Mining wireless gateway integrated management method and system

Abstract

The invention provides a mining wireless gateway integrated management method and system, and relates to the technical field of wireless gateways, and the method comprises the steps: interactively obtaining a tunnel topological structure of a target mine, and dividing the target mine into honeycomb region grids; k multi-mode sensor arrays are deployed; constructing a communication link between the multi-modal sensor array and the edge gateway; carrying out risk diffusion path integrated simulation by the diffusion risk analysis platform according to the multi-modal time sequence characteristics, and positioning a plurality of real-time risk partitions; directional activation of relay identification hierarchical topology is carried out; waking up the plurality of real-time relay identification sequences to carry out RFID coverage identification on the plurality of real-time risk partitions to obtain a plurality of real-time personnel distribution characteristics; and carrying out rescue resource dynamic allocation. According to the method and the device, the technical problem that the rescue efficiency is affected due to poor risk source identification accuracy of the wireless gateway caused by complex mine environment and frequent change is solved, and the mine rescue scheduling capability and efficiency are improved.

Description

Technical Field

[0001] This application relates to the technical field of wireless gateways, and particularly to an integrated management method and system for mine wireless gateways. Background Art

[0002] Mine wireless gateways achieve seamless connection and efficient collaboration among various sensors, devices, and communication systems in the mine environment through centralized management and distributed deployment. The dynamics and complexity of the mine environment require mine wireless gateways to be able to identify and locate multiple risk sources in real time, and at the same time be able to dynamically adjust the network architecture and monitoring mode to adapt to the changing working environment. Wireless gateways obtain data from different sensors and devices in real time and conduct comprehensive analysis, thereby realizing the monitoring and risk warning of the mine environment. However, problems such as complex factors in the mine environment (such as tortuous roadways and many obstacles), limited wireless signal propagation, and uneven signal coverage result in low accuracy of personnel positioning and risk source identification for existing mine wireless gateways, being unable to respond to sudden risks in a timely manner, and affecting the emergency response and rescue efficiency.

[0003] In summary, there is a technical problem in the prior art that due to the complex and frequently changing mine environment, the accuracy of risk source identification by wireless gateways is poor, further affecting the rescue efficiency. Summary of the Invention

[0004] The purpose of this application is to provide an integrated management method and system for mine wireless gateways to solve the technical problem in the prior art that due to the complex and frequently changing mine environment, the accuracy of risk source identification by wireless gateways is poor, further affecting the rescue efficiency.

[0005] In view of the above problems, this application provides an integrated management method and system for mine wireless gateways.

[0006] In a first aspect, the present application provides a method for integrated management of mine wireless gateways. The method for integrated management of mine wireless gateways is implemented through an integrated management system for mine wireless gateways. Among them, the method for integrated management of mine wireless gateways includes: interactively obtaining the roadway topological structure of a target mine, and dividing the target mine into cellular area grids according to the roadway topological structure; deploying K multi-modal sensor arrays in K grid partitions of the cellular area grids through monitoring coverage fitting; constructing communication links between the K multi-modal sensor arrays and K edge gateways, where the edge gateways are built-in with lightweight AI chips; the diffusion risk analysis platform performs an integrated simulation of the risk diffusion path according to the K multi-modal time series features transmitted back by the K edge gateways to locate multiple real-time risk partitions, where the K edge gateways are connected to the diffusion risk analysis platform; performing directional activation of the relay identification hierarchical topology according to the multiple real-time risk partitions, and outputting multiple real-time relay identification sequences; waking up the multiple real-time relay identification sequences to perform RFID coverage identification on the multiple real-time risk partitions to obtain multiple real-time personnel distribution characteristics; dynamically allocating rescue resources according to the multiple real-time personnel distribution characteristics.

[0007] Optionally, through multi-source data acquisition and fusion, the roadway topological structure is extracted, and a weighted roadway connectivity graph is generated according to the roadway topological structure; multiple intersection node coordinates and multiple node weight data of multiple roadway topological nodes are extracted from the weighted roadway connectivity graph; dynamic grid division is performed according to the multiple intersection node coordinates and multiple node weight data to divide the target mine into the cellular area grids.

[0008] Optionally, by parsing and processing the BIM model in the design stage of the target mine, an initial topological skeleton is generated, where the parsing and processing include the extraction of roadway centerlines, intersection point coordinates, and support structure parameters; after using an explosion-proof mobile laser scanner to collect point cloud data along the roadway of the target mine, the collected data is compressed through voxel grid filtering to obtain roadway laser point cloud data; by registering the roadway laser point cloud data and the initial topological skeleton, a roadway fusion three-dimensional model is obtained; the roadway centerline is extracted from the roadway fusion three-dimensional model to obtain the roadway topological structure; by injecting dynamic and static weights of nodes into the roadway topological structure, the construction of the weighted roadway connectivity graph is completed.

[0009] Optionally, based on the multiple node spatial positions of the multiple roadway topology nodes in the roadway topology structure, perform lithological stability analysis on the multiple node spatial positions to obtain multiple node static weights; according to the multiple topological edge spatial positions of the multiple roadway topology edges in the roadway topology structure, extract multiple roof displacement data from the associated geological exploration report; calculate the deformation rate of the multiple roof displacement data to obtain multiple topological edge dynamic weights; based on the multiple node spatial positions and multiple topological edge spatial positions, inject the multiple node static weights and multiple topological edge dynamic weights into the roadway topology structure mapping to complete the construction of the weighted roadway connection graph.

[0010] Optionally, locally call the associated geological exploration report of the target mine; guided by the multiple node spatial positions, extract multiple rock layer stability coefficients from the associated geological exploration report; perform local accident record calls based on the multiple node spatial positions to obtain multiple node accident frequencies; normalize the multiple rock layer stability coefficients and multiple node accident frequencies to obtain the multiple node static weights.

[0011] Optionally, according to the connection relationship between nodes and topological edges, extract multiple groups of topological edge dynamic weights of the multiple roadway topology nodes; fuse and calculate the multiple node static weights and multiple groups of topological edge dynamic weights through the entropy weight method to obtain the multiple node weight data; using the multiple roadway topology nodes as the grid starting point, dynamically correct the grid density according to the multiple node weight data, and divide the target mine into the honeycomb area grid, where the grid weight density deviation of the K grid partitions meets the preset deviation scale.

[0012] Optionally, according to the roadway topology structure and the K sensor deployment coordinates of the K multi-modal sensor arrays, perform spatio-temporal alignment mapping of the K multi-modal time series features to complete the configuration of the spatio-temporal data cube; based on the spatio-temporal data cube, synchronously execute physical-driven modeling and data-driven modeling to obtain a risk diffusion simulation model; perform risk diffusion localization by running the risk diffusion simulation model to obtain the multiple real-time risk partitions, where the multiple real-time risk partitions include a risk source partition and M risk diffusion partitions.

[0013] Optionally, extract the roadway three-dimensional boundary, roadway wind speed distribution, and roadway gas concentration gradient distribution from the spatio-temporal data cube; input the roadway three-dimensional boundary, roadway wind speed distribution, and roadway gas concentration gradient distribution into the fluid dynamics equation to construct a basic diffusion model and output a dynamic diffusion trajectory set; process the multi-modal time series distribution of the spatio-temporal data cube through a spatio-temporal graph neural network to output a risk probability distribution matrix; perform dynamic weighted fusion on the dynamic diffusion trajectory set and the risk probability distribution matrix to generate the risk diffusion simulation model.

[0014] Optionally, through signal coverage fitting, deploy the relay identification hierarchical topology in the target mine; perform identification energy consumption analysis on the coverage relationship between the risk source partition and M risk diffusion partitions according to the relay identification hierarchical topology to obtain a first real-time relay identification sequence and M real-time relay identification sequences; wherein, the first real-time relay identification sequence and the M real-time relay identification sequences constitute the multiple real-time relay identification sequences.

[0015] In a second aspect, the present application also provides a mine wireless gateway integrated management system for executing the mine wireless gateway integrated management method as described in the first aspect, wherein the mine wireless gateway integrated management system includes: a mine division processing module for interactively obtaining the roadway topology of the target mine and dividing the target mine into honeycomb area grids according to the roadway topology; a monitoring deployment processing module for deploying K multi-modal sensor arrays in K grid partitions in the honeycomb area grid through monitoring coverage fitting; a communication link construction module for constructing communication links between the K multi-modal sensor arrays and K edge gateways, wherein the edge gateways are built-in with lightweight AI chips; a risk diffusion simulation module for the risk diffusion analysis platform to perform an integrated simulation of the risk diffusion path according to the K multi-modal time series features transmitted back by the K edge gateways to locate multiple real-time risk partitions, wherein the K edge gateways are connected to the risk diffusion analysis platform; a relay identification activation module for directionally activating the relay identification hierarchical topology according to the multiple real-time risk partitions and outputting multiple real-time relay identification sequences; a coverage identification processing module for waking up the multiple real-time relay identification sequences to perform RFID coverage identification on the multiple real-time risk partitions to obtain multiple real-time personnel distribution characteristics; and a resource allocation processing module for dynamically allocating rescue resources according to the multiple real-time personnel distribution characteristics.

[0016] One or more technical solutions provided in the present application have at least the following beneficial effects:

[0017] Obtain the roadway topological structure of the target mine through interaction, and divide the target mine into honeycomb area grids according to the roadway topological structure; deploy K multi-modal sensor arrays in K grid partitions of the honeycomb area grids through monitoring coverage fitting; construct communication links between the K multi-modal sensor arrays and K edge gateways, where the edge gateways are built-in with lightweight AI chips; the diffusion risk analysis platform performs an integrated simulation of the risk diffusion path based on the K multi-modal time series features transmitted back by the K edge gateways to locate multiple real-time risk partitions, where the K edge gateways are connected to the diffusion risk analysis platform; perform directional activation of the relay identification hierarchical topology according to the multiple real-time risk partitions, and output multiple real-time relay identification sequences; wake up the multiple real-time relay identification sequences to perform RFID coverage identification on the multiple real-time risk partitions to obtain multiple real-time personnel distribution features; perform dynamic allocation of rescue resources according to the multiple real-time personnel distribution features. That is to say, by dividing the mine into multiple honeycomb area grids, deploying multi-modal sensor arrays to monitor the mine environment in real time, establishing a communication link in the wireless communication gateway to transmit the collected multi-modal time series feature data back to the edge gateway, performing an integrated simulation of the risk diffusion path on the multi-modal time series feature data, locating the risk partitions, through the directional activation of the relay identification hierarchical topology, waking up the relevant relay devices, determining the real-time personnel in the risk source and the risk diffusion direction, and performing dynamic allocation of rescue resources, it is ensured that in case of an emergency, rescue resources can be quickly and accurately allocated, improving the rescue scheduling ability and efficiency.

[0018] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically gives the specific implementation manners of the present application. It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become easily understandable through the following description. Brief Description of the Drawings

[0019] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings described below are only exemplary, and for those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0020] Figure 1 It is a schematic flowchart of the integrated management method of the mine wireless gateway of the present application;

[0021] Figure 2This is a schematic structural diagram of the integrated management system for mine wireless gateways in this application.

[0022] Explanation of reference numerals in the drawings: Mine division processing module 11, monitoring and deployment processing module 12, communication link construction module 13, risk diffusion simulation module 14, relay identification and activation module 15, coverage identification processing module 16, resource allocation processing module 17. Specific implementation mode

[0023] This application provides an integrated management method and system for mine wireless gateways, which solves the technical problem in the prior art that due to the complex and frequently changing mine environment, the accuracy of risk source identification by wireless gateways is poor, further affecting the rescue efficiency. By dividing the mine into multiple cellular area grids, deploying a multi-modal sensor array to monitor the mine environment in real time, establishing a communication link in the wireless communication gateway to transmit the collected multi-modal time-series feature data back to the edge gateway, performing an integrated simulation of the risk diffusion path for the multi-modal time-series feature data, locating the risk area, activating the relevant relay devices through the directional activation of the relay identification hierarchical topology, determining the real-time personnel in the risk source and the risk diffusion direction, and performing dynamic rescue resource allocation, it ensures that in case of an emergency, rescue resources can be quickly and accurately allocated, improving the rescue scheduling ability and efficiency.

[0024] Next, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited by the example embodiments described here. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of this application. Additionally, it should be noted that for the sake of description, only the parts related to this application are shown in the drawings, rather than all of them.

[0025] Example 1. Please refer to the attached Figure 1 , this application provides an integrated management method for mine wireless gateways. Among them, the integrated management method for mine wireless gateways is executed by an integrated management system for mine wireless gateways. The integrated management method for mine wireless gateways specifically includes the following steps:

[0026] S100: Interactively obtain the roadway topology of the target mine, and divide the target mine into cellular area grids according to the roadway topology.

[0027] Further, S100 of this application includes:

[0028] Through multi-source data collection and fusion, the roadway topological structure is extracted, and a weighted roadway connectivity graph is generated based on the roadway topological structure; multiple intersection node coordinates and multiple node weight data of multiple roadway topological nodes are extracted from the weighted roadway connectivity graph; dynamic grid division is performed based on the multiple intersection node coordinates and multiple node weight data to divide the target mine into the honeycomb area grids.

[0029] Specifically, through the BIM model in the design stage of the target mine and the point cloud data along the roadway of the target mine, a three-dimensional roadway fusion model is obtained, and the roadway centerline is extracted from the three-dimensional roadway fusion model to obtain a complete roadway topological structure. According to the multiple node static weights of multiple roadway topological nodes and the multiple topological edge dynamic weights of multiple roadway topological edges in the roadway topological structure, a weighted roadway connectivity graph is obtained. Nodes represent intersection points, end points or connection points in the roadway, reflecting the spatial layout and structure of the mine roadway. Edges represent the connections between two nodes, that is, the paths between two roadway segments. The weights are used to represent attributes such as the safety and stability of each roadway segment.

[0030] Multiple intersection node coordinates and multiple node weight data of multiple roadway topological nodes are extracted from the weighted roadway connectivity graph, that is, the position of each node and its corresponding weight. The intersection node coordinates represent the spatial position where the roadways in the mine intersect, and the node weight data is obtained by combining factors such as rock stratum stability, accident records, and geological stability, reflecting the safety risks in different positions of the mine.

[0031] According to the importance and risk level of the nodes, the grid density is adjusted to ensure that the grid density can accurately reflect the risk levels in different regions. For example, high-risk nodes will be assigned to denser grids for more refined monitoring; low-risk nodes will be assigned to larger and sparser grids. Through the grid density, the target mine is divided into honeycomb area grids, and each honeycomb grid represents an independent management area, which can be managed differently according to the risk characteristics of each area. The shape of each grid may be different, usually a polygon, and the specific shape is closely related to the topography and geomorphology of the mine, the roadway layout, and the risk distribution. Through multi-source data collection and fusion analysis, refined dynamic grid division is performed based on the weighted roadway connectivity graph to achieve efficient identification and emergency response of mine risks. Guided by the node weight data, the grid density can reflect the risk differences in different regions of the mine, ensuring that high-risk areas are monitored and intervened more precisely.

[0032] Furthermore, the present application further includes the following steps:

[0033] By parsing and processing the BIM model in the design stage of the target mine, an initial topological skeleton is generated, where the parsing and processing include the extraction of roadway centerlines, intersection coordinates, and support structure parameters; after using an explosion-proof mobile laser scanner to collect point cloud data along the roadway of the target mine, voxel grid filtering is used to compress the collected data to obtain roadway laser point cloud data; by registering the roadway laser point cloud data and the initial topological skeleton, a roadway fusion three-dimensional model is obtained; the roadway centerline is extracted from the roadway fusion three-dimensional model to obtain a roadway topological structure; by injecting node static and dynamic weights into the roadway topological structure, the construction of the weighted roadway connectivity graph is completed.

[0034] Furthermore, the present application also includes the following steps:

[0035] According to the multiple node spatial positions of the multiple roadway topological nodes in the roadway topological structure; performing lithological stability analysis on the multiple node spatial positions to obtain multiple node static weights; according to the multiple topological edge spatial positions of the multiple roadway topological edges in the roadway topological structure, extracting multiple roof displacement data from the associated geological exploration report; calculating the deformation rate of the multiple roof displacement data to obtain multiple topological edge dynamic weights; according to the multiple node spatial positions and multiple topological edge spatial positions, injecting the multiple node static weights and multiple topological edge dynamic weights in the mapping of the roadway topological structure to complete the construction of the weighted roadway connectivity graph.

[0036] Specifically, obtain the BIM model in the design stage of the target mine, that is, the detailed information of the target mine in the initial design. The BIM model contains detailed information about the mine design, such as the geometric shape of the roadway, the location of intersections, the setting of support structures, etc. Parse and process the BIM model in the design stage, and extract key information from the BIM model in the design stage to generate an initial topological skeleton. Specifically, the steps of the parsing and processing include the extraction of roadway centerlines, intersection coordinates, and support structure parameters.

[0037] By analyzing the geometric data in the BIM model, the centerline of the roadway is extracted, which represents the main channel trajectory of the mine roadway. Intersections are important nodes in the mine roadway structure, representing the connection points or convergence points of different roadway branches. Extracting the intersection coordinates can clearly mark the key connection positions in the roadway system. The support structure is an important part of the mine roadway, used to ensure the stability and safety of the roadway. Extract the detailed parameters of the support structure (such as support type, size, material, etc.). The initial topological skeleton is composed of the extracted roadway centerlines, intersections, and support structure information, forming the basic framework of the mine roadway.

[0038] Since the mine environment often comes with hazardous gases, extreme temperatures, or high humidity, explosion-proof mobile laser scanners are needed to ensure safe operation in such complex and dangerous mine environments while efficiently collecting three-dimensional spatial data of mine roadways. The laser scanner emits laser beams and measures their return times to calculate the distances to different points, generating a large amount of point cloud data that represents the geometric shape of the mine roadway, including the curves, inclination angles, spatial distribution of the roadway, and surrounding obstacles, etc. The explosion-proof mobile laser scanner can achieve high-precision, full-range scanning of the internal structure of the roadway without contacting the mine surface.

[0039] Voxel grid filtering is used to compress the collected point cloud data. Voxel grid filtering is a commonly used point cloud data dimensionality reduction technique that divides the point cloud space into uniform voxels (i.e., three-dimensional grid cells) and integrates the points within each voxel into a representative point. Voxel grid filtering is a point cloud data processing technique that divides the point cloud data into many small cubes (voxels), and then statistically processes the points within each voxel to achieve data compression and noise reduction.

[0040] By registering the point cloud data with the initial topological skeleton, the laser point cloud data of the roadway is matched and fused with the initial topological skeleton in the BIM model, generating a complete three-dimensional fusion model of the roadway. This integrated model combines the design information from the BIM model and the point cloud data collected from the actual situation, providing a more realistic and accurate display of the mine roadway structure. The core purpose of registration is to spatially align the laser point cloud data (real data collected by the explosion-proof laser scanner) with the initial topological skeleton (the roadway geometric information parsed from the BIM model). Since the laser point cloud data and the initial topological skeleton come from different data sources, one is theoretical data based on the design stage and the other is point cloud data collected from the actual environment, these two data sets usually have differences in position, scale, orientation, etc. Therefore, the registration process is very crucial to ensure that the two data sets can be accurately aligned in the same coordinate system.

[0041] During the registration process, point-to-point registration algorithms (such as the ICP algorithm, Iterative Closest Point algorithm) or feature point-based registration methods are usually adopted to align the laser point cloud data with the initial topological skeleton. The feature points corresponding to the laser point cloud data in the initial topological skeleton, such as the intersection points and end points of the roadway, are automatically identified and aligned through these feature points. During this process, if there are errors, the point cloud data is fine-tuned to better conform to the actual spatial layout of the roadway.

[0042] The registered laser point cloud data and the initial topological skeleton data will be combined to generate a three-dimensional model of roadway fusion, which not only retains the design structure of the roadway in the initial topological skeleton (such as the roadway centerline, intersection coordinates, and support structure, etc.), but also incorporates the actually collected point cloud data, providing more accurate geometric information of the roadway space and being able to truly reflect the actual roadway.

[0043] Extract the roadway centerline from the three-dimensional model of roadway fusion, which can accurately reflect the overall trend, curvature, and structural changes of the roadway. The extraction of the roadway centerline usually uses the minimum path algorithm or curve fitting method based on point cloud data to extract the central axis of the roadway from the point cloud data of the roadway. The minimum path algorithm determines the centerline by finding the shortest path from the roadway entrance to the exit. Usually, it starts from the starting point in the point cloud data and extends along the main channel of the roadway until the end of the roadway. The curve fitting algorithm will try to fit the centerline of the roadway into a smooth curve, which can more accurately capture the curved part of the roadway. It uses spline curves (such as B-spline or Bezier curves) to fit the key points in the point cloud data to ensure that the fitted curve is still smooth and conforms to the actual roadway shape when passing through the transition curve.

[0044] Based on the extracted roadway centerline, construct the topological structure of the roadway. The roadway topological structure adopts a node-edge topological network. Nodes represent the intersections, endpoints, or key positions of the roadway, such as the bifurcation, end point of the roadway, or the connection with other roadways; edges represent the connecting parts of the roadway, usually the straight or curved sections between adjacent nodes, connecting the various parts of the roadway into an overall network. The topological structure of the roadway is attached with other attributes, such as slope, length, bend radius, etc., which are used to identify the hazards or difficulties of different roadway parts, making the modeling of the mine roadway structure more accurate.

[0045] By performing lithological stability analysis and deformation rate calculation on the nodes and edges of the roadway topological structure, different weights are assigned to each node and edge, thus constructing a weighted roadway connectivity graph. According to the spatial positions of multiple roadway topological nodes in the roadway topological structure, lithological stability analysis is carried out. By locally calling the associated geological exploration report and accident records, combined with the actual geological conditions and historical accident data of the mine, static weights are assigned to each node of the roadway, reflecting the safety of each node in the mine.

[0046] According to the spatial positions of multiple topological edges of multiple roadway topological edges in the roadway topological structure, extract the roof displacement data corresponding to the position of each topological edge from the associated geological exploration report. Roof displacement refers to the displacement of the rock layer at the top of the roadway during the mining process of the mine. The geological exploration report usually includes a large amount of displacement measurement data, reflecting the deformation of the roadway roof within a certain period of time.

[0047] Calculate the deformation rate of the extracted roof displacement data to obtain the dynamic weight of each topological edge. The deformation rate reflects the deformation speed of the roadway roof, that is, the displacement of the rock formation per unit time. Calculate the difference in roof displacement at the same area at different time points. Calculate the ratio of displacement to time to obtain the deformation rate (unit: displacement / time). Perform similar calculations for all roadway topological edges to obtain the deformation rate of each topological edge. By calculating the deformation rate, assign a dynamic weight to each topological edge. For example, displacement data collected ten days apart, for topological edge 1 it is 2.5mm and 4.0mm, and for topological edge 2 it is 1.2mm and 1.8mm, then the deformation rate of topological edge 1 is 0.15mm / day, and the deformation rate of topological edge 2 is 0.06mm / day.

[0048] Map the static weights of multiple nodes at multiple node spatial positions and the dynamic weights of multiple topological edges at multiple topological edge spatial positions to the roadway topological structure to obtain a weighted roadway connectivity graph, which is used to represent the stability and risk distribution of the roadway network and accurately reflect the stability of different roadway segments.

[0049] Furthermore, the present application further includes the following steps:

[0050] Locally call the associated geological exploration report of the target mine; guided by the multiple node spatial positions, extract multiple rock formation stability coefficients from the associated geological exploration report; call the local accident records according to the multiple node spatial positions to obtain the accident frequencies of multiple nodes; by normalizing the multiple rock formation stability coefficients and the accident frequencies of multiple nodes, obtain the static weights of the multiple nodes.

[0051] Specifically, by locally calling the associated geological exploration report of the target mine, obtain the geological exploration information of the target mine, that is, the detailed information of the geological conditions of the target mine, including key geological information such as the rock formation distribution, rock type, and rock formation stability coefficient of the target mine. According to the multiple node spatial positions, that is, the spatial positions of each roadway topological node in the roadway topological structure, extract the rock formation stability coefficient corresponding to each node from the associated geological exploration report. The rock formation stability coefficient is a numerical value that measures the stability of the rock formation, usually calculated based on the physical and mechanical properties of the rock, and reflects the ability of the rock formation to maintain stability under specific conditions. The rock formation stability coefficient is a quantitative value used to evaluate the rock formation strength and deformation potential. A higher stability coefficient indicates that the rock formation in this area is more solid and stable, while a lower stability coefficient indicates that the rock formation in this area may be more fragile or prone to deformation. The multiple rock formation stability coefficients reflect the geological stability of the locations where the multiple nodes are located.

[0052] Based on the spatial positions of multiple nodes, local accident records are called to obtain historical accident data, including the types, frequencies, and affected areas of accidents that occurred in the target mine during different past time periods. The multiple node accident frequencies corresponding to the multiple node spatial positions are obtained to understand how many accidents have occurred historically in the node areas. Areas with higher accident frequencies usually indicate greater potential risks, which may be caused by poor rock formation conditions, poor ventilation, or other factors in these areas.

[0053] Through normalization, the rock formation stability coefficient and the accident frequency are converted into comparable quantitative values. Normalization is to convert data with different dimensions into the same standard scale, which can eliminate the dimensional differences between different data, enabling them to be comprehensively evaluated under the same standard. The normalized rock formation stability coefficient and accident frequency will assign a unified standardized value to each node, ensuring that these two factors have a reasonable influence when calculating the static weight. The static weights of multiple nodes represent the safety of each node in the roadway topological structure. Normalization is to scale the data to a similar value range and then remove the data units. The weight is calculated by taking the normalized value of the node as the numerator and the sum of the normalized values of all nodes as the denominator after obtaining the normalized values of all nodes.

[0054] Exemplarily, assume that from a geological exploration report, the rock formation stability coefficients and accident frequencies of two nodes A and B are obtained, and the static weights of nodes A and B are calculated as shown in Table 1:

[0055] Table 1 Static Weights of Nodes

[0056] node A B rock stratum stability coefficient 0.8 0.6 accident frequency 3 1 normalized value of rock stratum stability coefficient 0.8 0.6 normalized value of accident frequency 1.5 0.5 static weight of node 0.68 0.32

[0057] According to the data in the table, the normalized value of node A is 2.3, and the normalized value of node B is 1.1. Then the static weight of node A is 2.3 / 3.4, and the static weight of node B is 1.1 / 3.4. Through normalization, multiple rock formation stability coefficients and multiple node accident frequencies are converted into comparable static weights for evaluating the risk level of each node under static conditions.

[0058] Furthermore, the present application further includes the following steps:

[0059] Based on the connection relationship between the nodes and the topological edges, multiple groups of topological edge dynamic weights of the multiple roadway topological nodes are extracted; the multiple node static weights and the multiple groups of topological edge dynamic weights are fused and calculated by the entropy weight method to obtain the multiple node weight data; the multiple roadway topological nodes are used as the grid starting points, and according to the multiple node weight data, the grid density is dynamically corrected, and the target mine is divided into the honeycomb area grids, where the grid weight density deviation of the K grid partitions meets the preset deviation scale.

[0060] Specifically, according to the connection relationship between the nodes and topological edges in the roadway topological structure, multiple groups of topological edge dynamic weights of multiple roadway topological nodes are extracted. In the entropy weight method, first, the entropy value of each index is calculated, and then the weight distribution of each index is determined according to the magnitude of the entropy value. Finally, by integrating the static weight and the dynamic weight, the comprehensive risk weight of each node is obtained, that is, the weight data of multiple nodes. The entropy weight method is an objective weighting method that calculates weights based on the degree of variation between index data. In information theory, entropy is an index to measure information uncertainty, and the entropy weight method uses this principle to allocate weights.

[0061] Taking multiple roadway topological nodes as the starting point of gridification, the grid density is dynamically corrected according to the multiple node weight data, and the target mine is divided into honeycomb area grids. The density of the grid is adjusted according to the node weight data to ensure that the grid can accurately reflect the risk levels of different regions. Usually, the grid density corresponding to the areas with higher risks will increase accordingly, while the grid density in the areas with lower risks will be reduced, so that the high-risk areas can be monitored and responded to more precisely. According to the corrected grid density, the target mine is divided into multiple polygon areas with inconsistent shapes, that is, honeycomb area grids. The shape and size of the grid are dynamically adjusted according to the risk weight of the area, not necessarily regular polygons, but determined according to the actual mine topology and risk distribution.

[0062] To ensure the rationality and consistency of grid division, a preset grid density deviation scale is set. Under this scale, the density deviations of all grid areas must meet certain standards. Through this deviation scale control, it can be ensured that the grid division of the mine not only accurately reflects the risk distribution, but also avoids excessive concentration or sparsity of the grid density, ensuring that monitoring resources can be reasonably allocated. Through the dynamic correction of the grid density and the division of the honeycomb area grids, the risk levels of different areas of the target mine are accurately evaluated, thereby improving the accuracy of risk management.

[0063] S200: Deploy K multi-modal sensor arrays to K grid partitions in the honeycomb area grid through monitoring coverage fitting.

[0064] Specifically, the goal of monitoring coverage fitting is to ensure that each grid partition in the honeycomb area grid can effectively collect environmental monitoring data, so as to monitor and identify the risks of the mine in real time. Monitoring coverage fitting refers to the regional coverage analysis of each honeycomb area grid of the mine to ensure that the monitoring requirements of each grid partition are met. According to the actual situation of the mine, reasonably determine the monitoring requirements of each grid, ensure that potential risk sources in each area are covered, and configure appropriate sensor types, quantities and positions according to the risk characteristics of different areas to achieve the best monitoring effect.

[0065] Deploy corresponding multi-modal sensor arrays inside each grid partition, including but not limited to laser methane sensors, infrared thermal imagers, microseismic monitoring nodes, etc. The laser methane sensor is used to detect the methane gas concentration in the mine roadway, especially in the early stage of accidents such as gas leakage and fire, and can provide early warning signals. The infrared thermal imager is used to detect heat sources in the mine environment, especially areas with abnormal temperature changes. The microseismic monitoring node is used to monitor small vibrations and geological activities in the mine, especially during the mine exploitation process, where the risk of geological vibration or mine collapse cannot be ignored.

[0066] By deploying the multi-modal sensor arrays, ensure that different monitoring requirements in each grid area can be met, and comprehensive monitoring of the mine environment can be achieved. Due to the diversity of sensor types, each sensor can work independently or collaboratively to provide various types of data, such as gas concentration, temperature, vibration, etc., covering all aspects of the mine environment.

[0067] S300: Build communication links between the K multi-modal sensor arrays and the K edge gateways, where the edge gateways are built-in with lightweight AI chips.

[0068] Specifically, the K multi-modal sensor arrays are deployed in different areas of the mine. Each array collects mine environment data through different types of sensors (such as laser sensors, gas detection sensors, temperature sensors, etc.). Establish communication links between each multi-modal sensor array and its corresponding edge gateway to ensure that the data of the sensor array can be received, and use the AI chip to quickly process this data. The edge gateway is a network device located between the sensor array and the central monitoring system, responsible for collecting sensor data and performing preliminary processing and analysis. The lightweight AI chip is a chip integrated with artificial intelligence algorithms, which can perform real-time data processing and analysis on edge devices, perform preprocessing operations such as filtering and feature extraction on the original data, such as calculating the gradient change rate of the gas concentration, and only upload key indicators to the central platform to reduce bandwidth pressure. For example, the AI chip can calculate the gradient change rate of the gas concentration, analyze the diffusion trend of the gas in real time, and extract relevant key data (such as the change rate of the gas concentration, temperature change, microseismic events, etc.). By building an efficient communication link between the edge gateway and the multi-modal sensor array, real-time monitoring, analysis, and rapid response to the mine environment are achieved.

[0069] S400: The diffusion risk analysis platform performs an integrated simulation of the risk diffusion path based on the K multi-modal time series features returned by the K edge gateways to locate multiple real-time risk partitions, where the K edge gateways are connected to the diffusion risk analysis platform.

[0070] Furthermore, S400 of this application includes:

[0071] Perform spatio-temporal alignment mapping of the K multi-modal temporal features according to the roadway topological structure and the K sensor deployment coordinates of the K multi-modal sensor arrays, and complete the configuration of the spatio-temporal data cube; based on the spatio-temporal data cube, synchronously execute physical-driven modeling and data-driven modeling to obtain a risk diffusion simulation model; perform risk diffusion localization by running the risk diffusion simulation model to obtain the multiple real-time risk zones, where the multiple real-time risk zones include a risk source zone and M risk diffusion zones.

[0072] Extract the three-dimensional roadway boundary, roadway wind speed distribution, and roadway gas concentration gradient distribution from the spatio-temporal data cube; input the three-dimensional roadway boundary, roadway wind speed distribution, and roadway gas concentration gradient distribution into the hydrodynamic equation to construct a basic diffusion model and output a set of dynamic diffusion trajectories; process the multi-modal temporal distribution of the spatio-temporal data cube through a spatio-temporal graph neural network and output a risk probability distribution matrix; perform dynamic weighted fusion on the set of dynamic diffusion trajectories and the risk probability distribution matrix to generate the risk diffusion simulation model.

[0073] Specifically, the diffusion risk analysis platform is responsible for real-time monitoring, analysis, and localization of risks in the mine environment based on the data transmitted back from multiple edge gateways. Through integrated simulation of the risk diffusion paths of the K multi-modal temporal features transmitted back from the K edge gateways, it aims to quickly and accurately identify and locate the possible risk zones in the mine. The diffusion risk analysis platform will identify multiple areas in the mine that may be threatened according to the simulation results and demarcate these areas as real-time risk zones.

[0074] According to the roadway topological structure and the K sensor deployment coordinates of the K multi-modal sensor arrays, that is, match the spatial position and time of the multi-modal temporal features collected by each sensor array through spatio-temporal alignment mapping. By merging the temporal data at each time point and each sensor position, all the data sets are synthesized into a spatio-temporal data cube, including a time dimension, a spatial dimension, and a modal dimension. For example, the gas concentration (0.8%), temperature (35°C), and vibration acceleration (2.5g) at a certain area at time t are integrated into a vector [0.8, 35, 2.5], and together with its corresponding timestamp and spatial coordinates, they form a part of the spatio-temporal data cube.

[0075] Extract the three-dimensional boundary of the roadway, the wind speed distribution in the roadway, and the gas concentration gradient distribution in the roadway from the spatio-temporal data cube. The three-dimensional boundary of the roadway is the spatial form and structure of the mine roadway, which is crucial for simulating the gas diffusion path and the effectiveness of the ventilation system. Through the spatial data part of the spatio-temporal data cube, the geometric shape of the roadway can be extracted, including the width, height, and overall spatial layout of the roadway. The wind speed distribution in the roadway is an important factor affecting the diffusion speed and path of gases (such as gas) in the roadway. The data collected by sensors (such as wind speed sensors) can reflect the airflow pattern in the roadway, thereby affecting the gas diffusion path. The gas concentration gradient distribution in the roadway reflects the change of gas concentration in the roadway in space. The gas concentration data usually comes from gas sensors or gas detection equipment, which is used to evaluate the risk of gas leakage or accumulation. These data exist in the spatio-temporal data cube in a time series form.

[0076] Input the three-dimensional boundary of the roadway, the wind speed distribution in the roadway, and the gas concentration gradient distribution in the roadway into the hydrodynamic equation to construct a basic diffusion model. The hydrodynamic equation is used to simulate the diffusion behavior of gases (such as gas), and outputs a set of dynamic diffusion trajectories, which describe the diffusion process of gas in the mine roadway. The basic diffusion model considers the geometric shape of the roadway, the speed and direction of the air flow, and the concentration and gradient of the gas to simulate the diffusion process of gas in the roadway. The initial condition is the initial distribution of the gas concentration in the roadway. Usually, these initial concentration values are derived from the concentration data at the gas leakage point. The areas with higher concentrations are often the starting points of diffusion, and gas diffusion will start from these sources. The boundary conditions need to set the gas flow rate and concentration at the outlet (ventilation opening) and inlet (external air input point) of the roadway. According to the set initial conditions and boundary conditions, the hydrodynamic equation will be numerically solved. As time goes by, the concentration distribution, flow rate, and diffusion trajectory of the gas will change continuously. Through diffusion simulation, the change of gas concentration at each position in the roadway over time is obtained, and then the dynamic trajectory of gas diffusion is generated.

[0077] The set of dynamic diffusion trajectories includes time series, diffusion paths, concentration changes, etc. The time series is the spatial distribution of the gas corresponding to each time point, which can describe the spatio-temporal evolution during the gas diffusion process; the diffusion path is the propagation path of the gas in the roadway, showing how the gas diffuses to different areas of the roadway over time, and the concentration change is the change of the gas concentration at each position, which helps to judge the high-risk areas.

[0078] Spatio-temporal graph neural network is a deep learning model that combines graph neural network (GNN) with spatio-temporal data. It represents the relationships between data through a graph structure (graph nodes and graph edges), while considering the changes in time series data. In the risk assessment of mines, spatio-temporal graph neural network can capture the spatio-temporal relationships between nodes (such as different positions in roadways) and edges (such as the spatial dependencies between sensors), so as to conduct risk prediction and anomaly detection. Multimodal time series data of the mine are obtained through a spatio-temporal data cube, that is, each data point represents the sensor value at a specific time, specific location and specific modality. According to the topological structure of the mine, the positions of sensors in the roadway are regarded as the nodes of the graph, and the spatial relationships between sensors are regarded as the edges of the graph. Each sensor node will have spatio-temporal data associated with it.

[0079] The spatial relationships between nodes are captured through the convolutional layer in the graph neural network. For each node, the model updates its own state according to the states of its neighboring nodes. The spatio-temporal graph neural network combines the information of time and space, effectively propagates information through the graph structure, and thus captures the potential patterns in spatio-temporal changes. After being processed by multiple layers of spatio-temporal graph neural network, a high-risk probability distribution matrix is output, which represents the risk probabilities at different times and different positions, specifically manifested as the risk values of each grid or node. Exemplarily, the risk probability distribution matrix is shown in Table 2:

[0080] Table 2 Risk Probability Distribution Matrix

[0081]

[0082] The dynamic diffusion trajectory set is based on the hydrodynamic model of the mine environment, especially the diffusion model of harmful substances such as gas or methane. By inputting environmental parameters (such as the three-dimensional boundaries of roadways, wind speed distribution, methane concentration, etc.), the diffusion trajectories under different time and space conditions are calculated. The output of the dynamic diffusion trajectory set represents the propagation path and changes of harmful substances in the mine environment.

[0083] Based on the risk probability distribution matrix generated by the spatio-temporal graph neural network, a risk probability weight is assigned to each region and time period. A higher risk probability means a greater risk in that region and a higher weight needs to be assigned. Each trajectory in the dynamic diffusion trajectory set represents the propagation path of gas or other risk sources in space. A larger diffusion range or a higher concentration change may indicate a higher risk in that region, so corresponding weights can be set according to the scale and speed of diffusion. For example, if the risk of gas concentration in a certain region is high and the diffusion trajectory shows that this region is the core area of the risk source, the risk weighted value of this region will be higher. The specific weighted fusion formula is R(t,x) = α·P(t,x) + β·E(t,x), where R(t,x) is the comprehensive risk value at time t and position x, P(t,x) is the risk probability at time t and position x, E(t,x) is the diffusion trajectory at time t and position x, and α, β are the weight coefficients of the risk probability and the diffusion trajectory.

[0084] After dynamic weighted fusion, the finally obtained risk diffusion simulation model will simultaneously consider the influences of multiple factors such as gas concentration, wind speed, air flow, etc., and can accurately predict how risks spread and which regions will become potential high-risk areas under different time and space conditions.

[0085] Run the risk diffusion simulation model to conduct dynamic simulation on each position of the roadway and simulate the diffusion path and changes of the risk source. Risk dispersion positioning refers to calculating and simulating the diffusion path and influence range of the risk source to real-time identify and locate different risk regions in mines or other industrial environments.

[0086] According to the risk values (such as gas concentration, fire hazard index, temperature, etc.) and diffusion paths of the operation results, the target mine is divided into multiple hazardous areas, including the risk source partition and multiple risk diffusion partitions. The risk source partition refers to the origin area of the risk event. In the scenario of gas leakage, these areas are the sources of gas leakage. Usually, these areas have a higher risk level and stronger diffusion potential, so they need to be protected and emergency treated preferentially. The risk diffusion partition is the part where the risk spreads from the risk source partition to other areas over time. According to the diffusion trajectory, the risk will spread to other areas along with factors such as air flow and roadway shape. The simulation model divides these areas according to the size of the risk value and labels them as M risk diffusion partitions.

[0087] By simulating the diffusion trajectory and predicting the future risk dynamics, make emergency response preparations in advance according to the risk diffusion simulation model, effectively reduce the occurrence probability and losses of disasters, comprehensively improve the safety of the mine, and optimize resource allocation to ensure the sustainable operation of the mine and the safety of the staff.

[0088] S500: Perform directional activation of the relay identification hierarchical topology based on the multiple real-time risk zones, and output multiple real-time relay identification sequences.

[0089] Furthermore, S500 of the present application includes:

[0090] Deploy the relay identification hierarchical topology in the target mine through signal coverage fitting; perform identification energy consumption analysis on the coverage relationship between the risk source zone and M risk diffusion zones according to the relay identification hierarchical topology to obtain a first real-time relay identification sequence and M real-time relay identification sequences; wherein, the first real-time relay identification sequence and the M real-time relay identification sequences constitute the multiple real-time relay identification sequences.

[0091] Specifically, perform directional activation of the relay identification hierarchical topology according to multiple real-time risk zones. Based on the known risk sources and diffusion directions, combined with the coverage capabilities, locations, and communication link conditions of relay devices, generate a relay identification transfer sequence, specify the activation timing of each relay device, ensure that the network can cover key areas in a timely manner, and maintain a stable communication connection in a dynamic environment. According to the relay identification transfer sequence, activate relay devices one by one for RFID identification of risk sources and risk diffusion directions.

[0092] Deploy the relay identification hierarchical topology in the target mine through signal coverage fitting, determine which areas are critical for the deployment of relay nodes, and construct a hierarchical topology structure based on the actual communication requirements of the mine, that is, determine which areas need to deploy relay devices and the optimal deployment locations of relay devices. When deploying the relay identification hierarchical topology, identify key risk areas in the mine, such as gas leakage sources, fire sources, areas with unstable airflows, etc., and formulate a deployment plan for relay devices through comprehensive analysis of the signal propagation characteristics and environmental factors in these areas. According to the mine tunnel structure and layout, signal coverage capabilities, equipment power consumption limitations, etc. of the target mining area, determine the hierarchical topology structure, where each relay device is divided into different levels according to its deployment location, coverage area, and signal strength to ensure uniform coverage and efficient use of equipment.

[0093] The relay identification hierarchical topology may include a main relay node, secondary relay nodes, and auxiliary relay nodes. Among them, the main relay node is located in the core area of the mine, such as the main roadway or risk source area, and is responsible for handling important communication tasks; the secondary relay nodes are located in the peripheral area or diffusion zone of the mine and are responsible for supplementing the signal coverage of the main node and enhancing the extensiveness of the network; the auxiliary relay nodes are deployed in places with poor communication or signal blind spots to provide temporary signal coverage support.

[0094] According to the location information of the risk source partition and the risk diffusion partition, use the relay identification hierarchical topology to conduct identification energy consumption analysis. The risk source area (such as a gas leakage source, a fire source, etc.) is an area with relatively high danger in the mine, and it is necessary to ensure that these areas are always under effective signal coverage. The energy consumption analysis will consider the signal strength and communication requirements between the risk source area and the relay device to ensure that the relay device can provide optimal signal coverage with the lowest power consumption. The risk diffusion area is a potential danger area that spreads from the risk source area. During the risk diffusion process, the requirements for signal coverage will be continuously adjusted with time and the environment. The energy consumption analysis takes into account the direction, speed, and coverage area of the risk diffusion to help optimize the scheduling of relay devices to achieve the optimal energy efficiency.

[0095] By calculating the relationship between each relay device and the risk area, evaluate its signal transmission strength and energy consumption in different situations. According to the results of the energy consumption analysis, generate a real-time relay identification sequence, which determines the activation order of the relay devices to ensure that the entire mine area, especially the risk source area and the risk diffusion area, always maintains effective signal coverage. The first real-time relay identification sequence is usually the sequence of relay devices that are most prioritized to be activated in the risk source area of the mine, and the M real-time relay identification sequences are the relay device activation sequences generated according to the risk diffusion area.

[0096] Combine the first real-time relay identification sequence and the M real-time relay identification sequences to form multiple real-time relay identification sequences, which are used to guide the operation of the relay nodes during the risk identification and transmission process. By deploying the relay identification hierarchical topology and conducting identification energy consumption analysis, optimize the use of relay nodes, improve communication efficiency, and extend the network lifespan.

[0097] S600: Wake up the multiple real-time relay identification sequences to perform RFID coverage identification on the multiple real-time risk partitions, and obtain multiple real-time personnel distribution characteristics.

[0098] S700: Dynamically allocate rescue resources according to the multiple real-time personnel distribution characteristics.

[0099] Specifically, in the mine area, all staff will wear devices or cards with RFID tags. When the staff enters the risk partition, the RFID tags they wear will be scanned by the RFID readers of the relay nodes. The activated relay nodes will emit RFID signals and scan the RFID tags in the mine. The distance between the tag and the reader determines the transmission time and accuracy of the information, thus realizing accurate personnel positioning. Wake up the corresponding relay identification devices according to the changes in the risk source partition and the risk diffusion partition to ensure that the required risk areas can be covered and identified.

[0100] RFID coverage identification equips underground personnel with RFID (radio frequency identification) and quickly scans the personnel composition of a specific area. RFID usually consists of RFID tags (worn on personnel), RFID readers (installed in key locations in the mine), and a background data management system. When a person enters the signal coverage of the reader, the reader automatically identifies the tag and reads the information stored in it, such as the person's ID, department, entry time, etc., and transmits this data to the background for analysis and display.

[0101] Once the relay identification device is awakened, the target area is identified and located through RFID technology, and radio waves are used for contactless data transmission. By wearing RFID tags on mine workers, the distribution of personnel in the mine can be identified in real time. Each relay device will perform RFID identification on the area it covers and report the location data of the personnel. Through the joint work of RFID technology and relay equipment, the distribution of personnel in each real-time risk zone in the mine is obtained, the number, location, and activity of personnel in each area are monitored, and real-time personnel distribution characteristics are generated, including personnel location, personnel density, personnel flow trajectory, personnel status, etc., to accurately identify the number and distribution of personnel in the risk area, as well as the location of each person.

[0102] Generate rescue paths based on personnel location and status, dynamically allocate rescue resources, and ensure that resources are delivered on demand. First, evaluate the distribution of personnel in each risk zone and identify areas of emergency need. If there are many staff in the area, and the gas concentration is too high or the source of leakage is unstable, prioritize gas detectors, ventilation system repairers, and emergency evacuation teams. Prioritize rescue resources based on the risk area, risk level, and health status of the personnel. Analyze the resource needs of each area in real time, calculate the required amount of resources, and automatically dispatch the corresponding rescue teams and equipment. Guide the rescue team's actions through AR navigation to ensure that the response time is within 5 minutes.

[0103] Through the deployment of multimodal sensors (such as thermal imagers, microseismic monitoring nodes, gas concentration detectors, etc.), the on-site environmental data is obtained in real time. After the data is fed back to the central platform, it is combined with the real-time personnel distribution characteristics to help optimize resource allocation. During the rescue process, as the conditions of personnel and the environment change, rescue resources are dynamically adjusted and optimized. Once the trapped people are rescued, the rescue resources in that area are dispatched to other affected areas.

[0104] Before the actual rescue, based on historical data and real-time data, different resource allocation plans are simulated and evaluated, including time-cost analysis and risk assessment, to ensure the best rescue effect within the shortest time. Through the precise analysis of real-time personnel distribution characteristics and the dynamic scheduling of rescue resources, the mine emergency response system can minimize casualties and ensure the safety of mine personnel when disasters occur.

[0105] In summary, the integrated management method of the mine wireless gateway provided by this application has the following beneficial effects:

[0106] Obtain the roadway topology structure of the target mine through interaction, and divide the target mine into honeycomb area grids according to the roadway topology structure; deploy K multi-modal sensor arrays in K grid partitions in the honeycomb area grids through monitoring coverage fitting; construct communication links between the K multi-modal sensor arrays and K edge gateways, where the edge gateways are built-in with lightweight AI chips; the diffusion risk analysis platform performs an integrated simulation of the risk diffusion path according to the K multi-modal time series features transmitted back by the K edge gateways to locate multiple real-time risk partitions, where the K edge gateways are connected to the diffusion risk analysis platform; perform directional activation of the relay identification hierarchical topology according to the multiple real-time risk partitions, and output multiple real-time relay identification sequences; wake up the multiple real-time relay identification sequences to perform RFID coverage identification on the multiple real-time risk partitions to obtain multiple real-time personnel distribution characteristics; perform dynamic allocation of rescue resources according to the multiple real-time personnel distribution characteristics. That is to say, by dividing the mine into multiple honeycomb area grids, deploying multi-modal sensor arrays to monitor the mine environment in real time, establishing a communication link in the wireless communication gateway to transmit the collected multi-modal time series feature data back to the edge gateway, performing an integrated simulation of the risk diffusion path on the multi-modal time series feature data, locating the risk partitions, through the directional activation of the relay identification hierarchical topology, waking up the relevant relay devices, determining the real-time personnel in the risk source and the risk diffusion direction, and performing dynamic rescue resource allocation, ensuring that in case of an emergency, rescue resources can be quickly and accurately allocated, improving the rescue scheduling ability and efficiency.

[0107] Embodiment 2, based on the same inventive concept as the integrated management method of the mine wireless gateway in Embodiment 1, this application also provides an integrated management system of the mine wireless gateway. Please refer to the appendix Figure 2 The integrated management system of the mine wireless gateway includes:

[0108] The mine division processing module 11 is used to interactively obtain the roadway topological structure of the target mine and divide the target mine into honeycomb area grids according to the roadway topological structure; the monitoring and deployment processing module 12 is used to deploy K multi-modal sensor arrays in K grid partitions in the honeycomb area grids through monitoring coverage fitting; the communication link construction module 13 is used to construct communication links between the K multi-modal sensor arrays and K edge gateways, where the edge gateways are built-in with lightweight AI chips; the risk diffusion simulation module 14 is used for the risk diffusion analysis platform to perform an integrated simulation of risk diffusion paths based on K multi-modal time series features transmitted back by the K edge gateways to locate multiple real-time risk partitions, where the K edge gateways are connected to the risk diffusion analysis platform; the relay identification activation module 15 is used to perform directional activation of the relay identification hierarchical topology according to the multiple real-time risk partitions and output multiple real-time relay identification sequences; the coverage identification processing module 16 is used to wake up the multiple real-time relay identification sequences to perform RFID coverage identification on the multiple real-time risk partitions to obtain multiple real-time personnel distribution characteristics; the resource allocation processing module 17 is used to perform dynamic allocation of rescue resources according to the multiple real-time personnel distribution characteristics.

[0109] Furthermore, the mine division processing module 11 in the integrated management system of mine use wireless gateways is also used to: extract the roadway topological structure through multi-source data acquisition and fusion, and generate a weighted roadway connectivity graph according to the roadway topological structure; extract multiple intersection node coordinates and multiple node weight data of multiple roadway topological nodes from the weighted roadway connectivity graph; perform dynamic grid division according to the multiple intersection node coordinates and multiple node weight data to divide the target mine into the honeycomb area grids.

[0110] Furthermore, the mine division processing module 11 in the integrated management system of mine use wireless gateways is also used to: generate an initial topological skeleton by parsing the BIM model in the design stage of the target mine, where the parsing process includes the extraction of roadway centerlines, intersection point coordinates, and support structure parameters; after using an explosion-proof mobile laser scanner to collect point cloud data along the roadway of the target mine, perform data compression on the collected data through voxel grid filtering to obtain roadway laser point cloud data; obtain a roadway fusion three-dimensional model by registering the roadway laser point cloud data and the initial topological skeleton; extract the roadway centerline from the roadway fusion three-dimensional model to obtain the roadway topological structure; complete the construction of the weighted roadway connectivity graph by injecting dynamic and static weights of nodes into the roadway topological structure.

[0111] Furthermore, the mine division processing module 11 in the integrated management system of the mine-used wireless gateway is further configured to: based on the multiple node spatial positions of the multiple roadway topology nodes in the roadway topology structure; perform lithological stability analysis on the multiple node spatial positions to obtain multiple node static weights; based on the multiple topological edge spatial positions of the multiple roadway topology edges in the roadway topology structure, extract multiple roof displacement data from the associated geological exploration report; calculate the deformation rate of the multiple roof displacement data to obtain multiple topological edge dynamic weights; based on the multiple node spatial positions and the multiple topological edge spatial positions, perform weight injection of the multiple node static weights and the multiple topological edge dynamic weights in the mapping of the roadway topology structure, and complete the construction of the weighted roadway connection graph.

[0112] Furthermore, the mine division processing module 11 in the integrated management system of the mine-used wireless gateway is further configured to: locally call the associated geological exploration report of the target mine; guided by the multiple node spatial positions, extract multiple rock layer stability coefficients from the associated geological exploration report; call the local accident records according to the multiple node spatial positions to obtain multiple node accident frequencies; and obtain the multiple node static weights by normalizing the multiple rock layer stability coefficients and the multiple node accident frequencies.

[0113] Furthermore, the mine division processing module 11 in the integrated management system of the mine-used wireless gateway is further configured to: extract multiple groups of topological edge dynamic weights of the multiple roadway topology nodes according to the connection relationship between the nodes and the topological edges; fuse and calculate the multiple node static weights and the multiple groups of topological edge dynamic weights by the entropy weight method to obtain the multiple node weight data; using the multiple roadway topology nodes as the grid starting point, dynamically correct the grid density according to the multiple node weight data, and divide the target mine into the honeycomb area grid, where the grid weight density deviation of the K grid partitions meets the preset deviation scale.

[0114] Furthermore, the risk diffusion simulation module 14 in the integrated management system of the mine-used wireless gateway is further configured to: perform spatio-temporal alignment mapping of the K multi-modal time series features according to the roadway topology structure and the K sensor deployment coordinates of the K multi-modal sensor arrays to complete the configuration of the spatio-temporal data cube; based on the spatio-temporal data cube, synchronously execute physical-driven modeling and data-driven modeling to obtain a risk diffusion simulation model; perform risk diffusion positioning by running the risk diffusion simulation model to obtain the multiple real-time risk partitions, where the multiple real-time risk partitions include a risk source partition and M risk diffusion partitions.

[0115] Further, the risk diffusion simulation module 14 in the integrated management system of the mine wireless gateway is further configured to: extract the three-dimensional boundary of the roadway, the wind speed distribution of the roadway, and the gas concentration gradient distribution of the roadway from the spatio-temporal data cube; input the three-dimensional boundary of the roadway, the wind speed distribution of the roadway, and the gas concentration gradient distribution into the hydrodynamic equation to construct a basic diffusion model, and output a set of dynamic diffusion trajectories; process the multi-modal time series distribution of the spatio-temporal data cube through a spatio-temporal graph neural network, and output a risk probability distribution matrix; perform dynamic weighted fusion on the set of dynamic diffusion trajectories and the risk probability distribution matrix to generate the risk diffusion simulation model.

[0116] Further, the relay identification and activation module 15 in the integrated management system of the mine wireless gateway is further configured to: deploy the relay identification hierarchical topology in the target mine through signal coverage fitting; perform identification energy consumption analysis on the coverage relationship between the risk source partition and the M risk diffusion partitions according to the relay identification hierarchical topology, and obtain a first real-time relay identification sequence and M real-time relay identification sequences; wherein, the first real-time relay identification sequence and the M real-time relay identification sequences constitute the multiple real-time relay identification sequences.

[0117] The various embodiments in this specification are described in a progressive manner, and the key point of each embodiment is to illustrate the differences from other embodiments. The Figure 1 The integrated management method and specific examples of the mine wireless gateway in the first embodiment are equally applicable to the integrated management system of the mine wireless gateway in this embodiment. Through the detailed description of the integrated management method of the mine wireless gateway above, those skilled in the art can clearly know the integrated management system of the mine wireless gateway in this embodiment. Therefore, for the sake of brevity of the specification, it will not be described in detail here.

[0118] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0119] Obviously, for those skilled in the art, without departing from the principle of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the protection scope of the present application.

Claims

1. The integrated management method of mine wireless gateway is characterized by: include: interactively obtaining a tunnel topology structure of a target mine, and dividing the target mine into honeycomb area grids according to the tunnel topology structure; Deploy K multimodal sensor arrays to K grid partitions in the cellular area grid by monitoring coverage fitting; Establishing a communication link between the K multimodal sensor arrays and the K edge gateways, wherein the edge gateways have a built-in lightweight AI chip; The diffusion risk analysis platform performs integrated simulation of risk diffusion paths according to the K multimodal time series features transmitted back by the K edge gateways to locate multiple real-time risk partitions, wherein the K edge gateways are connected to the diffusion risk analysis platform; performing directed activation of relay identification hierarchical topology according to the multiple real-time risk partitions, and outputting multiple real-time relay identification sequences; Waking up the multiple real-time relay identification sequences to perform RFID coverage identification on the multiple real-time risk partitions to obtain multiple real-time personnel distribution features; Rescue resources are dynamically allocated according to the multiple real-time personnel distribution characteristics.

2. The integrated management method for mine wireless gateways according to claim 1, characterized in that: Interactively obtaining a tunnel topology structure of a target mine, and dividing the target mine into honeycomb area grids according to the tunnel topology structure, including: By collecting and fusing multi-source data, the lane topology structure is extracted, and a weighted lane connectivity graph is generated according to the lane topology structure; Extracting multiple intersection node coordinates and multiple node weight data of multiple lane topology nodes from the weighted lane connectivity graph; Dynamic grid division is performed according to the multiple cross-node coordinates and multiple node weight data to divide the target mine into the honeycomb area grids.

3. The integrated management method of mine wireless gateway according to claim 2, characterized in that: The tunnel topology structure is extracted by multi-source data collection and fusion, and a weighted tunnel connectivity graph is generated according to the tunnel topology structure, including: Generate an initial topological skeleton by analyzing the design phase BIM model of the target mine, wherein the analysis includes extracting the center line of the tunnel, the coordinates of the intersection, and the parameters of the support structure; After using an explosion-proof mobile laser scanner to collect point cloud data along the tunnel of the target mine, the collected data is compressed by voxel grid filtering to obtain tunnel laser point cloud data; By registering the lane laser point cloud data and the initial topological skeleton, a lane fusion three-dimensional model is obtained; Extracting the centerline of the lane from the fused three-dimensional model of the lane to obtain the lane topology structure; The construction of the weighted lane connectivity graph is completed by injecting dynamic and static node weights into the lane topology structure.

4. The integrated management method of mine wireless gateway according to claim 3, characterized in that: The construction of the weighted lane connectivity graph is completed by injecting dynamic and static node weights into the lane topology structure, including: According to a plurality of node spatial positions of the plurality of lane topology nodes in the lane topology structure; Performing lithologic stability analysis on the spatial positions of the multiple nodes to obtain static weights of the multiple nodes; extracting a plurality of roof displacement data from an associated geological exploration report according to a plurality of topological edge spatial positions of a plurality of tunnel topological edges in the tunnel topological structure; By calculating the deformation rate of the plurality of top plate displacement data, a plurality of topological edge dynamic weights are obtained; According to the spatial positions of the multiple nodes and the spatial positions of the multiple topological edges, weight injection of the multiple node static weights and the multiple topological edge dynamic weights is performed in the lane topology structure mapping to complete the construction of the weighted lane connectivity graph.

5. The integrated management method for mine wireless gateways according to claim 1, characterized in that: The diffusion risk analysis platform performs integrated simulation of risk diffusion paths according to the K multimodal time series features transmitted back by the K edge gateways to locate multiple real-time risk partitions, including: According to the lane topology and the K sensor deployment coordinates of the K multimodal sensor arrays, performing spatiotemporal alignment mapping of the K multimodal time series features to complete the configuration of the spatiotemporal data cube; Based on the spatiotemporal data cube, physically driven modeling and data driven modeling are synchronously performed to obtain a risk diffusion simulation model; The risk diffusion simulation model is run to locate risk diffusion, thereby obtaining the multiple real-time risk partitions, wherein the multiple real-time risk partitions include risk source partitions and M risk diffusion partitions.

6. The integrated management method for mine wireless gateways according to claim 5, characterized in that: Based on the spatiotemporal data cube, physical-driven modeling and data-driven modeling are performed simultaneously to obtain a risk diffusion simulation model, including: Extracting the three-dimensional boundary of the tunnel, the wind speed distribution of the tunnel, and the gas concentration gradient distribution of the tunnel from the spatiotemporal data cube; Input the three-dimensional boundary of the tunnel, the wind speed distribution of the tunnel, and the gas concentration gradient distribution of the tunnel into the fluid dynamics equation to construct a basic diffusion model, and output a dynamic diffusion trajectory set; Processing the multimodal time series distribution of the spatiotemporal data cube through a spatiotemporal graph neural network, and outputting a risk probability distribution matrix; The dynamic diffusion trajectory set and the risk probability distribution matrix are dynamically weighted and fused to generate the risk diffusion simulation model.

7. The integrated management method for mine wireless gateways according to claim 4, characterized in that: Performing lithologic stability analysis on the spatial positions of the multiple nodes to obtain multiple node static weights, including: Locally call the relevant geological exploration report of the target mine; Extracting multiple rock formation stability coefficients from the associated geological exploration report based on the spatial positions of the multiple nodes; Calling local accident records according to the spatial positions of the multiple nodes to obtain accident frequencies of the multiple nodes; The static weights of the multiple nodes are obtained by normalizing the multiple rock formation stability coefficients and the multiple node accident frequencies.

8. The integrated management method for mine wireless gateways according to claim 4, characterized in that: Dynamically meshing the target mine into the honeycomb area grids according to the plurality of cross node coordinates and the plurality of node weight data, including: Extracting multiple groups of topological edge dynamic weights of the multiple lane topological nodes according to the connection relationship between the nodes and the topological edges; The plurality of node weight data are obtained by fusing and calculating the plurality of node static weights and the plurality of groups of topological edge dynamic weights through an entropy weight method; A plurality of tunnel topology nodes are used as gridding starting points, and the grid density is dynamically corrected according to the plurality of node weight data, and the target mine is divided into the honeycomb area grids, wherein the grid weight density deviation of the K grid partitions satisfies a preset deviation scale.

9. The integrated management method of mine wireless gateway according to claim 5, characterized in that: Directed activation of relay identification hierarchical topology is performed according to the multiple real-time risk partitions, and multiple real-time relay identification sequences are output, including: Deploy the relay identification hierarchical topology in the target mine through signal coverage fitting; Performing identification energy consumption analysis on the coverage relationship between the risk source partition and the M risk diffusion partitions according to the relay identification hierarchical topology to obtain a first real-time relay identification sequence and M real-time relay identification sequences; The first real-time relay identification sequence and M real-time relay identification sequences constitute the multiple real-time relay identification sequences.

10. The integrated management system of mine wireless gateway is characterized by: Steps for implementing the integrated management method of a mine wireless gateway according to any one of claims 1 to 9, wherein the integrated management system of a mine wireless gateway comprises: A mine division processing module, used for interactively obtaining a tunnel topology structure of a target mine, and dividing the target mine into honeycomb area grids according to the tunnel topology structure; A monitoring deployment processing module, configured to deploy K multimodal sensor arrays on K grid partitions in the cellular area grid by monitoring coverage fitting; A communication link building module, used to build a communication link between the K multimodal sensor arrays and the K edge gateways, wherein the edge gateway has a built-in lightweight AI chip; A risk diffusion simulation module, used for the diffusion risk analysis platform to perform integrated simulation of risk diffusion paths according to the K multimodal time series features returned by the K edge gateways, so as to locate multiple real-time risk partitions, wherein the K edge gateways are connected to the diffusion risk analysis platform; A relay identification activation module, configured to perform directional activation of a relay identification hierarchical topology according to the multiple real-time risk partitions, and output multiple real-time relay identification sequences; A coverage identification processing module is used to wake up the multiple real-time relay identification sequences to perform RFID coverage identification on the multiple real-time risk partitions to obtain multiple real-time personnel distribution features; The resource allocation processing module is used to dynamically allocate rescue resources according to the multiple real-time personnel distribution characteristics.

Images