Emergency command method and system based on Internet of Things

Abstract

The invention provides an emergency command method and system based on the Internet of Things, and relates to the technical field of emergency command systems, and the method comprises the steps: obtaining Internet of Things monitoring data including underground pipeline information, accumulated water depth data and historical disaster information, and carrying out the alignment fusion to generate comprehensive situation data; based on this, determining a waterlogged road section and analyzing a sound signal thereof to identify an acoustic abnormal region; fusing the acoustic features and pipeline attributes by using a manifold learning algorithm to generate risk feature vectors; discretizing a geographic area of a ponding road section into grid cells, and distributing and inputting risk vectors into the cellular automaton model; the model carries out iterative calculation among grids based on a risk propagation rule to generate disaster prediction data; and finally, according to the data and historical information, a command instruction for emergency resource scheduling is generated through a resource scheduling optimization model. According to the invention, the accuracy of early warning and resource scheduling of urban inland inundation secondary disasters is improved.

Description

Technical Field

[0001] This application relates to the field of emergency command system technology, and in particular to an emergency command method and system based on the Internet of Things. Background Technology

[0002] With the acceleration of urbanization, emergency command systems are becoming increasingly important in disaster early warning and resource allocation. The Internet of Things (IoT) provides a new means for this purpose. IoT-based emergency command methods can aggregate various monitoring data in real time, supporting rapid response and scientific decision-making, and have broad application prospects in the field of urban public safety.

[0003] Existing technologies typically focus on acquiring single or multiple real-time data points, such as water depth and meteorological information, through sensor networks, and then visualizing this data on electronic maps. These methods usually also trigger alarms based on preset threshold rules and generate preliminary resource scheduling suggestions by matching against a contingency plan database.

[0004] However, the accuracy of early warning and decision-making in addressing the complex secondary disaster risks caused by urban flooding, such as road collapses or leakage of electrical facilities, is limited. Existing solutions lack effective in-depth fusion and dynamic risk coupling analysis of pipeline layout, geological conditions, and physical field anomaly signals from multi-source heterogeneous data in underground space; furthermore, their resource scheduling is mostly based on static plans, with limited adaptability to the dynamic evolution of disasters, resulting in a need to improve the foresight and targeting of emergency responses. Therefore, existing technologies suffer from insufficient technical capabilities for dynamic prediction and precise emergency command of complex risks associated with urban flooding secondary disasters. Summary of the Invention

[0005] This application provides an emergency command method and system based on the Internet of Things to solve the problem of low accuracy in early warning and resource allocation for secondary disasters caused by urban flooding in the existing technology.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides an emergency command method based on the Internet of Things, comprising:

[0007] Acquire IoT monitoring data, which includes underground pipeline information, water depth information, and historical disaster information;

[0008] The IoT monitoring data is aligned and fused to generate comprehensive situational data;

[0009] Based on the comprehensive situational data, waterlogged road sections are identified, and sound signals collected in the waterlogged road sections are acquired and analyzed to identify areas of acoustic anomaly.

[0010] Using a manifold learning algorithm, the acoustic features of the acoustic anomaly area and the pipeline attributes in the underground pipeline information are fused to generate a risk feature vector;

[0011] The geographical area corresponding to the waterlogged road section is discretized into multiple grid cells, and the risk feature vector is assigned to the corresponding grid cells and then input into the cellular automaton model. Based on the preset risk propagation rules, the cellular automaton model is driven to perform iterative calculations among the multiple grid cells to generate disaster prediction data.

[0012] Based on the disaster prediction data and the historical disaster information, a resource scheduling optimization model is used to generate command instructions, which are used for emergency resource scheduling.

[0013] Optionally, the step of generating command instructions using a resource scheduling optimization model based on the disaster prediction data and the historical disaster information includes:

[0014] By analyzing the disaster prediction data, at least one predicted disaster area and the corresponding disaster type and disaster impact level for each predicted disaster area can be obtained;

[0015] Based on the disaster type, find the corresponding historical cases from the historical disaster information;

[0016] Extract emergency resource information and historical response locations from the aforementioned historical cases;

[0017] The geographical extent of the predicted disaster area, the level of disaster impact, and the emergency resource information are input into the resource scheduling optimization model to generate a resource scheduling scheme.

[0018] Based on the resource scheduling scheme and the historical response locations, command instructions are generated.

[0019] Optionally, the step of inputting the geographical range of the predicted disaster area, the disaster impact level, and the emergency resource information into the resource scheduling optimization model to generate a resource scheduling scheme includes:

[0020] The geographical area of ​​the predicted disaster zone is divided into multiple demand points using the partitioning module of the resource scheduling optimization model.

[0021] The determination module of the resource scheduling optimization model determines the total amount, type, and distribution location of available resources based on the emergency resource information.

[0022] The resource scheduling optimization model's allocation module allocates resource demand priorities to each demand point based on the disaster impact level.

[0023] The calculation module of the resource scheduling optimization model uses the priority of resource demand as the weight, combines the total amount and type of available resources, takes the transportation time from the distribution location to each demand point as the cost constraint, and takes the shortest total resource delivery time as the objective to perform iterative calculations to determine the resource scheduling path, scheduling type and scheduling quantity from the distribution location of each available resource to each demand point.

[0024] The resource scheduling optimization model output module generates a resource scheduling scheme based on the resource scheduling path, the scheduling type, and the scheduling quantity.

[0025] Optionally, the step of using a manifold learning algorithm to fuse the acoustic features of the acoustic anomaly area with the pipeline attributes in the underground pipeline information to generate a risk feature vector includes:

[0026] The acoustic features and pipeline properties are used as the original data set;

[0027] The original data is reduced in dimensionality using a manifold learning algorithm to obtain low-dimensional data.

[0028] Using a pre-defined association rule base, association analysis is performed on the low-dimensional data to obtain effective association data;

[0029] The effective correlation data is combined with the water depth information to form an initial feature vector;

[0030] The initial feature vector is processed using a feature enhancement network to generate a risk feature vector.

[0031] Secondly, this application provides an Internet of Things-based emergency command system, comprising:

[0032] The acquisition module is used to acquire IoT monitoring data, which includes underground pipeline information, water depth information, and historical disaster information.

[0033] The alignment module is used to align and fuse the IoT monitoring data to generate comprehensive situational data;

[0034] The determination module is used to determine waterlogged road sections based on the comprehensive situation data, acquire sound signals collected in the waterlogged road sections, and analyze the sound signals to identify areas of acoustic anomalies.

[0035] The fusion module is used to fuse the acoustic features of the acoustic anomaly area and the pipeline attributes in the underground pipeline information using a manifold learning algorithm to generate a risk feature vector.

[0036] The allocation module is used to discretize the geographical area corresponding to the waterlogged road section into multiple grid cells, and after allocating the risk feature vector to the corresponding grid cells, input it into the cellular automaton model. Based on the preset risk propagation rules, the module drives the cellular automaton model to perform iterative calculations among the multiple grid cells to generate disaster prediction data.

[0037] The generation module is used to generate command instructions based on the disaster prediction data and the historical disaster information using a resource scheduling optimization model. The command instructions are used for emergency resource scheduling.

[0038] Thirdly, this application provides an electronic device, comprising:

[0039] Memory, used to store computer programs;

[0040] A processor for executing the computer program to implement the steps of the Internet of Things-based emergency command method as described in the first aspect above.

[0041] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the Internet of Things-based emergency command method described in the first aspect above.

[0042] The technical solution provided in this application has the following beneficial effects:

[0043] This application acquires IoT monitoring data containing various information, providing a comprehensive data foundation for subsequent analysis. Secondly, by aligning and fusing the data to generate comprehensive situational data, a unified description of the urban underground space status is achieved. Next, based on this situational data, waterlogged road sections are identified and sound signals are analyzed, helping to discover potential areas of abnormal soil structure. Then, a manifold learning algorithm is used to fuse acoustic features and pipeline attributes to generate risk feature vectors, thereby characterizing the composite risk of waterlogging, underground anomalies, and pipeline coupling. Furthermore, risk features are assigned to a grid and iteratively calculated using a cellular automata model, enabling dynamic simulation of the spatiotemporal evolution of disasters. Finally, based on disaster prediction data and historical information, an optimization model is used to generate command instructions, achieving the effect of improving the accuracy and timeliness of emergency resource scheduling.

[0044] Furthermore, this application also analyzes the predicted data to obtain the disaster area, type and level, then searches for historical cases based on the disaster type and extracts resource information and response locations, then inputs the predicted area, disaster level and resource information into the resource scheduling optimization model to generate a resource scheduling plan, and finally generates command instructions based on the plan and historical response locations.

[0045] Furthermore, by combining real-time forecasts with historical experience, the model generates specific and feasible scheduling plans, enabling emergency resources to be allocated to key locations more rationally and efficiently, thereby improving the scientific nature of command and decision-making and the pertinence of response actions.

[0046] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 A flowchart illustrating an IoT-based emergency command method provided in this application embodiment;

[0049] Figure 2 A schematic diagram illustrating a specific implementation of an emergency command method based on the Internet of Things (IoT) provided in this application embodiment;

[0050] Figure 3 This is a schematic diagram of the structure of an Internet of Things-based emergency command system provided in an embodiment of this application. Detailed Implementation

[0051] To address the problems of existing technologies, this application provides an emergency command method based on the Internet of Things (IoT). Its key lies in constructing a coherent chain from risk perception to evolutionary prediction and then to decision optimization through multi-stage data fusion and dynamic simulation technology. Specifically, the method first integrates pipeline information, water depth, and historical data to form a comprehensive data layer reflecting the overall situation of underground space. Then, it combines the analysis of sound signals collected in waterlogged sections to identify acoustic anomaly areas and capture early changes in the physical field. Based on this, a manifold learning algorithm is used to mine the deep correlation between acoustic features and pipeline attributes, generating feature vectors representing complex risks. Subsequently, by gridding the geographical area and applying a cellular automata model, this feature vector is iteratively calculated under the drive of preset risk propagation rules, thereby dynamically simulating the spatiotemporal evolution path and scope of secondary disasters such as collapses or electrical leaks, generating disaster prediction data. Finally, based on this dynamic prediction data and historical cases, a resource scheduling optimization model is used to generate command instructions.

[0052] Therefore, this method transforms previously scattered and static monitoring information into quantitative predictions of the dynamic coupling and evolution trend of risks, enabling emergency command to shift from passively responding to threshold alarms to actively intervening in the risk evolution process. This effectively solves the problems of insufficient dynamic prediction of complex risks and difficulty in forward-looking and accurate command in existing technologies.

[0053] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0054] The core of this application is to provide an emergency command method based on the Internet of Things, and a flowchart of one specific implementation is shown below. Figure 1 As shown, the method includes:

[0055] Step 101: Obtain IoT monitoring data, which includes underground pipeline information, water depth information, and historical disaster information.

[0056] In step 101, IoT monitoring data refers to the real-time or near-real-time data set collected and transmitted by various sensors and monitoring devices deployed in urban infrastructure. These data together constitute the digital perception basis for the status of specific urban areas, especially waterlogged areas and their surrounding environment.

[0057] The underground pipeline information mainly describes the spatial location, direction, type attributes, and burial depth of various pipelines buried under the road. The water depth information reflects the height of the water accumulation on the ground as measured in real time by equipment such as level sensors. The historical disaster information includes records of past flood-related disaster events. These records usually involve the location, type, degree of impact, and emergency response measures taken in relation to the disaster.

[0058] In this embodiment of the application, the required multi-source monitoring data is obtained by accessing and calling the city's deployed Internet of Things sensing network and related database system. Specifically, the pre-entered urban underground pipeline information is retrieved from the geographic information system database, the water depth information is received in real time from the liquid level sensing nodes deployed at flood-prone locations, and historical disaster information associated with the current waterlogged area is queried from the case database of the emergency management department.

[0059] Step 102: Align and fuse the IoT monitoring data to generate comprehensive situational data.

[0060] In step 102, the comprehensive situation data is the data set generated after fusion processing. This data set contains a three-dimensional description of the area surrounding a certain waterlogged section of road, which can simultaneously reveal the water depth at that location, the distribution of underground pipelines, and historical records of disasters occurring under similar conditions.

[0061] In this embodiment, the location information involved in the acquired IoT monitoring data, such as pipeline node coordinates and water accumulation sensor locations, is first uniformly converted to a standard geographic coordinate system. At the same time, the real-time collected water accumulation depth information and the occurrence time of historical disaster information are time-synchronized. Then, based on the unified spatiotemporal reference, the pipeline attributes, real-time water depth values, and historical disaster types describing the same or adjacent geographical locations are associated and bound. Finally, these associated data are organized and stored according to spatial location to generate comprehensive situational data that can comprehensively display the underground and surface conditions of water accumulation sections.

[0062] Step 103: Based on the comprehensive situational data, determine the waterlogged road sections and acquire the sound signals collected in the waterlogged road sections. Analyze the sound signals to identify areas with acoustic anomalies.

[0063] Among them, waterlogged road sections refer to continuous road sections where the water depth exceeds a preset depth threshold; sound signals refer to vibration wave signals propagating in the medium below the waterlogged road sections, which may originate from physical structural changes such as soil erosion or the formation of cavities; and acoustically abnormal areas refer to geographical locations identified by analyzing sound signals whose acoustic characteristics are significantly different from the normal soil background, suggesting the existence of underground structural anomalies.

[0064] In this embodiment, step 103 includes the following process:

[0065] Step 1031: Mark the road sections in the comprehensive situation data whose water depth value is greater than the preset depth threshold on the geographic layer, and designate them as waterlogged road sections.

[0066] In step 1031, the preset depth threshold is a pre-set value used to determine whether the surface water has reached a level that requires attention and detailed risk detection.

[0067] The embodiments of this application do not specifically limit the value of the preset depth threshold; it can be set according to the actual situation.

[0068] In this embodiment of the application, the water depth information contained in the comprehensive situation data is read and compared one by one with the preset depth threshold; then the geographical locations corresponding to all sensors with water depth values ​​greater than the threshold are connected and marked on the electronic map, thereby determining the waterlogged road sections that need to be monitored.

[0069] Step 1032: Deploy multiple acoustic sensors at preset intervals along the waterlogged section of road. The multiple acoustic sensors form a sensor array, and acquire sound signals generated at different locations below the waterlogged section of road in a synchronous acquisition manner through the sensor array.

[0070] In step 1032, the preset spacing is a fixed distance set in advance for deploying two adjacent acoustic sensors. The sensor array refers to a collection of multiple acoustic sensors that can work together and are arranged along the waterlogged section according to the preset spacing. The synchronous acquisition method means that each acoustic sensor in the sensor array acquires sound signals at the same time reference.

[0071] In this embodiment, the installation points of the acoustic sensors are planned according to the geographical range of the waterlogged road section and the preset spacing. Then, all the deployed acoustic sensors are connected to the same synchronous clock network through wired or wireless communication, so that they can start collecting sound signals from below the road at the same time. Finally, the sensor array collects and uploads the sound signal data from multiple channels.

[0072] Step 1033: Perform spectral transformation on the sound signal to obtain spectral data, calculate the energy value of the spectral data within a preset abnormal frequency band range, and identify key spectral components with energy values ​​greater than a preset energy threshold.

[0073] In step 1033, the spectrum data refers to the data set obtained after the spectrum transformation, which describes the energy or amplitude of the sound signal at various frequency points; the preset abnormal frequency band range refers to a specific frequency range that is pre-set based on historical experience or experimental data and is associated with the abnormality of the underground soil structure.

[0074] The preset energy threshold is a pre-set value used to determine whether the energy of a certain frequency component has reached an abnormal level; the key spectral component refers to the frequency component whose energy value exceeds the preset energy threshold.

[0075] In this embodiment, a fast Fourier transform is first performed on the sound signal of each channel acquired by the sensor array to convert the time-domain sound signal into spectral data. Then, for the converted spectral data, the total energy value within the preset abnormal frequency band is extracted and calculated. Next, the calculated energy value is compared with the preset energy threshold. If the energy value is greater than the threshold, it is considered that an anomaly has been detected within the frequency band, and the spectral components of the energy within the frequency band are marked as key spectral components.

[0076] Step 1034: Calculate the relative time difference between the sound signal and the different acoustic sensors in the sensor array.

[0077] In step 1034, the relative time difference refers to the time difference between the arrival of the same sound signal event, such as the acoustic emission generated by a soil rupture, at different locations of the acoustic sensors in the sensor array.

[0078] In this embodiment of the application, firstly, signal waveforms belonging to the same physical event are identified from the multi-channel sound signal data collected by the sensor array; then, the time delay between the arrival of the signal waveform at any two acoustic sensors in the array is accurately calculated using a cross-correlation algorithm, thereby obtaining a series of relative time differences.

[0079] Step 1035: Combining the key spectral components with the relative time difference, determine the acoustic anomaly area with soil structure changes below the waterlogged road section using the cross-location method.

[0080] In this embodiment of the application, the key spectral components identified in step 1033 are used as the basis for determining the existence of a valid acoustic emission event, and the multiple relative time differences calculated in step 1034 are used as inputs; then, based on the propagation speed model of sound waves in the medium, a time difference positioning algorithm is used to calculate the coordinates of the most likely underground source point that generated the sound signal event by solving a set of equations with the location of the acoustic sensor as the known point and the relative time difference as the constraint; finally, the geographical area where all the source points located in this way are located is defined as one or more of the acoustic anomaly areas.

[0081] This application analyzes sound signals collected from waterlogged road sections to not only identify anomalies from frequency characteristics, but also achieves precise spatial positioning by combining the relative time difference of the signals arriving at different sensors. This enables non-invasive and accurate identification and positioning of changes in underground soil structure, providing direct physical evidence for subsequent risk quantification.

[0082] Step 104: Using a manifold learning algorithm, fuse the acoustic features of the acoustic anomaly area with the pipeline attributes in the underground pipeline information to generate a risk feature vector.

[0083] Among them, the risk feature vector is a numerical feature representation that integrates multiple risk factors such as sound anomalies, pipeline information, and water depth, and is used to quantify the comprehensive risk level of secondary disasters occurring at a specific location.

[0084] In this embodiment, step 104 includes the following process:

[0085] Step 1041: Use the acoustic features and pipeline attributes as the original data set.

[0086] In step 1041, the original dataset is the overall dataset to be processed, which is formed by bringing together different types of data.

[0087] In this embodiment, acoustic features corresponding to the acoustic anomaly area are first extracted, such as dominant frequency and average signal strength; at the same time, underground pipeline information that overlaps with or is adjacent to the acoustic anomaly area is obtained, including pipeline type and burial depth; then the extracted acoustic features and the obtained pipeline attributes are combined to form the original data set for subsequent fusion processing.

[0088] Step 1042: Use the manifold learning algorithm to reduce the dimensionality of the original data to obtain low-dimensional data.

[0089] In step 1042, low-dimensional data refers to the new data representation obtained after dimensionality reduction of the original acoustic features and pipeline attributes through manifold learning algorithms. Its dimension is lower than that of the original data, and it is used to reflect the inherent, simplified structural relationship and coupling pattern between different sound anomaly patterns and different underground pipeline configurations in a unified low-dimensional space.

[0090] In this embodiment, the manifold learning algorithm, such as isometric mapping or local linear embedding algorithm, is applied to the original dataset. The algorithm constructs a low-dimensional manifold that reflects the inherent geometric structure of the data by analyzing the proximity relationships between the original data points, and maps each original data point to the corresponding position in this low-dimensional manifold space, thereby obtaining low-dimensional data that retains the key information of the original data structure.

[0091] Step 1043: Using a preset association rule base, perform association analysis on the low-dimensional data to obtain effective association data.

[0092] In step 1043, the preset association rule base is a pre-established set of rules containing multiple risk patterns. These rules define the risk association relationships corresponding to different combinations of acoustic features and different pipeline attributes. Valid association data refers to low-dimensional data points that are determined to meet the high-risk pattern after being filtered by the association rule base.

[0093] In this embodiment of the application, the low-dimensional data is input into the preset association rule base, and the acoustic features and pipeline attributes represented by each low-dimensional data point are combined and matched with various risk association rules predefined in the rule base; then, those data points that match the high-risk rules are selected, and these data points and their corresponding low-dimensional coordinates are extracted as valid association data.

[0094] Step 1044: Combine the effective correlation data with the water depth information to form an initial feature vector.

[0095] In step 1044, the initial feature vector refers to the intermediate feature representation formed by splicing the filtered valid correlation data with the water depth value at the corresponding location. It is used to comprehensively reflect the preliminary combination state of the three key risk factors—identified abnormal sound patterns, related pipeline attributes, and real-time water depth—under a specific geographical location.

[0096] In this embodiment of the application, the water depth value corresponding to the geographical location of the effective correlation data is first obtained from the comprehensive situation data; then, the low-dimensional coordinate values ​​of the effective correlation data and the water depth value are concatenated together in sequence to form a longer numerical sequence, which is the initial feature vector.

[0097] Step 1045: Process the initial feature vector using a feature enhancement network to generate a risk feature vector.

[0098] In step 1045, the feature enhancement network can be designed as an example of a feedforward neural network structure containing an input layer, at least one fully connected hidden layer, and an output layer. The number of neurons in the input layer is the same as the dimension of the initial feature vector, the hidden layer uses the ReLU activation function to introduce nonlinearity, and the number of neurons in the output layer is consistent with the dimension of the target risk feature vector. The training process of this network is as follows: first, a large number of historical disaster event cases and their corresponding initial feature vectors are collected as training samples, and disaster risk levels are labeled as supervision signals. Then, the network weight parameters are iteratively optimized through the backpropagation algorithm to minimize the difference loss between the risk feature vector output by the network and the true risk level, so that the trained network can learn from the initial features and extract patterns that are more discriminative for risk judgment.

[0099] It should be noted that the above structure is exemplary. This application does not impose specific limitations on the structural design of the layers and other components used in the internal structure of the feature enhancement network. The corresponding settings can be made according to the actual situation.

[0100] In this embodiment of the application, the feature enhancement network is trained using a large amount of historical disaster case data. The training objective is to enable the network to learn the complex mapping relationship from initial features to disaster risk. In application, the initial feature vector is input into the trained feature enhancement network. The network reconstructs and enhances the initial features through its internal multi-layer nonlinear transformation, and finally outputs a numerical vector with fixed dimensions that can better distinguish risk levels, namely the risk feature vector.

[0101] This application achieves early and accurate prediction and forward-looking emergency command of the hidden secondary disaster risks caused by urban flooding by multi-level fusion and dynamic simulation of sound signals, pipeline data and water depth.

[0102] Step 105: Discretize the geographical area corresponding to the waterlogged road section into multiple grid cells, and after assigning the risk feature vector to the corresponding grid cells, input it into the cellular automaton model. Based on the preset risk propagation rules, drive the cellular automaton model to perform iterative calculations among the multiple grid cells to generate disaster prediction data.

[0103] Among them, a grid cell is a number of small squares obtained by dividing the continuous geographical area corresponding to the waterlogged road section into a certain size. Each grid cell is the basic unit for simulation calculation in the cellular automaton model.

[0104] Cellular automata are discrete dynamic models that simulate the spatiotemporal evolution of complex systems through local interactions of a large number of simple individual cells. Furthermore, this application does not impose specific limitations on the structural design of the internal structure of cellular automata, such as the number of layers, which can be set according to the actual situation.

[0105] The logic of the pre-set risk propagation rules is based on the physical mechanism of secondary disasters caused by urban flooding. It quantifies geological conditions, hydraulic connections, and electrical connectivity into coefficients that affect the transmission of risk between spatial grids. The content may include: the geological coefficient is determined based on the geological structure information in the comprehensive situation data, which characterizes the ease or difficulty of soil instability or cavity expansion under different soil conditions; the hydraulic coefficient is determined based on real-time water depth and water flow direction information, which characterizes the driving strength of water infiltration or pressure gradient on the propagation of risk along the hydraulic path; and the electrical coefficient is determined based on the pipeline type and topological connection relationship in the underground pipeline distribution information, which characterizes the possibility of leakage risk being transmitted between energized facilities through the pipeline network.

[0106] Disaster prediction data is a quantitative result output by the cellular automata model after the simulation is completed, reflecting how disaster risks may be spatially distributed and evolve over a period of time in the future.

[0107] In this embodiment, step 105 includes the following process:

[0108] Step 1051: Assign the risk feature vector to one or more corresponding grid cells as the initial risk value of the grid cells.

[0109] In step 1051, the initial risk value refers to the risk quantification value carried by each grid cell at the start of the simulation, which is directly derived from the risk feature vector generated in step 104.

[0110] In this embodiment of the application, the specific geographical location corresponding to the risk feature vector is first determined; then, the grid cell to which it belongs is found based on the geographical location; finally, the comprehensive risk quantification value contained in the risk feature vector is assigned to the grid cell as the initial risk value of the grid cell.

[0111] Step 1052: Obtain the geological coefficient, hydraulic coefficient, and electrical coefficient from the risk propagation rules.

[0112] In step 1052, the geological coefficient is a parameter used to quantify the impact of soil or rock strata characteristics on the propagation of collapse risk, the hydraulic coefficient is a parameter used to quantify the impact of water flow on risk propagation, and the electrical coefficient is a parameter used to quantify the impact of underground pipeline electrical connectivity on the propagation of leakage risk.

[0113] In this embodiment of the application, the corresponding geological coefficients, hydraulic coefficients, and electrical coefficients are queried and extracted from the pre-built risk propagation rule base according to the specific geological type, water flow conditions, and pipeline network configuration of the current waterlogged road section. These coefficients will serve as key parameters for determining the intensity and direction of risk propagation in subsequent iterative calculations.

[0114] Step 1053: In the cellular automaton model, for each grid cell, calculate the risk value for the next iteration based on its own initial risk value, the initial risk values ​​of adjacent grid cells, and the geological, hydraulic, and electrical coefficients.

[0115] In this embodiment, for each grid cell, the cellular automata model first identifies its multiple directly adjacent grid cells, for example, using von Neumann neighborhood or Moore neighborhood; then, according to a preset risk propagation formula, which takes the current grid cell's own risk value, the risk values ​​of all its adjacent grid cells, and the geological coefficient, hydraulic coefficient, and electrical coefficient as input variables, calculates a new value representing the risk level of the grid cell at the next moment through weighted summation or other function operations. This new value is the risk value for the next iteration.

[0116] Step 1054: The cellular automaton model updates the state of all grid cells based on the calculated risk value, completing one iteration.

[0117] In this embodiment of the application, after calculating the risk value for the next iteration for all grid cells in step 1053, the cellular automaton model will use these newly calculated risk values ​​to uniformly and simultaneously replace the current risk values ​​of all grid cells, thereby completing a global state update, which marks the end of a complete iteration calculation.

[0118] Step 1055: Repeat the update process multiple times until the risk value of all grid cells reaches a stable state or the preset number of iterations is reached.

[0119] In this embodiment of the application, the distribution of grid cell risk values ​​after one iteration update in step 1054 is taken as the new initial state, and the calculation and update process of steps 1053 and 1054 is repeated again. This process is executed repeatedly. When the risk value change of all grid cells is less than a small threshold in multiple consecutive iterations, the model is considered to have reached a stable state and the iteration stops. Alternatively, when the number of iterations reaches a preset maximum number, the iteration is forcibly stopped regardless of whether it is stable or not.

[0120] Step 1056: Generate disaster prediction data based on the risk value distribution of all grid cells after multiple iterations.

[0121] In this embodiment of the application, after the model iteration stops, the risk value of each grid cell obtained at the end of the simulation period represents the potential disaster risk level of that location. By organizing and visualizing the risk values ​​of all grid cells according to their geographical location, a risk distribution map can be generated. This map and the risk level information of each location contained therein constitute the disaster prediction data.

[0122] This application uses a cellular automata model to simulate the dynamic propagation process of risk in a spatial grid, transforming static risk point assessment into a quantitative prediction of the spatiotemporal evolution trend of risk, thereby enabling a forward-looking judgment of the possible spread range and path of disasters.

[0123] Step 106: Based on the disaster prediction data and the historical disaster information, generate command instructions using the resource scheduling optimization model. The command instructions are used for emergency resource scheduling.

[0124] Command instructions are directives that include specific resource allocation plans and execution objectives, used to guide emergency responders to deliver specific resources to specific locations at specific times.

[0125] The resource scheduling optimization model can include multiple functional modules in its structural design. For example, the partitioning module is responsible for spatial geometry calculation, the determination module is responsible for database query and filtering, the allocation module is responsible for calculating priority weights based on rules, the calculation module, as the core optimization engine, can be implemented using a software library of linear programming solvers or heuristic algorithms, and the output module is responsible for formatting and encapsulating the results.

[0126] As an integrated system, the "training" process of this model mainly involves the optimization of the rules or parameters of the computation module. Specifically, it involves collecting a large number of historical successful scheduling cases and their corresponding input data such as demand points, resources, and road conditions, as well as the final actual delivery time as the optimization target. Using mathematical optimization methods or machine learning techniques, the key parameters affecting scheduling decisions in the model are iteratively adjusted so that the scheduling scheme output by the model can be closer to the historical optimal solution in simulation tests, thereby continuously optimizing its decision quality.

[0127] It should be noted that the above structure is exemplary. This application does not impose specific limitations on the structural design of the layers and other components used in the internal structure of the resource scheduling optimization model. The corresponding settings can be made according to the actual situation.

[0128] A command instruction is a specific set of actionable orders that clearly specifies the type and quantity of resources to be dispatched, their origin and destination, and the recommended transportation route.

[0129] In this embodiment, step 106 includes the following process, such as... Figure 2 As shown:

[0130] Step 1061: Analyze the disaster prediction data to obtain at least one predicted disaster area and the disaster type and disaster impact level corresponding to each predicted disaster area.

[0131] In step 1061, the predicted disaster area is the geographical area marked with a high risk in the disaster prediction data; the disaster type refers to the specific category of disaster that is predicted to occur, such as road collapse or power facility leakage; the disaster impact level is a level used to represent the potential severity of the disaster, which is quantified according to the risk value.

[0132] In this embodiment, the risk value and geographical location of each grid cell in the disaster prediction data are first read. Then, a threshold is set according to the level of the risk value, and adjacent grid cells with risk values ​​exceeding the threshold are aggregated together to form one or more continuous predicted disaster areas. Next, the disaster type is determined according to the main factors that generate the risk of the area. Finally, the disaster impact level of the predicted disaster area is determined according to the average or maximum risk value in the area and the preset level classification rules.

[0133] In practical applications, for example, disaster prediction data shows that the risk value of an area composed of 12 grid cells all exceeds 0.7. After aggregation, a predicted disaster area A is formed. According to the rules used in the cellular automata simulation, the risk in this area is mainly caused by water accumulation coupled to power lines, so the disaster type is determined to be leakage risk. The average risk value of this area is 0.85. According to the preset rule of "above 0.8 is level three, and 0.6 to 0.8 is level two", the disaster impact level is determined to be level three.

[0134] Step 1062: Based on the disaster type, find the corresponding historical cases from the historical disaster information.

[0135] In step 1062, historical cases are real disaster events that have occurred in the past and have been recorded and archived, along with their handling processes.

[0136] In this embodiment of the application, the disaster type obtained in step 1061 is used as the query keyword. A retrieval operation is performed in the database storing historical disaster information to filter out all historical records that match or partially match the disaster type field with the keyword. These records constitute the corresponding historical case set.

[0137] In practical applications, for a predicted disaster area A with a disaster type of "leakage risk", all historical disaster event records marked as "leakage" or "electricity-related" in the past five years are retrieved from the database, such as Case C1 and Case C2.

[0138] Step 1063: Extract emergency resource information and historical response locations from the historical cases.

[0139] In step 1063, the emergency resource information describes the types and quantities of specific rescue or disposal materials, equipment and personnel called up in historical cases, and the historical response location is the specific location where these resources were ultimately delivered or deployed in historical cases.

[0140] In this embodiment of the application, for each historical case found in step 1062, the fields of "resources used" and "disposal location" in its record are read; the descriptions of resource types and quantities extracted from multiple cases are summarized and deduplicated to form an emergency resource information list containing multiple resource options; at the same time, the historical response location corresponding to each case is recorded.

[0141] In practical applications, "50 pairs of insulating gloves, 10 voltage detectors, and 2 emergency vehicles" were extracted from historical case C1, with the response location being "the intersection of Renmin Road and Guangming Street"; "30 pairs of insulating boots, 20 warning signs, and 1 generator" were extracted from case C2, with the response location being "the east gate of Zhongshan Park"; the summarized emergency resource information list includes six types of resources and their corresponding quantities: insulating gloves, voltage detectors, emergency vehicles, insulating boots, warning signs, and generators, and the historical response locations include the two specific locations mentioned above.

[0142] Step 1064: Input the geographical range of the predicted disaster area, the disaster impact level, and the emergency resource information into the resource scheduling optimization model to generate a resource scheduling scheme.

[0143] Step 1064 may specifically include the following steps:

[0144] A1: The geographical area of ​​the predicted disaster zone is divided into multiple demand points through the partitioning module of the resource scheduling optimization model.

[0145] In step A1, the demand point is a specific location point representing local resource demand, formed by discretizing a continuous disaster area in the resource scheduling optimization model.

[0146] In this embodiment of the application, the partitioning module divides the geographical area of ​​the predicted disaster area according to a preset grid size, such as a square grid with a side length of 100 meters; the center point of each grid is defined as a demand point, and the set of all demand points is used to represent the overall resource demand distribution of the area.

[0147] In practical applications, the disaster area A is predicted to be an approximately rectangular area, 300 meters long and 200 meters wide. The partitioning module divides it into 6 grids with a side length of 100 meters, and takes the center coordinates of each grid as the demand point, thus obtaining the demand points D1, D2, D3, D4, D5, and D6.

[0148] A2: The determination module of the resource scheduling optimization model determines the total amount, type, and distribution location of available resources based on the emergency resource information.

[0149] In step A2, available resources refer to resources that can be immediately mobilized from various emergency material warehouses, rescue team bases, and other locations. The distribution location is the geographical location of the warehouses or team bases where these available resources are currently stored.

[0150] In this embodiment of the application, the determination module first accesses the urban emergency resource management database to query the real-time inventory and warehouse location of various types of resources; then, it compares the query results with the emergency resource information list extracted in step 1063, filters out the resource types mentioned in the list that have an inventory greater than zero, their specific quantities and warehouse locations, and forms the final list of available resources.

[0151] In practical application, database queries revealed that warehouse W1 currently has 80 pairs of insulating gloves and 15 voltage detectors; warehouse W2 has 3 emergency vehicles and 40 pairs of insulating boots; and warehouse W3 has 50 warning signs. Generator inventory is zero. Therefore, the identified available resource types include insulating gloves, voltage detectors, emergency vehicles, insulating boots, and warning signs, with total quantities of 80, 15, 3, 40, and 50 respectively, distributed in warehouses W1, W2, W3.

[0152] A3: The allocation module of the resource scheduling optimization model allocates resource demand priority to each demand point according to the disaster impact level.

[0153] In step A3, resource demand priority is a numerical weight used to represent the urgency of different demand points for resources. The higher the priority, the more urgently resources need to be obtained.

[0154] In this embodiment of the application, the allocation module sets a basic priority coefficient for the entire predicted disaster area based on the disaster impact level determined in step 1061; then, it fine-tunes the basic coefficient by combining the specific risk value of the grid where each demand point is located in the disaster prediction data; the higher the risk value of the demand point, the greater the calculated resource demand priority value.

[0155] In practical applications, the disaster impact level of disaster area A is predicted to be level three, with a corresponding basic priority coefficient of 1.0; the risk value of the grid where demand point D1 is located is 0.9, so its resource demand priority is 1.0×(0.9 / 0.85)≈1.06; similarly, the priorities of D2 to D6 are calculated.

[0156] A4: The calculation module of the resource scheduling optimization model uses the resource demand priority as the weight, combines the total amount and type of available resources, takes the transportation time from the distribution location to each demand point as the cost constraint, and takes the shortest total resource delivery time as the objective to perform iterative calculations to determine the resource scheduling path, scheduling type and scheduling quantity from the distribution location of each available resource to each demand point.

[0157] In step A4, the transportation time is the time required to drive from the resource distribution location to the demand point, calculated based on real-time or predicted traffic conditions; the total resource delivery time is the weighted sum of the transportation times of all scheduled resources from their origin to their respective destinations, and the weight is the priority of resource demand.

[0158] The resource scheduling path specifies which batch of resources is transported from which warehouse to which demand point, while the scheduling type and scheduling quantity specify what resources are being transported along this path and how many are transported.

[0159] In this embodiment, the calculation module takes as input all the demand points and their priorities, all available resource distribution locations and their resource types and quantities, and the transportation time matrix between each location obtained in the above steps; then, the module constructs the resource scheduling problem as a linear programming or heuristic optimization problem with the goal of minimizing the total resource delivery time; by iteratively solving the optimization problem, it finally outputs a set of specific scheduling instructions. This set of instructions clarifies how much of each type of resource should be taken from which warehouse and delivered to which one or more demand points, while ensuring that the total resource limit is not exceeded and the demand of high priority points is met as much as possible.

[0160] In practical applications, the calculation module solves the following solution: transport 50 pairs of insulating gloves and 10 voltage detectors from warehouse W1 to demand point D1, and transport 30 pairs of insulating gloves and 5 voltage detectors to demand point D2; transport 2 emergency vehicles and 25 pairs of insulating boots from warehouse W2 to D1, and transport 1 emergency vehicle and 15 pairs of insulating boots to D3; transport 30 warning signs from warehouse W3 to D1, and transport 20 warning signs to D2.

[0161] A5: The resource scheduling scheme is constructed through the output module of the resource scheduling optimization model based on the resource scheduling path, the scheduling type, and the scheduling quantity.

[0162] In this embodiment, the output module formats and organizes the specific scheduling instructions obtained by the calculation module to generate a text or structured data document with a clear structure that contains detailed scheduling paths, resource types and quantities. This document is the resource scheduling scheme.

[0163] In practical applications, the generated resource scheduling plan is a table with columns including "departure warehouse", "target demand point", "resource type", "scheduled quantity" and "estimated transportation time".

[0164] Step 1065: Generate command instructions based on the resource scheduling scheme and the historical response locations.

[0165] In this embodiment of the application, the resource scheduling scheme generated in step A5 is combined with the historical response location information extracted in step 1063; based on the location of the resource delivery demand point in the resource scheduling scheme and with reference to the nearby historical response locations, a specific and easily identifiable final delivery location or assembly point is determined for the personnel or vehicles performing the scheduling task; finally, an action order containing complete information such as "scheduling content", "departure point", "destination", and "contact person" is generated, namely the command instruction.

[0166] In practical application, the resource scheduling plan requires the resources to be delivered to the demand point D1, which is geographically close to the historical response location "intersection of Renmin Road and Guangming Street". Therefore, in the generated command instructions, "intersection of Renmin Road and Guangming Street" is clearly designated as the resource receiving point. The instructions are: "Please have warehouse W1 deliver 50 pairs of insulating gloves and 10 voltage detectors to the intersection of Renmin Road and Guangming Street within 30 minutes. Contact person: Captain Li".

[0167] In this embodiment, after step 106, the following process is also included:

[0168] B1: Obtain real-time traffic information and resource location information.

[0169] In step B1, real-time traffic information describes the current traffic conditions of various roads in the city, such as whether there is congestion and traffic speed; resource location information refers to the real-time geographical location of emergency vehicles or teams performing dispatch tasks, which is usually obtained through vehicle-mounted GPS devices.

[0170] In this embodiment of the application, the current average vehicle speed or congestion level information of the main roads is obtained by accessing the real-time traffic data interface of the urban traffic management department; at the same time, the real-time location coordinates transmitted by the GPS device on the dispatched emergency vehicle are obtained through the Internet of Things platform.

[0171] In practical applications, the route from warehouse W1 to the intersection of Renmin Road and Guangming Street is currently congested on Zhongshan Road, with an average speed of 20 kilometers per hour. At the same time, the real-time location of trucks that have departed from warehouse W1 is obtained, such as their latitude and longitude coordinates.

[0172] B2: Input the command instructions, the real-time traffic information, and the resource location information into the path dynamic planning model. The path dynamic planning model generates a dynamic scheduling path based on a traffic congestion prediction algorithm. The dynamic scheduling path indicates the optimal path from the resource location to the target delivery location in the command instructions.

[0173] In step B2, the path dynamic programming model can include a road network graph construction module, a cost calculation module, and a path search module in its structural design. The road network graph construction module constructs a topology graph with urban road intersections as nodes and road segments as edges. The cost calculation module dynamically assigns a cost weight representing travel time to each edge by combining real-time traffic information and congestion prediction algorithms. The path search module uses an improved Dijkstra algorithm as the core searcher.

[0174] The training process of this model mainly focuses on its internal congestion prediction algorithm. By collecting historical traffic flow data, weather data, time data, and corresponding actual travel times, it uses time series prediction models such as Long Short-Term Memory networks for supervised learning. The goal is to train the network parameters with the aim of minimizing the error between the predicted travel time and the actual travel time, so that the model can more accurately predict future traffic conditions and calculate more reasonable dynamic routes.

[0175] Traffic congestion prediction algorithms can be designed using time series prediction models based on long short-term memory networks. This model takes historical traffic flow data, real-time traffic flow data, weather conditions, day of the week, and holiday markers as input features. It processes time dependencies through its internal forget gate, input gate, and output gate structure, and finally outputs a predicted value for the average vehicle speed or congestion level of each road segment in a specific future time period. For example, if the input features include the vehicle speed sequence of a road segment in the past hour, the current rainfall intensity, and whether it is a weekday evening rush hour, the trained long short-term memory network model will calculate and output a prediction that the average vehicle speed of that road segment will drop to 20 kilometers per hour in the next 30 minutes, which corresponds to a yellow congestion level.

[0176] It should be noted that the above structure is exemplary. This application does not impose specific limitations on the structural design of the internal structure of the path dynamic programming model and the traffic congestion prediction algorithm, such as the layers used. The corresponding settings can be made according to the actual situation.

[0177] Dynamic dispatching routes refer to the optimal driving routes from the current location of emergency resources to the designated target delivery location, calculated and dynamically updated in real time by a dynamic path planning model during the emergency resource dispatching process, based on real-time changes in traffic conditions, vehicle locations, and predicted future congestion. These routes are not fixed but are adjusted according to dynamic changes in the traffic environment to ensure that resources can be delivered in the shortest time or with the highest efficiency.

[0178] In this embodiment, the path dynamic programming model takes the target delivery location in the command instruction as the endpoint, the real-time coordinates of the resource location information as the starting point, and the future road conditions output by the real-time traffic information and congestion prediction algorithm as the road weights. Then, the model uses the Dijkstra algorithm or the A* search algorithm to search for a path from the starting point to the endpoint in the road network graph with the shortest overall travel time. This path is the dynamic scheduling path.

[0179] In practical applications, the model predicted that the congestion on Zhongshan Road would worsen in 10 minutes. Therefore, the calculated dynamic scheduling route avoided Zhongshan Road and instead detoured via Jianshe Road, which had less traffic. Although the distance increased by 2 kilometers, the total time was expected to be shortened by 8 minutes.

[0180] B3: Send the dynamic scheduling path and the command instructions to the corresponding mobile emergency terminal, and receive and record the confirmation information and status feedback information from the mobile emergency terminal through the task confirmation and feedback module deployed on the mobile emergency terminal.

[0181] In step B3, the mobile emergency terminal refers to a smart handheld device or vehicle-mounted terminal assigned to frontline emergency personnel or installed on emergency vehicles; the task confirmation and feedback module is a software program running on the mobile emergency terminal, used to realize the functions of receiving instructions, confirming replies, and reporting execution status.

[0182] Confirmation information refers to a feedback signal sent by the operator through the terminal interface or automatically by the terminal to the command center after the mobile emergency terminal receives the command instructions or dynamic dispatch path issued by the command center, indicating that the instructions have been successfully received and acknowledged.

[0183] Status feedback information refers to the real-time status data updates that mobile emergency terminals send to the command center periodically or manually during the execution of instructions, based on specific events such as vehicle departure, arrival at a midway point, encountering obstacles, or completion of a task. These updates are used to report the current execution progress, resource location, vehicle status, or problems encountered.

[0184] In this embodiment, the command center pushes updated command instructions containing detailed information on the dynamic scheduling route to the corresponding vehicle terminal via a wireless communication network. After receiving the instructions, the task confirmation and feedback module on the mobile emergency terminal automatically pops up a prompt and asks the driver to confirm. After the driver clicks to confirm, the terminal automatically sends a confirmation message of "instruction received" to the command center. During transportation, the terminal periodically or according to events, such as reaching a midway point or encountering obstacles, automatically reports its location and status, such as "departed", "arrived at Jianshe Road", "expected to arrive in 5 minutes".

[0185] This application combines dynamic disaster prediction with historical experience, real-time traffic and resource locations, and uses an optimization model for calculation and dynamic adjustment to ultimately generate specific and actionable instructions and achieve closed-loop feedback in the execution process. This enables intelligent command of the entire process of emergency resources from demand prediction to precise and efficient delivery.

[0186] Figure 3 A schematic diagram of the structure of an Internet of Things-based emergency command system provided in this application embodiment is shown below. Figure 3 As shown, the system includes:

[0187] The acquisition module 31 is used to acquire IoT monitoring data, which includes underground pipeline information, water depth information, and historical disaster information.

[0188] Alignment module 32 is used to align and fuse the IoT monitoring data to generate comprehensive situational data.

[0189] The determination module 33 is used to determine the waterlogged road section based on the comprehensive situation data, acquire the sound signal collected in the waterlogged road section, and analyze the sound signal to identify acoustic anomaly areas.

[0190] The fusion module 34 is used to fuse the acoustic features of the acoustic anomaly area and the pipeline attributes in the underground pipeline information using a manifold learning algorithm to generate a risk feature vector.

[0191] The allocation module 35 is used to discretize the geographical area corresponding to the waterlogged road section into multiple grid cells, and after allocating the risk feature vector to the corresponding grid cells, input it into the cellular automaton model. Based on the preset risk propagation rules, the cellular automaton model is driven to perform iterative calculations among the multiple grid cells to generate disaster prediction data.

[0192] The generation module 36 is used to generate command instructions based on the disaster prediction data and the historical disaster information using a resource scheduling optimization model. The command instructions are used for emergency resource scheduling.

[0193] The IoT-based emergency command system of this application embodiment is used to implement the aforementioned IoT-based emergency command method. Therefore, the specific implementation of the IoT-based emergency command system can be found in the embodiment section of the IoT-based emergency command method above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0194] This application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described Internet of Things-based emergency command methods.

[0195] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described Internet of Things-based emergency command methods.

[0196] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0197] The embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in the above-described emergency command method embodiments.

[0198] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0199] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in one or more embodiments of this specification are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of related data must comply with relevant laws, regulations and standards, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0200] The above provides a detailed description of an emergency command method and system based on the Internet of Things (IoT) provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. An emergency command method based on the Internet of Things, characterized in that, include: Acquire IoT monitoring data, which includes underground pipeline information, water depth information, and historical disaster information; The IoT monitoring data is aligned and fused to generate comprehensive situational data; Based on the comprehensive situational data, waterlogged road sections are identified, and sound signals collected in the waterlogged road sections are acquired and analyzed to identify areas of acoustic anomaly. Using a manifold learning algorithm, the acoustic features of the acoustic anomaly area and the pipeline attributes in the underground pipeline information are fused to generate a risk feature vector; The geographical area corresponding to the waterlogged road section is discretized into multiple grid cells, and the risk feature vector is assigned to the corresponding grid cells and then input into the cellular automaton model. Based on the preset risk propagation rules, the cellular automaton model is driven to perform iterative calculations among the multiple grid cells to generate disaster prediction data. Based on the disaster prediction data and the historical disaster information, a resource scheduling optimization model is used to generate command instructions, which are used for emergency resource scheduling.

2. The method according to claim 1, characterized in that, The step of generating command instructions based on the disaster prediction data and the historical disaster information using a resource scheduling optimization model includes: By analyzing the disaster prediction data, at least one predicted disaster area and the corresponding disaster type and disaster impact level for each predicted disaster area can be obtained; Based on the disaster type, find the corresponding historical cases from the historical disaster information; Extract emergency resource information and historical response locations from the aforementioned historical cases; The geographical extent of the predicted disaster area, the level of disaster impact, and the emergency resource information are input into the resource scheduling optimization model to generate a resource scheduling scheme. Based on the resource scheduling scheme and the historical response locations, command instructions are generated.

3. The method according to claim 2, characterized in that, The step of inputting the geographical range of the predicted disaster area, the disaster impact level, and the emergency resource information into the resource scheduling optimization model to generate a resource scheduling scheme includes: The geographical area of ​​the predicted disaster zone is divided into multiple demand points using the partitioning module of the resource scheduling optimization model. The determination module of the resource scheduling optimization model determines the total amount, type, and distribution location of available resources based on the emergency resource information. The resource scheduling optimization model's allocation module allocates resource demand priorities to each demand point based on the disaster impact level. The calculation module of the resource scheduling optimization model uses the priority of resource demand as the weight, combines the total amount and type of available resources, takes the transportation time from the distribution location to each demand point as the cost constraint, and takes the shortest total resource delivery time as the objective to perform iterative calculations to determine the resource scheduling path, scheduling type and scheduling quantity from the distribution location of each available resource to each demand point. The resource scheduling optimization model output module generates a resource scheduling scheme based on the resource scheduling path, the scheduling type, and the scheduling quantity.

4. The method according to claim 1, characterized in that, The method utilizes a manifold learning algorithm to fuse the acoustic features of the acoustic anomaly area with the pipeline attributes in the underground pipeline information to generate a risk feature vector, including: The acoustic features and pipeline properties are used as the original data set; The original data is dimensionality reduced using a manifold learning algorithm to obtain low-dimensional data; Using a pre-defined association rule base, association analysis is performed on the low-dimensional data to obtain effective association data; The effective correlation data is combined with the water depth information to form an initial feature vector; The initial feature vector is processed using a feature enhancement network to generate a risk feature vector.

5. The method according to claim 1, characterized in that, The process involves assigning the risk feature vector to corresponding grid cells and inputting it into a cellular automaton model. Based on a preset risk propagation rule, the cellular automaton model is driven to perform iterative calculations among the multiple grid cells to generate disaster prediction data, including: The risk feature vector is assigned to one or more corresponding grid cells as the initial risk value of the grid cells; Geological coefficients, hydraulic coefficients, and electrical coefficients are obtained from the aforementioned risk propagation rules; In the cellular automata model, for each grid cell, the risk value of the next iteration is calculated based on its own initial risk value, the initial risk values ​​of adjacent grid cells, and the geological, hydraulic, and electrical coefficients. The cellular automaton model updates the state of all grid cells based on the calculated risk value, completing one iteration; The update process is repeated multiple times until the risk value of all grid cells reaches a stable state or the preset number of iterations is reached. Disaster prediction data is generated based on the risk value distribution of all grid cells after multiple iterations.

6. The method according to claim 1, characterized in that, The process of determining waterlogged road sections based on the comprehensive situational data, acquiring sound signals collected in the waterlogged road sections, and analyzing the sound signals to identify areas of acoustic anomalies includes: On the geographic layer, road sections in the comprehensive situation data where the water depth value is greater than a preset depth threshold are marked as waterlogged road sections. Multiple acoustic sensors are deployed at preset intervals along the flooded road section. The multiple acoustic sensors form a sensor array, and the sensor array is used to acquire sound signals generated at different locations below the flooded road section in a synchronous acquisition manner. The sound signal is subjected to spectral transformation to obtain spectral data, and the energy value of the spectral data within a preset abnormal frequency band is calculated. Key spectral components with energy values ​​greater than a preset energy threshold are identified. Calculate the relative time difference between the arrival of the sound signal at different acoustic sensors in the sensor array; By combining the key spectral components with the relative time difference, the acoustic anomaly region with soil structure changes beneath the waterlogged road section is determined using the cross-location method.

7. The method according to claim 1, characterized in that, After generating command instructions, the following is also included: Obtain real-time traffic and resource location information; The command instructions, the real-time traffic information, and the resource location information are input into the path dynamic planning model. The path dynamic planning model generates a dynamic scheduling path based on a traffic congestion prediction algorithm. The dynamic scheduling path indicates the optimal path from the resource location to the target delivery location in the command instructions. The dynamic scheduling path and the command instructions are sent to the corresponding mobile emergency terminal, and the confirmation information and status feedback information from the mobile emergency terminal are received and recorded through the task confirmation and feedback module deployed on the mobile emergency terminal.

8. An emergency command system based on the Internet of Things, characterized in that, include: The acquisition module is used to acquire IoT monitoring data, which includes underground pipeline information, water depth information, and historical disaster information. The alignment module is used to align and fuse the IoT monitoring data to generate comprehensive situational data; The determination module is used to determine waterlogged road sections based on the comprehensive situation data, acquire sound signals collected in the waterlogged road sections, and analyze the sound signals to identify areas of acoustic anomalies. The fusion module is used to fuse the acoustic features of the acoustic anomaly area and the pipeline attributes in the underground pipeline information using a manifold learning algorithm to generate a risk feature vector. The allocation module is used to discretize the geographical area corresponding to the waterlogged road section into multiple grid cells, and after allocating the risk feature vector to the corresponding grid cells, input it into the cellular automaton model. Based on the preset risk propagation rules, the module drives the cellular automaton model to perform iterative calculations among the multiple grid cells to generate disaster prediction data. The generation module is used to generate command instructions based on the disaster prediction data and the historical disaster information using a resource scheduling optimization model. The command instructions are used for emergency resource scheduling.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the Internet of Things-based emergency command method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of the Internet of Things-based emergency command method as described in any one of claims 1 to 7.

Images