Emergency rescue wireless communication network topology prediction method based on deep learning

Abstract

The embodiment of the invention provides an emergency rescue wireless communication network topology prediction method based on deep learning. According to the method, an emergency rescue wireless communication network system is constructed, initial system parameters are generated, and the emergency rescue wireless communication network system comprises a plurality of nodes; acquiring node positions, environment information and inter-node signal strength in real time, and updating node states in the emergency rescue wireless communication network system according to the node positions, the environment information and the inter-node signal strength; the node state of the current time step is input into the trained prediction model to be analyzed and processed, a prediction result is output, the prediction model is composed of a graph convolutional network and a generative adversarial network, and the prediction result is used for indicating the topology connection condition of the emergency rescue wireless communication network in the next time step; and determining a corresponding emergency rescue strategy in the target rescue area according to the prediction result. According to the method, the reliability of a communication network is improved, and efficient and orderly development of rescue actions is guaranteed.

Description

Technical Field

[0001] The present application relates to the field of wireless communication technology, and in particular to a method for predicting the topology of an emergency rescue wireless communication network based on deep learning. Background Art

[0002] When responding to major natural disasters like earthquakes and landslides, the efficiency and reliability of communication networks are crucial for rescue operations. Existing communication networks rely on fixed base stations and pre-set communication nodes, making them difficult to adapt to the dynamic terrain changes and personnel movements caused by disasters. Network failures often occur due to equipment damage or node loss. Furthermore, existing communication network models often assume static topology and lack the ability to model node mobility and environmental interference in real time. This makes it impossible to warn of sudden link disruptions, which in turn hinders rescue operations.

[0003] Therefore, a new emergency rescue wireless communication network model is urgently needed to provide key technical support to ensure the efficient and orderly implementation of the emergency search and rescue process. Summary of the Invention

[0004] This application provides an emergency rescue wireless communication network topology prediction method based on deep learning to solve the above problems.

[0005] In a first aspect, the present application provides a method for predicting emergency rescue wireless communication network topology based on deep learning, the method comprising:

[0006] Constructing an emergency rescue wireless communication network system and generating initial system parameters, wherein the emergency rescue wireless communication network system includes a plurality of nodes, and the initial system parameters include: an initial position of the node, a unique identifier of the node, and an initial moving speed of the node;

[0007] Acquire node locations, environmental information, and inter-node signal strengths in real time, and update node states in the emergency rescue wireless communication network system based on the node locations, environmental information, and inter-node signal strengths. The environmental information includes a topographic map and obstacle map of the target rescue area, and the node states include the motion state and connection relationship of the nodes.

[0008] The node state at the current time step is input into a trained prediction model for analysis and processing, and a prediction result is output. The prediction model is composed of a graph convolutional network and a generative adversarial network. The prediction result is used to indicate the topological connection status of the emergency rescue wireless communication network at the next time step;

[0009] Determine the corresponding emergency rescue strategy within the target rescue area based on the prediction results.

[0010] Optionally, updating the node status in the emergency rescue wireless communication network system according to the node location, environmental information, and inter-node signal strength includes:

[0011] A node mobility model is constructed based on the node's current location and the terrain map and obstacle map of the target rescue area, and the motion state of the node in the emergency rescue wireless communication network system is determined based on the node mobility model;

[0012] The communication status between nodes is determined according to the signal strength between nodes, and the connection relationship between nodes in the emergency rescue wireless communication network system is determined according to the communication status between nodes.

[0013] Optionally, constructing a node movement model based on the node's current location and a terrain map and obstacle map of a target rescue area includes:

[0014] Determine the target location of the node within the target rescue area according to the preset rescue mission;

[0015] Determine the node's moving direction based on the node's current position and target position;

[0016] Determine the terrain and obstacle factors affecting the node's movement speed based on the terrain map and obstacle map of the target rescue area;

[0017] Determine the node movement speed based on terrain influence factors, obstacle influence factors and the node's initial movement speed;

[0018] A node movement model is constructed based on the node's movement direction and node movement speed.

[0019] Optionally, the step of inputting the node state at the current time step into a trained prediction model for analysis and processing, and outputting a prediction result, includes:

[0020] Constructing an adjacency matrix based on the connection relationship of nodes in the emergency rescue wireless communication network system at the current time step;

[0021] Inputting the adjacency matrix into a graph convolutional network, the graph convolutional network processes the input adjacency matrix through multi-layer convolution operations and outputs node features;

[0022] The output node features are processed by generating adversarial networks to obtain the connection probability between nodes;

[0023] The topological connection status of the emergency rescue wireless communication network in the next time step is determined according to the connection probability between nodes.

[0024] Optionally, before inputting the node state at the current time step into the trained prediction model for analysis and processing, the method further includes:

[0025] Constructing a generator and a discriminator of a generative adversarial network, and optimizing the parameters of the generator through a backpropagation algorithm, wherein the generator is used to generate connection relationships between nodes, and the discriminator is used to determine the difference between the connection relationships between nodes generated by the generator and the actual connection relationships between nodes;

[0026] The generator and the discriminator are subjected to adversarial training until the generator and the discriminator reach a dynamic equilibrium.

[0027] Optionally, before inputting the node state at the current time step into the trained prediction model for analysis and processing, the method further includes:

[0028] Construct a hyperparameter set for the prediction model and determine the value range of each parameter in the hyperparameter set;

[0029] Use uniform distribution to randomly determine parameter configurations from the hyperparameter set, and use cross-validation to determine the accuracy of the prediction model under each set of parameter configurations;

[0030] The set of parameter configurations with the highest accuracy is taken as the optimal parameter configuration of the prediction model.

[0031] Optionally, determining the communication status between nodes according to the signal strength between the nodes includes:

[0032] Determine whether the signal strength between nodes is greater than a preset signal strength threshold;

[0033] If so, it is determined that a communication link exists between the nodes;

[0034] If not, it is determined that there is no communication link between the nodes

[0035] In a second aspect, the present application provides a device for predicting emergency rescue wireless communication network topology based on deep learning, comprising:

[0036] A construction module is used to construct an emergency rescue wireless communication network system and generate initial system parameters, wherein the emergency rescue wireless communication network system includes multiple nodes, and the initial system parameters include: an initial position of the node, a unique identifier of the node, and an initial moving speed of the node;

[0037] A first processing module is configured to obtain node positions, environmental information, and inter-node signal strengths in real time, and update node states in the emergency rescue wireless communication network system based on the node positions, environmental information, and inter-node signal strengths, wherein the environmental information includes a topographic map and an obstacle map of the target rescue area, and the node states include a motion state and a connection relationship of the nodes;

[0038] A second processing module is configured to input the node state at the current time step into a trained prediction model for analysis and processing, and output a prediction result, wherein the prediction model is composed of a graph convolutional network and a generative adversarial network. The prediction result is used to indicate the topological connection status of the emergency rescue wireless communication network at the next time step;

[0039] The determination module is used to determine the corresponding emergency rescue strategy in the target rescue area according to the prediction results.

[0040] In a third aspect, the present application provides a deep learning-based emergency rescue wireless communication network topology prediction device, comprising:

[0041] Memory;

[0042] processor;

[0043] wherein the memory stores computer-executable instructions;

[0044] The processor executes the computer-executable instructions stored in the memory to implement the emergency rescue wireless communication network topology prediction method based on deep learning as described in the above-mentioned first aspect and various possible implementation methods of the first aspect.

[0045] In a fourth aspect, the present application provides a computer storage medium having a computer program stored thereon, which is executed by a processor to implement the deep learning-based emergency rescue wireless communication network topology prediction method as described in the first aspect and various possible implementation methods of the first aspect.

[0046] The present application provides a method for predicting the topology of an emergency rescue wireless communication network based on deep learning. The method constructs an emergency rescue wireless communication network system and generates initial system parameters, wherein the emergency rescue wireless communication network system includes multiple nodes; obtains node positions, environmental information, and inter-node signal strengths in real time, and updates the node status in the emergency rescue wireless communication network system according to the node positions, environmental information, and inter-node signal strengths; inputs the node status of the current time step into the trained prediction model for analysis and processing, and outputs the prediction result, wherein the prediction model is composed of a graph convolutional network and a generative adversarial network, and the prediction result is used to indicate the topological connection status of the emergency rescue wireless communication network in the next time step; and determines the corresponding emergency rescue strategy in the target rescue area based on the prediction result. This method improves the reliability of the communication network and ensures the efficient and orderly implementation of the rescue operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0048] Figure 1A flowchart of a method for predicting emergency rescue wireless communication network topology based on deep learning provided in an embodiment of the present application;

[0049] Figure 2 A schematic diagram of the structure of a deep learning-based emergency rescue wireless communication network topology prediction device provided in an embodiment of the present application;

[0050] Figure 3 A schematic diagram of the structure of a deep learning-based emergency rescue wireless communication network topology prediction device provided in an embodiment of the present application.

[0051] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0052] To make the objectives, technical solutions, and advantages of this application more clear, the technical solutions in this application will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0053] The terms "first," "second," "third," "fourth," and so forth (if any) in the description and claims of the present invention and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or sequential sequence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of the present invention described herein can be implemented in sequences other than those illustrated or described herein.

[0054] In the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described in this application as "exemplary" or "for example" should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0055] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0056] Figure 1The present invention provides a flowchart of a method for predicting the topology of an emergency rescue wireless communication network based on deep learning. Figure 1 As shown, the method for predicting the topology of an emergency rescue wireless communication network based on deep learning provided by this embodiment includes:

[0057] S1: Build an emergency rescue wireless communication network system and generate initial system parameters.

[0058] The emergency rescue wireless communication network system consists of multiple nodes. A node can represent a rescuer or a mobile device (such as a handheld terminal or vehicle-mounted device). Each node has a certain level of communication capability and can send, receive, and process information within the network. Initial system parameters include the node's initial position, unique node identifier, and initial node movement speed.

[0059] It is understandable that each rescuer node has a unique identification ID i , and has the current position P i =(x i ,y i ) and movement speed V i The node's moving path is determined by the location history H i (t) Records.

[0060] Specifically, each node is assigned a unique identification ID i The initial position coordinates of the nodes are generated by simulating the human behavior patterns based on the investigation of relevant data in the rescue area. The present invention introduces an improved Reynolds group behavior model to simulate the actual formation of rescue personnel to avoid excessive gathering or dispersion. The coordinate range of the rescue area is set to [X min ,X max ]×[Y min ,Y max ], the initial position P of node i i =(x i ,y i ) is given by the following formula:

[0061] x i =X min +(X max -X min )×rand()

[0062] y i =Y min +(Y max -Y min )×rand()

[0063] Among them, X min Represents the minimum value of the rescue area on the X axis, Xmax Represents the maximum value of the rescue area X axis, Y min Represents the minimum value of the Y axis of the rescue area, Y max Represents the maximum value of the Y axis of the rescue area, and rand() means returning a random number in the range of [0,1].

[0064] The node's moving speed v i The initialization data depends on the possible speed range of the rescuer and is expressed by the following formula:

[0065] v i =V min +(V max -V min )×rand()

[0066] Among them, V min is the minimum moving speed of the rescuer, V max Represents the maximum movement speed of the rescuer.

[0067] S2: Obtain node location, environmental information, and inter-node signal strength in real time, and update the node status in the emergency rescue wireless communication network system based on the node location, environmental information, and inter-node signal strength.

[0068] The environmental information includes the terrain map and obstacle map of the target rescue area, and the node status includes the motion status and connection relationship of the node.

[0069] Understandably, positioning technology and communication protocols are used to continuously obtain the specific geographic coordinates of each node in the target rescue area within the emergency rescue wireless communication network system, thereby understanding the real-time spatial distribution of nodes. Environmental information primarily includes a topographic map and an obstacle map of the target rescue area. The topographic map shows geographic features such as altitude, slope, and landforms; the obstacle map indicates the distribution of objects within the area, such as buildings, trees, and mountains, that may obstruct wireless signal transmission. Environmental information can be obtained through pre-stored geographic databases, on-site mapping equipment, or drone aerial photography. Signal strength detection modules installed on nodes measure the strength of wireless signals between nodes in real time. Signal strength reflects the quality and stability of the communication link between nodes. Node motion states describe the movement of nodes in space, including parameters such as speed, direction, and acceleration. Node connectivity refers to the communication connections between nodes in the network. This connectivity can be represented using a network topology diagram, where nodes represent communication devices and edges represent communication links between nodes.

[0070] Specifically, updating the node status in the emergency rescue wireless communication network system according to the node location, environmental information, and inter-node signal strength includes the following steps:

[0071] S21: Constructing a node mobility model according to the current location of the node and the terrain map and obstacle map of the target rescue area, and determining the motion state of the node in the emergency rescue wireless communication network system according to the node mobility model.

[0072] The node movement model is constructed based on the node's current location and the terrain map and obstacle map of the target rescue area, including the following steps:

[0073] S211: Determine the target position of the node within the target rescue area according to the preset rescue mission.

[0074] The target location refers to the specific location where the node needs to arrive or be deployed within the target rescue area in order to complete the preset rescue mission.

[0075] S212: Determine the moving direction of the node according to the current position and the target position of the node.

[0076] It can be understood that the current position coordinates and target coordinates of the node are obtained through the positioning system, and the direction vector of the node pointing from the current position to the target position is calculated according to the coordinates.

[0077] In an optional embodiment, the moving direction of the node satisfies the following formula:

[0078]

[0079] Among them, θ i is the moving direction of the node, α is the target attraction coefficient, β is the obstacle repulsion coefficient, d i is the distance from the node to the i-th obstacle, is the direction vector of the target position, is the direction vector of the current position, is the direction vector of the i-th obstacle.

[0080] S213: Determine terrain influencing factors and obstacle influencing factors of the node movement speed according to the terrain map and obstacle map of the target rescue area.

[0081] Among them, terrain influencing factors include terrain slope, and obstacle influencing factors include obstacle density.

[0082] It is understandable that the movement speed of the node will be affected by the terrain and obstacles. According to the actual situation, the movement speed will be dynamically adjusted to adapt to different terrain and obstacle distribution.

[0083] S214: Determine the node movement speed according to the terrain influence factor, the obstacle influence factor and the node initial movement speed.

[0084] Specifically, the slope value is compared with the preset slope value, and the terrain slope coefficient is determined according to the comparison result; the number of obstacles per unit area is compared with the preset number of obstacles, and the obstacle density coefficient is determined according to the comparison result.

[0085] Multiply the terrain slope influence coefficient and the obstacle density coefficient to obtain the total influence coefficient, and multiply the node's initial moving speed by the total influence coefficient to obtain the node's moving speed under the current terrain and obstacle conditions.

[0086] S215: Constructing a node movement model according to the movement direction and movement speed of the node.

[0087] Among them, the node mobility model is used to describe the node's movement path under the influence of terrain and obstacles.

[0088] Specifically, at each time step, the position of the node is updated according to its speed and direction, and the new position of the node is calculated as follows:

[0089] Δp=v i ×[cos(θ i ),sin(θ i )] T

[0090]

[0091] in, and are the positions of the nodes at time t and t+1 respectively, and T represents transposition.

[0092] S22: Determine the communication status between nodes according to the signal strength between the nodes, and determine the connection relationship between the nodes in the emergency rescue wireless communication network system according to the communication status between the nodes.

[0093] As you can understand, signal strength is generally inversely proportional to distance: closer distances result in higher signal strength, while farther distances result in lower signal strength. Signal strength between nodes is measured in real time using signal measurement equipment or the signal measurement capabilities of existing communication equipment. The system processes and analyzes the received signal strength data to understand the communication quality between nodes.

[0094] In practical applications, the received signal strength indicator (RSSI) can be used to quantify the signal strength. The RSSI represents the ratio of the received signal power to the reference power.

[0095] Specifically, it is determined whether the signal strength between nodes is greater than a preset signal strength threshold. If the signal strength between nodes is greater than the preset signal strength threshold, it is considered that there is a communication link between the nodes, that is, the nodes can communicate normally. If the signal strength between nodes is less than the preset signal strength threshold, it is considered that there is no communication link between the nodes, that is, the nodes cannot communicate.

[0096] Furthermore, if the nodes can communicate normally with each other, there will be a corresponding node connection in the emergency rescue wireless communication network system; if the nodes cannot communicate with each other, there will be no corresponding node connection in the emergency rescue wireless communication network system.

[0097] S3: Input the node state of the current time step into the trained prediction model for analysis and processing, and output the prediction result.

[0098] Among them, the prediction model is composed of a graph convolutional network and a generative adversarial network, and the prediction results are used to indicate the topological connection status of the emergency rescue wireless communication network in the next time step.

[0099] In an optional embodiment, before inputting the node state at the current time step into the trained prediction model for analysis and processing, the method further includes:

[0100] Construct the generator and discriminator of the generative adversarial network, and optimize the parameters of the generator through the back-propagation algorithm; perform adversarial training on the generator and discriminator until the generator and discriminator reach a dynamic balance.

[0101] The generator is used to generate connection relationships between nodes, and the discriminator is used to determine the difference between the connection relationships between nodes generated by the generator and the real connection relationships between nodes.

[0102] As you can understand, the generator is a neural network model whose input is a random noise vector and whose output is a sample with the same dimensions as the real data. The structure of the generator can be designed based on the specific task, for example, it can be a fully connected neural network, a convolutional neural network, or a recurrent neural network. The discriminator is also a neural network model whose input is a real sample or a sample generated by the generator, and its output is a probability value indicating the probability that the input sample is a real sample.

[0103] During training, the generator's parameters are first fixed, and only the discriminator is trained. Real samples are labeled as positive, and samples generated by the generator are labeled as negative. Backpropagation and gradient descent are used to optimize the discriminator's parameters, enabling it to distinguish between real and generated samples as accurately as possible. The discriminator's loss function can adopt the binary cross-entropy loss function. The discriminator's parameters are then fixed, and the generator is trained. The generator's goal is to ensure that the discriminator misclassifies its generated samples as real, that is, to maximize the discriminator's output probability for the generated samples. Backpropagation and gradient descent are also used to optimize the generator's parameters, making the samples it generates more realistic.

[0104] When the accuracy of the discriminator reaches a preset accuracy value (for example, 50%), it is considered that the generator and the discriminator have reached a dynamic balance, and a trained generative adversarial network is obtained.

[0105] In an optional embodiment, before inputting the node state at the current time step into the trained prediction model for analysis and processing, the method further includes:

[0106] Construct a hyperparameter set for the prediction model and determine the value range of each parameter in the hyperparameter set; use uniform distribution to randomly determine the parameter configuration from the hyperparameter set, and determine the accuracy of the prediction model under each set of parameter configurations through cross-validation; and use the set of parameter configurations with the highest accuracy as the optimal parameter configuration for the prediction model.

[0107] As you can understand, a random search strategy is used to optimize the model's hyperparameters to explore the optimal model configuration. First, the model's hyperparameter set is defined, and the value range of each parameter is set. The number of searches or the size of the search space is specified to ensure sufficient search while avoiding over-computation. Furthermore, a uniform distribution is used to randomly select parameter configurations from the defined parameter space. The randomly selected parameters satisfy the following formula:

[0108] θ selected =uniform(θ min ,θ max )

[0109] Among them, θ min and θ max They represent the minimum and maximum values ​​of the parameters respectively, and uniform represents uniform distribution.

[0110] Cross-validation is used to evaluate the performance of the model under each set of parameter configurations to obtain the average accuracy. The accuracy calculation formula is as follows:

[0111]

[0112] Among them, K is the number of cross-validation folds, which is 10, Correct Predictions k Indicates the number of samples correctly predicted in the k-fold cross validation, TotalPredictions k Represents the total number of samples in the k-fold cross validation.

[0113] Based on the cross-validation results, record the best-performing parameter configuration. Compare the accuracy of different configurations and select the best-performing parameter set. After all random searches are complete, select the parameter configuration with the highest average accuracy for final model training and deployment.

[0114] Specifically, the node state at the current time step is input into the trained prediction model for analysis and processing, and the prediction result is output, including the following steps:

[0115] S31: Constructing an adjacency matrix according to the connection relationship of the nodes in the emergency rescue wireless communication network system at the current time step.

[0116] Among them, each element in the adjacency matrix represents the connection relationship between the corresponding nodes.

[0117] S32: Input the adjacency matrix into the graph convolutional network. The graph convolutional network processes the input adjacency matrix through multi-layer convolution operations and outputs node features.

[0118] It can be understood that in the graph convolutional network, through multi-layer convolution operations, the neighborhood information of the node is gradually aggregated to generate a node feature representation with higher semantics. The output node feature includes not only the node's own attributes, but also the influence of surrounding nodes.

[0119] S33: Process the output node features through a generative adversarial network to obtain the connection probability between nodes.

[0120] It can be understood that the node features are input into the trained generative adversarial network, and the node features are analyzed and processed by the generator and discriminator of the generative adversarial network to obtain the node connection probability.

[0121] S34: Determine the topological connection status of the emergency rescue wireless communication network at the next time step according to the connection probability between the nodes.

[0122] It can be understood that the connection relationship between nodes in the emergency rescue wireless communication network at the next time step is determined according to the node connection probability, and then the topological connection status of the emergency rescue wireless communication network is determined according to the connection relationship between the nodes.

[0123] S4: Determine a corresponding emergency rescue strategy within the target rescue area based on the prediction results.

[0124] Understanding the changing trends in the topology of emergency rescue wireless communication networks allows for targeted response measures. For example, by calculating the eigenvalues ​​and eigenvectors of the predicted node connection matrix to assess connection stability and reliability, the predictions can be used to adjust resource allocation, divide rescue areas, and schedule personnel actions to address potential risks and challenges. Furthermore, the predicted connection probability information can be used to determine the optimal resource allocation plan, improving resource utilization efficiency.

[0125] The embodiment of the present application provides a method for predicting the topology of an emergency rescue wireless communication network based on deep learning. The method constructs an emergency rescue wireless communication network system and generates initial system parameters. The emergency rescue wireless communication network system includes multiple nodes; obtains node positions, environmental information, and inter-node signal strengths in real time, and updates the node states in the emergency rescue wireless communication network system according to the node positions, environmental information, and inter-node signal strengths; inputs the node states of the current time step into a trained prediction model for analysis and processing, and outputs a prediction result. The prediction model is composed of a graph convolutional network and a generative adversarial network. The prediction result is used to indicate the topological connection status of the emergency rescue wireless communication network at the next time step; and determines the corresponding emergency rescue strategy within the target rescue area based on the prediction result. This method improves the reliability of the communication network and ensures the efficient and orderly implementation of rescue operations.

[0126] Figure 2 This is a schematic diagram of the structure of the emergency rescue wireless communication network topology prediction device based on deep learning provided in the embodiment of the present application. Figure 2 As shown, the deep learning-based emergency rescue wireless communication network topology prediction device 200 provided in this embodiment includes:

[0127] A construction module 201 is used to construct an emergency rescue wireless communication network system and generate initial system parameters, wherein the emergency rescue wireless communication network system includes multiple nodes, and the initial system parameters include: an initial position of a node, a unique identifier of a node, and an initial moving speed of a node;

[0128] The first processing module 202 is configured to obtain node positions, environmental information, and inter-node signal strengths in real time, and update node states in the emergency rescue wireless communication network system based on the node positions, environmental information, and inter-node signal strengths, wherein the environmental information includes a topographic map and an obstacle map of the target rescue area, and the node states include a motion state and a connection relationship of the nodes;

[0129] The second processing module 203 is used to input the node state of the current time step into the trained prediction model for analysis and processing, and output a prediction result, wherein the prediction model is composed of a graph convolutional network and a generative adversarial network, and the prediction result is used to indicate the topological connection status of the emergency rescue wireless communication network at the next time step;

[0130] The determination module 204 is configured to determine a corresponding emergency rescue strategy within the target rescue area according to the prediction result.

[0131] The deep learning-based emergency rescue wireless communication network topology prediction device provided in this embodiment can execute the deep learning-based emergency rescue wireless communication network topology prediction method provided in the above method embodiment. Its implementation principle and technical effects are similar, and this embodiment will not be repeated here.

[0132] Figure 3 This is a schematic diagram of the structure of the emergency rescue wireless communication network topology prediction device based on deep learning provided in the embodiment of the present application. Figure 3 As shown, an embodiment of the present application provides an emergency rescue wireless communication network topology prediction device based on deep learning. The emergency rescue wireless communication network topology prediction device 300 based on deep learning includes: a receiver 301, a transmitter 302, a processor 303 and a memory 304.

[0133] Receiver 301, for receiving instructions and data;

[0134] Transmitter 302, used to send instructions and data;

[0135] Memory 304, for storing computer-executable instructions;

[0136] Processor 303 is configured to execute computer-executable instructions stored in memory 304 to implement the steps of the method for predicting the topology of an emergency rescue wireless communication network based on deep learning in the above-described embodiment. For details, please refer to the relevant description in the embodiment of the method for predicting the topology of an emergency rescue wireless communication network based on deep learning.

[0137] Optionally, the memory 304 may be independent or integrated with the processor 303 .

[0138] When the memory 304 is independently provided, the electronic device further includes a bus for connecting the memory 304 and the processor 303 .

[0139] An embodiment of the present application also provides a computer storage medium, which stores computer execution instructions. When a processor executes the computer execution instructions, it implements the deep learning-based emergency rescue wireless communication network topology prediction method performed by the above-mentioned deep learning-based emergency rescue wireless communication network topology prediction device.

[0140] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As is well known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable, and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media generally embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0141] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0142] It should be understood that the present application is not limited to the exact structure described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for predicting emergency rescue wireless communication network topology based on deep learning, characterized in that: The method comprises: Constructing an emergency rescue wireless communication network system and generating initial system parameters, wherein the emergency rescue wireless communication network system includes a plurality of nodes, and the initial system parameters include: an initial position of the node, a unique identifier of the node, and an initial moving speed of the node; Acquire node locations, environmental information, and inter-node signal strengths in real time, and update node states in the emergency rescue wireless communication network system based on the node locations, environmental information, and inter-node signal strengths. The environmental information includes a topographic map and obstacle map of the target rescue area, and the node states include the motion state and connection relationship of the nodes. The node state at the current time step is input into a trained prediction model for analysis and processing, and a prediction result is output. The prediction model is composed of a graph convolutional network and a generative adversarial network. The prediction result is used to indicate the topological connection status of the emergency rescue wireless communication network at the next time step; Determine the corresponding emergency rescue strategy within the target rescue area based on the prediction results.

2. The method according to claim 1, characterized in that The updating of the node status in the emergency rescue wireless communication network system according to the node location, environmental information and inter-node signal strength includes: A node mobility model is constructed based on the node's current location and the terrain map and obstacle map of the target rescue area, and the motion state of the node in the emergency rescue wireless communication network system is determined based on the node mobility model; The communication status between nodes is determined according to the signal strength between nodes, and the connection relationship between nodes in the emergency rescue wireless communication network system is determined according to the communication status between nodes.

3. The method according to claim 2, characterized in that The node movement model is constructed according to the current location of the node and the terrain map and obstacle map of the target rescue area, including: Determine the target location of the node within the target rescue area according to the preset rescue mission; Determine the node's moving direction based on the node's current position and target position; Determine the terrain and obstacle factors affecting the node's movement speed based on the terrain map and obstacle map of the target rescue area; Determine the node movement speed based on terrain influence factors, obstacle influence factors and the node's initial movement speed; A node movement model is constructed based on the node's movement direction and node movement speed.

4. The method according to claim 1, wherein The node state at the current time step is input into the trained prediction model for analysis and processing, and the prediction result is output, including: Constructing an adjacency matrix based on the connection relationship of nodes in the emergency rescue wireless communication network system at the current time step; Inputting the adjacency matrix into a graph convolutional network, the graph convolutional network processes the input adjacency matrix through multi-layer convolution operations and outputs node features; The output node features are processed by generating adversarial networks to obtain the connection probability between nodes; The topological connection status of the emergency rescue wireless communication network in the next time step is determined according to the connection probability between nodes.

5. The method according to claim 1, wherein Before inputting the node state at the current time step into the trained prediction model for analysis and processing, the method further includes: Constructing a generator and a discriminator of a generative adversarial network, and optimizing the parameters of the generator through a backpropagation algorithm, wherein the generator is used to generate connection relationships between nodes, and the discriminator is used to determine the difference between the connection relationships between nodes generated by the generator and the actual connection relationships between nodes; The generator and the discriminator are subjected to adversarial training until the generator and the discriminator reach a dynamic equilibrium.

6. The method according to claim 1, characterized in that Before inputting the node state at the current time step into the trained prediction model for analysis and processing, the method further includes: Construct a hyperparameter set for the prediction model and determine the value range of each parameter in the hyperparameter set; Use uniform distribution to randomly determine parameter configurations from the hyperparameter set, and use cross-validation to determine the accuracy of the prediction model under each set of parameter configurations; The set of parameter configurations with the highest accuracy is taken as the optimal parameter configuration of the prediction model.

7. The method according to claim 2, characterized in that Determining the communication status between nodes according to the signal strength between the nodes includes: Determine whether the signal strength between nodes is greater than a preset signal strength threshold; If so, it is determined that a communication link exists between the nodes; If not, it is determined that no communication link exists between the nodes.

8. A device for predicting emergency rescue wireless communication network topology based on deep learning, characterized in that: The device comprises: A construction module is used to construct an emergency rescue wireless communication network system and generate initial system parameters, wherein the emergency rescue wireless communication network system includes multiple nodes, and the initial system parameters include: an initial position of the node, a unique identifier of the node, and an initial moving speed of the node; A first processing module is configured to obtain node positions, environmental information, and inter-node signal strengths in real time, and update node states in the emergency rescue wireless communication network system based on the node positions, environmental information, and inter-node signal strengths, wherein the environmental information includes a topographic map and an obstacle map of the target rescue area, and the node states include a motion state and a connection relationship of the nodes; A second processing module is configured to input the node state at the current time step into a trained prediction model for analysis and processing, and output a prediction result, wherein the prediction model is composed of a graph convolutional network and a generative adversarial network. The prediction result is used to indicate the topological connection status of the emergency rescue wireless communication network at the next time step; The determination module is used to determine the corresponding emergency rescue strategy in the target rescue area according to the prediction results.

9. A deep learning-based emergency rescue wireless communication network topology prediction device, characterized in that: The device comprises: Memory; processor; wherein the memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the emergency rescue wireless communication network topology prediction method based on deep learning as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that The computer storage medium stores computer execution instructions, which, when executed by a processor, are used to implement the deep learning-based emergency rescue wireless communication network topology prediction method as described in any one of claims 1 to 7.

Images