Star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery method, device and equipment and storage medium

Abstract

The application discloses a kind of star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery method, device, equipment and storage medium, the method includes: based on the global routing information of target area to construct star-ground fusion network model, when receiving task request, based on the task priority of task request and star-ground fusion network model generates target routing strategy, based on target routing strategy, each node of communication link is called to carry out communication response to task request, and the communication response process of communication link is monitored, when there is fault in communication link, fault self-recovery analysis is carried out through target fault self-recovery analysis model, generates fault self-recovery link, and each node in fault self-recovery link is called to carry out communication response to task request, to realize the communication fault self-recovery in communication response process, to effectively improve the link stability and fault tolerance of unmanned area ad hoc network, realize the multi-objective balance in communication response process, improve the fault self-recovery speed when communication response failure.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of communication, in particular to a kind of star-ground fusion hybrid ad hoc network routing protocol and fault self-healing method, device, equipment and storage medium. BACKGROUND

[0002] With the continuous development of various communication network technologies, global communication conditions are increasingly perfect, but there are still problems of insufficient communication area coverage in large areas of uninhabited areas. The communication blind area of uninhabited area needs reliable coverage to ensure the diversity of communication scenarios in uninhabited areas. Uninhabited area communication scenarios include border uninhabited area safety control, communication blind area online monitoring, field exploration and resource exploitation, post-disaster emergency response and rescue, environmental monitoring and protection, intelligent agriculture and precision agriculture monitoring, and unmanned aerial vehicle inspection and monitoring. In the above uninhabited area communication scenarios, communication can be realized through star-ground fusion hybrid ad hoc network, but reliable communication guarantee is often lacking in actual communication environment.

[0003] Star-ground fusion hybrid ad hoc network faces problems such as short endurance time, high delay, poor link stability, and easy data loss in actual application. Especially in the dynamic and complex environment of uninhabited areas and the changing network conditions, the communication system is difficult to ensure the stable return of data, and lacks sufficient flexibility, stability and fault tolerance. The optimization target of current research on communication optimization and fault processing of star-ground fusion hybrid ad hoc network is single, and there is a lack of effective mechanism for flexible balance between energy consumption and delay. Conventional routing protocols or algorithms usually only judge whether the link is available according to the current measurement value, and the link stability is less considered in future changes. In actual environment, it is easy to appear chain break or continuous packet loss in the middle of the network, which reduces the communication stability. Traditional ad hoc network has problems of high energy consumption, long delay, unstable link and insufficient fault self-healing ability, which leads to short network endurance time and unstable communication quality in uninhabited area information monitoring, and network paralysis caused by node failure. SUMMARY

[0004] The main purpose of the present application is to provide a kind of star-ground fusion hybrid ad hoc network routing protocol and fault self-healing method, device, equipment and storage medium, to solve the problems of high energy consumption, long delay, unstable link and insufficient fault self-healing ability of traditional ad hoc network in prior art, which leads to short network endurance time and unstable communication quality in uninhabited area information monitoring, and network paralysis caused by node failure.

[0005] To achieve the above purpose, the present application provides a kind of star-ground fusion hybrid ad hoc network routing protocol and fault self-healing method, the method comprises the following steps:

[0006] construct a star-ground fusion network model based on global routing information of a target region, the global routing information comprising network topology information of the target region, the star-ground fusion network model being composed of a plurality of nodes in the target region and a communication topology between the nodes, the nodes comprising ad hoc network nodes and gateway nodes, the gateway nodes being connected with a plurality of ad hoc network nodes, and the ad hoc network nodes being connected with at least one ad hoc network node;

[0007] generate a routing strategy according to the star-ground fusion network model, the routing strategy comprising a reactive routing strategy and a proactive routing strategy;

[0008] generate a target routing strategy based on a task priority of the task request and the routing strategy upon receiving the task request, the target routing strategy comprising a communication link for determining a communication response;

[0009] communicate a response to the task request by nodes of the communication link based on the target routing strategy, and monitor a communication response process of the communication link;

[0010] input the network topology information, node features of nodes in the communication link, and link features of the communication link into a target fault self-healing analysis model for fault self-healing analysis upon detecting a fault in the communication link, generate a fault self-healing link, and communicate a response to the task request by nodes in the fault self-healing link to realize communication fault self-healing in a communication process.

[0011] Optionally, the generating a target routing strategy based on a task priority of the task request and the routing strategy upon receiving the task request comprises:

[0012] obtain a task state and a task priority of the task request upon receiving the task request;

[0013] if the task state of the task request is a routine state or the task priority of the task request is lower than a priority threshold, use the reactive routing strategy as the target routing strategy;

[0014] if the task state of the task request is an emergency state or the task priority of the task request is not lower than the priority threshold, use the proactive routing strategy as the target routing strategy, and monitor whether the task request is restored from the emergency state to the routine state;

[0015] if the task request is restored from the emergency state to the routine state, use the reactive routing strategy as the target routing strategy;

[0016] The reactive routing strategy includes generating a plurality of candidate paths based on the star-ground fusion network model, and taking a path with the lowest data transmission energy consumption in the candidate paths as a communication link for communication response.

[0017] The active routing strategy includes generating a plurality of candidate paths based on the star-ground fusion network model, and taking a path with the optimal comprehensive evaluation result of link stability in the candidate paths as a communication link for communication response.

[0018] Optionally, the generating the routing strategy according to the star-ground fusion network model comprises:

[0019] generating a plurality of candidate paths based on the star-ground fusion network model, and determining energy consumption receiving degrees of nodes in the candidate paths:

[0020]

[0021] wherein, the energy consumption receiving degree of a node , the residual energy of a node , the initial energy of a node ;

[0022] determining transmission energy consumption of a sending node data of each node in the candidate paths:

[0023]

[0024] wherein, the transmission energy consumption of the sending node data between a node and a node , an energy consumption coefficient, a data amount to be transmitted, a time coefficient, the time coefficient representing an activity degree of the node in a current time window;

[0025] determining a path selection weight of each candidate path based on the energy consumption receiving degree and the transmission energy consumption:

[0026]

[0027] wherein, the path selection weight;

[0028] determining a path with the lowest data transmission energy consumption in the candidate paths based on the path selection weight, taking the path with the lowest data transmission energy consumption as a communication link for communication response, and generating a reactive routing strategy based on the communication link.

[0029] Optionally, the generating the routing strategy according to the star-ground fusion network model comprises:

[0030] generating a plurality of candidate paths based on the star-ground fusion network model, and evaluating current link quality of each candidate path to obtain a current link quality evaluation result:

[0031]

[0032] wherein, represents the current link quality evaluation result, represents a historical data weight, represents a historical link quality evaluation result, represents a current link quality comprehensive score, which is obtained by weighting based on a data packet reception ratio, a smoothed data packet reception rate, a stability factor and a signal-to-noise ratio of the candidate path;

[0033] evaluating transmission anti-interference capability of each node in the candidate path to obtain an anti-interference safety degree:

[0034]

[0035] wherein, represents the anti-interference safety degree of transmission data between each node, represents a maximum anti-interference degree that can be tolerated between nodes, represents an interference degree between each node;

[0036] determining a prediction factor based on a link quality change speed of each candidate link:

[0037]

[0038]

[0039] wherein, represents the prediction factor, represents the link quality change speed, represents a fastest speed of historical link quality change, represents a time link quality evaluation result, represents a time link quality evaluation result, represents a time interval for calculating the change speed;

[0040] comprehensively evaluating link stability of each candidate path according to the current link quality evaluation result, the anti-interference safety degree and the prediction factor of each candidate path to obtain a link stability comprehensive evaluation result:

[0041]

[0042] wherein, represents a comprehensive evaluation result of link stability, , and represents a weight factor;

[0043] based on the comprehensive evaluation result of link stability, a path with the optimal comprehensive evaluation result of link stability in the candidate paths is taken as a communication link for communication response, and an active routing strategy is generated based on the communication link.

[0044] Optionally, before the network topology information, node features of each node in the communication link, and link features of the communication link are input into the target fault self-healing analysis model for fault self-healing analysis, the method further comprises:

[0045] configuring reinforcement learning parameters, the reinforcement learning parameters comprising state information, actions, and rewards, the state information comprising current neighbor node information of a node, a comprehensive evaluation result of link stability, a historical fault flag, and residual energy, the actions comprising generating a fault self-healing link based on the state information, and the rewards comprising giving positive feedback or negative feedback based on a communication response result of the fault self-healing path;

[0046] constructing a value function based on the reinforcement learning parameters, the value function comprising:

[0047]

[0048] wherein, represents a learning rate, represents a discount factor, represents an immediate reward of taking an action in a state , represents an optimal value of taking an action in a state , represents a current value of taking an action in a state , represents an updated value of taking an action in a state ;

[0049] constructing an original fault self-healing analysis model based on the value function;

[0050] constructing a loss function based on a time-difference strategy, and training the original fault self-healing analysis model based on the loss function to obtain a target fault self-healing analysis model, the loss function comprising:

[0051]

[0052] wherein, represents a loss function, the loss function is updated by back propagation and gradient descent iteration current network parameters , represents the current strategy network in the original fault self-healing analysis model training process, represents the current network parameters of the current strategy network, represents the target strategy network, represents the target network parameters of the target strategy network.

[0053] Optionally, the value function is constructed based on the reinforcement learning parameters, comprising:

[0054] Based on the network topology information, a graph structure is generated, the graph structure comprising a plurality of nodes and edges connecting between nodes;

[0055] According to the reinforcement learning parameters, an initial feature is configured to the graph structure, the initial feature comprising an initial node feature and an initial edge feature, the initial node feature comprising a residual energy and a historical fault flag of the node, and the edge feature comprising a link stability comprehensive evaluation result and a link time delay information;

[0056] Based on the initial feature and the graph structure, an initial graph neural network is generated;

[0057] Based on a message passing mechanism, a feature update is performed on each node and each edge in the initial graph neural network, to obtain a target graph neural network:

[0058]

[0059]

[0060] wherein, represents a node the node feature of the first layer, represents a node and a node the edge feature of the first layer, represents a set of neighbor nodes of the node , represents a node weight matrix of the first layer, represents an activation function, represents a node the node feature of the first layer, represents a neighbor node the first node features of a layer, representing a node with a node in a first edge features of a layer, representing a first edge weight matrix of a layer;

[0061] The reinforcement learning parameter is embedded into a vector by the target graph neural network to obtain a reinforcement learning embedding vector, and a value function is constructed based on the reinforcement learning embedding vector.

[0062] Optionally, the monitoring of the communication response process of the communication link comprises:

[0063] Monitoring the response state of the neighbor node after the neighbor node sends a communication request from each node in the communication link in the communication response process;

[0064] Monitoring the comprehensive evaluation result of the link stability of the communication link in the communication response process;

[0065] Monitoring the energy alarm information of each node in the communication link in the communication response process, the energy alarm information including the remaining energy of the node being lower than a preset energy threshold.

[0066] In addition, to achieve the above-mentioned purpose, the application further provides a star-ground hybrid ad hoc network routing protocol and fault self-healing device, which comprises:

[0067] A star-ground hybrid model construction module is configured to construct a star-ground hybrid network model based on global routing information of a target area, wherein the global routing information comprises network topology structure information of the target area, and the star-ground hybrid network model is composed of multiple nodes in the target area and communication topology structures between the nodes, wherein the nodes comprise ad hoc network nodes and gateway nodes, the gateway nodes are connected with multiple ad hoc network nodes, and the ad hoc network nodes are connected with at least one ad hoc network node.

[0068] A routing strategy generation module is configured to generate a routing strategy according to the star-ground hybrid network model, wherein the routing strategy comprises a reactive routing strategy and a proactive routing strategy.

[0069] A routing response module is configured to generate a target routing strategy based on a task priority of a task request and the routing strategy when the task request is received, wherein the target routing strategy comprises a communication link for determining a communication response.

[0070] A communication response module is configured to perform a communication response on the task request based on the target routing strategy by calling each node of the communication link, and monitor a communication response process of the communication link.

[0071] The fault self-healing module is configured to input the network topology information, the node features of each node in the communication link, and the link features of the communication link into a target fault self-healing analysis model to perform fault self-healing analysis when detecting that the communication link has a fault, generate a fault self-healing link, and call each node in the fault self-healing link to perform a communication response to the task request to realize communication fault self-healing in a communication process.

[0072] In addition, to achieve the above object, the present application further provides a satellite-ground hybrid ad hoc network routing protocol and fault self-healing device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the satellite-ground hybrid ad hoc network routing protocol and fault self-healing method as described above.

[0073] In addition, to achieve the above object, the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the satellite-ground hybrid ad hoc network routing protocol and fault self-healing method as described above.

[0074] In addition, to achieve the above object, the present application further provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the steps of the satellite-ground hybrid ad hoc network routing protocol and fault self-healing method as described above.

[0075] The application constructs a star-ground fusion network model based on global routing information of a target area, the global routing information includes network topology structure information of the target area, the star-ground fusion network model is composed of multiple nodes in the target area and a communication topology structure between the nodes, the nodes include ad hoc network nodes and gateway nodes, the gateway nodes are connected with multiple ad hoc network nodes, and the ad hoc network nodes are connected with at least one ad hoc network node; a routing strategy is generated according to the star-ground fusion network model, the routing strategy includes a reactive routing strategy and a proactive routing strategy; when a task request is received, a target routing strategy is generated based on a task priority of the task request and the routing strategy, the target routing strategy includes a communication link for determining a communication response; the nodes of the communication link are called to perform a communication response to the task request based on the target routing strategy, and a communication response process of the communication link is monitored; when a fault of the communication link is detected, the network topology structure information, node characteristics of the nodes in the communication link and link characteristics of the communication link are input to a target fault self-healing analysis model for fault self-healing analysis, a fault self-healing link is generated, and the nodes in the fault self-healing link are called to perform a communication response to the task request, so as to realize communication fault self-healing in the communication response process; since the star-ground fusion network model of the target area is constructed, the reactive routing strategy and the proactive routing strategy are generated based on the star-ground fusion network model, so that a suitable routing strategy is selected for different task scenarios, multi-target optimization in the communication response process is realized, the end-to-end delay of a burst task is significantly reduced while the overall energy consumption of the network is relatively low, the adaptability of the network to multiple types of services is improved, the communication response process is monitored, the target fault self-healing analysis model is used for fault self-healing analysis when the communication link has a fault, and the nodes in the fault self-healing link are called to perform a communication response to the task request, so as to realize rapid detection and self-healing decision of a communication fault, the standby path can be selected and seamlessly switched in time when a link or node has a fault, and the fault self-healing speed and processing efficiency of ad hoc network communication in an unmanned area are improved. BRIEF DESCRIPTION OF DRAWINGS

[0076] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, for those skilled in the art, other drawings can also be obtained without creative labor.

[0077] Figure 1 It is a structure schematic diagram of a star-ground fusion hybrid ad hoc network routing protocol and fault self-healing equipment related to a hardware running environment of an embodiment of the present application.

[0078] Figure 2Flow chart of the first embodiment of the star-ground fusion mixed ad hoc network routing protocol and fault self-recovery method of the present application;

[0079] Figure 3 Topological structure diagram of the star-ground fusion network model in the first embodiment of the star-ground fusion mixed ad hoc network routing protocol and fault self-recovery method of the present application;

[0080] Figure 4 Flow chart of the link stability comprehensive evaluation in the first embodiment of the star-ground fusion mixed ad hoc network routing protocol and fault self-recovery method of the present application;

[0081] Figure 5 Flow chart of the second embodiment of the star-ground fusion mixed ad hoc network routing protocol and fault self-recovery method of the present application;

[0082] Figure 6 Structure block diagram of the first embodiment of the star-ground fusion mixed ad hoc network routing protocol and fault self-recovery device of the present application.

[0083] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0084] It should be understood that the specific embodiments described herein are merely intended to explain the present application and not to limit the present application.

[0085] Reference Figure 1 , Figure 1 Structure diagram of the star-ground fusion mixed ad hoc network routing protocol and fault self-recovery device of the hardware running environment involved in the embodiment scheme of the present application.

[0086] As Figure 1As shown in the figure, the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery device can include a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to realize the connection communication between the components. The user interface 1003 can include a display screen, an input unit such as a keyboard, and can also include a standard wired interface and a wireless interface. The network interface 1004 can optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (WI-FI) interface). The memory 1005 can be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk memory. The memory 1005 can also be a storage device independent of the aforementioned processor 1001.

[0087] Those skilled in the art can understand that Figure 1 The structure shown in the figure does not constitute a limitation on the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery device, and can include more or fewer components than the figure, or combine certain components, or different component arrangements.

[0088] As Figure 1 As shown in the figure, the memory 1005 as a computer readable storage medium can include an operating system, a network communication module, a user interface module, and a star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery program.

[0089] In Figure 1 In the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery device shown in the figure, the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 in the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery device can be arranged in the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery device, and the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery device calls the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery program stored in the memory 1005 through the processor 1001, and executes the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery method provided by the embodiment of the application.

[0090] The embodiment of the application provides a star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery method, which refers to Figure 2 , Figure 2A flowchart of a first embodiment of a star-ground fusion hybrid ad hoc network routing protocol and fault self-healing method.

[0091] In this embodiment, the star-ground fusion hybrid ad hoc network routing protocol and fault self-healing method comprises the following steps:

[0092] Step S10: Construct a star-ground fusion network model based on global routing information of a target area.

[0093] It should be noted that the present embodiment proposes a star-ground fusion hybrid ad hoc network high-reliability routing protocol and fault self-healing method to solve the problems of high energy consumption, long time delay, unstable link, and insufficient fault self-healing capability of traditional ad hoc networks in the unmanned area communication scenario, which leads to short network endurance time, unstable communication quality, network paralysis caused by node failure, and other problems in unmanned area information monitoring. The scheme flexibly combines reactive routing and proactive routing to provide a low-energy transmission method for regular tasks, and quickly switches to a low-latency path when a burst or high-priority task occurs. At the same time, it cooperates with multi-dimensional link stability evaluation and prediction to identify potential unstable links in real time and avoid them in advance, significantly improving the reliability and efficiency of communication.

[0094] Among them, the unmanned area communication scenario applied in the present embodiment can be border unmanned area safety management and control, online monitoring of communication blind area, field exploration and resource exploitation, post-disaster emergency response and rescue, environmental monitoring and protection, intelligent agriculture and precision agriculture monitoring, and unmanned aerial vehicle inspection and monitoring, etc.

[0095] It should be understood that the execution subject of the present embodiment can be a computing service device with data processing, network communication, and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or a terminal electronic device capable of realizing the above functions. Hereinafter, a star-ground fusion hybrid ad hoc network routing protocol and fault self-healing terminal device (hereinafter referred to as a terminal device) will be taken as an example to illustrate the present embodiment and the following embodiments.

[0096] It should be noted that the global routing information includes network topology structure information of the target area, the star-ground fusion network model is composed of multiple nodes in the target area and the communication topology structure between the nodes, the nodes include ad hoc network nodes and gateway nodes, the gateway nodes are connected with multiple ad hoc network nodes, and the ad hoc network nodes are connected with at least one ad hoc network node.

[0097] It should be noted that the target area can be an unmanned area, such as a desert area, a virgin forest area, and the like. The self-organizing network node can be a sensor data collection node, and the main function thereof is to collect data of sensors in the surrounding area. The gateway node is a central node in which a wide area network terminal is deployed, and the main function thereof is to collect data of the self-organizing network nodes in a larger area, and to transmit the data back to the monitoring center through a public network or a low-orbit satellite Internet.

[0098] It can be understood that, since the unmanned area needs wide-range communication coverage, the embodiment constructs a star-ground fusion network model based on global routing information of the target area, with reference to Figure 3 , Figure 3 FIG. 1 is a schematic diagram of a topology structure of a star-ground fusion network model in an embodiment. The star-ground fusion network model can include three layers: the first layer network is a self-organizing network node that collects sensor data in a surrounding local area; the second layer is a self-organizing network node that performs multi-hop transmission on the collected sensor data, and transmits the sensor data to a gateway node in which a wide area network terminal is installed, and a self-organizing network node is placed at every unit distance interval, and the maximum transmission distance of the self-organizing network node is two unit distances; and the third layer network is a gateway node that transmits the collected data back to a monitoring center after preliminary preprocessing.

[0099] In some embodiments, a terminal device can adopt a star-ground fusion three-layer architecture (sensor layer, self-organizing network layer, and gateway center layer) to construct a star-ground fusion network model, which can reduce the use of satellite links in local multi-hop transmission, and can rely on satellites to achieve global coverage at critical moments. The star-ground fusion network model can realize coverage of a communication blind area in an unmanned area, and can rely on satellite connection when a ground network is blocked or nodes are excessively sparse, to further guarantee seamless data transmission.

[0100] Step S20: generating a routing strategy according to the star-ground fusion network model.

[0101] It should be noted that the routing strategy includes a reactive routing strategy and a proactive routing strategy.

[0102] The reactive routing strategy can be applied to daily tasks or routine tasks, and the self-organizing network power consumption is the lowest under the premise of meeting the data transmission requirements.

[0103] The proactive routing strategy is suitable for emergency tasks, tasks with high priority, or urgent tasks, and needs to reduce transmission delay and ensure the stability of the link.

[0104] In some embodiments, the connection of the sustainable power supply device at the gateway node in the network does not consider energy consumption, and a gateway device can be placed at both ends of each ad hoc network link for data backhaul, so that each ad hoc network node not at the two ends of the chain network has at least 4 links to choose from. Therefore, the terminal device can generate a reactive routing strategy based on the transmission energy consumption of node data and the energy consumption acceptance, and select the path with the lowest energy consumption for communication response when dealing with daily tasks.

[0105] In some embodiments, the terminal device generates an active routing strategy according to the link stability score, estimated delay, and residual energy of the node when facing an emergency task or a high-priority task, thereby ensuring the stability of the link during the communication response process and reducing the transmission delay.

[0106] Further, in order to effectively reduce the communication energy consumption in the communication response process of the reactive routing strategy, the above step S20 can include:

[0107] Step S201: generating a plurality of candidate paths based on the star-ground fusion network model, and determining the energy consumption acceptance of each node in the candidate paths;

[0108] Step S202: determining the transmission energy consumption of the node data transmitted by each node in the candidate paths;

[0109] Step S203: determining the path selection weight of each candidate path based on the energy consumption acceptance and the transmission energy consumption;

[0110] Step S204: determining the path with the lowest data transmission energy consumption in the candidate paths based on the path selection weight, taking the path with the lowest data transmission energy consumption as the communication link for communication response, and generating a reactive routing strategy based on the communication link.

[0111] In some embodiments, the terminal device first performs network initialization, and the node local node ID, energy, position, and other information are backhauled to the gateway in a hop-by-hop manner.

[0112] The specific approach is: the gateway sends an access data packet to the neighbor nodes on both sides, the data packet contains the node ID, energy status, timestamp, and other information, the node receives the data and sends an acknowledgement, and stores the information in the routing table. The next-hop neighbor node will further forward the data packet following the same process until all nodes have obtained the information. If there is a large difference between the node local information data and the data in the data packet, the data is stored in the routing table, and if there is no large difference, the data is discarded directly. Through this process, all nodes will maintain the neighbor node information within their one-hop maximum communication range, and establish a reverse path to the gateway.

[0113] Secondly, energy evaluation is carried out, and the node selects, according to the energy status of each node in the routing table, to avoid low-energy nodes as much as possible and select high-energy nodes for forwarding.

[0114] Node i Transmit data to the node j The energy receiving degree of the node is defined as:

[0115]

[0116] Among them, The energy receiving degree of the node is defined as: The energy receiving degree of the node is defined as: The residual energy of the node is defined as: The residual energy of the node is defined as: The residual energy of the node is defined as: The residual energy of the node is defined as:

[0117] The energy consumption of transmitting data of the node is:

[0118]

[0119] Among them, The energy consumption of transmitting data of the node is: The energy consumption of transmitting data of the node is: The energy consumption coefficient is defined as: The energy consumption coefficient is related to hardware characteristics (such as transmission power, antenna efficiency) and environmental factors (such as channel attenuation), and in some embodiments, the terminal device can determine the energy consumption coefficient based on the hardware characteristics and environmental factors of the node; The amount of data to be transmitted is defined as: The time coefficient is defined as: i The time coefficient is defined as:

[0120] The path selection weight is defined as:

[0121]

[0122] Among them, The path selection weight is defined as:

[0123] Finally, the energy state of the node will be updated, and the reaction routing protocol stipulates that if the energy of the node is lower than the threshold value (for example, 5% of the power), the node will send an alarm information to the neighbor node, and the neighbor node will update the information of the routing table in time after receiving the alarm information.

[0124] Further, in order to ensure communication stability and transmission timeliness in the communication response process of emergency tasks or high-priority tasks, the step S20 can include:

[0125] Step S211: generating a plurality of candidate paths based on the star-ground fusion network model, and evaluating the current link quality of each candidate path to obtain a current link quality evaluation result;

[0126] Step S212: evaluating the transmission anti-interference ability of each node in the candidate path to obtain an anti-interference safety degree;

[0127] Step S213: determining a prediction factor based on the link quality change speed of each candidate link;

[0128] Step S214: comprehensively evaluating the link stability of each candidate path according to the current link quality evaluation result, the anti-interference safety degree and the prediction factor of the candidate path to obtain a link stability comprehensive evaluation result;

[0129] Step S215: taking the path with the optimal link stability comprehensive evaluation result in the candidate path as the communication link for communication response based on the link stability comprehensive evaluation result, and generating an active routing strategy based on the communication link.

[0130] It should be noted that the terminal device can use the active routing strategy for communication response when facing a burst task, an emergency task or a high-priority task, to preferentially reduce transmission delay and ensure the stability of the link.

[0131] In some embodiments, the terminal device first updates the global routing information. Network initialization is performed in step 1 to realize the perception of network topology. In the active routing, the routing update message is broadcasted or multi-hop propagated in the network at a fixed time interval or when a large network topology change occurs, so that each node can master the change of network topology in real time and correct the global routing table in time.

[0132] Secondly, the priority path is calculated. The node residual energy, link stability score and estimated delay are included in the weighted calculation to construct a cost function. The Dijkstra algorithm is used to calculate the reachable paths of the whole network based on the cost function and mark the path in the routing table. When a burst task is detected, the system automatically finds and enables the corresponding "optimal path" table item without additional waiting for the path discovery process like reactive routing, thereby greatly shortening the end-to-end delay.

[0133] Finally, the time delay sensitivity and energy balance are designed in parallel. When the burst task volume is small, the node energy remaining is considered while calculating the optimal path to avoid high-priority data long-term occupation of some low-energy nodes. When the burst task volume is large, the energy consideration weight is reduced, and the link stability and time delay are mainly considered to ensure fast data return.

[0134] In some embodiments, with reference to Figure 4 , Figure 4 Figure 1 is a flowchart of a link stability comprehensive evaluation process in an embodiment. The terminal device can evaluate the link stability in multi-dimensional time by combining past, present and future information, using the following steps:

[0135] Step 1: Multi-dimensional link quality evaluation

[0136] The link quality estimator based on fuzzy logic constructs a comprehensive score of link stability, which is used to predict the link state and determine whether the link is likely to be disconnected or unstable, providing a basis for decision-making for subsequent route selection.

[0137] First, the link quality is evaluated by using the link quality estimator based on fuzzy logic. It contains four indicators: packet reception ratio (PRR), smoothed packet reception ratio (SPRR), stability factor (SF) and signal-to-noise ratio (SNR). The specific formula is as follows:

[0138] The packet reception ratio (PRR) is calculated according to the following formula:

[0139]

[0140] The smoothed packet reception ratio (SPRR) is calculated according to the following formula:

[0141]

[0142] wherein, is a historical weight factor, The larger the value, the more significant the historical data.

[0143] The stability factor (SF) reflects the degree of link fluctuation, which is calculated as the inverse of the standard deviation of PRR. The signal-to-noise ratio is directly measured by the physical layer and normalized to the range of 0-100.

[0144] The current link quality evaluation formula is:

[0145]

[0146] wherein, represents the current link quality evaluation result, represents the historical data weight, denotes the historical link quality evaluation result, denotes the current link quality comprehensive score, which is obtained by weighting based on the data packet reception ratio, the smoothed data packet reception rate, the stability factor and the signal-to-noise ratio of the candidate path.

[0147] Secondly, the unmanned area environment is complex, the distance between nodes is long, and the interference in the channel changes at any time. In this case, the concept of safety degree is defined to evaluate the transmission capacity of the node at the moment. The evaluation of the transmission capacity can be the evaluation of the anti-interference safety degree of the data transmission between nodes. Referring to the following formula:

[0148]

[0149] wherein, denotes the anti-interference safety degree of the transmission data between nodes, denotes the maximum anti-interference degree that can be tolerated between nodes, denotes the interference degree between nodes. The anti-interference safety degree indicates that the connection between nodes will become difficult, and the increase in safety degree will lead to the disconnection of the link between nodes and require more time.

[0150] Step 2: Link quality change prediction:

[0151] The communication environment between nodes is constantly changing, but the change often forms a continuous process, so the change trend of the next time period can be predicted by calculating the instantaneous change trend of the link quality. The calculation formula of the instantaneous speed of the change of the link quality between nodes and the prediction factor is as follows:

[0152]

[0153]

[0154] wherein, denotes the prediction factor, denotes the change speed of the link quality, denotes the fastest speed of the historical change of the link quality, denotes the time link quality evaluation result, denotes the link quality evaluation result at time denotes the link quality evaluation result at time denotes the time interval for calculating the change speed.

[0155] The system dynamically adjusts the routing according to the stability score to avoid using unstable links and preferentially selects reliable links to ensure efficient transmission of data.

[0156] Step 3: link stability comprehensive calculation, link stability can be regarded as the comprehensive of the past, present and future three state qualities of the link, so as to perform link stability comprehensive evaluation, referring to the following formula:

[0157]

[0158] wherein, represents the link stability comprehensive evaluation result, , and represent weight factors, the final stability depends on the minimum value of the link set, the greater the minimum value, the more stable the link.

[0159] Step S30: when receiving the task request, generating a target routing strategy based on the task priority of the task request and the routing strategy.

[0160] It should be noted that the target routing strategy includes determining a communication link for responding to communication.

[0161] In some embodiments, the target routing strategy can be a hybrid routing strategy combining the proactive routing strategy and the reactive routing strategy. The terminal device can dynamically select a suitable routing strategy according to the type of the task (routine task and emergency task). The reactive routing is used to reduce the communication energy consumption of the routine task, and the proactive routing is used to cope with the emergency task to ensure the minimum delay and link stability. Therefore, after executing the emergency task, the proactive routing strategy can be switched to the reactive routing strategy to reduce the communication energy consumption, thereby realizing multi-objective optimization.

[0162] In some embodiments, the terminal device can timely reduce the occupation of high-priority traffic to low-power nodes through an energy threshold mechanism; when the node is reconnected, a small-scale test and observation period management strategy is adopted to gradually increase the traffic sharing ratio of the node. This can effectively prolong the overall network life and quickly integrate the node into the available topology after the node recovers, thereby fully utilizing network resources.

[0163] Further, in order to realize multi-objective optimization in the communication response process and realize adaptive routing response in a dynamic communication scenario, the above step S30 can include:

[0164] Step S301: when receiving a task request, obtaining the task state and task priority of the task request;

[0165] Step S302: if the task state of the task request is a routine state or the task priority of the task request is lower than a priority threshold, the reactive routing strategy is used as the target routing strategy;

[0166] Step S303: If the task state of the task request is an emergency state or the task priority of the task request is not lower than a priority threshold, the proactive routing strategy is taken as a target routing strategy, and whether the task request is recovered from the emergency state to a routine state is monitored;

[0167] Step S304: When it is detected that the task request is recovered from the emergency state to the routine state, the reactive routing strategy is taken as the target routing strategy.

[0168] It should be noted that the reactive routing strategy includes generating a plurality of candidate paths based on the star-ground fusion network model, and taking a path with the lowest data transmission energy consumption in the candidate paths as a communication link for communication response; and the proactive routing strategy includes generating a plurality of candidate paths based on the star-ground fusion network model, and taking a path with the optimal comprehensive evaluation result of link stability in the candidate paths as a communication link for communication response.

[0169] In a specific implementation, the terminal device can dynamically select a suitable routing strategy according to the type of the task (routine task and burst task). The reactive routing is used to reduce the communication energy consumption of the routine task, and the proactive routing is used to cope with the burst task and ensure the minimum delay and link stability.

[0170] In some embodiments, first, the task is identified and classified. The system classifies the data stream to be sent into a corresponding category based on business priority, data emergency degree and other meta information, and identifies and labels the task type.

[0171] Secondly, the reactive and proactive routing is maintained. For routine tasks, a routing request is initiated only when data needs to be sent, path discovery and establishment are completed, the overhead of periodic topology update of the whole network is reduced, and thus the node energy consumption is reduced. For high-priority or burst-prone tasks, global routing information is broadcasted (or multi-hop transmitted) at fixed time intervals or triggered, so that each node has a more comprehensive network topology, and can quickly look up the optimal or suboptimal path when a burst task occurs, thereby reducing the end-to-end delay.

[0172] It should be noted that the embodiment can adopt a hybrid routing response strategy, that is, an active routing and reactive routing switching mechanism is designed, including: after detecting a high-priority task or a burst request, the system immediately enables the optimal path maintained by the active routing to avoid additional delay caused by re-flooding to find a path in the reactive routing; when the task ends or the network returns to the daily state, the system can return to the reactive routing as the main one to reduce the energy and bandwidth overhead of global update. If the network topology changes little, the link is stable, and the amount of conventional data is large, the reactive routing can become the dominant one; if frequent link failures, rapid energy decay of nodes, or multiple emergency data streams are sensed, the system will temporarily increase the update frequency of the active routing or switch to the active priority to ensure timeliness and reliability. In the hybrid strategy, a threshold is set, when the residual energy of a local node or a key forwarding node is lower than a certain level, the occupation of the local node or the key forwarding node in the high-priority task is automatically reduced, and other nodes with more sufficient energy are preferentially selected for forwarding, thereby prolonging the overall network lifetime.

[0173] Step S40: Based on the target routing strategy, each node of the communication link responds to the task request, and the communication response process of the communication link is monitored.

[0174] In a specific implementation, the terminal device generates one or more communication links based on the target routing strategy, calls each node in the communication link to respond to the task request, and monitors the communication response process.

[0175] For example, the target routing strategy is a hybrid routing strategy, in a first task execution time window, it is determined that the task request is a daily task, a first communication link is generated based on a reactive routing strategy for the purpose of reducing communication energy consumption, and each node of the first communication link responds to the task request; in a second task execution time window, it is determined that the task request is changed to an emergency task, a second communication link is generated based on an active routing strategy for the purpose of link stability and transmission timeliness, and each node of the second communication link responds to the task request; in a third task execution time window, the emergency task ends, and the task is changed to a daily task, a third communication link is generated based on a reactive routing strategy for the purpose of reducing communication energy consumption, and each node of the third communication link responds to the task request.

[0176] Further, in order to accurately detect faults and perform fault self-healing when there is a risk of fault in the communication link, and improve the fault self-healing response capability, the above step S40 can include:

[0177] Step S401: Monitor the response state of the neighbor node after each node in the communication link sends a communication request to the neighbor node in the communication response process;

[0178] Step S402: monitoring the link stability comprehensive evaluation result of the communication link in the communication response process;

[0179] Step S403: monitoring the energy alarm information of each node in the communication link in the communication response process, the energy alarm information including that the residual energy of the node is lower than a preset energy threshold.

[0180] In some embodiments, link fault detection, the node sends a communication request to the communication link, if there is no response for three times in succession, it indicates that the link is faulty, when the link stability comprehensive evaluation result is less than a threshold value or the energy alarm information of the neighbor node is received, the terminal device determines that the link has potential faults. Once the link is detected to be in a fault state, fault self-healing analysis is performed through a target fault self-healing analysis model, a fault self-healing link is generated, and the communication link is quickly switched.

[0181] Step S50: when it is detected that the communication link has faults, inputting the network topology structure information, the node characteristics of each node in the communication link, and the link characteristics of the communication link into a target fault self-healing analysis model for fault self-healing analysis, generating a fault self-healing link, and calling each node in the fault self-healing link to perform communication response to the task request to realize communication fault self-healing in the communication process.

[0182] In some embodiments, the interrupt device can input the network topology, node energy, link state and other information into a graph neural network for high-dimensional embedding, and realize adaptive decision of the fault switching strategy by combining reinforcement learning, so that the optimal alternative path can be quickly switched to when the node fails or the link is faulty. In a complex, large-scale and frequently changing environment network, the routing reconstruction time can be significantly shortened, the network fault tolerance and recovery ability can be enhanced, and self-healing in a short time can be realized.

[0183] The embodiment is based on global routing information of a target area to construct a star-ground fusion network model, the global routing information includes network topology information of the target area, the star-ground fusion network model is composed of multiple nodes in the target area and a communication topology structure between the nodes, the nodes include ad hoc network nodes and gateway nodes, the gateway nodes are connected with multiple ad hoc network nodes, and the ad hoc network nodes are connected with at least one ad hoc network node; a routing strategy is generated according to the star-ground fusion network model, the routing strategy includes a reactive routing strategy and a proactive routing strategy; when a task request is received, a target routing strategy is generated based on a task priority of the task request and the routing strategy, the target routing strategy includes a communication link for determining a communication response; the nodes of the communication link are called to perform a communication response to the task request based on the target routing strategy, and a communication response process of the communication link is monitored; when it is detected that the communication link has a fault, the network topology information, node characteristics of the nodes in the communication link and link characteristics of the communication link are input to a target fault self-healing analysis model for fault self-healing analysis, a fault self-healing link is generated, and the nodes in the fault self-healing link are called to perform a communication response to the task request, so that a communication fault self-healing in the communication response process is realized; since the embodiment constructs a star-ground fusion network model of a target area, generates a reactive routing strategy and a proactive routing strategy based on the star-ground fusion network model, selects a suitable routing strategy for different task scenarios, realizes multi-target optimization in a communication response process, ensures that the overall energy consumption of the network is relatively low, significantly reduces the end-to-end delay of a burst task, improves the adaptability of the network to multiple types of services, monitors the communication response process, performs fault self-healing analysis through a target fault self-healing analysis model when the communication link has a fault, and calls the nodes in the fault self-healing link to perform a communication response to the task request, realizes rapid detection and self-healing decision of a communication fault, can select a backup path in time and perform seamless switching when a link or node has a fault, and improves the fault self-healing speed and processing efficiency of ad hoc network communication in an unmanned area.

[0184] Reference Figure 5 , Figure 5 The flowchart of the second embodiment of the star-ground fusion hybrid ad hoc network routing protocol and fault self-healing method of the application is shown.

[0185] Based on the first embodiment, before the step S50, the embodiment further includes the following steps.

[0186] Step S501: configuring reinforcement learning parameters.

[0187] It should be noted that the reinforcement learning parameters include state information, actions and rewards, the state information includes current neighbor node information of the node, link stability comprehensive evaluation result, historical failure flag and residual energy, the actions include generating a fault self-healing link based on the state information, and the rewards include giving positive feedback or negative feedback based on the communication response result of the fault self-healing path.

[0188] In some embodiments, the state can include current neighbor node information of the node, link stability evaluation situation, historical failure information, residual capacity of the node itself and the like. The action can be a link decision that the node can select, and the backup path is updated in real time in the path library; the reward is positive feedback given after communication success and meeting the requirements after switching the path, and negative feedback is given after serious packet loss, failure or timeout.

[0189] Step S502: constructing a value function based on the reinforcement learning parameters.

[0190] It should be noted that the action mainly realizes short-time path switching, and the action set can include: switching between the next hop links that can be selected at the current node; selecting an alternative route in the backup path library; traffic re-distribution or load adjustment; maintaining the existing path (retaining action in the case of no failure or good performance).

[0191] The reward refers to the reward value determined by detecting the communication quality of the link after performing the path switching or fault repair action. Positive reward value is given after normal communication and meeting the predetermined requirements after switching the path, and the system gives a punitive reward (negative number) after serious packet loss, failure aggravation or timeout. The reward mechanism can be expressed as:

[0192]

[0193] wherein, and are adjustable parameters for balancing the needs of timely recovery and stable operation.

[0194] It should be noted that the embodiment can realize the path selection of fault self-healing through reinforcement learning, and the path selection strategy based on the value function reinforcement learning is adopted for short-time path switching caused by node failure or link problem. That is is the comprehensive return that can be obtained by performing the action (e.g. switching to a candidate link or backup path) in the state The value function includes:

[0195]

[0196] wherein, denotes a learning rate, denotes a discount factor, represents the optimal value of the action under the state of taking action under the state of taking action under the state of taking action under the state of taking action under the state of taking action under the state

[0197] Further, in order to accurately encode the multi-dimensional state characteristics in the star-ground fusion ad hoc network, thereby efficiently constructing the value function and improving the fault self-healing performance, the step S502 can include:

[0198] Step S5021: generating a graph structure based on network topology information;

[0199] Step S5022: configuring an initial feature to the graph structure according to the reinforcement learning parameter;

[0200] Step S5023: generating an initial graph neural network based on the initial feature and the graph structure;

[0201] Step S5024: updating the features of each node and each edge in the initial graph neural network based on a message passing mechanism to obtain a target graph neural network;

[0202] Step S5025: converting the reinforcement learning parameter into an embedding vector through the target graph neural network to obtain a reinforcement learning embedding vector, and constructing a value function based on the reinforcement learning embedding vector.

[0203] It should be noted that the graph structure includes a plurality of nodes and edges connected between the nodes. The initial feature includes an initial node feature and an initial edge feature. The initial node feature includes the residual energy and the historical fault flag of the node, and the edge feature includes the link stability comprehensive evaluation result and the link time delay information.

[0204] It should be noted that, since the state dimension in the ad hoc network can be large (the number of nodes is large, and the link features are rich), the graph neural network (GNN) is used in the embodiment to embed and encode the multi-dimensional state characteristics in the star-ground fusion ad hoc network.

[0205] For the network topology at time , the initial feature of the node can be represented as:

[0206]

[0207] The edge features in the network, such as link quality and time delay, can be denoted as:

[0208]

[0209] Through a series of message passing and updating operations of GNN layers, the high-dimensional vector of the node and the global can be obtained.

[0210] The output of the GNN is taken as the state representation of reinforcement learning, denoted as , and the trainable model is used to approximate , and the formula is:

[0211]

[0212] , wherein including the parameters of the GNN itself and the parameters of the possible additional decision layer (such as a small MLP). In the graph neural network, the node features and edge features are updated through the message passing mechanism:

[0213]

[0214]

[0215] , wherein denotes the node feature of the node in the first layer, denotes the edge feature between the node and the node in the first layer, denotes the edge feature between the node and the node in the first layer, denotes the neighbor node set of the node , denotes the node weight matrix of the first layer, denotes the activation function, denotes the node feature of the node in the first layer, denotes the node feature of the neighbor node in the first layer, denotes the edge feature between the node and the node in the first layer, denotes the edge feature between the node and the node in the first layer, denotes the edge feature between the node and the node in the first layer,

[0216] Step S503: constructing an original fault self-healing analysis model based on the value function.

[0217] It should be noted that the specific process of fault self-healing is as follows: fault detection – if a node energy alarm, link failure, or severe packet loss is detected, the reinforcement learning decision for fault self-healing is triggered; state acquisition – the current network graph structure is encoded to obtain a state vector. Action selection, based on The estimated value is used to select the optimal action (e.g., switching to an alternative path); a reward / penalty feedback system is implemented, providing positive / negative rewards based on the switching result, and then updating the algorithm. The value (i.e.) Value); continuous iteration: through multiple failure scenarios and path switching processes, the value function is continuously iterated. The value enables the system to converge to a better short-term switching strategy.

[0218] Step S504: Construct a loss function based on the time-series difference strategy, and train the original fault self-healing analysis model based on the loss function to obtain the target fault self-healing analysis model.

[0219] It should be noted that whenever an action is performed... And receive instant rewards and the next state Then, according to the temporal difference (TD) update principle, the following loss function is minimized, wherein the loss function includes:

[0220]

[0221] in, This represents the loss function, which iteratively updates the current network parameters through backpropagation and gradient descent. , This represents the current policy network during the training of the original fault self-healing analysis model. This represents the current network parameters of the current policy network. Represents the target policy network. The target network parameters, representing the target policy network, can be iteratively updated through backpropagation and gradient descent. This allows GNNs to gradually learn to select the optimal short-term switching action under different fault scenarios.

[0222] In some embodiments, the reinforcement learning path switching process is roughly as follows: fault information or low-energy alarms are written into the corresponding node / edge features for graph updating; the embedding vector of the current graph structure is extracted. And calculate the values ​​in each set of optional actions. The highest value is used for GNN forward inference; the highest value is selected. The value is executed; corresponding positive and negative rewards are given based on the data transmission results, and then updated. The network parameters are set and iterated continuously.

[0223] In some embodiments, link fault detection, the node sends a communication request to the communication link, such as no response for three times in succession, indicating that the link is faulty, when the link stability comprehensive evaluation result Less than the threshold value or receiving the energy warning information of the neighbor node, the terminal device determines that the link has potential failure. Once the link is detected as a fault state, the corresponding action is considered not selectable in the reinforcement learning decision, and a very low Value is given. At the same time, the optimal next action is selected in the standby link, and the specific formula is:

[0224]

[0225] Flow table and group table update, in the flow table pre-installed in the gateway or local node, the main path is updated to The corresponding forwarding rule, realizing "local one-key" switching, reducing the time consumption of re-routing or flooding.

[0226] The embodiment also includes fault switching feedback and experience playback. If the data is successfully transmitted after switching, the system gives a positive reward to update The value; if it is still packet loss, give negative reward reinforcement learning, guide the subsequent iteration to select a better link. Save the fault and switching process (state-action-reward-new state) to the experience playback pool for the next round of training batch sampling, and improve the continuous adaptation ability of the algorithm to fault scenarios.

[0227] The embodiment also proposes a standby path selection mechanism. The standby path selection is to deal with the link communication failure caused by the interruption of the link between nodes or the damage of the network node.

[0228] For each link, a corresponding standby path is generated. The current link Remove the topology network, and use the shortest path algorithm to calculate the bypass Select the standby path The total cost of the standby communication link is not higher than twice the initial optimal path, that is:

[0229]

[0230] The embodiment also includes anti-loop check. Check whether the backup path forms a loop. If there is a loop, remove the backup path. If there is no loop, keep the backup path and update the backup path library.

[0231] The embodiment also includes standby path table maintenance. Sort by cost and reinforcement learning score, and update the standby path table regularly.

[0232] The embodiment also includes node failure determination. If the node fails to communicate for three times in succession or the link stability The link is judged to be invalid. If the node If all links in the set fail, then the node is considered to be faulty, or the node... A node is considered to have insufficient energy to support any single communication link. Invalid.

[0233] This embodiment also includes: topology reconstruction after node failure. After a node fails, neighboring nodes first perform instantaneous switching according to the paths in the backup path table to ensure seamless communication, while the gateway broadcasts the node's information. Failure message, updating node Values ​​and neural network parameters.

[0234] This embodiment also includes a node self-healing mechanism, which uses an exponential smoothing coefficient when a node recovers from a fault and comes back online. To synthesize the most recent observations:

[0235]

[0236] When link stability is continuous If all windows are above the lower bound threshold for link stability, then the reinforcement learning function of the link should be improved. The value is used to revert to the available topological set.

[0237] At the node Preliminary assessment indicates recovery, but it is not yet fully confirmed whether it meets high-load reliability requirements. This invention can be verified using small-scale test traffic.

[0238]

[0239] in, For the target total amount of data, Indicates only allocation ×100% of the traffic goes to the node If the packet loss rate, latency, and other link stability metrics of the test traffic consistently meet expectations, the traffic share will be gradually increased. Based on the test verification results, the system can gradually increase the traffic share and adjust it within the Agent's... Update nodes in the value table The value of an action is determined until it is treated exactly the same as a normal node.

[0240] When node Once the network is confirmed to have recovered during the testing phase, it will switch back to the node, either by rerouting the original path or, in some scenarios, by sharing traffic with the node.

[0241] The gateway broadcasts a notification to re-add the mapping rules, ensuring that subsequent data can be successfully distributed to the nodes. To prevent nodes If a secondary failure occurs for unknown reasons, this embodiment can also set an observation period. The link is monitored based on an observation period. During this observation period the gateway continuously monitors the link metrics; if everything is fine it finally confirms the "full switchover", if it degrades again it re-reduces its value or cancels the switchover.

[0242] The embodiment configures a reinforcement learning parameter, constructs a value function based on the reinforcement learning parameter, constructs an original fault self-healing analysis model based on the value function, constructs a loss function according to a time difference strategy, and trains the original fault self-healing analysis model based on the loss function to obtain a target fault self-healing analysis model, so that fault self-healing in a communication fault scenario is realized through reinforcement learning, dynamic switching of a communication link is realized, routing reconstruction time is significantly shortened, network fault tolerance and recovery capability are enhanced, and self-healing within a short time is realized.

[0243] In addition, the embodiment of the present application also provides a computer readable storage medium, and a satellite-ground fusion hybrid ad hoc network routing protocol and fault self-healing program is stored on the computer readable storage medium. The satellite-ground fusion hybrid ad hoc network routing protocol and fault self-healing program is executed by a processor to realize the steps of the satellite-ground fusion hybrid ad hoc network routing protocol and fault self-healing method described above.

[0244] The computer readable storage medium provided in the present application may, for example, be a U disk, but is not limited to an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, system, or device, or any combination of the above. More specific examples of the computer readable storage medium can include, but are not limited to, an electrical connection with one or more conductive wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present embodiment, the computer readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer readable storage medium can be transmitted by any suitable medium, including but not limited to an electrical wire, an optical cable, an RF (Radio Frequency: radio frequency), and the like, or any suitable combination of the above.

[0245] ​The computer readable storage medium can be included in the star-ground hybrid ad hoc network routing protocol and fault self-recovery device, or can exist independently without being assembled into the star-ground hybrid ad hoc network routing protocol and fault self-recovery device.

[0246] In addition, the embodiment of the present application further provides a computer program product including a star-ground hybrid ad hoc network routing protocol and fault self-recovery program, which realizes the steps of the star-ground hybrid ad hoc network routing protocol and fault self-recovery method when executed by a processor.

[0247] The computer program product embodiment of the present application is basically the same as the above-mentioned star-ground hybrid ad hoc network routing protocol and fault self-recovery method, and will not be repeated here.

[0248] Reference Figure 6 , Figure 6 The structure block diagram of a first embodiment of the star-ground hybrid ad hoc network routing protocol and fault self-recovery device of the present application is shown in the figure.

[0249] As Figure 6 shown, the star-ground hybrid ad hoc network routing protocol and fault self-recovery device provided by the embodiment of the present application includes:

[0250] A star-ground hybrid model construction module 10 is configured to construct a star-ground hybrid network model based on global routing information of a target area, wherein the global routing information includes network topology structure information of the target area, and the star-ground hybrid network model is composed of a plurality of nodes in the target area and a communication topology structure between the nodes, the nodes include ad hoc network nodes and gateway nodes, the gateway nodes are connected with a plurality of ad hoc network nodes, and the ad hoc network nodes are connected with at least one ad hoc network node.

[0251] A routing strategy generation module 20 is configured to generate a routing strategy according to the star-ground hybrid network model, wherein the routing strategy includes a reactive routing strategy and a proactive routing strategy.

[0252] A routing response module 30 is configured to generate a target routing strategy based on a task priority of a task request and the routing strategy when receiving the task request, wherein the target routing strategy includes a communication link for determining a communication response.

[0253] A communication response module 40 is configured to perform a communication response on the task request by nodes of the communication link based on the target routing strategy, and monitor a communication response process of the communication link.

[0254] The fault self-healing module 50 is configured to input the network topology information, the node features of each node in the communication link, and the link features of the communication link into a target fault self-healing analysis model for fault self-healing analysis when detecting that the communication link has a fault, generate a fault self-healing link, and call each node in the fault self-healing link to perform a communication response to the task request, so as to realize communication fault self-healing in the communication response process.

[0255] The embodiment constructs a star-ground fusion network model based on global routing information of a target region, the global routing information including network topology information of the target region, the star-ground fusion network model being composed of a plurality of nodes in the target region and a communication topology between the nodes, the nodes including ad hoc network nodes and gateway nodes, the gateway nodes being connected with a plurality of ad hoc network nodes, and the ad hoc network nodes being connected with at least one ad hoc network node; generates a routing strategy based on the star-ground fusion network model, the routing strategy including a reactive routing strategy and a proactive routing strategy; generates a target routing strategy based on a task priority of a task request and the routing strategy when receiving the task request, the target routing strategy including a communication link for determining a communication response; calls each node of the communication link to perform a communication response to the task request based on the target routing strategy, and monitors a communication response process of the communication link; inputs the network topology information, the node features of each node in the communication link, and the link features of the communication link into a target fault self-healing analysis model for fault self-healing analysis when detecting that the communication link has a fault, generates a fault self-healing link, and calls each node in the fault self-healing link to perform a communication response to the task request, so as to realize communication fault self-healing in the communication response process; since the embodiment constructs a star-ground fusion network model of a target region, generates a reactive routing strategy and a proactive routing strategy based on the star-ground fusion network model, and thus selects a suitable routing strategy for different task scenarios, realizes multi-target optimization in a communication response process, significantly reduces end-to-end delay of a burst task while ensuring that the overall energy consumption of the network is relatively low, improves the adaptability of the network to multiple types of services, detects and analyzes a fault through a target fault self-healing analysis model when a communication link has a fault, and calls each node in a fault self-healing link to perform a communication response to a task request, so as to realize rapid detection and self-healing decision of a communication fault, select a backup path in time and perform seamless switching when a link or node has a fault, and improve the fault self-healing speed and processing efficiency of ad hoc network communication in an unmanned area.

[0256] The star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery device provided by the application can solve the technical problems of the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery. Compared with the prior art, the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery device provided by the application has the same beneficial effects as the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery method provided by the above-mentioned embodiments, and other technical features in the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery device are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here.

[0257] It should be understood that the above is only an example, and does not constitute any limitation on the technical solutions of the application. In specific applications, those skilled in the art can set it up according to the needs, and the application does not limit it.

[0258] It should be noted that the above-described workflow is only illustrative and does not limit the scope of protection of the application. In actual application, those skilled in the art can select part or all of them to achieve the purpose of the embodiment scheme according to actual needs, which is not limited here.

[0259] In addition, technical details not described in detail in this embodiment can be referred to the star-ground fusion hybrid ad hoc network routing protocol and fault self-recovery method provided by any embodiment of the application, which will not be repeated here.

[0260] It should be noted that in this paper, the term "includes", "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or includes elements inherent to such process, method, article or system. Without more limitations, the element defined by the sentence "includes a" does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0261] The above-mentioned application embodiment serial number is only for description, not representing the advantages and disadvantages of the embodiments.

[0262] Through the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned example method can be realized by means of software and a necessary general hardware platform, and of course, it can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, or an optical disk) and includes a number of instructions for causing an end device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method described in each embodiment of the present application.

[0263] The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent flow transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied to other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A satellite-terrestrial hybrid mixed ad hoc network routing protocol and fault self-healing method, characterized in that, The method comprises: constructing a star-ground fusion network model based on global routing information of a target area, the global routing information comprising network topology information of the target area, the star-ground fusion network model being composed of a plurality of nodes in the target area and a communication topology between the nodes, the nodes comprising ad hoc network nodes and gateway nodes, the gateway nodes being connected with a plurality of ad hoc network nodes, and the ad hoc network nodes being connected with at least one ad hoc network node; generating a routing strategy according to the star-ground fusion network model, the routing strategy comprising a reactive routing strategy and a proactive routing strategy; generating a target routing strategy based on a task priority of a task request and the routing strategy upon receiving the task request, the target routing strategy comprising a communication link for responding to the task request; communicating the task request by nodes of the communication link based on the target routing strategy and monitoring a communication response process of the communication link; detecting a fault in the communication link, inputting the network topology information, node characteristics of the nodes in the communication link and link characteristics of the communication link into a target fault self-healing analysis model for fault self-healing analysis, generating a fault self-healing link, and communicating the task request by nodes in the fault self-healing link to realize communication fault self-healing in the communication process. 2.The satellite-terrestrial hybrid mixed ad hoc network routing protocol and fault self-healing method of claim 1, wherein, The generating of the target routing strategy based on the task priority of the task request and the routing strategy upon receiving the task request comprises: obtaining a task state and a task priority of the task request upon receiving the task request; if the task state of the task request is a routine state or the task priority of the task request is lower than a priority threshold, taking the reactive routing strategy as the target routing strategy; if the task state of the task request is an emergency state or the task priority of the task request is not lower than the priority threshold, taking the proactive routing strategy as the target routing strategy and monitoring whether the task request is restored from the emergency state to the routine state; if the task request is restored from the emergency state to the routine state, taking the reactive routing strategy as the target routing strategy; wherein the reactive routing strategy comprises generating a plurality of candidate paths based on the star-ground fusion network model, and taking a path with the lowest data transmission energy consumption in the candidate paths as the communication link for responding to the task request; the proactive routing strategy comprises generating a plurality of candidate paths based on the star-ground fusion network model, and taking a path with the optimal link stability comprehensive evaluation result in the candidate paths as the communication link for responding to the task request. 3.The satellite-terrestrial hybrid mesh network routing protocol and fault self-healing method of claim 2, wherein, The generating of the routing strategy according to the star-ground fusion network model comprises: generating a plurality of candidate paths based on the star-ground fusion network model and determining an energy consumption receiving degree of each node in the candidate paths; wherein, represents the energy consumption reception of a node , represents the residual energy of a node , represents the initial energy of a node ; determining a transmission energy consumption of transmission node data of each node in the candidate paths; wherein, denotes a node transmits node data to a node , the transmission energy consumption of the node data transmitted between the nodes, denotes an energy consumption coefficient, denotes the amount of data to be transmitted, denotes a time coefficient, the time coefficient representing the activity of the node within the current time window; determining a path selection weight of each candidate path based on the energy consumption receiving degree and the transmission energy consumption; wherein, denotes a path selection weight; determine a path with the lowest data transmission energy consumption in the candidate paths based on the path selection weight, and take the path with the lowest data transmission energy consumption as a communication link for communication response, and generate a reactive routing strategy based on the communication link. 4.The satellite-terrestrial hybrid mesh network routing protocol and fault self-healing method of claim 3, wherein, The generating a routing strategy according to the star-ground fusion network model comprises: generating a plurality of candidate paths based on the star-ground fusion network model, and evaluating current link quality of each candidate path to obtain a current link quality evaluation result: wherein, represents a current link quality evaluation result, represents a historical data weight, represents a historical link quality evaluation result, represents a current link quality comprehensive score, which is obtained by weighting based on a data packet receiving ratio, a smoothed data packet receiving rate, a stability factor and a signal-to-noise ratio of the candidate path; evaluating transmission anti-interference capability of each node in the candidate paths to obtain an anti-interference safety degree: wherein, represents the degree of security against interference in the transmission of data between the nodes, represents the maximum degree of interference that can be tolerated between the nodes, represents the degree of interference between the nodes; determining a prediction factor based on a link quality change speed of each candidate link: wherein, denotes a prediction factor, denotes a link quality change speed, denotes a fastest speed of historical link quality change, denotes a time link quality assessment result, denotes a time link quality assessment result, denotes a time interval for calculating the change speed; comprehensively evaluating link stability of each candidate path according to the current link quality evaluation result, the anti-interference safety degree and the prediction factor of each candidate path to obtain a link stability comprehensive evaluation result: wherein, represents the link stability comprehensive evaluation result, , and represents the weight factor; taking a path with the optimal link stability comprehensive evaluation result in the candidate paths as a communication link for communication response based on the link stability comprehensive evaluation result, and generating a proactive routing strategy based on the communication link. 5.The satellite-terrestrial hybrid mesh network routing protocol and fault self-healing method according to any one of claims 1 to 4, wherein, Before the inputting the network topology structure information, node features of each node in the communication link and link features of the communication link into the target fault self-healing analysis model for fault self-healing analysis, the method further comprises: configuring reinforcement learning parameters, wherein the reinforcement learning parameters comprise state information, actions and rewards, the state information comprises current neighbor node information of a node, a link stability comprehensive evaluation result, a historical fault flag and residual energy, the actions comprise generating a fault self-healing link based on the state information, and the rewards comprise giving positive feedback or negative feedback based on a communication response result of the fault self-healing path; constructing a value function based on the reinforcement learning parameters, wherein the value function comprises: wherein, denotes a learning rate, denotes a discount factor, denotes an immediate reward for taking action in state denotes an immediate reward for taking action in state denotes an optimal value for taking action in state denotes an optimal value for taking action in state denotes a corresponding current value for taking action in state denotes an updated value for taking action in state constructing an original fault self-healing analysis model based on the value function; constructing a loss function according to a time difference strategy, and training the original fault self-healing analysis model based on the loss function to obtain a target fault self-healing analysis model, wherein the loss function comprises: wherein, represents a loss function, the loss function updates the current network parameters by back propagation and gradient descent iteration , represents the current policy network in the original fault self-healing analysis model training process, represents the current network parameters of the current policy network, represents the target policy network, represents the target network parameters of the target policy network. 6.The satellite-terrestrial hybrid mesh network routing protocol and fault self-healing method of claim 5, wherein, the constructing a value function based on the reinforcement learning parameters comprises: generating a graph structure based on network topology structure information, wherein the graph structure comprises a plurality of nodes and edges connecting between the nodes; configuring initial features to the graph structure according to the reinforcement learning parameters, wherein the initial features comprise initial node features and initial edge features, the initial node features comprise residual energy and a historical fault flag of a node, and the edge features comprise a link stability comprehensive evaluation result and link time delay information; generating an initial graph neural network based on the initial features and the graph structure; updating features of each node and each edge in the initial graph neural network based on a message passing mechanism to obtain a target graph neural network: wherein, representing a node a first layer of node features, representing a node a neighbor node a first layer of edge features, representing a node a set of neighbor nodes, representing a node weight matrix for a first layer, representing an activation function, representing a node a first layer of node features, representing a neighbor node a first layer of node features, representing a node a neighbor node a first layer of edge features, representing an edge weight matrix for a first layer; performing embedding vector conversion on the reinforcement learning parameters through the target graph neural network to obtain a reinforcement learning embedding vector, and constructing a value function based on the reinforcement learning embedding vector. 7.The satellite-terrestrial hybrid mesh network routing protocol and fault self-healing method according to any one of claims 1 to 4, wherein, The monitoring a communication response process of the communication link comprises: monitoring a response state of a neighbor node after each node in the communication link sends a communication request to the neighbor node in the communication response process; monitor a comprehensive evaluation result of link stability of the communication link in a communication response process; monitor energy alarm information of each node in the communication link in a communication response process, the energy alarm information including that the remaining energy of a node is lower than a preset energy threshold.

8. A satellite-terrestrial hybrid mixed ad hoc network routing protocol and fault self-healing device, characterized in that, The star-ground fusion hybrid ad hoc network routing protocol and fault self-healing device comprises: a star-ground fusion model construction module, configured to construct a star-ground fusion network model based on global routing information of a target region, the global routing information including network topology structure information of the target region, the star-ground fusion network model being composed of a plurality of nodes in the target region and a communication topology structure between the nodes, the nodes including ad hoc network nodes and gateway nodes, the gateway nodes being connected with a plurality of ad hoc network nodes, and the ad hoc network nodes being connected with at least one ad hoc network node; a routing strategy generation module, configured to generate a routing strategy according to the star-ground fusion network model, the routing strategy including a reactive routing strategy and a proactive routing strategy; a routing response module, configured to, when a task request is received, generate a target routing strategy based on a task priority of the task request and the routing strategy, the target routing strategy including a communication link for communication response; a communication response module, configured to, based on the target routing strategy, call each node in the communication link to perform communication response on the task request, and monitor a communication response process of the communication link; a fault self-healing module, configured to, when detecting that the communication link has a fault, input network topology structure information, node characteristics of each node in the communication link, and link characteristics of the communication link into a target fault self-healing analysis model to perform fault self-healing analysis, generate a fault self-healing link, and call each node in the fault self-healing link to perform communication response on the task request, so as to realize communication fault self-healing in a communication process.

9. A satellite-terrestrial hybrid mesh network routing protocol and fault self-healing device, characterized in that, The star-ground fusion hybrid ad hoc network routing protocol and fault self-healing device comprises a memory, a processor, and a star-ground fusion hybrid ad hoc network routing protocol and fault self-healing program stored in the memory and executable on the processor, the star-ground fusion hybrid ad hoc network routing protocol and fault self-healing program being configured to implement the star-ground fusion hybrid ad hoc network routing protocol and fault self-healing method in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a star-ground fusion hybrid ad hoc network routing protocol and fault self-healing program, the star-ground fusion hybrid ad hoc network routing protocol and fault self-healing program being executed by a processor to implement the star-ground fusion hybrid ad hoc network routing protocol and fault self-healing method in any one of claims 1 to 7.

Images