the larger the reduction, to more accurately represent the congestion level of the path.

[0059] In one embodiment of this application, adjusting the weights of edges based on real-time pedestrian flow data, when the pedestrian flow of a certain path is detected to be greater than a preset flow threshold, reduces the weight of the corresponding edge of the path based on the first step length. Specifically, this includes: obtaining the real-time pedestrian flow value of the current path and obtaining the preset flow threshold for that path; calculating the deviation between the real-time pedestrian flow and the threshold; and dynamically determining the first step length based on the deviation. The first step length is positively correlated with the deviation; that is, the greater the deviation, the greater the reduction in weight, thus more accurately reflecting the inhibitory effect of the actual congestion status of the path on the spread of facility heat.

[0060] Specifically, exponential or logarithmic functions are used to map the relationship between the degree of deviation and the step size. That is, the first step size is:

[0061] Where a and b are adjustment coefficients, ensuring that the step size increases non-linearly when congestion worsens, thus enabling rapid response to extreme congestion situations. ΔQ represents the deviation of real-time pedestrian flow from the threshold. ΔW is the first step length.

[0062] The methods for determining the adjustment coefficient include: collecting data pairs under different congestion levels during the community trial operation phase or in historical databases; real-time pedestrian flow along the route and changes in the actual travel cost of the route (e.g., the proportion of travel time extension), which can be regarded as the weighted dependent variable ΔWt under ideal conditions.

[0063] For each acquisition time, a data point (ΔQi, ΔWt) is formed. Using optimization algorithms such as least squares, the function ε-fits onto the above sample set.

[0064] By continuously adjusting the values ​​of a and b, the sum of squared errors between the predicted ΔW and the actual ΔWt is minimized for all sample points; the final convergent values ​​of a and b are the desired values.

[0065] Through the aforementioned technical solution, this application monitors pedestrian flow in real time and dynamically adjusts the weights of edges in the graph neural network, particularly reducing their weights when path congestion occurs. This enables the facility popularity assessment model to more accurately represent the actual usage status of community facilities and the efficiency of path passage. This ensures that in the subsequent generation of facility guidance and wayfinding schemes, a more realistic and real-time popularity level and spatial popularity distribution can be used to generate more efficient and adaptable guidance and wayfinding schemes that better reflect the actual situation on site, thereby improving the accuracy of guidance and wayfinding and the user experience.

[0066] In one embodiment of this application, based on heat level and spatial heat distribution, and with the goal of meeting data collection coverage and real-time requirements, an electronic identification information collection strategy is generated under the constraints of community physical space and network topology, including: A multi-objective optimization function is constructed with the goal of maximizing data collection coverage and minimizing construction costs; The heat level is converted into a weight requirement for data collection accuracy, and the spatial heat distribution is converted into a candidate area for the deployment of electronic markers. Based on candidate regions, under community physical space constraints including power supply and installation location permits, and network topology constraints including communication distance and bandwidth limitations, a multi-objective optimization function is solved to determine the specific deployment location and number of electronic markers. Based on the determined locations of electronic identification points, the optimal data transmission path and communication protocol are planned within the existing network topology. Based on the facility's popularity level, dynamic information collection rules, including collection frequency and data upload cycle, are generated for each electronic tag point, resulting in the final electronic tag information collection strategy.

[0067] In this embodiment, a multi-objective optimization function is constructed based on heat level and spatial heat distribution, aiming to maximize data collection coverage and minimize construction costs. This multi-objective optimization function aims to simultaneously optimize multiple conflicting objectives, such as maximizing data collection coverage and minimizing construction costs. The multi-objective optimization function quantifies these objectives through a mathematical model and seeks a solution that achieves the optimal balance among all objectives. In practical applications, a weighted sum method can be used, assigning a weight to each objective to combine them into a single optimization objective; or an ε-constraint method can be used, where one objective is the primary optimization objective, while other objectives are transformed into constraints.

[0068] Subsequently, the heat level is converted into a weighted requirement for data collection accuracy, and the spatial heat distribution is converted into candidate areas for electronic tag deployment. Converting the facility's heat level into a weighted requirement for data collection accuracy aims to ensure higher precision and reliability in data collection from frequently active, densely populated areas within the community. For example, facilities with high heat levels can be mapped to higher weights, requiring more frequent electronic tag collection and finer data granularity; while facilities with low heat levels correspond to lower weights. This conversion can be achieved through a pre-defined mapping table, where the corresponding weight is directly obtained by looking up the table based on different heat levels; or by defining a continuous function that takes heat values ​​as input and outputs corresponding weights, such as using the Sigmoid function or a linear function for mapping. Simultaneously, converting the community's spatial heat distribution into candidate areas for electronic tag deployment aims to prioritize the deployment of limited electronic tag resources in areas with frequent pedestrian activity and high data value. For example, a heat threshold can be set, and areas with heat density greater than the heat threshold can be designated as candidate areas; or a clustering algorithm can be used to divide the community into several regional clusters based on the characteristics of spatial heat distribution, with each cluster representing a potential candidate area.

[0069] Based on this, and considering candidate regions, a multi-objective optimization function is solved under constraints including power supply, permitted installation locations within the community's physical space, and network topology constraints including communication distance and bandwidth limitations, to determine the specific deployment locations and number of electronic markers. Solving this multi-objective optimization function to determine the specific deployment locations and number of electronic markers requires comprehensive consideration of multiple factors such as data collection coverage, construction costs, physical space constraints, and network topology constraints. During the solution process, heuristic algorithms, such as genetic algorithms, can be used to iteratively search for solutions that satisfy all constraints and optimize the objectives. Community physical space constraints refer to the actual physical limitations that must be considered during the deployment of electronic identification points. These constraints include, but are not limited to, the accessibility of power supply to ensure stable operation of electronic identification points; the permissibility of installation locations, such as avoiding deployment in areas where installation is prohibited or poses safety hazards; and factors such as building structure, aesthetic requirements, and obstructions. These constraints directly affect the feasible deployment range and number of electronic identification points. Network topology constraints refer to the network communication limitations that need to be considered during the deployment of electronic identification points and data transmission. These constraints include, but are not limited to, communication distance limitations to ensure effective communication between electronic identification points and data aggregation points or central servers; bandwidth limitations to ensure real-time data transmission and throughput; and factors such as signal interference, network latency, and security requirements. These constraints determine the data transmission efficiency and reliability between electronic identification points and between electronic identification points and backend systems.

[0070] Furthermore, based on the determined locations of electronic identification points, the optimal data transmission path and communication protocol are planned within the existing network topology. Planning the optimal data transmission path and communication protocol aims to ensure that the data collected by the electronic identification points can be transmitted to the backend processing system efficiently and reliably. The optimal path can be planned based on the network topology, using shortest path algorithms (such as Dijkstra's algorithm) to select the path with the lowest latency and most sufficient bandwidth. The choice of communication protocol needs to consider factors such as data volume, real-time requirements, network environment, and security. For example, for lightweight data with high real-time requirements, MQTT or CoAP protocols can be selected; for scenarios with large data volumes and high reliability requirements, HTTP / 2 or TCP / IP protocols can be selected.

[0071] Finally, based on the facility's popularity level, dynamic information collection rules, including collection frequency and data upload cycle, are generated for each electronic tag point, resulting in the final electronic tag information collection strategy. The purpose of generating these dynamic information collection rules is to enable electronic tag points to adaptively adjust their data collection behavior according to the facility's real-time popularity level, thereby optimizing resource consumption while ensuring data quality. For example, for facilities with high popularity levels, a higher collection frequency and a shorter data upload cycle can be set to obtain more real-time status information; while for facilities with low popularity levels, the collection frequency can be appropriately reduced and the upload cycle extended. This dynamic adjustment can be achieved through a preset rule engine, triggering different collection strategies based on the popularity level. The final electronic tag information collection strategy is a comprehensive solution that details the deployment location, number, data transmission method, and dynamic information collection rules of the electronic tag points.

[0072] Through the above technical solution, this application constructs a multi-objective optimization function aimed at maximizing data collection coverage and minimizing construction costs. This allows for a comprehensive balance between economic benefits and data acquisition scope, avoiding resource waste or data blind spots. By converting facility heat levels into weighted requirements for data collection accuracy and combining this with spatial heat distribution to generate candidate areas for electronic marker deployment, data collection resources can be intelligently directed towards high-value, high-demand areas, improving data quality and relevance in key areas. In solving the multi-objective optimization function, community physical space constraints such as power supply and installation location permits, as well as network topology constraints such as communication distance and bandwidth limitations, are fully considered, ensuring the generated electronic marker deployment scheme is highly feasible and operable in a real-world environment. Furthermore, by planning the optimal data transmission path and communication protocol based on the determined electronic marker locations, the real-time performance and reliability of data transmission are effectively guaranteed, reducing communication latency. More importantly, information collection rules, including collection frequency and data upload cycle, are dynamically generated based on the facility popularity level. This allows the data collection process to adapt to the real-time flow of people in the community, avoiding unnecessary resource consumption and ensuring that sufficiently detailed and real-time facility data can be obtained during peak hours.

[0073] In one embodiment of this application, facility heat level is converted into a weighted requirement for data collection accuracy, and spatial heat distribution is converted into candidate areas for electronic marker deployment, including: Based on the rate of change of real-time pedestrian flow and the mapping relationship between the preset facility heat level and the basic collection frequency, the actual collection frequency of each electronic marker is adjusted. Kernel density estimation is performed on the spatial heat distribution to generate a heat probability density surface. Regions on the heat probability density surface that are greater than a preset threshold are identified as core candidate regions for the deployment of electronic markers, while regions that are less than a preset threshold but greater than a minor threshold are identified as edge candidate regions.

[0074] In this embodiment, the rate of change of real-time pedestrian flow refers to the speed or trend of increase or decrease in pedestrian flow within the community per unit time. It can characterize the drastic degree of dynamic change in pedestrian flow and is a key indicator for judging the fluctuation of facility popularity. For example, it can be obtained by calculating the ratio of the difference in pedestrian flow at adjacent time points to the time interval by continuously collected pedestrian flow data (such as per minute or per second).

[0075] The preset mapping relationship between facility popularity levels and basic data collection frequencies is a predefined rule or lookup table used to associate different facility popularity levels with corresponding basic data collection frequencies, providing an initial frequency benchmark based on facility popularity levels for data collection at electronic markers. For example, a piecewise function or discrete mapping table can be used to map high popularity levels to 1 second / time, medium popularity levels to 5 seconds / time, and low popularity levels to 30 seconds / time. By adjusting the actual collection frequency of each electronic marker, the preset basic collection frequency can be corrected based on real-time dynamic information (people flow change rate) to adapt to actual scenario requirements. For example, the basic collection frequency can be multiplicatively or additively adjusted based on the people flow change rate; when the change rate exceeds a threshold, the frequency is increased by N times or by M times / second, thus outputting the optimal actual collection frequency.

[0076] Furthermore, kernel density estimation of the aforementioned spatial heat distribution is a nonparametric statistical method used to estimate the probability density function of a random variable. By placing a kernel function (such as a Gaussian kernel) around each data point and then superimposing these kernel functions, the overall density distribution can be smoothly estimated. This method can transform discrete spatial heat data into a continuous, smooth heat probability density surface, more accurately representing the continuous spatial distribution and concentration trend of heat. For example, a Gaussian kernel function can be selected, and the smoothing degree can be controlled by adjusting the bandwidth parameter to calculate the heat probability density surface from the spatial heat distribution data.

[0077] The heat probability density surface is a continuous surface obtained after kernel density estimation, representing the probability or density of heat occurrence at each spatial point within a community. Its height represents the heat density of that area, providing a direct and accurate representation of the heat concentration in different areas of the community, and offering a quantitative basis for subsequent candidate area delineation. For example, the physical space of the community can be divided into grids, and the heat probability density value at the center point of each grid can be calculated to form a two-dimensional density map. The preset threshold and secondary threshold are two pre-defined numerical boundaries used to delineate the heat probability density surface. Their function is to discretize the continuous heat probability density surface into regions of different importance, guiding the hierarchical deployment of electronic markers. For example, based on historical data analysis or expert experience, a high value can be set as the preset threshold (e.g., density value greater than 0.8), and a medium-low value as the secondary threshold (e.g., density value greater than 0.3 but less than 0.8). Alternatively, a clustering algorithm (such as K-means) can be used to cluster the heat probability density values, and the optimal threshold can be determined based on the clustering results. Through this division, regions on the heat probability density surface greater than the preset threshold are identified as core candidate areas for electronic marker deployment, representing core hotspot areas with high population density and frequent facility use within the community; regions less than the preset threshold but greater than the secondary threshold are identified as edge candidate areas, representing secondary hotspot areas with relatively concentrated population density and active facility use within the community.

[0078] Through the above technical solution, this application adjusts the actual collection frequency based on the rate of change of real-time pedestrian flow and a preset mapping relationship. Using real-time changing data as input, and combining preset rules to dynamically optimize weight requirements, it ensures that the collection frequency can adapt to pedestrian flow fluctuations, improving the real-time performance and accuracy of data collection. Simultaneously, it generates a heat probability density surface by kernel density estimation of spatial heat distribution, and divides core and edge candidate areas based on thresholds. Statistical modeling methods are used to accurately characterize the spatial distribution, providing hierarchical candidate area definitions, facilitating efficient use of hotspot information in subsequent deployment strategies, and optimizing coverage and cost efficiency. The entire solution emphasizes the specific input methods of real-time change rate and kernel density estimation, achieving intelligent and refined transformation processes. This enables a more effective response to pedestrian flow fluctuations when generating electronic identification information collection strategies, optimizing data collection coverage and cost efficiency, thereby improving the real-time matching degree between facility guidance and orientation schemes and the actual on-site conditions.

[0079] In one embodiment of this application, solving a multi-objective optimization function to determine the specific deployment locations and number of electronic markers includes: The multi-objective particle swarm optimization algorithm is used to solve the problem, where the position of the particle represents a marker point deployment scheme. During the iterative optimization process, a deployment priority reward factor is given to the particle positions that fall into the core candidate region, and a secondary reward factor is given to the particle positions that fall into the edge candidate region. Through Pareto front analysis, the optimal deployment scheme set that simultaneously maximizes coverage and minimizes cost is output.

[0080] In this embodiment, the Multi-Objective Particle Swarm Optimization (MOPSO) algorithm is used for solving the problem, where the particle positions represent a deployment scheme for marker points. MOPSO is a heuristic optimization algorithm based on swarm intelligence, simulating the foraging behavior of flocks of birds and finding the optimal solution through the movement of particles in the search space. In multi-objective optimization problems, MOPSO can simultaneously optimize multiple conflicting objectives, not just a single one. Here, the particle positions represent a combination of specific deployment locations and quantities of electronic marker points in the community's physical space. This algorithm effectively explores complex deployment spaces, avoids getting trapped in local optima, and thus improves solution efficiency and solution quality. The MOPSO algorithm stores non-dominated solutions by maintaining an external archive and uses Pareto dominance to update the individual best position (pBest) and global best position (gBest) of particles. Each particle updates its position and velocity based on its current velocity, pBest, and gBest, where the velocity update formula typically includes an inertial weight, a cognitive component (tending towards pBest), and a social component (tending towards gBest).

[0081] During iterative optimization, particle positions falling into the core candidate region are given a deployment priority reward factor, while particle positions falling into the edge candidate region are given a secondary reward factor. The reward factor is a mechanism used in the optimization algorithm to assign different rewards based on the region (core or edge candidate region) of the particle's location during iteration, guiding the algorithm to prioritize the exploration and optimization of deployment schemes in specific areas. Core candidate regions typically correspond to densely populated, high-intensity areas, requiring higher data collection accuracy and coverage; edge candidate regions correspond to less important areas. By introducing reward factors, the optimization algorithm can pursue overall optimality while considering the importance of different regions, thus generating deployment schemes that better meet practical needs. Reward factors can be integrated into the fitness function of the optimization algorithm. For example, when calculating the fitness value of a particle (i.e., a deployment scheme), if the scheme deploys marker points in the core candidate region, its fitness value will be increased by a larger positive value (deployment priority reward factor); if deployed in the edge candidate region, it will be increased by a smaller positive value (secondary reward factor). Thus, when selecting and updating the optimal solution, the algorithm will tend to include schemes with more core region deployment points. Reward factors can also be implemented by adjusting the behavioral parameters of particles. For example, for particles that fall into the core candidate region, their exploration ability or convergence speed can be increased, enabling them to find better deployment points in these regions more quickly; while for particles that fall into the edge candidate region, a lower exploration weight can be given, prioritizing cost-effectiveness while ensuring basic coverage.

[0082] Pareto front analysis outputs a set of optimal deployment schemes that simultaneously maximize coverage and minimize cost. Pareto front analysis is a core concept in multi-objective optimization, used to identify a set of non-dominated solutions. In multi-objective optimization problems, there is usually no single optimal solution that simultaneously optimizes all objectives. The Pareto front (or Pareto optimal solution set) is the set of all Pareto optimal solutions, where any one solution cannot improve any other objective without worsening at least one objective. Pareto front analysis provides a series of trade-offs that balance maximizing data acquisition coverage and minimizing construction costs, offering decision-makers choices. After each iteration of the MOPSO algorithm, or after the algorithm finishes running, all generated non-dominated solutions are filtered. Specifically, for any two solutions A and B, if A is superior to or equal to B on all objectives, and strictly superior to B on at least one objective, then A is said to dominate B. Solutions on the Pareto front are those that are not dominated by any other solution. Pareto front analysis not only outputs a set of nondominated solutions, but also allows for further visualization of these solutions, such as using scatter plots to illustrate the trade-off between coverage and cost. Decision-makers can then select the most suitable option from this set of optimal deployment plans based on their actual needs and preferences—for example, choosing the option with the highest coverage within an acceptable cost range, or the option with the lowest cost while meeting minimum coverage requirements.

[0083] Through the aforementioned technical solution, this application employs a multi-objective particle swarm optimization algorithm for solving the problem. This algorithm can efficiently explore and generate diverse electronic tag deployment schemes under complex community physical space and network topology constraints, avoiding the limitations of traditional methods that may get stuck in local optima. During the iterative optimization process, by assigning deployment priority reward factors to particle positions falling into core candidate areas and secondary reward factors to particle positions falling into edge candidate areas, the algorithm can intelligently identify and prioritize resource allocation to core areas with high population density and data demand, while also considering coverage of edge areas. This maximizes data collection coverage while effectively controlling construction costs. Finally, through Pareto front analysis, an optimal deployment scheme set that simultaneously maximizes coverage and minimizes cost is output, providing a multi-dimensional and optimized selection space for subsequent electronic tag information collection strategies.

[0084] In one embodiment of this application, the historical real-time pedestrian flow status of the community corresponding to the initial facility data asset collection time is obtained. The initial facility data asset is then cleaned and verified using an anomaly detection algorithm to obtain valid facility data assets, including: By associating the community's historical real-time pedestrian flow status with the initial facility data assets at the time of initial facility data asset collection, the associated data is obtained. Based on the associated data, threshold verification is performed on the initial facility data assets. According to the physical limits of sensors or the distribution of historical data, the data range is determined, and data exceeding the data range is identified and marked as the first abnormal data assets. Logical association verification is performed on the data that passes the threshold verification. Based on the preset spatial geometric relationship or physical conservation law between electronic identification points, the logical consistency between the data is verified, and data that violates the logical relationship is identified and marked as the second abnormal data asset. Spatiotemporal clustering verification is performed on the data that has passed the logical association verification. Based on the spatiotemporal clustering algorithm or the outlier detection algorithm within the sliding window, the distribution pattern of the data in the time and space dimensions is analyzed, and outliers that do not conform to the conventional distribution pattern are identified and marked as third anomalous data assets. The first, second, and third anomalous data assets are processed, and the remaining data assets are identified as valid facility data assets.

[0085] In this embodiment, the historical real-time pedestrian flow status of the community at the time of initial facility data asset collection is associated with the initial facility data assets to obtain associated data. Specifically, the initial facility data assets (e.g., pedestrian flow and dwell time perceived by a certain electronic marker at a certain moment) can be associated with the historical real-time pedestrian flow status of the community as a whole within the same time period (e.g., total pedestrian flow at the community entrance, average pedestrian density in each area, etc.) through timestamp matching. Alternatively, the data of a specific electronic marker can be associated with the historical pedestrian flow pattern data of the area where the marker is located through geographic location information to form a comprehensive dataset including location, time, current status, and historical background.

[0086] Based on correlated data, threshold verification is performed on initial facility data assets. The data range is determined according to the physical limits of sensors or historical data distribution, and data exceeding this range is identified and marked as the first abnormal data asset. This threshold verification step effectively identifies and filters out invalid data that exceeds physically or statistically reasonable ranges, improving the efficiency and accuracy of data cleaning. For example, hard upper and lower limit thresholds can be set based on the physical measurement range of sensors (such as infrared sensors and cameras), such as the flow of people being unlikely to be negative, or a single measurement value not exceeding the sensor's maximum counting capacity. Alternatively, dynamic thresholds can be set based on historical data statistical analysis, calculating the mean, standard deviation, or percentiles of the data. For example, if the flow of people to a facility during a specific time period is typically between 10 and 50, data exceeding this range is considered abnormal.

[0087] Next, logical correlation verification is performed on the data that passes the threshold verification. Based on the preset spatial geometric relationships or physical conservation laws between electronic markers, the logical consistency between the data is verified, and data that violates the logical relationships is identified and marked as second anomalous data assets. This step aims to detect whether there are logical contradictions between the data, ensure the rationality of the data at the physical level, and avoid logical errors caused by equipment failure or environmental interference. For example, verification can be based on spatial geometric relationships. If the entrance traffic flow data of an area is much greater than the exit traffic flow data, and there are no other entrances or exits in that area, then there may be a logical inconsistency.

[0088] Subsequently, spatiotemporal clustering verification is performed on the data that has passed the logical association check. Based on spatiotemporal clustering algorithms or outlier detection algorithms within a sliding window, the distribution patterns of the data in the temporal and spatial dimensions are analyzed to identify and mark outliers that do not conform to the conventional distribution patterns as third-generation anomalous data assets. This step can identify outliers that do not conform to the conventional distribution patterns in the temporal and spatial dimensions, capturing more complex anomalies, such as sudden events or local sensor failures. Specifically, the DBSCAN (Density-Based Spatial Clustering with Noise) algorithm can be used to cluster data points in time and space, identifying points that do not belong to any dense cluster as outliers.

[0089] Finally, the first, second, and third anomalous data assets are processed, and the remaining data assets are identified as valid facility data assets. This comprehensive processing step aims to ensure that the final output of valid facility data assets is of high quality, reliable, and complete. Processing methods may include directly deleting all data assets marked as anomalous, or repairing or interpolating the anomalous data. For example, for minor anomalous data, historical averages, adjacent time point data, or machine learning models can be used for predictive imputation to preserve data integrity.

[0090] Through the aforementioned technical solutions, this application provides rich spatiotemporal contextual information for anomaly detection by associating initial facility data assets with the real-time historical pedestrian flow status of the community at the time of collection. This allows subsequent verification processes to more comprehensively consider the dynamic background of the data, avoiding misjudgments that may result from isolated data analysis. Secondly, the phased verification mechanism, including threshold verification based on sensor physical limits or historical data distribution, can quickly and effectively identify and filter out first-order anomalous data assets that clearly exceed reasonable limits, ensuring the basic validity of the data. Building upon this, logical association verification, based on the spatial geometric relationships or physical conservation laws between electronic markers, further detects the inherent logical consistency between data, thereby identifying second-order anomalous data assets that violate physical laws or spatial logic, ensuring the physical rationality of the data. Furthermore, through spatiotemporal clustering verification, utilizing spatiotemporal clustering algorithms or outlier detection algorithms within a sliding window, the distribution patterns of data in the temporal and spatial dimensions are analyzed in depth. This allows for the capture of third-order anomalous data assets that do not conform to conventional patterns, such as sudden or localized data anomalies, thereby enhancing the depth and adaptability of anomaly detection. Ultimately, comprehensive processing of these three types of anomalous data assets ensured that the final obtained valid facility data assets possessed higher reliability, completeness, and accuracy. This effectively solved the problem that traditional single-anomaly detection methods could not comprehensively identify various types of data anomalies, further improving the real-time matching degree between facility guidance and orientation schemes and actual on-site conditions, and optimizing the refined operation effect of community facilities.

[0091] In one embodiment of this application, effective facility data assets are input into a path planning decision model. Within the boundaries of community operational constraints, a multi-objective optimization problem is solved to generate a facility guidance and orientation scheme, including: Feature extraction is performed on effective facility data assets to obtain feature data representing the real-time population flow and facility availability status of the community; Based on the popularity level of the facility, a multi-objective guidance function is constructed, including guidance efficiency, path balance, and crowd evacuation time. Based on the community's physical spatial layout and real-time pedestrian flow, extract the community's operational constraint boundaries, including maximum path capacity, one-way direction, and restricted areas. The characteristic data, multi-objective guiding function, and community operation constraint boundary are input into the path planning decision model to solve for the Pareto optimal solution set that makes the multi-objective guiding function optimal within the community operation constraint boundary. From the Pareto optimal solution set, the final facility guidance scheme is selected based on the preset decision preferences.

[0092] In this embodiment, feature extraction is performed on the effective facility data assets to extract key information valuable for route planning and guidance from the cleaned and verified raw data. This feature data can intuitively or abstractly represent the real-time pedestrian flow status within the community, such as population density, flow direction, and dwell time in various areas, as well as the availability status of facilities, such as their open / closed status, current occupancy rate, and queuing situation. In terms of implementation, Principal Component Analysis (PCA) can be used to automatically learn and extract low-dimensional and representative feature vectors from the high-dimensional effective facility data assets.

[0093] Based on this, a multi-objective guidance function is constructed according to the facility's popularity level, including guidance efficiency, path balance, and crowd evacuation time. This is to comprehensively consider multiple conflicting or mutually influential objectives when generating guidance and orientation schemes, thereby achieving a more comprehensive and optimized guidance effect. Guidance efficiency aims to minimize the time or distance for users to reach the target facility; path balance aims to avoid excessive congestion on specific paths, distributing pedestrian flow across multiple available paths; and crowd evacuation time focuses on ensuring the rapid and safe evacuation of people in emergency situations. The facility's popularity level serves as a weight or priority factor, influencing the importance of each objective in the function. In terms of construction, a weighted sum method can be used, linearly combining the various objective functions through preset weights to form a single comprehensive objective function. The weights can be dynamically adjusted based on the facility's popularity level, community manager preferences, or real-time needs. Alternatively, an ε-constraint method can be used, optimizing one objective as the primary objective while transforming other objectives into constraints, setting acceptable upper or lower limits. For example, the main optimization goal is to improve guidance efficiency, while also requiring path balance and crowd evacuation time to meet certain thresholds.

[0094] Meanwhile, based on the community's physical spatial layout and real-time pedestrian flow, extracting community operational constraint boundaries, including maximum path capacity, one-way traffic, and restricted areas, ensures that the generated guidance and orientation scheme is feasible and safe in the actual community environment. These constraints characterize the physical limitations and operational management regulations of community infrastructure. Maximum path capacity limits the maximum number of people passing through a path per unit time; one-way traffic specifies the direction of passage for a particular passage; and restricted areas clearly define areas where passage is prohibited. In terms of extraction methods, physical spatial layout (such as road width and building structure) and static rules (such as one-way streets and restricted areas) can be digitized and encoded into edge attributes or node attributes in a graph structure based on Geographic Information System (GIS) data and community management regulations.

[0095] Subsequently, the feature data, multi-objective guiding function, and community operation constraint boundary are input into the path planning decision model. The goal is to find a Pareto optimal solution set within the community operation constraint boundary that optimizes the multi-objective guiding function. This aims to find a series of guiding schemes that achieve the best balance among the various objectives while satisfying all practical constraints. No solution in the Pareto optimal solution set can improve any objective without worsening at least one other objective. The multi-objective evolutionary algorithm, Multi-Objective Particle Swarm Optimization (MOPSO), can be used in the solution process.

[0096] Finally, the final facility guidance scheme is selected from the Pareto optimal solution set based on preset decision preferences. This is to choose the scheme that best meets the current actual needs or management strategy from multiple Pareto optimal solutions. The Pareto optimal solution set usually includes multiple non-dominated solutions, which have different trade-offs between different objectives, thus requiring a final selection based on specific preferences. In terms of selection methods, a priority ranking can be preset, and each scheme in the Pareto solution set can be evaluated to select the scheme with the highest score. For example, in emergency evacuation scenarios, the scheme with the shortest evacuation time can be prioritized; in daily guidance, the scheme with the highest guidance efficiency can be prioritized. Simultaneously, an interactive decision support interface can be provided, allowing community managers to dynamically adjust their decision preferences based on real-time situations or specific events (such as large-scale events or emergencies) and select the most suitable scheme from the Pareto solution set in real time.

[0097] Through the aforementioned technical solution, this application constructs a multi-objective guidance function that comprehensively considers guidance efficiency, path balance, and crowd evacuation time based on the facility's popularity level. This ensures that the generated guidance scheme not only pursues rapid user arrival but also considers overall community traffic flow and safety in emergency situations, avoiding local optima and potential risks that may result from single-objective optimization. Simultaneously, by accurately extracting the operational constraint boundaries formed by the community's physical spatial layout and real-time pedestrian flow, it ensures that all generated guidance schemes strictly comply with the community's actual physical limitations and management regulations, thereby avoiding the generation of infeasible or unsafe paths. Finally, these factors are input into the path planning decision model to solve for the Pareto optimal solution set, which is then filtered according to preset decision preferences. This ensures that the final facility guidance scheme can achieve multi-objective balanced optimization in complex and ever-changing environments while meeting specific management needs, improving the matching degree, practicality, and safety of the guidance scheme with the actual site conditions.

[0098] In one embodiment of this application, the guidance effect of the facility guidance scheme is evaluated based on a preset data asset utility evaluation index, and attribution analysis is used to identify the data feature dimensions that contribute the most to the generation of the facility guidance scheme, including: Extract primary utility metrics from effective facility data assets to reflect their quality. These primary utility metrics include data completeness, accuracy, and timeliness. By analyzing the correlation between effective facility data assets and facility guidance and orientation schemes, a second utility indicator reflecting the value of data is calculated; Input the first utility index and the second utility index into a preset utility aggregation function to calculate the data asset utility index for each data asset. The utility index of all data assets is aggregated to evaluate the overall data foundation quality of the facility guidance and guidance scheme and assess its guiding effect. Based on the second utility metric, all effective facility data assets involved in the calculation are ranked to identify the data feature dimensions that contribute the most to the final solution.

[0099] In this embodiment, the first utility metric aims to quantify the inherent quality attributes of the data asset itself, ensuring that the data is reliable and trustworthy before it is used. For example, data integrity can be assessed by checking for null values, missing fields, or incomplete data entries in the data records. For example, for pedestrian flow data, it can be checked whether there are corresponding number records for all timestamps. Data accuracy can be assessed by comparing with known true values, cross-validating data from different sources, or using statistical methods (such as outlier detection). For example, comparing the pedestrian flow data collected by sensors with manual counting or video analysis results. Data timeliness can be assessed by calculating the time difference between the data generation time and the data usage time, or by setting a threshold for the data validity period. For example, pedestrian flow data is considered highly timely within 5 minutes of generation, and its timeliness decreases after 10 minutes.

[0100] The second utility metric aims to assess the actual value of data assets in generating the final facility guidance and orientation plan. For example, it can be established by tracing the use of data assets in the route planning decision model, recording which pedestrian flow data and facility status data were adopted by the model and influenced the final route selection or facility recommendation. Another approach is to quantify the contribution of data assets to the plan's optimization objectives (such as improved guidance efficiency, improved route balance, and reduced crowd dispersal time). For instance, if a piece of pedestrian flow data enabled the model to successfully avoid congested areas, that data would have a high second utility metric.

[0101] The utility aggregation function aims to combine the intrinsic quality and extrinsic value of data to generate a unified and comprehensive data asset utility index. For example, a linear weighted summation method can be used, where the utility index equals the weighted sum of the first utility index and the second utility index, where the weights represent the degree of importance attached to data quality and data value.

[0102] The utility indices of all data assets are aggregated to assess the overall data infrastructure quality of the facility guidance program. This overall data infrastructure quality assessment is a macro-level evaluation of the combined utility level of all data assets supporting the facility guidance program. For example, the average or median of the utility indices of all data assets can be calculated as an indicator of overall data infrastructure quality. Alternatively, a weighted average can be calculated based on the importance or frequency of use of the data assets.

[0103] Based on the second utility index, all valid facility data assets involved in the calculation are ranked to identify the data feature dimensions that contribute the most to the final solution. Identifying the most contributing data feature dimensions aims to quantify the value contribution of data assets to the solution and identify the data types or attributes that have the greatest impact on the guiding scheme. For example, all data assets can be directly ranked in descending order according to the second utility index, and then the common feature dimensions included in the top-ranked data assets can be analyzed, such as pedestrian flow data in a specific area or real-time status data of a specific facility.

[0104] Through the aforementioned technical solution, this application integrates the first and second utility indicators using a pre-defined utility aggregation function to calculate the utility index of each data asset, thus providing a comprehensive evaluation standard that considers both the intrinsic quality and extrinsic value of the data. Based on this, summarizing the utility indices of all data assets allows for a macro-level assessment of the overall data foundation quality of the facility guidance scheme, providing a global perspective for judging the scheme's reliability and optimization direction. Furthermore, ranking all valid facility data assets involved in the calculation based on the second utility indicator accurately identifies the data feature dimensions that contribute the most to the final scheme. This enables the feedback learning mechanism to focus on optimizing data sources or data types that have the greatest impact on the guidance effect, thereby improving the accuracy and efficiency of feedback learning. Ultimately, this allows the facility guidance scheme to more accurately and effectively match the real-time operational status of the community, improving user experience and community management efficiency.

[0105] In one embodiment of this application, based on the results of the guidance effect evaluation and the dimensions of data features, a preset feedback learning mechanism is used to adjust the configuration parameters of the electronic identification information collection strategy and the internal parameters of the path planning decision model, including: The discrepancy between the expected and actual effects of the computing facility guidance and orientation program; Vectorize the deviation, the features of the data feature dimension, and the current community state to obtain the state vector; The state vector is input into a preset feedback learning mechanism to obtain the first parameter adjustment amount of the electronic identification information collection strategy; The configuration parameters of the electronic identification information collection strategy are updated based on the adjustment amount of the first parameter; The state vector is input into the feedback learning mechanism again to obtain the adjustment of the second parameter of the path planning decision model. The internal parameters of the path planning decision model are updated based on the adjustment amount of the second parameter.

[0106] In this embodiment, the deviation between the expected effect and the actual guidance effect of the facility guidance scheme is calculated. This deviation calculation aims to quantify the gap between the facility guidance scheme and the preset target in actual operation. The expected effect refers to the performance indicators that the scheme should achieve under ideal conditions, based on model predictions or historical data analysis, while the actual guidance effect refers to the real performance obtained through real-time monitoring and data feedback after the scheme is implemented. For example, the mean squared error or mean absolute error between the two is calculated.

[0107] Subsequently, the bias, data feature dimensions, and current community state are vectorized to obtain a state vector. Feature engineering techniques can be used to encode discrete features one-hot and normalize or standardize continuous features, then all processed features are concatenated into a fixed-length vector.

[0108] Next, the state vector is input into a pre-defined feedback learning mechanism to obtain the first parameter adjustment amount for the electronic tag information collection strategy. The feedback learning mechanism is an intelligent algorithm or model capable of self-adjustment and optimization based on the system's current state and objectives. It receives the state vector as input and, through internal learning logic, outputs parameter adjustment suggestions for the electronic tag information collection strategy. The first parameter adjustment amount is specifically used to optimize parameters in the data collection process, such as collection frequency and data upload cycle. This feedback learning mechanism can employ machine learning models, such as neural networks or support vector machines, trained on historical data to learn the mapping relationship between the state vector and the optimal parameter adjustment amount.

[0109] Based on the adjustment of the first parameter, the system updates the configuration parameters of the electronic tag information collection strategy. This step applies the adjustment of the first parameter output by the feedback learning mechanism to the electronic tag information collection strategy to dynamically change its operating configuration. Configuration parameters include, but are not limited to, the collection frequency of each electronic tag point, the data upload cycle, and the data transmission protocol. The system can directly apply the adjustment to the corresponding parameters; for example, if the adjustment indicates a 10% increase in the collection frequency, the current collection frequency is multiplied by 1.1. Alternatively, the system can use the configuration management module to convert the adjustment into specific configuration instructions and issue them to each electronic tag device or data collection service, enabling it to collect data according to the new parameters.

[0110] Based on this, the system again inputs the state vector into the feedback learning mechanism to obtain the second parameter adjustment amount for the path planning decision model. This step demonstrates the feedback learning mechanism's ability to independently and collaboratively optimize the parameters of different modules within the system. After obtaining the adjustment amount for the data acquisition strategy, the system again uses the same state vector to generate the parameter adjustment amount for the path planning decision model through the feedback learning mechanism. The second parameter adjustment amount is used to optimize the internal logic of the path planning decision model, such as the weights of each objective function in the model and the relaxation factors of constraints. The feedback learning mechanism can include multiple output layers or independent sub-models, each responsible for generating the parameter adjustment amount for different modules; alternatively, it can also adjust the parameters of one module first, observe its effect, and then adjust the parameters of another module based on the updated state vector, using a time series or iterative approach.

[0111] Finally, the internal parameters of the path planning decision model are updated based on the second parameter adjustment. This step applies the second parameter adjustment output by the feedback learning mechanism to the path planning decision model to dynamically change its internal operating logic and decision preferences. Internal parameters may include the weights of each objective in the multi-objective optimization function (such as the weights of guidance efficiency, path balance, and crowd evacuation time), constraint thresholds, the number of iterations of the algorithm, or the learning rate. The system can directly modify the corresponding parameter values ​​in the path planning decision model configuration file and reload the model or make it dynamically effective; alternatively, it can call the model's parameter update function through the model management interface to apply the adjustment to the model's internal weights, biases, or hyperparameters.

[0112] Through the aforementioned technical solution, this application achieves adaptive adjustment of parameters in the electronic identification information collection strategy and the route planning decision model. This not only enhances the real-time responsiveness and adaptability of the entire guidance system to changes in the community environment, but also, through continuous feedback learning, enables the generated facility guidance scheme to more accurately match the actual situation on site, improving guidance effectiveness and user experience. Simultaneously, this dynamic optimization mechanism provides strong technical support for the refined operation and management of community facilities, ensuring the system's continuous and efficient operation in complex and ever-changing environments.

[0113] Corresponding to the activity community facility guidance method based on electronic building identification in the above embodiment, Figure 2 This is a structural block diagram of an activity community facility guidance and wayfinding system based on electronic building signage, provided as an embodiment of this application. For ease of explanation, only the parts relevant to the embodiment of this application are shown. References Figure 2The activity community facility guidance system 20 based on electronic building identification includes: a data acquisition module 21, a popularity assessment module 22, a strategy generation module 23, an asset construction module 24, a cleaning and verification module 25, a route planning module 26, an evaluation and analysis module 27, and a feedback and adjustment module 28.

[0114] Among them, the data acquisition module 21 is used to obtain the initial real-time pedestrian flow status in the community through a preset basic acquisition cycle; The heat assessment module 22 is used to input the initial real-time pedestrian flow status and community physical space layout data into the facility heat assessment model to calculate the facility's heat level and spatial heat distribution. The strategy generation module 23 is used to generate electronic identification information collection strategies based on heat level and spatial heat distribution, with the goal of maximizing data collection coverage, minimizing collection latency and construction costs, under the constraints of community physical space and network topology. The asset construction module 24 is used to collect data based on the electronic identification information collection strategy to obtain initial facility data assets. The cleaning and verification module 25 is used to obtain the real-time historical population flow status of the community corresponding to the time of initial facility data asset collection, and to clean and verify the initial facility data assets through an anomaly detection algorithm to obtain valid facility data assets. The route planning module 26 is used to input effective facility data assets into the route planning decision model, solve a multi-objective optimization problem within the boundaries of community operation constraints, and generate facility guidance and orientation schemes. The evaluation and analysis module 27 is used to evaluate the guidance effect of the facility guidance scheme based on the preset data asset utility evaluation indicators, and to use attribution analysis to identify the data feature dimensions that contribute the most to the generation of the facility guidance scheme. The feedback adjustment module 28 is used to adjust the configuration parameters of the electronic identification information collection strategy and the internal parameters of the path planning decision model based on the results and data feature dimensions of the guidance effect evaluation and using a preset feedback learning mechanism.

[0115] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of the modules in the aforementioned device embodiments, for example... Figure 2 The functions of the data acquisition module 21, heat assessment module 22, strategy generation module 23, asset construction module 24, cleaning and verification module 25, path planning module 26, evaluation and analysis module 27, and feedback adjustment module 28 are shown.

[0116] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), but it may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0117] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0118] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store device type information.

[0119] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation methods described in any embodiment of the activity community facility guidance and orientation method based on electronic building identification provided in the embodiments of this application, or they can execute the implementation methods of the electronic devices described in the embodiments of this application, which will not be repeated here.

[0120] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0121] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., provided on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0122] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0123] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic devices and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0124] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces or units, or it may be an electrical, mechanical, or other form of connection.

[0125] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0126] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0127] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for guiding and directing activity community facilities based on electronic building signage, characterized in that, include: The initial real-time pedestrian flow status within the community is obtained through a preset basic collection cycle; The initial real-time pedestrian flow status and community physical space layout data are input into the facility heat assessment model to calculate the facility heat level and spatial heat distribution; Based on the aforementioned heat level and spatial heat distribution, with the goal of maximizing data collection coverage and minimizing collection latency and construction costs, an electronic identification information collection strategy is generated under the constraints of community physical space and network topology. Data is collected based on the electronic identification information collection strategy to obtain initial facility data assets; The community's historical real-time pedestrian flow status corresponding to the time of initial facility data asset collection is obtained, and the initial facility data asset is cleaned and verified through an anomaly detection algorithm to obtain valid facility data assets. The effective facility data assets are input into the path planning decision model, and within the boundaries of community operation constraints, a multi-objective optimization problem is solved to generate a facility guidance and orientation scheme. Based on preset data asset utility evaluation indicators, the guidance effect of the facility guidance scheme is evaluated, and attribution analysis is used to identify the data feature dimensions that contribute the most to the generation of the facility guidance scheme. Based on the results of the guidance effect evaluation and the data feature dimensions, the configuration parameters of the electronic identification information collection strategy and the internal parameters of the path planning decision model are adjusted using a preset feedback learning mechanism.

2. The method for guiding and directing activity community facilities based on electronic building signage according to claim 1, characterized in that, The step of inputting the initial real-time pedestrian flow status and community physical space layout data into the facility heat assessment model to calculate the facility's heat level and spatial heat distribution includes: Collect basic data on the real-time operation status of the community, including pedestrian flow, crowd density, and individual dwell time. The basic data is input into the facility popularity assessment model, which calculates the initial popularity value of the facility based on the weighted aggregation of the basic data; The initial heat value is matched to a preset heat level system to obtain the facility heat level; By combining the spatial location information in the aforementioned basic data with the physical spatial layout of the community, a spatial heat distribution is generated.

3. The method for guiding and directing activity community facilities based on electronic building signage according to claim 2, characterized in that, The facility popularity assessment model is a graph neural network model that integrates community physical space constraints. The basic data is input into the facility popularity assessment model, which calculates the initial popularity value of the facilities based on a weighted aggregation of the basic data, including: The various facilities and their connecting paths within the community are constructed as a spatial topology graph, where nodes represent facilities and edges represent the physical connections between facilities. The basic data is embedded into the spatial topology graph as the initial features of each node; By utilizing the message passing mechanism of graph neural networks, feature information of neighboring nodes is aggregated among nodes, and edge weights are introduced during the aggregation process. After message passing and feature updating through a multi-layer graph neural network, the initial heat value of each facility node is output.

4. The method for guiding and directing activity community facilities based on electronic building signage according to claim 3, characterized in that, Also includes: The weights of the edges are adjusted based on real-time pedestrian traffic data. When the pedestrian traffic on a certain path is detected to be greater than a preset traffic threshold, the weights of the corresponding edges on the path are reduced based on the first step length.

5. A method for guiding and directing activity community facilities based on electronic building signage according to claim 2, characterized in that, Based on the heat level and spatial heat distribution, and aiming to meet the requirements of data collection coverage and real-time performance, an electronic identification information collection strategy is generated under the constraints of community physical space and network topology, including: A multi-objective optimization function is constructed with the goal of maximizing data collection coverage and minimizing construction costs; The heat level is converted into a weight requirement for data collection accuracy, and the spatial heat distribution is converted into a candidate area for the deployment of electronic markers. Based on the candidate regions, under the constraints of community physical space including power supply and installation location permission, and network topology constraints including communication distance and bandwidth limitations, the multi-objective optimization function is solved to determine the specific deployment location and number of electronic identification points. Based on the determined locations of electronic identification points, the optimal data transmission path and communication protocol are planned within the existing network topology. Based on the facility's popularity level, dynamic information collection rules, including collection frequency and data upload cycle, are generated for each electronic identification point, resulting in the final electronic identification information collection strategy.

6. A method for guiding and directing activity community facilities based on electronic building signage according to claim 5, characterized in that, The step of converting the facility heat level into a weighted requirement for data collection accuracy and converting the spatial heat distribution into candidate areas for electronic marker deployment includes: Based on the rate of change of real-time pedestrian flow and the preset mapping relationship between the facility heat level and the basic collection frequency, the actual collection frequency of each electronic marker is adjusted. Kernel density estimation is performed on the spatial heat distribution to generate a heat probability density surface. Regions on the heat probability density surface that are greater than a preset threshold are identified as core candidate regions for electronic marker deployment, and regions that are less than the preset threshold but greater than a minor threshold are identified as edge candidate regions.

7. A method for guiding and directing activity community facilities based on electronic building signage according to claim 6, characterized in that, Solving the multi-objective optimization function to determine the specific deployment locations and number of electronic markers includes: The multi-objective particle swarm optimization algorithm is used to solve the problem, where the position of the particle represents a marker point deployment scheme. During the iterative optimization process, a deployment priority reward factor is given to the particle positions that fall into the core candidate region, and a secondary reward factor is given to the particle positions that fall into the edge candidate region. Through Pareto front analysis, the optimal deployment scheme set that simultaneously maximizes coverage and minimizes cost is output.

8. A method for guiding and directing activity community facilities based on electronic building signage according to claim 1, characterized in that, The process of inputting the effective facility data assets into the path planning decision model, solving a multi-objective optimization problem within the boundaries of community operation constraints, and generating facility guidance and orientation schemes includes: Feature extraction is performed on effective facility data assets to obtain feature data representing the real-time population flow and facility availability status of the community; Based on the popularity level of the facilities, a multi-objective guidance function is constructed, including guidance efficiency, path balance, and crowd evacuation time. Based on the community's physical spatial layout and real-time pedestrian flow, extract the community's operational constraint boundaries, including maximum path capacity, one-way direction, and restricted areas. The feature data, the multi-objective guiding function, and the community operation constraint boundary are input into the path planning decision model to solve for the Pareto optimal solution set that makes the multi-objective guiding function optimal within the community operation constraint boundary. From the Pareto optimal solution set, the final facility guidance scheme is selected according to the preset decision preferences.

9. A method for guiding and directing activity community facilities based on electronic building signage according to claim 1, characterized in that, Based on the results of the guidance effect evaluation and the data feature dimensions, and utilizing a preset feedback learning mechanism, the configuration parameters of the electronic identification information collection strategy and the internal parameters of the path planning decision model are adjusted, including: The discrepancy between the expected and actual guidance effects of the computing facility guidance scheme; The deviation, the features of the data feature dimension, and the current community state are vectorized to obtain a state vector; The state vector is input into a preset feedback learning mechanism to obtain the first parameter adjustment amount of the electronic identification information collection strategy; The configuration parameters of the electronic identification information collection strategy are updated based on the adjustment amount of the first parameter; The state vector is input into the feedback learning mechanism again to obtain the second parameter adjustment of the path planning decision model; The internal parameters of the path planning decision model are updated based on the adjustment amount of the second parameter.

10. A wayfinding system for activity community facilities based on electronic building signage, characterized in that, include: The data acquisition module is used to obtain the initial real-time population flow status within the community through a preset basic collection cycle; The heat assessment module is used to input the initial real-time pedestrian flow status and community physical space layout data into the facility heat assessment model to calculate the facility's heat level and spatial heat distribution. The strategy generation module is used to generate an electronic identification information collection strategy based on the heat level and spatial heat distribution, with the goal of maximizing data collection coverage and minimizing collection latency and construction costs, under the constraints of community physical space and network topology. The asset construction module is used to collect data based on the electronic identification information collection strategy to obtain initial facility data assets. The cleaning and verification module is used to obtain the real-time historical population flow status of the community corresponding to the time of collection of the initial facility data assets, and to clean and verify the initial facility data assets through an anomaly detection algorithm to obtain valid facility data assets. The route planning module is used to input the effective facility data assets into the route planning decision model, solve the multi-objective optimization problem within the boundaries of community operation constraints, and generate facility guidance and orientation schemes. The evaluation and analysis module is used to evaluate the guidance effect of the facility guidance scheme based on preset data asset utility evaluation indicators, and to use attribution analysis to identify the data feature dimensions that contribute the most to the generation of the facility guidance scheme. The feedback adjustment module is used to adjust the configuration parameters of the electronic identification information collection strategy and the internal parameters of the path planning decision model based on the results of the guidance effect evaluation and the data feature dimensions, using a preset feedback learning mechanism.

Images