Intelligent label dynamic guidance interaction management system
By constructing a smart signage dynamic guidance and interactive management system, and utilizing data collection and path optimization modules, weak connection areas are identified and optimized, enabling adaptive guidance of the smart signage system. This solves the path deviation and congestion problems of the existing system during peak traffic hours, and improves navigation accuracy and user experience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-19
- Publication Date
- 2026-03-24
AI Technical Summary
Existing smart signage systems lack the ability to perceive and adjust to changes in the on-site environment in real time, resulting in an inability to effectively divert or guide traffic during peak hours or emergencies. The recommended routes deviate from the actual traffic flow, and the lack of a global route optimization and linkage mechanism can easily lead to route conflicts or local congestion.
A smart signage dynamic guidance and interactive management system is constructed, including a data acquisition module, a navigation grid division module, a weak area identification module, a parameter assignment module, and a path optimization module. Through real-time pedestrian flow data analysis, weak connection areas are identified and guidance constraint parameters are assigned. The weights of path nodes are iteratively optimized to achieve adaptive path guidance.
It enables dynamic response and global coordination of pedestrian behavior, improving navigation accuracy and traffic efficiency, reducing path conflicts and congestion, and enhancing user experience.
Smart Images
Figure CN121725656A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of signage management technology, specifically to a smart signage dynamic guidance and interactive management system. Background Technology
[0002] The management system is a digital platform that provides real-time path guidance, information display, and interactive services for public places (such as airports, shopping malls, hospitals, parks, subway stations, etc.). The Smart-Signage system integrates screen display, sensors, network communication, location awareness, and content management platform to achieve real-time push, precise targeting, and user behavior analysis of information content, thereby realizing dynamic guidance and interactive management functions.
[0003] The existing technology has the following drawbacks: 1. The existing management system uses preset fixed paths and display content, lacking the ability to perceive and adjust to changes in the on-site environment in real time. This results in the inability to effectively divert or guide traffic during peak hours or emergencies, affecting traffic efficiency and user experience. At the same time, it lacks the means to model the characteristics of crowd behavior and cannot analyze the direction, density and trend of crowd flow based on real-time data, causing the recommended path to deviate from the actual movement and reducing navigation accuracy. 2. Since sign nodes usually operate independently and lack a unified path optimization and linkage mechanism, it is impossible to dynamically coordinate guidance strategies on a global scale, which can easily lead to path conflicts or local congestion.
[0004] Based on this, the present invention proposes a smart signage dynamic guidance and interactive management system, which constructs an intelligent interactive management mechanism centered on pedestrian behavior, realizing the transformation and upgrading of the signage system from a static guidance tool to an adaptive guidance platform. Summary of the Invention
[0005] The purpose of this invention is to provide a smart signage dynamic guidance and interactive management system to address the shortcomings of the prior art.
[0006] To achieve the above objectives, the present invention provides the following technical solution: a smart signage dynamic guidance and interactive management system, comprising a data acquisition module, a navigation grid division module, a weak area identification module, a parameter assignment module, a path optimization module, and an output module; Data acquisition module: Collects real-time pedestrian flow data and behavioral data, and constructs a probability density model of pedestrian flow direction; Navigation grid division module: Constructs a local coordinate system for each smart sign node in the site, dividing the entire building space into a non-uniform navigation grid; Weak Area Identification Module: Analyzes channel connectivity in the navigation mesh, identifies abnormal channel areas, and marks them as weakly connected areas; Parameter assignment module: Assigns guiding constraint parameters to each path node in the weakly connected region; Path optimization module: Calculates the congestion coefficient of weakly connected regions, determines whether the congestion coefficient exceeds the coefficient threshold, and if it does, iteratively corrects the weights of path nodes in weakly connected regions based on the guiding constraint parameters. Output module: When the congestion coefficient of all weakly connected areas is not higher than the coefficient threshold, the current iteratively optimized global path guidance scheme is output to all smart signage displays.
[0007] In a preferred embodiment, the parameters assign the module to receive a set of weakly connected regions. and its weak area score For each path node located in the weak region Generate vectors: The initial priority of nodes is set using a logarithmic probability function. The weight decay coefficient is calculated based on the real-time congestion sensitivity of nodes. Based on the fluctuation pattern of pedestrian flow in different time periods, time period weights are introduced to calculate the time window adaptation coefficient of nodes. Generate a triplet of guiding constraint parameters based on the initial priority, weight decay coefficient, and time window adaptation coefficient of the path nodes. .
[0008] In a preferred embodiment, a logarithmic probability function is used to set the initial priority of nodes. , represented as: , in the formula, This indicates the score of the weak region where the node is located. This is the threshold offset. Used to control the slope of the curve; Path node Generate guiding constraint parameter triplet , represented as: , This is the weight decay coefficient. This is the time window adaptation coefficient.
[0009] In a preferred embodiment, the weak area identification module reads the navigation mesh and identifies each passable edge. Calculate weighted betweenness centrality; Based on real-time pedestrian flow prediction, the expected pedestrian flow per unit time is calculated for each edge. Combined with static traffic capacity Define the congestion risk coefficient when This indicates that demand is overloaded. These are potential congestion areas; Extract the minimum effective passage width of the structurally restricted area With longitudinal slope Construct a geometric decay function; The three types of factors—weighted betweenness centrality, congestion risk coefficient, and geometric decay function—are linearly normalized and combined to form the weak area score. When the weak area score is greater than or equal to the diagnostic threshold, a passable edge is determined. The channel in question is a weakly connected region.
[0010] In a preferred embodiment, the weak region identification module targets each access edge. Calculate weighted betweenness centrality: ,in, For passage edge Weighted betweenness centrality, For node pairs The number of shortest paths between them. For the process of The number of paths, The edge weights are oriented by preference.
[0011] In a preferred embodiment, the navigation grid partitioning module reads the probability density model of pedestrian flow direction. ,in Indicates the local azimuth angle. Calculate the spatial coordinates of each smart sign node. The dominant pedestrian flow direction vector: A density tensor is defined over a continuous domain to adjust the shape and size of the mesh cells; Under density tensor constraints, for node sets Anisotropic Delaunay triangulation is performed to obtain an initial triangulation network, followed by topology optimization iteration.
[0012] In a preferred embodiment, the topology optimization iteration includes long-side contraction, short-side merging, and directional consistency adjustment; Long side contraction: If a certain edge geodetic length Greater than the second length threshold If so, insert the midpoint and triangulate again; Short edge merging: If a certain edge geodetic length Less than the first length threshold Merge nodes and update adjacent connections; Orientation consistency adjustment: the angle between the detection unit normal vector and the continuous orientation field. ,like , If the master axis is aligned with the tolerance angle, then the elements are rotated or re-divided to obtain a non-uniform and anisotropic navigation mesh. , where nodes Includes grid vertices and sign nodes, edges Indicates a passageway.
[0013] In a preferred embodiment, the data acquisition module constructs a probability density model of pedestrian flow direction to describe the statistical distribution of pedestrian movement direction within the area; Let a certain area Within the time window Internal observation The trajectory of the crowd is denoted as follows: Calculate the main direction unit vector within the region. ; A probability density model of pedestrian flow direction is established using the directional kernel density estimation method. Used to describe a person in that area at an angle The probability distribution of movement; Construct a spatial access relationship diagram ,in Represents a set of POI nodes. Represents the set of paths between connected nodes, each edge Assign weights , representing the density of people moving from POI node i to POI node j per unit time; Based on the direction probability density model Spatial Travel Map Generate a probability density model of pedestrian flow direction that includes global and local movement trends.
[0014] In a preferred embodiment, a direction kernel density estimation method is used to establish a probability density model of pedestrian flow direction. The expression is: ,in, For the first The direction and angle of the trajectory To smooth bandwidth, For kernel function, denoted by azimuth, and n represents the number of crowd trajectories.
[0015] In a preferred embodiment, the principal direction unit vector within the acquired region is calculated, expressed as: In the formula, The main direction is the unit vector, and n represents the number of crowd trajectories. This represents the direction vector of the i-th crowd trajectory. Let represent the norm of the direction vector of the i-th crowd trajectory.
[0016] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention constructs a local coordinate system for each smart sign node in a location through a navigation grid partitioning module, dividing the entire building space into a non-uniform navigation grid. A weak-connection identification module analyzes channel connectivity within the navigation grid, identifying and marking abnormal channel areas as weakly connected regions. A parameter assignment module assigns guidance constraint parameters to each path node in the weakly connected regions. A path optimization module calculates the congestion coefficient of the weakly connected regions and determines whether the congestion coefficient exceeds a threshold. If it does, the weights of the path nodes in the weakly connected regions are iteratively adjusted based on the guidance constraint parameters. When the congestion coefficients of all weakly connected regions are found to be below the threshold, the output module outputs the currently iteratively optimized global path guidance scheme to all smart signs. This management system systematically enhances existing smart signage systems in terms of data perception, dynamic modeling, path optimization, and distributed collaboration. It not only solves core problems such as the inability of traditional technologies to dynamically respond, lack of global coordination, and delayed information updates, but also constructs an intelligent interactive management mechanism centered on pedestrian flow, realizing the transformation and upgrading of the signage system from a static guidance tool to an adaptive guidance platform. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.
[0018] Figure 1 This is a flowchart of the present invention.
[0019] Figure 2 This is a diagram of the architecture of the present invention. Detailed Implementation
[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0021] Example: Please refer to Figure 1-2 As shown in the figure, the intelligent signage dynamic guidance and interactive management system described in this embodiment includes a data acquisition module, a navigation grid division module, a weak area identification module, a parameter assignment module, a path optimization module, and an output module; Data Acquisition Module: Acquires architectural structure diagrams, functional area division information, and real-time pedestrian flow and behavior data collected via cameras, Wi-Fi, Bluetooth, or pedestrian sensors. This data is analyzed to extract the main direction of pedestrian movement, pedestrian density distribution, and the passage relationships between points of interest (POIs), constructing a pedestrian flow direction probability density model. This model is then sent to the navigation grid division module.
[0022] The navigation grid division module constructs a local coordinate system for each smart sign node in the site based on a pedestrian flow direction probability density model. Using these local coordinate systems as a foundation, the entire building space is divided into a non-uniform navigation grid, forming a path guidance map with directional preferences. Reflecting actual pedestrian flow and spatial accessibility, the navigation grid is then sent to the weak area identification module.
[0023] Weak Area Identification Module: Analyzes channel connectivity within the navigation grid to identify abnormal (prone to congestion, narrow spaces, weak diversion capacity) channel areas, such as stairwells and narrow corridors. These abnormal channel areas are marked as weakly connected areas and become key monitoring targets for subsequent dynamic management. The weakly connected areas are then sent to the parameter assignment module and the path optimization module.
[0024] The parameter assignment module assigns initial guidance constraint parameters to each path node in the weakly connected region. These parameters include path display priority, weight decay rate (reducing guidance priority when congestion trends occur), and time window adaptation coefficient. These parameters determine the change strategy of the recommended path by the smart signage under specific conditions. The guidance constraint parameters are then sent to the path optimization module.
[0025] The route optimization module calculates the congestion coefficient of weakly connected areas as a dynamic indicator of traffic pressure. This coefficient provides real-time feedback on the traffic status of each area, directly influencing the judgment criteria for subsequent route guidance adjustments, specifically determining whether the congestion coefficient exceeds a threshold. If it does, the weights of path nodes in the weakly connected areas are iteratively adjusted according to preset guidance constraint parameters, increasing the priority of recommended alternative or detour routes. The content displayed on smart signs is updated in real-time based on this, achieving proactive pedestrian flow diversion and route optimization. The congestion coefficient and optimization results are then sent to the output module.
[0026] Output module: When the congestion coefficient of all weakly connected areas is detected to be no higher than the coefficient threshold, the currently iteratively optimized global path guidance scheme is output to all smart signs. The smart signs synchronously update their display content, showing on-site users the optimal navigation route, estimated travel time, accessibility warnings, or emergency avoidance information, ensuring information consistency and response speed.
[0027] The workflow of the management system is as follows: Acquire architectural structural diagrams and functional area divisions of the site, as well as real-time pedestrian flow and behavior data collected via cameras, Wi-Fi, Bluetooth, or pedestrian sensors. Analyze this data to extract the main direction of pedestrian movement, pedestrian density distribution, and the passage relationships between points of interest (POIs), and construct a probability density model of pedestrian flow direction.
[0028] Based on a pedestrian flow direction probability density model, a local coordinate system is constructed for each smart sign node in the site. Using these local coordinate systems as a foundation, the entire building space is divided into a non-uniform navigation grid, forming a path guidance map with directional preferences. This reflects the actual pedestrian flow and spatial accessibility.
[0029] Analyze channel connectivity within the navigation grid to identify anomalous channel areas (prone to congestion, narrow spaces, weak diversion capacity), such as stairwells and narrow corridors. Mark these anomalous channel areas as weakly connected areas for focused monitoring in subsequent dynamic management.
[0030] Each path node in a weakly connected region is assigned initial guidance constraint parameters, including path display priority, weight decay rate (reducing guidance priority when congestion trends occur), and time window adaptation coefficient. These parameters determine the change strategy of the smart signage's recommended path under specific conditions.
[0031] The congestion coefficient of weakly connected areas is calculated as a dynamic indicator to measure traffic pressure. This coefficient provides real-time feedback on the traffic status of each area and directly affects the judgment criteria for subsequent route guidance adjustments, determining whether the congestion coefficient exceeds a threshold. If it does, the weights of path nodes in the weakly connected areas are iteratively adjusted according to preset guidance constraint parameters, increasing the priority of recommended alternative or detour routes. The content displayed on smart signage is updated in real-time based on this, achieving proactive pedestrian flow diversion and route optimization.
[0032] When the congestion coefficient of all weakly connected areas is detected to be no higher than the coefficient threshold, the currently iteratively optimized global path guidance scheme is output to all smart signs. The smart signs simultaneously update their display content, showing on-site users the optimal navigation route, estimated travel time, accessibility warnings, or emergency avoidance information, ensuring information consistency and response speed.
[0033] This application constructs a local coordinate system for each smart sign node in a location through a navigation grid division module, dividing the entire building space into a non-uniform navigation grid. A weak-connection area identification module analyzes channel connectivity within the navigation grid, identifying and marking abnormal channel areas as weakly connected areas. A parameter assignment module assigns guidance constraint parameters to each path node in the weakly connected areas. A path optimization module calculates the congestion coefficient of the weakly connected areas and determines whether the congestion coefficient exceeds a threshold. If it does, the weights of the path nodes in the weakly connected areas are iteratively adjusted based on the guidance constraint parameters. When the congestion coefficients of all weakly connected areas are found to be below the threshold, the output module outputs the currently iteratively optimized global path guidance scheme to all smart signs. This management system systematically enhances existing smart signage systems in terms of data perception, dynamic modeling, path optimization, and distributed collaboration. It not only solves core problems such as the inability of traditional technologies to dynamically respond, lack of global coordination, and delayed information updates, but also constructs an intelligent interactive management mechanism centered on pedestrian flow, realizing the transformation and upgrading of the signage system from a static guidance tool to an adaptive guidance platform.
[0034] The data acquisition module obtains the building structure diagram and functional area division information of the venue, as well as real-time pedestrian flow and behavior data collected through cameras, Wi-Fi, Bluetooth, or pedestrian sensors. This data is analyzed to extract the main direction of pedestrian movement, pedestrian density distribution, and the passage relationships between points of interest (POIs), constructing a pedestrian flow direction probability density model. This model is then sent to the navigation grid division module.
[0035] The data acquisition module, as a foundational subsystem of the smart signage dynamic guidance and interaction management system, is responsible for building a realistic and computable foundation for crowd behavior modeling. This module mainly includes the following key steps, forming a complete scene perception and data transformation process: First, the system accesses the site's architectural structure data and functional area division information, including floor plans, passageways, staircases, elevators, entrance and exit locations, and semantic tags for each functional area (such as shops, waiting areas, and service counters). This structured data provides a static framework for subsequent spatial modeling and path accessibility analysis.
[0036] Secondly, multimodal sensing devices such as cameras, Wi-Fi probes, Bluetooth beacons, and people counting sensors distributed throughout the venue are used to collect real-time people flow and behavioral data. This data includes parameters such as changes in the number of people in different areas per unit time, individual movement trajectories, dwell time, and movement speed. To ensure the real-time nature and accuracy of the data, the system employs edge computing nodes to perform rapid preprocessing and anonymization of the raw video stream or signal data, achieving efficient computation while ensuring privacy compliance.
[0037] Next, based on the collected pedestrian trajectory and density data, the system constructs a probability density model of pedestrian movement directions to describe the statistical distribution of pedestrian movement directions within a specific area. This model is based on spatial units to construct a local direction field. Let a certain spatial region be considered. Within the time window Internal observation The trajectory of the crowd is denoted as follows: Then the principal direction unit vector within this region... It can be calculated using the following formula: Based on this, a probability density model of pedestrian flow direction is established using the directional kernel density estimation method. Used to describe a person in that area at an angle The probability distribution of movement. The formula is as follows: ,in, For the first The direction and angle of the trajectory To smooth bandwidth, The kernel function (commonly a von Mises or Gaussian kernel) smooths the local clustering characteristics of pedestrian flow, resulting in a continuous probability distribution. Furthermore, to represent the connectivity and traffic intensity between different Points of Interest (POIs), the system constructs a spatial traffic relationship graph. ,in Represents a set of POI nodes. Represents the set of paths between connected nodes, each edge Assign weights , representing the population density moving from POI node i to POI node j per unit time.
[0038] Finally, based on the directional probability density model Spatial Travel Map The system generates a probability density model of pedestrian flow direction that includes global and local movement trends, and inputs this model as a data structure and parameter set into the navigation grid partitioning module. This process ensures that the subsequent navigation network construction has a high degree of fit to actual pedestrian behavior, thus laying a data foundation for achieving efficient and adaptive path guidance.
[0039] The navigation grid division module constructs a local coordinate system for each smart sign node in the site based on a pedestrian flow direction probability density model. Using these local coordinate systems as a foundation, the entire building space is divided into a non-uniform navigation grid, forming a path guidance map with directional preferences. Reflecting actual pedestrian flow and spatial accessibility, the navigation grid is then sent to the weak area identification module.
[0040] The system first reads the "probability density model of pedestrian flow direction" output by the data acquisition module. ,in Indicates the local azimuth angle. Let be the spatial coordinates of each smart sign node. (common Calculate the expected vector of the directional distribution within the neighborhood of this node: ,in, From angle The corresponding unit vector represents the node Center, radius The local neighborhood, The probability density model for human flow direction, where Indicates the local azimuth angle. Let be the spatial coordinate point, then, with As a local Axis, combined with the architectural horizontal plane method and orthogonalization process, generates a three-dimensional local coordinate base. This basis provides directional constraints for subsequent anisotropic mesh generation. For nodes The dominant direction vector of pedestrian flow; From angle The corresponding unit vector; Represented by node Center, radius The local neighborhood.
[0041] To make the grid cells finer along the main pedestrian flow direction and coarser in the vertical direction, the system defines a density tensor in the continuous domain: ,in For unit tensors, Control the baseline grid size, Controlling the intensity of directional anisotropy To interpolate from each The resulting continuous directional field, express The transpose of . This tensor is used as a Riemannian metric in subsequent mesh generation algorithms to adjust the shape and size of mesh cells. The smaller the value, the denser the grid. As a direction-weighted factor, it determines the tendency for subdivision along the d direction. Anisotropic metric tensor used for mesh generation.
[0042] Further utilize the pedestrian flow density function (Depend on (Marginalization) Constructing the guiding potential field: , among which, when At higher levels, The smaller node size prompts sampling algorithms (improved Poisson-Disk or Lloyd iterations) to generate more nodes in high-traffic areas, thereby improving path resolution. The sampling results form the initial node set. ,in, The density of people per unit area This is a density weighting coefficient that controls the density level in high-density areas. Let be the sampling potential function for the node.
[0043] Local measurement Under constraints, for the node set Anisotropic Delaunay triangulation is performed to obtain the initial triangulated network. Then, topology optimization iterations are applied. Long side contraction: If a certain edge geodetic length Greater than the second length threshold If so, insert the midpoint and triangulate again; Short edge merging: If a certain edge geodetic length Less than the first length threshold Merge nodes and update adjacent connections; Directional consistency adjustment: Adjusting the element normal vector and local... The included angle If the test is successful, , If the principal axis is aligned with the tolerance angle, then the elements are rotated or re-divided to ensure the principal axis of the mesh is aligned with the direction of pedestrian flow. This results in a non-uniform, anisotropic navigation mesh. , where nodes Includes grid vertices and sign nodes, edges Indicate the passable passage and input the directional preference weights. This is for use in subsequent path searches.
[0044] The system will assign each edge Related attribute collection ,in For directional preference weights, This refers to the static passable capacity. The geometric length is specified. All attributes are serialized using Protocol-Buffer or GeoISON format, written to the buffer along with the complete topology structure, and sent to the weak connection region identification module for bottleneck analysis. At this point, the navigation mesh partitioning module completes the semantic discretization of the space, ensuring that subsequent algorithms can perform weak connection region identification and dynamic guidance decisions while balancing computational efficiency and behavioral realism.
[0045] The weak connectivity identification module analyzes channel connectivity within the navigation grid, identifying abnormal (prone to congestion, narrow spaces, weak diversion capacity) channel areas, such as stairwells and narrow corridors. These abnormal channel areas are marked as weak connectivity areas and become key monitoring targets for subsequent dynamic management. The weak connectivity areas are then sent to the parameter assignment module and the path optimization module.
[0046] The weak area identification module first reads the navigation mesh. For each passage edge Calculate weighted betweenness centrality: ,in, For node pairs The number of shortest paths between them. For the process of The number of paths, The weights of edges are determined by their direction preference. Edges with high betweenness centrality play a more critical role in converging traffic in the global path, and can become bottlenecks if they are blocked.
[0047] Based on real-time pedestrian flow prediction, the expected pedestrian flow per unit time is calculated for each edge. Combined with static traffic capacity (Derived from factors such as channel width and permissible traffic speed), define the congestion risk coefficient: ,when This indicates that demand is overloaded; if This is a potential congestion area.
[0048] For structurally restricted areas such as stairwells, corners, and narrow corridors, extract their minimum effective passage width. With longitudinal slope Construct the geometric decay function: ,in This is an empirical coefficient. If... Decrease or increase slope A tendency toward 1 amplifies risk assessment; conversely, a tendency toward 0.
[0049] The three types of factors—weighted betweenness centrality, congestion risk coefficient, and geometric decay function—are linearly normalized and combined to form the weak zone score: Among them, weight It is trained using historical operational data. Let the system diagnostic threshold be set. ,when When, that is, determine the edge The passage in question is located in a weakly connected region. Furthermore, to avoid spatial fragmentation, a connected component expansion algorithm is used to aggregate adjacent high-risk edges into contiguous weak blocks.
[0050] A unique ID is assigned to the identified weak blocks, and the field is written into the grid metadata: weak_zone={ , , , }. `weak_zone` is the weak block ID, where each... The item represents the average value of indicators within the block, facilitating rapid retrieval and parameterization operations by subsequent modules. Finally, the weak area identification module sends the updated navigation grid containing these markers to the parameter assignment module and the path optimization module, enabling real-time monitoring of bottleneck channels and prior input for dynamic guidance strategies.
[0051] The parameter assignment module assigns initial guidance constraint parameters to each path node in the weakly connected region. These parameters include path display priority, weight decay rate (reducing guidance priority when congestion trends occur), and time window adaptation coefficient. These parameters determine the change strategy of the recommended path by the smart signage under specific conditions, and the guidance constraint parameters are sent to the path optimization module.
[0052] First, receive the set of weak connection regions output by the weak connection region identification module. and its weak area score For each path node located in the weak region The following characteristics are statistically analyzed: static capacity Node degree Average congestion risk (take adjacent edge) The weighted average of the vector and geographic geometric parameters (width, slope, etc.) constitute a vector: ,in, This is the mean of the geometric decay function. This vector provides input for subsequent parameter calculations. To ensure that bottleneck nodes remain recommended for visibility under normal traffic conditions, while rapidly reducing guidance strength during congestion, the system uses a log-odds function to set the initial priority of nodes. , in the formula This indicates the score of the weak region where the node is located. This is the threshold offset. Controls the slope of the curve. Function description: When At a lower level (minor risk), When the value is close to 1, the node maintains a high level of guidance and exposure; when the risk increases, Rapidly decreasing allows for dynamic traffic diversion.
[0053] A weight decay coefficient is assigned to each node based on its real-time congestion sensitivity: ,in and These represent the minimum and maximum decay rates configured for the system, respectively. The fractional terms in this formula ensure that under low load conditions... near To maintain path stability; when hour, This ensures that the priority is rapidly reduced under high pressure, enabling rapid diversion.
[0054] Considering the fluctuation patterns of pedestrian flow in different time periods, time period weights are introduced. (Derived from historical statistics). The time window fitness coefficient of a node is defined as: ,in Adjusting time-period sensitivity, This allows low-risk nodes to maintain their parameters for a longer period during off-peak hours. Function Description: Determines how often the system triggers the next round of parameter updates: peak periods big, Smaller, more frequent updates; conversely, during off-peak periods, fewer unnecessary adjustments.
[0055] Ultimately, it becomes a node. Generate guiding constraint parameter triples: And write it to the distributed parameter cache. The module uses a publish-subscribe mechanism to... The data is pushed to the path optimization module, thus decoupling it from the real-time path calculation kernel. The path optimization module operates based on the following criteria at runtime: The node priority is exponentially decayed, and at intervals... When the deadline is reached or a significant change in congestion is detected, parameters are retrieved again to complete closed-loop control.
[0056] Through the above process, the parameter assignment module establishes a guiding constraint system for weakly connected regions that reflects risk differences and can adapt to changes in time and load, ensuring that subsequent path optimization strategies are both accurate and stable.
[0057] The route optimization module calculates the congestion coefficient of weakly connected areas as a dynamic indicator of traffic pressure. This coefficient provides real-time feedback on the traffic status of each area, directly influencing the judgment criteria for subsequent route guidance adjustments, specifically determining whether the congestion coefficient exceeds a threshold. If it does, the weights of path nodes in the weakly connected areas are iteratively adjusted according to preset guidance constraint parameters, increasing the priority of recommended alternative or detour routes. The content displayed on smart signs is updated in real-time based on this, achieving proactive pedestrian flow diversion and route optimization. The congestion coefficient and optimization results are then sent to the output module.
[0058] The module first receives the set of weak connection edges from the weak region identification module. And instantaneous passenger flow from the data collection layer (persons / second). The static capacity provided by the module is determined by the parameters. With node weight decay coefficient Define the congestion coefficient for weakly connected edges: ,in For the index of the join node. When and hour, This indicates an early overload warning, facilitating proactive system response. Parameter Description Real-time via edge The flow of people; The maximum designed throughput of this passage at a safe walking speed; The congestion sensitivity coefficient of a node / edge is given by the parameter assignment module; : Comprehensive congestion coefficient, measuring the level of traffic pressure. Step 2: Threshold determination and risk classification system sets three-stage thresholds. .when This indicates entry into the congestion control domain. Further, the Sigmoid function is used to calculate a continuous risk score: ,in This represents the slope coefficient. Risk score. It plays a gain role in subsequent weight updates, making weight adjustments more sensitive to high-pressure edges. For edges exceeding the threshold, the system adjusts their guiding priority according to an exponential decay model. ,in Display weights for paths. This represents the global decay rate. Function description: The more severe the congestion, the higher the decay rate. The closer the value is to 1, the faster the decay; conversely, for mild congestion, the decay is slower to maintain guidance stability. At the same time, the system considers alternative path edges. in accordance with: , to perform symmetrical lifting, where To increase the coefficient, ensure the overall weight is maintained, and encourage diversion.
[0059] After the weight matrix is updated, the module is in the navigation grid. Executing the dynamic Dijkstra-A* composite algorithm, with linear incremental complexity Internally, find the latest optimal path for each origin-end pair. Path cost function: Taking into account geometric distance with decayed weights This naturally avoids highly congested edges. After the algorithm is completed, the system generates a personalized recommended route and estimated travel time for each sign node. .
[0060] The route optimization module pushes update packages to the front-end rendering service via a message queue. These packages include the latest congestion heatmap, recommended route ID sequences for sign nodes, and corresponding text / graphic navigation instructions, traffic warnings, and detour suggestions. The front-end automatically adapts its display based on the hardware resolution and theme template, achieving seamless refresh on the device side (<200ms).
[0061] The module will , The calculated results are recorded in a time-series database and sent to the output module via a gRPC streaming interface. The output module then triggers control dashboards, maintenance alerts, and historical trend analyses, providing a complete data chain for subsequent decision-making. The path optimization module completes congestion detection, weight correction, and path recalculation within milliseconds, achieving adaptive traffic distribution and global navigation optimization for smart signage.
[0062] Specifically: The following are examples of human-computer interaction application scenarios in this application: When a user approaches the smart signage device, the signage activates the screen through its built-in near-field sensing module (such as infrared sensing or camera recognition) and prompts: "Please touch the screen for personalized navigation".
[0063] Users can directly select frequently used destination options (such as "waiting room", "restroom" and "luggage") on the touchscreen, or enter a specific gate number (such as "gate B18") via the on-screen keyboard or voice recognition.
[0064] After receiving input, the system invokes the route optimization module to calculate the optimal travel route based on the current pedestrian density and congestion level. This is then displayed in real-time via smart signage: a dynamic route map (dynamic arrow navigation), current and target location markers, estimated walking time and congestion level, and alternative route suggestions (e.g., the main route is highly congested).
[0065] If the system detects that the user's selected route passes through a less congested area, a helpful suggestion will pop up on the screen: "The main passage is currently quite congested. We recommend that you take the east passage to the waiting hall. It will only take 1 minute longer, but it will be smoother."
[0066] Users can choose to "accept the recommendation" or "stick to the original route," and the system will provide real-time feedback and update the route plan.
[0067] To accommodate diverse passenger needs, the system supports switching between Chinese, English, and graphic symbols. If a wheelchair user or visually impaired user is detected, the system can automatically switch to accessible pathway guidance (elevators instead of stairs, voice navigation, etc.).
[0068] Users can scan the QR code generated on the screen to synchronize the route plan to the mobile app / mini-program and receive real-time prompts during the journey, such as: "The passage ahead is congested, please pay attention to your speed" or "You are about to reach the destination and turn right."
[0069] The interactive advantages and technical value of this application are as follows: Real-time and personalized: Based on the system's dynamic modeling of pedestrian flow, it provides guidance solutions that best fit the actual situation on site for each interaction.
[0070] Distributed collaboration: Different sign nodes work together to ensure that the subsequent path guidance is consistent and globally optimal no matter where the user operates on the screen.
[0071] Behavioral learning feedback: The system can record user behavior (such as whether they accept the recommended path) for iterative optimization of the guidance strategy.
[0072] Enhanced experience and efficiency: Compared to static signage, this interactive method greatly improves the adaptability of spatial navigation and user satisfaction.
[0073] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.
[0074] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.
Claims
1. A smart signage dynamic guidance and interactive management system, characterized in that: It includes a data acquisition module, a navigation grid division module, a weak area identification module, a parameter assignment module, a path optimization module, and an output module; Data acquisition module: Collects real-time pedestrian flow data and behavioral data, and constructs a probability density model of pedestrian flow direction; Navigation grid division module: Constructs a local coordinate system for each smart sign node in the site, dividing the entire building space into a non-uniform navigation grid; Weak Area Identification Module: Analyzes channel connectivity in the navigation mesh, identifies abnormal channel areas, and marks them as weakly connected areas; Parameter assignment module: Assigns guiding constraint parameters to each path node in the weakly connected region; Path optimization module: Calculates the congestion coefficient of weakly connected regions, determines whether the congestion coefficient exceeds the coefficient threshold, and if it does, iteratively corrects the weights of path nodes in weakly connected regions based on the guiding constraint parameters. Output module: When the congestion coefficient of all weakly connected areas is not higher than the coefficient threshold, the current iteratively optimized global path guidance scheme is output to all smart signage displays.
2. The intelligent signage dynamic guidance and interactive management system according to claim 1, characterized in that: The parameters are assigned to the module to receive a set of weak connection regions. and its weak area score For each path node located in the weak region Generate vectors: The initial priority of nodes is set using a logarithmic probability function. The weight decay coefficient is calculated based on the real-time congestion sensitivity of nodes. Based on the fluctuation pattern of pedestrian flow in different time periods, time period weights are introduced to calculate the time window adaptation coefficient of nodes. Generate a triplet of guiding constraint parameters based on the initial priority, weight decay coefficient, and time window adaptation coefficient of the path nodes. .
3. The intelligent signage dynamic guidance and interactive management system according to claim 2, characterized in that: The initial priority of nodes is set using a logarithmic probability function. , is represented as: , in the formula, This indicates the score of the weak region where the node is located. This is the threshold offset. Used to control the slope of the curve; Path node Generate guiding constraint parameter triplet , is represented as: , This is the weight decay coefficient. This is the time window adaptation coefficient.
4. The intelligent signage dynamic guidance and interactive management system according to claim 3, characterized in that: The weak area identification module reads the navigation grid and identifies each passage edge. Calculate weighted betweenness centrality; Based on real-time pedestrian flow prediction, the expected pedestrian flow per unit time is calculated for each edge. Combined with static traffic capacity Define the congestion risk coefficient when This indicates that demand is overloaded. These are potential congestion areas; Extract the minimum effective passage width of the structurally restricted area With longitudinal slope Construct a geometric decay function; The three types of factors—weighted betweenness centrality, congestion risk coefficient, and geometric decay function—are linearly normalized and combined to form the weak area score. When the weak area score is greater than or equal to the diagnostic threshold, a passable edge is determined. The channel in question is a weakly connected region.
5. The intelligent signage dynamic guidance and interactive management system according to claim 4, characterized in that: The weak area identification module targets each passage edge. Calculate weighted betweenness centrality: ,in, For passage edge Weighted betweenness centrality, For node pairs The number of shortest paths between them. For the process of The number of paths, The edge weights are oriented by preference.
6. The intelligent signage dynamic guidance and interactive management system according to claim 5, characterized in that: The navigation grid division module reads the probability density model of pedestrian flow direction. ,in Indicates the local azimuth angle. Calculate the spatial coordinates of each smart sign node. The dominant pedestrian flow direction vector: A density tensor is defined over a continuous domain to adjust the shape and size of the mesh cells; Under density tensor constraints, for node sets Anisotropic Delaunay triangulation is performed to obtain an initial triangulation network, followed by topology optimization iteration.
7. The intelligent signage dynamic guidance and interactive management system according to claim 6, characterized in that: The application of topology optimization iteration includes long-side contraction, short-side merging, and directional consistency adjustment; Long side contraction: If a certain edge geodetic length Greater than the second length threshold If so, insert the midpoint and triangulate again; Short edge merging: If a certain edge geodetic length Less than the first length threshold Merge nodes and update adjacent connections; Orientation consistency adjustment: the angle between the detection unit normal vector and the continuous orientation field. ,like , If the master axis is aligned with the tolerance angle, then the elements are rotated or re-divided to obtain a non-uniform and anisotropic navigation mesh. , where nodes Includes grid vertices and sign nodes, edges Indicates a passageway.
8. The intelligent signage dynamic guidance and interactive management system according to claim 7, characterized in that: The data acquisition module constructs a probability density model of pedestrian flow direction, which is used to describe the statistical distribution of the movement direction of people in the area; Let a certain area Within the time window Internal observation The trajectory of the crowd is denoted as follows: Calculate the main direction unit vector within the region. ; A probability density model of pedestrian flow direction is established using the directional kernel density estimation method. Used to describe a person's angle in this area The probability distribution of movement; Construct a spatial access relationship diagram ,in Represents a set of POI nodes. Represents the set of paths between connected nodes, each edge Assign weights , representing the density of people moving from POI node i to POI node j per unit time; Based on the direction probability density model Spatial Travel Map Generate a probability density model of pedestrian flow direction that includes global and local movement trends.
9. The intelligent signage dynamic guidance and interactive management system according to claim 8, characterized in that: A probability density model of pedestrian flow direction is established using the directional kernel density estimation method. The expression is: ,in, For the first The direction and angle of the trajectory To smooth bandwidth, For kernel function, denoted by azimuth, and n represents the number of crowd trajectories.
10. The intelligent signage dynamic guidance and interactive management system according to claim 9, characterized in that: The principal direction unit vector within the obtained region is calculated using the following expression: In the formula, The main direction is the unit vector, and n represents the number of crowd trajectories. This represents the direction vector of the i-th crowd trajectory. Let represent the norm of the direction vector of the i-th crowd trajectory.