Park emergency navigation edge cooperation system and method based on building structure situation awareness

Abstract

The invention provides a park emergency navigation edge cooperation system and method based on building structure situation awareness, and relates to the technical field of path planning. The safety situation of the building structure is quantified for the first time and serves as a core optimization target which is the same as the passing time to be fused into path planning, personnel are prevented from being guided to a high-risk area of the structure fundamentally, and the essential safety of evacuation is remarkably improved. By exchanging decision factors among edge nodes and executing distributed multi-target planning, cross-regional path real-time collaboration under centerless scheduling is realized, coordination of a global path network in safety and efficiency is guaranteed, and robustness and autonomous decision ability of a system when a local network fails are improved.

Description

Technical Field

[0001] This invention relates to the field of path planning technology, and in particular to an emergency navigation edge collaboration system and method for parks based on building structure situation awareness. Background Technology

[0002] Current emergency navigation systems in industrial parks mostly adopt a centralized path planning architecture, relying on the transmission of multi-source sensor data for unified path planning. This results in high decision-making delays and a lack of local autonomy. In disaster scenarios, the system is prone to failure when the network is interrupted or bandwidth drops suddenly. Furthermore, existing systems fail to dynamically couple the building's structural safety status with path planning, making it difficult to cope with real-time changes in disaster scenarios. Summary of the Invention

[0003] The purpose of this invention is to provide an edge-coordinated emergency navigation system and method for parks based on building structure situation awareness, so as to improve the robustness and autonomous decision-making ability of the system in the event of local network failure, and at the same time improve the safety of evacuation routes.

[0004] In a first aspect, the present invention provides an edge-coordinated emergency navigation method for a park based on building structure situation awareness, comprising: real-time collection of building structure deformation data, environmental status data, and personnel distribution data through a multi-source sensor network deployed within the park, and spatiotemporal alignment processing of the collected data to obtain multi-source perception data; each edge node performing feature extraction processing on the multi-source perception data within its jurisdiction to obtain a local decision factor set; wherein, the local decision factor set includes risk factors for quantifying risks within the area; the risk factors include: a path safety factor determined based on building structure deformation data, and a passage efficiency factor determined based on environmental status data and personnel distribution data; each edge node exchanging its local decision factor set with other edge nodes through an inter-node communication protocol to obtain a global decision factor set; each edge node, based on the global decision factor set, executing a distributed multi-objective path planning algorithm with the parallel optimization objectives of maximizing path safety and minimizing passage time, to generate a navigation path coordinated with its neighboring edge nodes.

[0005] In an optional implementation, the collected data is subjected to spatiotemporal alignment processing, including: constructing a unified spatiotemporal coordinate system, unifying the time reference of the collected building structure deformation data, environmental status data, and personnel distribution data to a preset standard time, and unifying the spatial reference to the local coordinates of the park buildings; setting a sliding time window, and using a dynamic time warping algorithm within the sliding time window to compensate for the time sequence deviation of the data collected by sensors with different sampling frequencies, thereby obtaining multi-source sensing data.

[0006] In an optional implementation, each edge node performs feature extraction processing on the multi-source sensing data within its jurisdiction to obtain a local decision factor set, including: processing the building structural deformation data in the multi-source sensing data using a temporal neural network model to obtain the structural failure rate of each building within a specified future time period, and determining the path safety factor of the road segment to which each building belongs based on the structural failure rate; wherein, the higher the structural failure rate, the lower the path safety factor; constructing an environmental risk field based on the environmental state data in the multi-source sensing data to obtain an environmental impact coefficient reflecting the degree of hazard of smoke and temperature; identifying congested areas based on the personnel distribution data in the multi-source sensing data using a density clustering algorithm to obtain a personnel density coefficient reflecting traffic resistance; and calculating the traffic efficiency factor of each road segment by combining the environmental impact coefficient and the personnel density coefficient using a preset traffic speed model.

[0007] In an optional implementation, each edge node, based on a global decision factor set, executes a distributed multi-objective path planning algorithm with the parallel optimization objectives of maximizing path safety and minimizing travel time to generate a navigation path that is consistent with its neighboring edge nodes. This includes: constructing a multi-objective optimization function with the objectives of maximizing path safety and minimizing travel time; solving the multi-objective optimization function based on the distributed multi-objective path planning algorithm and the global decision factor set to obtain a set of non-dominated path solutions; determining candidate navigation paths from the set of non-dominated path solutions using a fuzzy decision method; and when a candidate navigation path determined by an edge node conflicts with a candidate navigation path planned by its neighboring nodes, initiating a consensus negotiation sub-process based on security proof, where relevant nodes exchange evidence supporting their path security and reach a consensus through quantitative evaluation to obtain the navigation path.

[0008] In an optional implementation, after generating a navigation path that is consistent with its neighboring edge nodes, the method further includes: monitoring the path status in real time during the guidance process based on the navigation path; and when a local path blockage is detected, triggering the affected edge nodes to update their local decision factor sets and replan the navigation path.

[0009] In an optional implementation, after generating a navigation path that is consistent with its adjacent edge nodes, the method further includes: converting the navigation path into multimodal navigation instructions; the multimodal navigation instructions include: lighting control instructions for physical layer guidance and mobile navigation instructions for digital layer guidance; sending the lighting control instructions to the intelligent emergency light array in the park to control the intelligent emergency lights to transmit navigation direction information through specific light intensity changes; and pushing the mobile navigation instructions to the mobile terminals of personnel in the park to provide real-time navigation guidance to personnel through the display interface of the mobile terminals.

[0010] In an optional implementation, the method further includes: the cloud platform receiving decision log data uploaded by each edge node to construct an emergency navigation scenario library for the park; based on the data in the scenario library, using reinforcement learning algorithms to optimize the model parameters in the multi-objective path planning algorithm; and distributing the optimized model parameters to each edge node to update the local multi-objective path planning algorithm of each edge computing node.

[0011] Secondly, this invention provides a park emergency navigation edge collaboration system based on building structure situation awareness, comprising: multiple edge nodes and a multi-source sensor network deployed within the park; the multi-source sensor network is used to collect building structure deformation data, environmental status data, and personnel distribution data in real time, and performs spatiotemporal alignment processing on the collected data to obtain multi-source perception data; each edge node is used to perform feature extraction processing on the multi-source perception data within its jurisdiction to obtain a local decision factor set; wherein, the local decision factor set includes risk factors for quantifying risks within the area; the risk factors include: a path safety factor determined based on building structure deformation data, and a passage efficiency factor determined based on environmental status data and personnel distribution data; each edge node is also used to exchange its local decision factor set with other edge nodes through an inter-node communication protocol to obtain a global decision factor set; each edge node is also used to execute a distributed multi-objective path planning algorithm based on the global decision factor set, with the goal of maximizing path safety and minimizing passage time in parallel optimization, to generate a navigation path that is coordinated with its neighboring edge nodes.

[0012] Thirdly, the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the park emergency navigation edge collaboration method based on building structure situation awareness as described in any of the foregoing embodiments.

[0013] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the edge collaboration method for emergency navigation in a park based on building structure situation awareness as described in any of the foregoing embodiments.

[0014] This invention extracts path safety factors and traffic efficiency factors from multi-source sensing data, and for the first time quantifies the structural safety status of buildings and integrates it into path planning as a core optimization objective of equal importance to travel time. This fundamentally avoids guiding personnel to structurally high-risk areas and significantly improves the inherent safety of evacuation. By exchanging decision factors among edge nodes and executing distributed multi-objective planning, real-time cross-regional path coordination is achieved without centralized scheduling. This ensures the consistency of the global path network in terms of safety and efficiency, and also improves the system's robustness and autonomous decision-making capabilities in the event of local network failures. Attached Figure Description

[0015] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 A flowchart of an edge-coordinated emergency navigation method for a park based on building structure situation awareness is provided in an embodiment of the present invention; Figure 2 A flowchart for generating a navigation path that is consistent with its adjacent edge nodes is provided in an embodiment of the present invention; Figure 3 A functional module diagram of a park emergency navigation edge collaboration system based on building structure situation awareness is provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0017] 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, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0018] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0019] The following detailed description of some embodiments of the present invention is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0020] Example 1 This invention addresses the core shortcomings of existing emergency navigation systems for industrial parks, such as over-reliance on centralized architecture, disconnect between path planning and building structural safety, and weak cross-regional collaboration capabilities. It proposes an integrated edge collaboration solution encompassing "perception-assessment-collaboration-decision-execution." The core idea is to bring traditional cloud-based intelligence down to the network edge, enabling edge computing nodes deployed across various areas of the park to make real-time autonomous decisions based on local multi-source situational awareness (especially building structural safety status). Through a lightweight collaboration mechanism, a globally consistent, safe, and efficient evacuation path network is formed without centralized scheduling.

[0021] Figure 1 A flowchart of an edge-coordinated emergency navigation method for a park based on building structure situation awareness is provided for an embodiment of the present invention, as shown below. Figure 1 As shown, the method specifically includes the following steps: Step S102 involves collecting building structure deformation data, environmental status data, and personnel distribution data in real time through a multi-source sensor network deployed in the park, and performing spatiotemporal alignment processing on the collected data to obtain multi-source sensing data.

[0022] The method provided in this invention relies on a multi-source sensor network deployed within the park. Specifically, high-precision strain sensors (such as fiber optic grating sensors and piezoelectric ceramic sensors) are deployed at key structural points of the building (such as load-bearing beams and column nodes) to monitor deformation and micro-strain in real time and collect building structural deformation data; infrared thermal imagers, gas sensors, and lidar (LiDAR) are deployed in passageways and open areas to monitor environmental conditions (temperature, smoke concentration, obstacles) and obtain environmental condition data; and UWB (ultra-wideband), Wi-Fi probes, cameras, and other devices are used to acquire personnel distribution data.

[0023] To address the fusion error caused by inconsistent spatiotemporal benchmarks and differences in sampling frequencies in the original data, and to provide highly consistent and fusionable input data for subsequent data processing, spatiotemporal alignment processing is performed on the original multi-source data (building structure deformation data, environmental status data, and personnel distribution data) after obtaining the original multi-source data, thereby obtaining multi-source sensing data.

[0024] In step S104, each edge node performs feature extraction processing on the multi-source sensing data within its jurisdiction to obtain a local decision factor set.

[0025] The local decision factor set includes risk factors used to quantify risks within the region; the risk factors include: path safety factors determined based on building structure deformation data, and traffic efficiency factors determined based on environmental status data and personnel distribution data.

[0026] Edge nodes refer to edge computing units with independent computing capabilities deployed in sub-regions of a park (such as buildings, floors, or functional areas), responsible for local data processing and preliminary decision generation. In this embodiment of the invention, feature extraction processing is not a simple numerical transformation, but rather relies on a specific model to map physical quantities to a risk dimension.

[0027] Specifically, path safety factors are extracted based on spatiotemporally aligned building structural deformation time-series data (such as strain, displacement, and vibration modes), and traffic efficiency factors are extracted by combining spatiotemporally aligned environmental state data and personnel distribution data. The local decision factor set is essentially a lightweight risk summary information package, containing two dimensions: path safety factors and traffic efficiency factors. It retains key risk characteristics while reducing communication overhead, making it suitable as a basic input for distributed collaboration.

[0028] In step S106, each edge node exchanges its local decision factor set with other edge nodes through the inter-node communication protocol to obtain the global decision factor set.

[0029] In this embodiment of the invention, the inter-node communication protocol refers to a lightweight communication mechanism suitable for resource-constrained and low-latency scenarios. Optionally, the inter-node communication protocol is the MQTT-SN (MQTT for Sensor Networks) protocol.

[0030] Edge nodes do not transmit raw sensor data; they only exchange local decision factor sets, significantly reducing bandwidth consumption and privacy risks. Obtaining the global decision factor set can be understood as acquiring a complete copy of the entire campus data, or as each edge node constructing an extended decision view covering its planned path range based on neighbor broadcast information, supporting cross-regional path coordination. This data exchange mechanism between edge nodes achieves a balance between "local autonomy" and "regional collaboration." Even if some links between edge nodes are interrupted, multi-hop forwarding can maintain a certain range of collaborative capabilities, enhancing system robustness.

[0031] In step S108, each edge node, based on the global decision factor set, executes a distributed multi-objective path planning algorithm with the parallel optimization objectives of maximizing path safety and minimizing travel time, to generate a navigation path that is consistent with its neighboring edge nodes.

[0032] In this embodiment of the invention, path planning is not a single-objective optimization, but a dual-objective trade-off: maximizing path safety allows for prioritizing the avoidance of high-risk areas such as structural instability and high-temperature smoke accumulation; minimizing travel time effectively reduces the total evacuation time and improves overall efficiency. The distributed multi-objective path planning algorithm refers to each edge node independently solving for its local optimal path without a central scheduler. Furthermore, this embodiment of the invention also considers the path schemes of its neighboring edge nodes, requiring that the navigation paths of adjacent edge nodes be coordinated and consistent to ensure that there are no navigation path conflicts.

[0033] This invention extracts path safety factors and traffic efficiency factors from multi-source sensing data, and for the first time quantifies the structural safety status of buildings and integrates it into path planning as a core optimization objective of equal importance to travel time. This fundamentally avoids guiding personnel to structurally high-risk areas and significantly improves the inherent safety of evacuation. By exchanging decision factors among edge nodes and executing distributed multi-objective planning, real-time cross-regional path coordination is achieved without centralized scheduling. This ensures the consistency of the global path network in terms of safety and efficiency, and also improves the system's robustness and autonomous decision-making capabilities in the event of local network failures.

[0034] In an optional implementation, step S102 above, which involves performing spatiotemporal alignment processing on the collected data, specifically includes the following steps: Step S1021: Construct a unified spatiotemporal coordinate system, unifying the time reference of the collected building structure deformation data, environmental status data, and personnel distribution data to a preset standard time, and unifying the spatial reference to the local coordinates of the park buildings.

[0035] Specifically, this step is a prerequisite for the fusion of multi-source heterogeneous data. It aims to eliminate the spatiotemporal benchmark differences caused by the dispersed and diverse deployment of sensors, and to ensure that subsequent feature extraction and collaborative analysis are comparable and consistent.

[0036] Constructing a unified spatiotemporal coordinate system refers to establishing a globally consistent reference framework that includes a unified time axis and spatial coordinate system: Unified Time Base: All sensor devices are synchronized to the same standard time source (i.e., the "preset standard time") via a network time synchronization protocol (such as NTP or PTP), such as UTC+8 Beijing time. This eliminates timing misalignment issues caused by local clock drift and startup delays of each device, ensuring that the timestamps of data reported by different nodes are horizontally comparable.

[0037] Unified Spatial Reference System: The physical locations of all sensors are mapped to the same georeferenced system—the "local coordinates of the park buildings." This coordinate system typically uses a fixed point within the park (such as the main control center or the origin of the first floor of a building) as its origin, and employs Cartesian three-dimensional rectangular coordinates (x, y, z) to represent spatial locations. Coordinate registration is then completed using the building's BIM model. For example, the location of a fiber optic strain sensor is labeled with the local coordinates of its corresponding load-bearing beam, and the UWB tag positioning results are also converted to this same coordinate system, thus achieving precise spatial correspondence.

[0038] Optionally, the multi-source data mentioned above must complete the above mapping before being connected to the edge node: 1. Associate the building structure deformation data with the spatial coordinates of specific components; 2. Bind the environmental status data (such as smoke concentration and temperature) to the spatial location of the monitoring point; 3. Map the personnel distribution data to the real-time coordinate trajectory of dynamically moving individuals in the coordinate system.

[0039] The above data processing flow solves the spatiotemporal fragmentation problem that is common in multi-source sensing systems, and provides a geometric and temporal alignment basis for cross-modal data association, which is a prerequisite for achieving accurate situational awareness.

[0040] Step S1022: Set a sliding time window, and within the sliding time window, use a dynamic time warping algorithm to compensate for the timing deviation of the data collected by sensors with different sampling frequencies to obtain multi-source sensing data.

[0041] Specifically, this step focuses on addressing the non-uniformity and asynchronicity of multi-source data over time, especially the information mismatch problem in scenarios where high-frequency and low-frequency sensors coexist.

[0042] Different sampling frequencies of sensors generate different amounts of data at the same time, and there is a slight offset in the event triggering time. Direct splicing will lead to timing misalignment. For example, the sampling rate of fiber optic grating sensors can reach 1kHz, the sampling rate of vibration accelerometers is above 100Hz, the sampling rate of infrared thermal imagers is 30fps, the sampling rate of UWB positioning modules is 10-50Hz, and the sampling rate of gas sensors is 1-10Hz.

[0043] Therefore, this embodiment of the invention sets a sliding time window and uses a dynamic time warping algorithm within the sliding time window to compensate for timing deviations in data collected by sensors with different sampling frequencies. Setting a sliding time window means defining a fixed-length time interval (e.g., 500ms) as the basic time unit for current data processing. This window continuously scrolls forward over time, incorporating only the most recently arrived data segment within the window for fusion calculation.

[0044] In this embodiment of the invention, the Dynamic Time Warping (DTW) algorithm is used to align response curves of similar events from different sensors (such as the initial temperature rise curve of a fire and the structural thermal expansion signal), or to assist in cross-modal data fusion (such as aligning the trend of video frame changes with the trend of smoke concentration increase). The algorithm operates on local data segments within a sliding time window, with controllable computational cost, making it suitable for deployment at edge nodes. The output is a synchronized data sequence after time elastic alignment, significantly reducing information fusion errors caused by differences in sampling frequency.

[0045] The final output multi-source sensing data refers to a structured data set that has completed spatiotemporal dual-dimensional calibration and has the following characteristics: 1. The same timestamp corresponds to the same physical moment; 2. The same spatial coordinate corresponds to the same geographical location; 3. Different modal data can be overlaid and analyzed on the same spatiotemporal network (such as risk heat map generation, path weight assignment, etc.).

[0046] This step effectively overcomes the information loss and noise amplification problems caused by traditional hard interpolation or downsampling methods, and improves the accuracy and robustness of multi-source data fusion, especially suitable for application scenarios of capturing transient features during disaster evolution.

[0047] In an optional implementation, step S104 above, in which each edge node performs feature extraction processing on the multi-source sensing data within its jurisdiction to obtain a local decision factor set, specifically includes the following steps: Step S1041: The building structure deformation data in the multi-source sensing data is processed using a temporal neural network model to obtain the structural failure rate of each building within a specified time period in the future, so as to determine the path safety factor of the road segment to which each building belongs based on the structural failure rate; wherein, the higher the structural failure rate, the lower the path safety factor.

[0048] Specifically, temporal neural network models refer to deep learning architectures suitable for modeling dynamic changes in time series, such as Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), or One-Dimensional Convolutional Neural Network (1D-CNN). The input data of the model is spatiotemporally aligned temporal data of building structural deformation, such as strain change sequences collected by fiber optic grating sensors or component displacement sequences fed back by displacement sensors.

[0049] The model learns historical deformation patterns (such as linear creep, accelerated degradation, and periodic fluctuations) to identify the current degradation stage of the structure and predicts the probability of local failure (such as yielding, fracture, or instability) within a specified timeframe (e.g., 3–5 minutes), i.e., the structural failure rate. The structural failure rate, as a key risk indicator, is further converted into a path safety factor for the corresponding spatial path segment. The conversion relationship is negatively correlated: a higher structural failure rate leads to a lower path safety factor. Specifically, a nonlinear decay function (such as exponential decay or sigmoid mapping) can be used for normalization, with the output value typically in the range [0, 1], facilitating subsequent optimization algorithms. The path safety factor is ultimately used as a weight term in the path planning objective function, directly influencing the preference for evacuation route selection.

[0050] Step S1042: Construct an environmental risk field based on the environmental state data in the multi-source sensing data to obtain environmental impact coefficients that reflect the degree of hazard from smoke and temperature.

[0051] Environmental status data includes real-time information collected by devices such as infrared thermal imagers, lidar, gas sensors, and temperature and humidity sensors. It mainly covers smoke concentration and its diffusion direction, areas of abnormally high local temperatures, and the accumulation of toxic and harmful gases. Constructing an environmental risk field involves generating a two-dimensional or three-dimensional spatial risk distribution map covering the entire area based on the aforementioned discrete observation point data, using interpolation or physical modeling methods. Optionally, spatial interpolation techniques such as inverse distance weighting (IDW) and Kriging interpolation can be used, or simplified fluid dynamics models (such as the finite volume method) can be combined to simulate smoke diffusion trends, thus constructing a directional and dynamically evolving vector risk field.

[0052] Within this risk field, each geographic grid cell is assigned a comprehensive risk score, from which an environmental impact coefficient is derived. This coefficient reflects the overall hazard level of the environment in which a particular route segment is located; a higher value indicates that the area is less suitable for passage. Specifically, weighted combination rules can be set according to hazard types, such as high temperature weight + smoke concentration weight + toxic gas level weight. The environmental impact coefficient supports dynamic updates over time to adapt to the disaster spread process.

[0053] Step S1043: Based on the personnel distribution data in the multi-source sensing data, the congested areas are identified through a density clustering algorithm to obtain the personnel density coefficient that reflects traffic resistance.

[0054] Personnel distribution data is sourced from UWB / BLE fusion positioning systems, video surveillance AI analysis modules, or other high-precision indoor positioning methods, providing real-time location coordinates for individuals or groups within the park. Density clustering algorithms are suitable for unsupervised clustering methods that discover densely populated areas of arbitrary shapes; optionally, improved density peak clustering (DPC) or DBSCAN algorithms can be used. These algorithms can automatically identify congested areas with high personnel density without pre-setting the number of clusters. Adaptive bandwidth kernel density estimation is supported to handle density variations at different spatial scales (e.g., corridor bottlenecks vs. open spaces), outputting several high-density clusters and their boundary ranges. Finally, the personnel density coefficient is calculated based on the identification results. The personnel density coefficient is defined as the population density per unit area, or the normalized relative congestion (e.g., the ratio relative to the maximum carrying capacity of the passageway).

[0055] Step S1044: Combine the environmental impact coefficient and the personnel density coefficient to calculate the traffic efficiency factor of each road segment using a preset traffic speed model.

[0056] The preset traffic speed model is an empirical or semi-physical mathematical model used to estimate the actual average movement speed of people on a specific road segment. Its input variables include: environmental impact coefficient (reflecting the influence of the external environment on the willingness to move, such as dense smoke reducing visibility and thus slowing down the pace) and personnel density coefficient (reflecting the degree of mutual interference between internal pedestrian flows, such as narrowing the stride and increasing the pauses under high density).

[0057] Optionally, the expression for the preset traffic speed model is: ,in, Indicates the expected traffic speed of the road segment. This indicates the free-flow velocity of a road segment under ideal conditions. The environmental impact function is typically a decay function, such as exponential decay. This indicates that the more severe the environmental hazard, the lower the effective passage speed. When the environmental impact coefficient... When a certain threshold is exceeded, The value may be close to 0, indicating that the road section is impassable. This represents the first preset coefficient. The congestion impact function is typically related to the population density coefficient. Related to the square or other powers, such as This reflects the nonlinear speed decrease caused by congestion. This represents the second preset coefficient.

[0058] For each road segment, the corresponding environmental impact coefficient and population density coefficient are substituted into the above model to calculate the expected traffic speed for that road segment. Finally, the expected travel speed is converted into a traffic efficiency factor. Optionally, the expected travel speed is normalized to keep the traffic efficiency factor and the path safety factor on similar orders of magnitude, facilitating subsequent multi-objective optimization calculations. A higher traffic efficiency factor indicates a faster expected travel time through the road segment.

[0059] In one alternative implementation, such as Figure 2 As shown, in step S108 above, each edge node, based on the global decision factor set, executes a distributed multi-objective path planning algorithm with the parallel optimization objectives of maximizing path safety and minimizing travel time, to generate a navigation path that is coordinated with its neighboring edge nodes. Specifically, this includes the following steps: Step S1081: Construct a multi-objective optimization function with the objectives of maximizing path safety and minimizing travel time.

[0060] Specifically, the goal of maximizing route safety is to guide evacuees away from high-risk areas of building structures. Its mathematical form is typically maximizing the overall route safety score. Specifically, for a route consisting of several segments... Its safety target value Calculated for all road segments on the path Path safety factor Weighted sum: .in, It is the quantization value determined by step S1041 (the higher the value, the safer). This is the weight of the road segment in the global path (which can be set based on segment length, importance, etc.). The goal is to make... Make it as large as possible.

[0061] Minimizing travel time: This objective aims to achieve rapid evacuation. Its mathematical form is typically minimizing the estimated total travel time. Specifically, for a path... Its time target value Calculated for all road segments Estimated travel time sum: .in, , For the length of the road segment, The expected traffic speed is derived from the traffic efficiency factor in step S1044. It is dynamically changing due to environmental factors and traffic congestion. The goal is to make... As small as possible.

[0062] These two objectives are often in conflict. The safest path may require a detour, increasing time; the fastest path may pass through high-risk or densely populated areas, reducing safety. Therefore, there is no single solution that is absolutely optimal in both respects, but rather a set of Pareto optimal solutions.

[0063] Step S1082: Based on the distributed multi-objective path planning algorithm and the global decision factor set, solve the multi-objective optimization function to obtain the non-dominated path solution set.

[0064] Distributed multi-objective path planning algorithms refer to the independent optimization algorithms run by each edge node without intervention from a central scheduler, based on its own acquired global decision factor set (i.e., incorporating risk information from itself and neighboring regions). Typical implementations include improved NSGA-II and MOEA / D evolutionary optimization algorithms, supporting efficient exploration of multi-objective solution spaces in complex graph structures. In the graph model, nodes represent spatial locations (such as passage intersections), and edges represent traversable path segments, each carrying two attributes: path safety factor and travel time cost. Within a finite number of iterations, the algorithm generates a set of candidate paths that cannot be completely dominated by each other, i.e., the non-dominated path solution set (Pareto-optimal Set). The non-dominated path solution set contains multiple candidate paths, each representing a different combination of safety-efficiency preferences.

[0065] Step S1083: Determine candidate navigation paths from the non-dominated path solution set using fuzzy decision-making method.

[0066] Fuzzy decision-making is an intelligent selection mechanism applicable to multi-criteria decision-making, capable of handling trade-offs between objectives that are difficult to quantify precisely. First, an evaluation index system is defined, typically including the overall safety score of the route, estimated travel time, and risk volatility (standard deviation of safety factors on the route). Then, fuzzy membership functions are used to normalize each index into a satisfaction score within the [0, 1] interval. Next, an adjustable weight vector is introduced to reflect strategy preferences under different disaster levels. For example, in the initial warning stage, emphasis is placed on travel efficiency to accelerate evacuation; in the severe danger stage, safety weights are significantly increased to ensure survival priority. Finally, a fuzzy comprehensive evaluation method (such as TOPSIS combined with fuzzy weighting) is used to calculate the comprehensive evaluation value of each route, and the route with the highest score is selected as the candidate navigation route.

[0067] Step S1084: When a candidate navigation path determined by an edge node conflicts with the candidate navigation paths planned by its neighboring nodes, a consensus negotiation sub-process based on security proof is initiated. The relevant nodes exchange evidence supporting the security of their paths and reach a consensus through quantitative evaluation to obtain the navigation path.

[0068] Typical scenarios of path conflicts include, for example, two adjacent nodes guiding people in their respective areas to converge at the same exit, exceeding the exit's capacity; node A recommending evacuation to the east, while node B simultaneously blocks the eastward passage for control operations; multiple recommended paths creating opposing crowd flows at an intersection, triggering a stampede risk.

[0069] Once such a conflict is detected (which can be identified by comparing information such as the endpoint, direction, and traffic prediction of candidate paths broadcast by neighboring nodes), a consensus negotiation sub-process based on security proofs is immediately triggered: 1. Evidence Exchange: Upon detecting conflicting nodes, a negotiation process is immediately initiated. This involves not simply exchanging the final path, but rather exchanging the key evidence supporting their path selection decisions—namely, proof of safety. This includes quantitative evidence such as detailed path safety factors for each critical segment of the path, underlying structural failure rate prediction data, and local snapshots of the environmental risk field.

[0070] 2. Quantitative Assessment: After receiving evidence from the other party, each node uses an objective quantitative assessment model to reassess the conflict path. A core assessment metric can be "security entropy": .in It can be the "risk probability distribution" of each segment of the path derived from the structural failure rate. Entropy value The lower the entropy value, the more concentrated the risk is on a few high-probability road segments (the path safety is more predictable, but potentially more dangerous); the higher the entropy value, the more dispersed the risk distribution (the path may be more balanced and safer). An objective comparison is made by calculating safety entropy and other indicators for both paths.

[0071] 3. Reaching Consensus: Based on the quantitative evaluation results, nodes reach a consensus by following pre-defined consensus rules. For example, the rule might be "choose a path with lower security entropy (i.e., less secure) to circumvent the restrictions, or negotiate a compromise path using a third party." This process borrows the idea of ​​Proof-of-Stake (PoS) in blockchain, but here it proves "security." The negotiation process has strict time limits (e.g., 200ms) to ensure real-time performance.

[0072] Once a consensus is reached, the relevant nodes adjust their candidate paths, ultimately generating a navigation path that seamlessly connects between areas, has no logical conflicts, and has been verified through secure consensus. This ensures that, from the perspective of the entire park, the evacuation guidance network is globally consistent and efficiently coordinated.

[0073] In one optional implementation, after generating a navigation path consistent with its adjacent edge nodes, the present invention further includes the following steps: Step S201: During the guidance process based on the navigation path, the path status is monitored in real time.

[0074] Step S202: When local path congestion is detected, the affected edge nodes are triggered to update their local decision factor sets and replan the navigation path.

[0075] The above steps aim to achieve continuous, closed-loop health monitoring of the planned and executed navigation paths. Its core objective is to ensure the effectiveness and safety of navigation commands remain up-to-date in the context of dynamically evolving disaster environments, thereby upgrading the system from static, one-off planning to dynamic, continuous protection.

[0076] The driving logic behind monitoring is that disaster scenarios are highly uncertain. Routes initially considered safe and unobstructed may become unusable within a short period due to fire spread, secondary collapses, accumulation of toxic gases, or sudden gatherings of people. Therefore, completing the planning and issuing instructions is only the beginning; a continuous perception and feedback mechanism for the execution environment must be established, which is a prerequisite for achieving dynamic adaptive navigation.

[0077] Optionally, the focus of monitoring is not the entire park, but rather currently active navigation paths and their adjacent buffer zones, to reflect the targeted and cost-effectiveness of monitoring. The following monitoring dimensions can be selected: 1. Physical accessibility: Using video analytics, lidar, or infrared beam sensors deployed on key passageways, determine in real time whether there are physical obstructions (such as falling objects or collapses) on the path.

[0078] 2. Environmental risk level: Utilize gas sensors and temperature sensors deployed along the route to continuously collect data and determine whether the smoke concentration and temperature exceed the safe passage threshold.

[0079] 3. Structural safety status: Strain sensor data from the path-related building structure parts are continuously input to determine whether new severe deformations occur or whether the danger warning threshold is reached.

[0080] 4. Personnel execution status: Analyze whether the actual flow of people is moving on the expected path through UWB positioning system or camera, and whether there are abnormal stops, turns, or abnormal increases in density. These phenomena may be indirect signals that there are obstacles ahead that are not directly perceived.

[0081] By performing real-time fusion analysis of monitoring data from the aforementioned different dimensions at the edge nodes, a comprehensive health status indicator for each active path is formed. The system sets clear status judgment thresholds. When monitoring data indicates that any dimension of a path segment (such as temperature > 70°C, sudden change in deformation rate, or video detection of obstruction) exceeds its safety or passability threshold, the path segment is determined to be in an abnormal or failed state.

[0082] Once an edge node detects a partial blockage or a surge in risk on a navigation path within its jurisdiction, it immediately initiates an emergency replanning process autonomously. Specifically, it updates its local decision factor set and then returns to execute steps S106-S108 above to replan the navigation path.

[0083] Optionally, the computational scope of the replanning is strictly limited to the affected local topology network. Centered on the failure point or blockage point, it expands outward by a limited number of hops (e.g., 2-3 intersections) to form a local replanning subgraph, while the paths in other parts of the global network remain unchanged. On the local replanning subgraph, an incremental optimization algorithm suitable for fast single-path search is adopted. Compared with global multi-objective evolutionary algorithms that require multiple iterations, this type of algorithm can find a locally optimal alternative path from the current blocked position to bypass the obstacle, reconnect to the original global path, or reach a new exit in a very short time (e.g., within 50ms).

[0084] Because the scope of replanning is strictly limited to the jurisdiction of a single or a few neighboring nodes, the required cross-node collaborative communication is minimized. Typically, a new consensus can be reached simply by quickly synchronizing the locally updated decision factors and new local path schemes with one or two directly affected neighboring nodes, avoiding large-scale global coordination overhead.

[0085] In one optional implementation, after generating a navigation path consistent with its adjacent edge nodes, the present invention further includes the following steps: Step S301: Convert the navigation path into multimodal navigation instructions; the multimodal navigation instructions include: light control instructions for physical layer guidance and mobile terminal navigation instructions for digital layer guidance.

[0086] Step S302: Send the lighting control command to the intelligent emergency light array in the park to control the intelligent emergency lights to transmit navigation direction information through specific light intensity changes.

[0087] Step S303: Push the mobile navigation instructions to the mobile terminals of people in the park to provide real-time navigation guidance to people through the display interface of the mobile terminals.

[0088] The above steps are the key encoding and interface conversion links from system decision results to execution instructions in the physical and digital worlds. Its core objective is to transform the abstract path sequence (a series of coordinate points or road segment IDs) generated by the algorithm into a composite instruction set that can be accurately executed by different types of terminal devices and can be intuitively understood by evacuees.

[0089] Evacuated individuals may be in various states (familiar with the environment, possessing a mobile phone, or having obstructed vision), and the disaster site environment is complex (potentially power outages, unstable networks, or smoke). A single guidance method (such as relying solely on mobile phones or signs) carries the risk of failure. Therefore, it is essential to provide redundant, complementary, and adaptable multimodal guidance for different scenarios.

[0090] Physical layer instruction generation: Analyze the planned navigation path and extract its key turning points, direction of travel, and path continuity information. Encode this information into a set of instruction protocols suitable for controlling lighting equipment. For example, generate a series of lighting control instructions that include target luminaire ID, desired light intensity level, change frequency (e.g., PWM duty cycle), color coding (e.g., green for passage, red for danger), and timing relationships. The core idea is to transform the static path line into a dynamic "light flow" or "arrow sequence" distributed along the path.

[0091] Digital layer instruction generation: Also based on path information, this generates data packages suitable for mobile applications (Apps) or mini-programs. This typically includes: a detailed sequence of coordinates for the path, augmented reality (AR) anchor information for each turning point (e.g., "Turn left in 10 meters"), estimated arrival time, and possible risk warning text. This information is then packaged into mobile navigation instructions in standard JSON or Protobuf format.

[0092] The constraint for instruction generation is that the guidance information conveyed by the physical layer and the digital layer is strictly synchronized in space and time, pointing in the same direction and indicating the same path, so as to avoid confusion for personnel.

[0093] In one optional embodiment, the present invention further includes the following steps: In step S401, the cloud platform receives decision log data uploaded by each edge node to build an emergency navigation scenario database for the park.

[0094] Specifically, the cloud platform receives anonymized decision log data uploaded by various edge nodes. This data includes three main elements: the situation on-site at the time, the system's decision-making process, and the actual evacuation effect. By aggregating, cleaning, and structuring the massive logs, the cloud platform builds a panoramic "decision-feedback" scenario library, integrating scattered local experiences from various nodes and time periods into a systematic knowledge asset that can be mined and learned globally.

[0095] Step S402: Based on the data in the scenario library, a reinforcement learning algorithm is used to optimize the model parameters in the multi-objective path planning algorithm.

[0096] The cloud platform models historical decision data from the scenario database as a reinforcement learning problem, with the current situation as the state, the planned path as the action, and the actual safety and efficiency of the path as the reward. By running reinforcement learning algorithms (such as PPO), the system automatically analyzes which decisions under which situations will yield higher long-term rewards, thereby discovering better decision-making strategies than existing rules or models. The output of this process is optimization suggestions or new parameters for key parameters (such as the weights of safety and efficiency, and the thresholds of the risk assessment model) in the path planning algorithm deployed on edge nodes.

[0097] Step S403: The optimized model parameters are sent to each edge node to update the local multi-objective path planning algorithm of each edge computing node.

[0098] The cloud platform distributes the optimized model parameters to all edge computing nodes within the park via a secure communication channel. Upon receiving the new parameters, each node performs a hot update or silent switch within its local path planning algorithm. This process typically employs privacy-preserving technologies such as federated learning to ensure that data remains local and only model parameters are exchanged. This allows all edge nodes to synchronously acquire globally validated, superior algorithmic capabilities without leaking sensitive local data.

[0099] The above steps constitute a complete intelligent closed loop of "edge execution - cloud learning - global evolution". This allows the system to move beyond relying solely on pre-set fixed rules and continuously learn from each real-world emergency, optimizing its judgment and decision-making capabilities. This endows the system with powerful potential for continuous evolution, enabling it to better adapt to the structural characteristics of different parks and cope with unforeseen and complex disaster scenarios that may arise in the future.

[0100] Example 2 This invention also provides a park emergency navigation edge collaboration system based on building structure situation awareness. This system is mainly used to execute the park emergency navigation edge collaboration method based on building structure situation awareness provided in Embodiment 1 above. The system provided by this invention will be described in detail below.

[0101] Figure 3 A functional module diagram of a park emergency navigation edge collaboration system based on building structure situation awareness is provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the system mainly includes: multiple edge nodes 11 and a multi-source sensor network 12 deployed in the park.

[0102] Multi-source sensor networks are used to collect data on building structural deformation, environmental conditions, and personnel distribution in real time, and perform spatiotemporal alignment processing on the collected data to obtain multi-source sensing data.

[0103] Each edge node is used to perform feature extraction processing on multi-source sensing data within its jurisdiction to obtain a local decision factor set. The local decision factor set includes risk factors used to quantify risks within the region. The risk factors include: path safety factors determined based on building structure deformation data, and traffic efficiency factors determined based on environmental status data and personnel distribution data.

[0104] Each edge node is also used to exchange its local decision factor set with other edge nodes through an inter-node communication protocol in order to obtain the global decision factor set.

[0105] Each edge node is also used to execute a distributed multi-objective path planning algorithm based on a global decision factor set, with the goal of maximizing path safety and minimizing travel time in parallel optimization, in order to generate a navigation path that is consistent with its neighboring edge nodes.

[0106] This invention extracts path safety factors and traffic efficiency factors from multi-source sensing data, and for the first time quantifies the structural safety status of buildings and integrates it into path planning as a core optimization objective of equal importance to travel time. This fundamentally avoids guiding personnel to structurally high-risk areas and significantly improves the inherent safety of evacuation. By exchanging decision factors among edge nodes and executing distributed multi-objective planning, real-time cross-regional path coordination is achieved without centralized scheduling. This ensures the consistency of the global path network in terms of safety and efficiency, and also improves the system's robustness and autonomous decision-making capabilities in the event of local network failures.

[0107] Optionally, multi-source sensor networks are specifically used for: A unified spatiotemporal coordinate system is constructed, unifying the time reference of the collected building structure deformation data, environmental status data, and personnel distribution data to a preset standard time, and unifying the spatial reference to the local coordinates of the buildings in the park.

[0108] A sliding time window is set, and a dynamic time warping algorithm is used within the sliding time window to compensate for the timing deviation of data collected by sensors with different sampling frequencies, so as to obtain multi-source sensing data.

[0109] Optionally, edge nodes are specifically used for: A temporal neural network model is used to process building structural deformation data from multi-source sensing data to obtain the structural failure rate of each building within a specified time period in the future. The path safety factor of the road segment to which each building belongs is determined based on the structural failure rate; the higher the structural failure rate, the lower the path safety factor.

[0110] An environmental risk field is constructed based on environmental state data from multi-source sensing data, and an environmental impact coefficient reflecting the degree of hazard from smoke and temperature is obtained.

[0111] Based on the population distribution data in the multi-source sensing data, the density clustering algorithm is used to identify congested areas and obtain the population density coefficient that reflects traffic resistance.

[0112] By combining the environmental impact coefficient and the population density coefficient, the traffic efficiency factor of each road segment is calculated using a preset traffic speed model.

[0113] Optionally, edge nodes are specifically used for: Construct a multi-objective optimization function with the goals of maximizing path safety and minimizing travel time.

[0114] Based on the distributed multi-objective path planning algorithm and the global decision factor set, the multi-objective optimization function is solved to obtain the non-dominated path solution set.

[0115] Candidate navigation paths are determined from the solution set of non-dominated paths using fuzzy decision-making.

[0116] When a candidate navigation path determined by an edge node conflicts with the candidate navigation paths planned by its neighboring nodes, a consensus negotiation subprocess based on security proof is initiated. The relevant nodes exchange evidence supporting the security of their paths and reach a consensus through quantitative evaluation to obtain the navigation path.

[0117] Optionally, edge nodes are also used for: During the guidance process based on the navigation path, the path status is monitored in real time.

[0118] When local path congestion is detected, the affected edge nodes are triggered to update their local decision factor sets and replan the navigation path.

[0119] Optionally, edge nodes are also used for: The navigation path is converted into multimodal navigation instructions; the multimodal navigation instructions include: lighting control instructions for physical layer guidance and mobile navigation instructions for digital layer guidance.

[0120] Lighting control commands are sent to the intelligent emergency light array within the park to control the intelligent emergency lights to transmit navigation direction information through specific changes in light intensity.

[0121] Mobile navigation instructions are pushed to the mobile devices of people within the park to provide real-time navigation guidance through the mobile device's display interface.

[0122] Optionally, the system may also include: a cloud platform; The cloud platform receives decision log data uploaded by each edge node to build an emergency navigation scenario database for the park.

[0123] Based on data from the scenario library, a reinforcement learning algorithm is used to optimize the model parameters in the multi-objective path planning algorithm.

[0124] The optimized model parameters are distributed to each edge node to update the local multi-objective path planning algorithm of each edge computing node.

[0125] Example 3 See Figure 4 This invention provides an electronic device, which includes a processor 60, a memory 61, a bus 62, and a communication interface 63. The processor 60, the communication interface 63, and the memory 61 are connected via the bus 62. The processor 60 is used to execute executable modules, such as computer programs, stored in the memory 61.

[0126] The memory 61 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 63 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc.

[0127] Bus 62 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0128] The memory 61 is used to store programs. After receiving an execution instruction, the processor 60 executes the program. The method executed by the apparatus defined by the process disclosed in any of the foregoing embodiments of the present invention can be applied to the processor 60 or implemented by the processor 60.

[0129] Processor 60 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 60 or by instructions in software form. Processor 60 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 61. Processor 60 reads the information in memory 61 and, in conjunction with its hardware, completes the steps of the above method.

[0130] The computer program product of the emergency navigation edge collaboration system and method for parks based on building structure situation awareness provided in this embodiment of the invention includes a computer-readable storage medium storing non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the methods described in the preceding method embodiments. For specific implementation, please refer to the method embodiments, which will not be repeated here.

[0131] In addition, the functional units in the various embodiments of the present invention 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.

[0132] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0133] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0134] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of this invention is in use. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention. In addition, the terms "first," "second," "third," etc., are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0135] Furthermore, terms such as "horizontal," "vertical," and "sag" do not imply that components must be absolutely horizontal or suspended, but rather that they can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal relative to "vertical," and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0136] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0137] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A campus emergency navigation edge collaboration method based on building structure situation awareness, characterized in that, include: By deploying a multi-source sensor network within the park, data on building structural deformation, environmental conditions, and personnel distribution are collected in real time. The collected data is then spatiotemporally aligned to obtain multi-source sensing data. Each edge node performs feature extraction processing on multi-source sensing data within its jurisdiction to obtain a local decision factor set; wherein, the local decision factor set includes risk factors for quantifying risks within the region; the risk factors include: a path safety factor determined based on building structure deformation data, and a traffic efficiency factor determined based on environmental status data and personnel distribution data; Each edge node exchanges its local decision factor set with other edge nodes through an inter-node communication protocol in order to obtain the global decision factor set. Each edge node, based on the global decision factor set, executes a distributed multi-objective path planning algorithm with the parallel optimization objectives of maximizing path safety and minimizing travel time, in order to generate a navigation path that is consistent with its neighboring edge nodes.

2. The edge-coordinated emergency navigation method for industrial parks based on building structure situation awareness as described in claim 1, characterized in that, The collected data undergoes spatiotemporal alignment processing, including: A unified spatiotemporal coordinate system is constructed, and the time reference of the collected building structure deformation data, environmental status data and personnel distribution data is unified to the preset standard time, and the spatial reference is unified to the local coordinates of the buildings in the park. A sliding time window is set, and a dynamic time warping algorithm is used within the sliding time window to compensate for the timing deviation of data collected by sensors with different sampling frequencies, thereby obtaining the multi-source sensing data.

3. The edge-coordinated emergency navigation method for parks based on building structure situation awareness according to claim 1, characterized in that, Each edge node performs feature extraction processing on the multi-source sensing data within its jurisdiction to obtain a local decision factor set, including: The building structural deformation data in the multi-source sensing data is processed using a temporal neural network model to obtain the structural failure rate of each building within a specified time period in the future, and the path safety factor of the road segment to which each building belongs is determined based on the structural failure rate; wherein, the higher the structural failure rate, the lower the path safety factor. An environmental risk field is constructed based on the environmental state data in the multi-source sensing data to obtain an environmental impact coefficient that reflects the degree of hazard from smoke and temperature. Based on the personnel distribution data in the multi-source sensing data, a density clustering algorithm is used to identify congested areas and obtain a personnel density coefficient that reflects traffic resistance. Combining the environmental impact coefficient and the population density coefficient, the traffic efficiency factor for each road segment is calculated using a preset traffic speed model.

4. The edge-coordinated emergency navigation method for industrial parks based on building structure situation awareness as described in claim 1, characterized in that, Each edge node, based on the global decision factor set, executes a distributed multi-objective path planning algorithm with the parallel optimization objectives of maximizing path safety and minimizing travel time, to generate a navigation path that is consistent with its neighboring edge nodes, including: Construct a multi-objective optimization function that aims to maximize path safety and minimize travel time; Based on the distributed multi-objective path planning algorithm and the global decision factor set, the multi-objective optimization function is solved to obtain the non-dominated path solution set. Candidate navigation paths are determined from the set of non-dominated paths using a fuzzy decision-making method. When a candidate navigation path determined by an edge node conflicts with the candidate navigation paths planned by its neighboring nodes, a consensus negotiation sub-process based on security proof is initiated. The relevant nodes exchange evidence supporting the security of their paths and reach a consensus through quantitative evaluation to obtain the navigation path.

5. The edge-coordinated emergency navigation method for parks based on building structure situation awareness according to claim 1, characterized in that, After generating a navigation path that is consistent with its neighboring edge nodes, the following steps are also included: During the guidance process based on the navigation path, the path status is monitored in real time; When local path congestion is detected, the affected edge nodes are triggered to update their local decision factor sets and replan the navigation path.

6. The edge-coordinated emergency navigation method for parks based on building structure situation awareness according to claim 1, characterized in that, After generating a navigation path that is consistent with its neighboring edge nodes, the following steps are also included: The navigation path is converted into multimodal navigation instructions; the multimodal navigation instructions include: lighting control instructions for physical layer guidance and mobile navigation instructions for digital layer guidance; The lighting control command is sent to the intelligent emergency light array in the park to control the intelligent emergency lights to transmit navigation direction information through specific light intensity changes; The mobile navigation instructions are pushed to the mobile terminals of people in the park to provide real-time navigation guidance through the display interface of the mobile terminals.

7. The edge-coordinated emergency navigation method for parks based on building structure situation awareness according to claim 1, characterized in that, Also includes: The cloud platform receives decision log data uploaded by each edge node to build an emergency navigation scenario database for the park; Based on the data in the scenario library, a reinforcement learning algorithm is used to optimize the model parameters in the multi-objective path planning algorithm; The optimized model parameters are distributed to each edge node to update the local multi-objective path planning algorithm of each edge computing node.

8. A campus emergency navigation edge collaboration system based on building structure situation awareness, characterized in that, include: Multiple edge nodes and a multi-source sensor network deployed within the park; The multi-source sensor network is used to collect building structure deformation data, environmental status data, and personnel distribution data in real time, and to perform spatiotemporal alignment processing on the collected data to obtain multi-source sensing data. Each edge node is used to perform feature extraction processing on multi-source sensing data within its jurisdiction to obtain a local decision factor set; wherein, the local decision factor set includes risk factors for quantifying risks within the region; the risk factors include: a path safety factor determined based on building structure deformation data, and a traffic efficiency factor determined based on environmental status data and personnel distribution data; Each edge node is also used to exchange its local decision factor set with other edge nodes through the inter-node communication protocol in order to obtain the global decision factor set; Each edge node is also used to execute a distributed multi-objective path planning algorithm based on the global decision factor set, with the goal of maximizing path safety and minimizing travel time in parallel, in order to generate a navigation path that is consistent with its neighboring edge nodes.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, characterized in that, When the processor executes the computer program, it implements the edge collaboration method for emergency navigation in a park based on building structure situation awareness, as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, which, when executed by a processor, implement the edge collaboration method for emergency navigation in a park based on building structure situation awareness as described in any one of claims 1 to 7.

Images