A wireless sensor network node security state evaluation method and system

Abstract

The application discloses a kind of wireless sensor network node security state evaluation method and system, it is related to network security technical field, the present application obtains multiple groups of virtual neighbors by each wireless sensor node, and judges based on virtual adjacent coefficient Xx, so that network can flexibly cope with node loss or link failure problem.When node fails, it can be replaced by virtual neighbor in time to replace the lost node, avoids the problem that network interruption is caused by node loss, improves the fault tolerance and reliability of network.In the process of node online state updating, the present application sets timeout duration T, and starts maintenance mechanism, dynamically updates the online state of each node in task path, which avoids the situation that path updating is lagged due to network topology change, makes the network path more flexible, timely response to node failure or fault change.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of network security, in particular to a wireless sensor network node security state evaluation method and system. BACKGROUND

[0002] As an important part of the Internet of Things technology, wireless sensor networks have been widely applied in intelligent management, environmental monitoring, intelligent transportation, smart cities and other fields. The core idea is to complete environmental data collection, analysis and transmission through the cooperation of a large number of sensor nodes. The role of wireless sensor network nodes is extremely important in these systems, which undertakes tasks such as data transmission and network coordination. However, with the increase in the number of nodes and environmental complexity, how to ensure the reliability of inter-node communication and reduce distortion and delay in the data transmission process has become the key to the stability and efficiency optimization of wireless sensor networks.

[0003] Currently, the node security state evaluation of wireless sensor networks mainly relies on the comprehensive analysis of multi-dimensional information such as the health status of each node, communication quality and network topology structure. However, the traditional security evaluation method usually ignores the influence of dynamic interaction between nodes and fails to effectively consider the complexity and real-time changes in wireless sensor networks. For example, in an environment with frequent changes in network topology, the online status of some nodes and virtual neighbor relationships may change frequently, and existing methods lack adaptability to such dynamics. Therefore, when there are problems such as loss of relay nodes, unstable paths in the network, the traditional method often cannot make accurate judgments in real time, resulting in problems such as task path delay, data loss or transmission failure. And when selecting virtual neighbors with better signal quality and path quality as new relay nodes, although these nodes may be superior in signal quality and path quality, the selection of these nodes does not necessarily mean that they are better in the entire data transmission path, so the accurate selection of replacement nodes also needs to be improved. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a wireless sensor network node security state evaluation method and system, which solves the problems in the background art.

[0005] To achieve the above purpose, the present application is implemented by the following technical scheme: a wireless sensor network node security state evaluation method, comprising the following steps,

[0006] S1, based on the real-time monitoring of the sensor group on the network running information in the wireless sensor network, and according to the network topology in the wireless sensor network, determining the task path;

[0007] S2, based on the sensor group in S1 real-time monitoring of the operation, get the history set, and according to the history set, for each wireless sensor node seeks a plurality of virtual neighbors, analysis of the virtual proximity coefficient Xx, if the virtual proximity coefficient Xx exceeds the set threshold, then set the preliminary replacement relay node;

[0008] S3 set timeout duration T, start maintenance mechanism, update the online state of each wireless sensor node in the task path, based on the online state, start updating the routing table in the wireless sensor network;

[0009] S4, in the process of updating the routing table in the wireless sensor network, according to the replacement relay node, analyze the replacement effect of each replacement relay node, and combine the trained network security prediction model to fit and obtain the path transmission distortion index Zs;

[0010] S5, according to the value of the path transmission distortion index Zs, select the final relay node to reset the task path and complete the secure communication between each wireless sensor node.

[0011] Preferably, S1 specific steps include:

[0012] S11, the sensor group is used to collect the network running information in the wireless sensor network in advance, wherein the network running information includes the self running state data of each wireless sensor node in the wireless sensor network and the running state data between each wireless sensor node, the self running state data of each wireless sensor node includes the total number of connections Ls of wireless sensor node i with other wireless sensor nodes in the past time i , the independent neighbor group Z of wireless sensor node i i , the delay time YC of the kth wireless sensor node in the independent neighbor group k , the bandwidth DK of the kth wireless sensor node in the independent neighbor group k , the packet loss rate DZ of the kth wireless sensor node in the independent neighbor group k , the average connection time LS of wireless sensor node in the history period avg , and the loss frequency SP of each wireless sensor node; the running state data between each wireless sensor node includes the connection times Ls of each wireless sensor node with other wireless sensor nodes and the independent neighbor group Z corresponding to the kth wireless sensor node in the independent neighbor group of wireless sensor node i ik ; the network running information is dimensionless processed based on the method of Z-score standardization.

[0013] S12, each wireless sensor node in the wireless sensor network collects network running information of wireless sensor nodes directly communicating with it by listening to the wireless channel to establish physical adjacency relationship, and forms the topology of the wireless sensor network through multiple groups of physical adjacency relationship.

[0014] Preferably, the specific steps of S2 include:

[0015] S21, the network running information obtained in real time in S11 is stored periodically to generate a history set, which includes self running state data of each wireless sensor node in the wireless sensor network and running state data between each wireless sensor node in a historical period;

[0016] S22, based on the history set obtained in S21, the wireless sensor nodes directly communicating with each wireless sensor node in the historical period are determined to establish an independent neighbor group for each wireless sensor node, and the degree of virtual adjacency relationship between each wireless sensor node in each group of physical adjacency relationship in the wireless sensor network is analyzed according to the establishment of the independent neighbor group for each wireless sensor node to respectively calculate the virtual neighbor probability P(p i →p j ), distance prediction difference Jj, similarity value Ss(i, k) and path quality value Lc;

[0017] S221, based on time series analysis method and in combination with the history set, the probability of becoming virtual adjacency relationship between each wireless sensor node is predicted to obtain the virtual neighbor probability P(p i →p j ), which is specifically obtained according to the following formula:

[0018]

[0019] In the formula, P(p i →p j ) is the virtual neighbor probability, which represents the probability of wireless sensor node i becoming virtual adjacency relationship with wireless sensor node j after time t; i and j are the numbers of wireless sensor nodes in the wireless sensor network; Ls ij is the connection times of wireless sensor node i and wireless sensor node j in the past time window; Ls i is the total connection number of wireless sensor node i with other wireless sensor nodes in the past time; f(Δt) is a function, which represents the influence of time difference on virtual adjacency relationship;

[0020] S222, for the mobility of the wireless sensor node, predicting the distance between each wireless sensor node and the wireless sensor nodes in its corresponding independent neighbor group, and obtaining a distance prediction difference Jj based on the predicted distance between each wireless sensor node and the wireless sensor nodes in its corresponding independent neighbor group;

[0021] S223, determining the similarity between the independent neighbor group of each wireless sensor node and the independent neighbor group based on each wireless sensor node and its corresponding independent neighbor group, to obtain a similarity value Js(i, k), specifically according to the following formula:

[0022]

[0023] In the formula, Js(i, k) is the similarity value, which represents the similarity between nodes by comparing the intersection and union of the independent neighbor group of the wireless sensor node i and the independent neighbor group corresponding to the kth wireless sensor node in it. i Z ik Z i ∩Z ik |Z i and Z ik |Z i ∪Z ik |Z i and Z ik all physical adjacent nodes;

[0024] S224, based on the multi-hop path existing in the task path, evaluating the path quality state between each wireless sensor node and the wireless sensor nodes in its independent neighbor group to obtain a path quality value Lc, specifically according to the following formula:

[0025]

[0026] In the formula, G is the selected number, YC k DK k DZ k is the packet loss rate of the selected kth wireless sensor node in the independent neighbor group; k is the number of wireless sensor nodes in the independent neighbor group.

[0027] Preferably, the specific steps of S2 further include:

[0028] S23, based on the virtual neighbor probability P(p i →p j), distance prediction difference value Jj, similarity value Js(i, k) and path quality value Lc, the degree of virtual adjacency relationship between each wireless sensor node in each group of physical adjacency relationship in the wireless sensor network is analyzed to construct a virtual proximity coefficient Xx, which is obtained by the following formula:

[0029]

[0030] In the formula, α1, α2, α3 and α4 are weights, and specific values are set by the user according to the situation;

[0031] S24, if the virtual proximity coefficient Xx exceeds the set threshold value, the corresponding wireless sensor node in the independent neighbor group is preliminarily set as a relay node to be replaced, and after statistics, a replacement list is established, which includes a plurality of groups of relay nodes to be replaced corresponding to the corresponding wireless sensor node.

[0032] Preferably, S3 includes the following specific steps:

[0033] S31, based on the history set, the stability of the connection of each wireless sensor node in the historical period is dynamically analyzed to extract the loss frequency SP of each wireless sensor node, and a dynamic independent timeout period T is set for each wireless sensor node, wherein the independent timeout period T corresponding to each wireless sensor node is obtained as follows:

[0034] T = LS avg + (μ * LS σ ) * K;

[0035] In the formula, LS avg is the average connection time of the wireless sensor node in the historical period; LS σ is the standard deviation of the historical connection time of the wireless sensor node, μ is an adjustment factor; and K is a coefficient.

[0036] S311, the value of the coefficient K is determined by the value of the loss frequency SP of each wireless sensor node, when the value of the loss frequency SP of the corresponding wireless sensor node exceeds the average loss frequency SP avg , it indicates that the connection of the corresponding wireless sensor node fluctuates frequently, and at this time the value of the coefficient K is set to 2; when the value of the loss frequency SP of the corresponding wireless sensor node does not exceed the average loss frequency SP avg , it indicates that the connection of the corresponding wireless sensor node fluctuates within a normal range, and at this time the value of the coefficient K is set to 1.

[0037] Preferably, S3 further includes the following specific steps:

[0038] S32, according to S31, dynamically setting an independent timeout period T for each wireless sensor node, starting a maintenance mechanism for each wireless sensor node in the task path, each wireless sensor node sharing information through broadcasting according to the independent timeout period T dynamically set by itself with the directly communicating wireless sensor nodes having a physical adjacent relationship, when the corresponding wireless sensor node does not share information through broadcasting within the independent timeout period T dynamically set by itself, it indicates that the current wireless sensor node has a malfunction, updates the current wireless sensor node to be offline, and marks it as an offline physical node; when the corresponding wireless sensor node shares information through broadcasting within the independent timeout period T dynamically set by itself, update the current wireless sensor node to be online and do not mark it, at this time the wireless sensor node not marked will continue to transmit information according to the task path;

[0039] S33, according to the replacement list to be replaced in S24, extracting the replacement list to be replaced in S32, based on the replacement list to be replaced of the offline physical node, starting the routing table updating operation of each wireless sensor node in the task path.

[0040] Preferably, S4 includes the following specific steps:

[0041] S41, in the process of updating the routing table in the wireless sensor network, according to the offline physical node, determining the bandwidth condition of each replacement relay node in the replacement list to be replaced of the offline physical node, obtaining the bandwidth difference ΔDc, specifically:

[0042] ΔDc g = DK g - DK;

[0043] In the formula, ΔDc g is the difference between the bandwidth of the gth replacement relay node in the replacement list to be replaced of the offline physical node and the bandwidth of the offline physical node, DK g is the bandwidth of the gth replacement relay node in the replacement list to be replaced of the offline physical node, and DK is the bandwidth of the offline physical node.

[0044] Preferably, S4 further includes the following specific steps:

[0045] S42, analyze the replacement effect of the replacement relay node in the replacement list to be replaced of the offline physical node, and construct a network security prediction model according to the deep learning technology, and after dimensionless processing, fit the output path transmission distortion index Zs from the output end of the network security prediction model, the path transmission distortion index Zs is obtained by the following formula:

[0046]

[0047] In the formula, Delta CZ is the increment of transmission delay, Delta TZ is the increment of hop count, Delta DC is the bandwidth difference, Lambda is the penalty factor, SP is the loss frequency, rho is the exponent of the penalty factor, eta is the attenuation coefficient, LS max is the longest time of the history of the corresponding to-be-replaced relay node connection, is the attenuation factor, exp(*) is the exponential function with e as the base, and I(ADc<0) is the indicator function.

[0048] Preferably, the specific steps of S5 include the following steps:

[0049] S51, based on the manner of acquiring the path transmission distortion index Zs in S42, the path transmission distortion index Zs of each to-be-replaced relay node relative to the offline physical node in the to-be-replaced list of the offline physical node is acquired respectively, and through a statistical algorithm, the to-be-replaced relay node with the minimum path transmission distortion index Zs value is updated, which is recorded as a replacement relay node, the replacement relay node is replaced by the corresponding offline physical node, and the task path resetting operation is realized.

[0050] A wireless sensor network node security state evaluation system includes a present situation determination subsystem, a relay analysis subsystem, a mechanism setting subsystem and a screening subsystem.

[0051] The present situation determination subsystem is used for monitoring network operation information in the wireless sensor network in real time according to a sensor group, and determining a task path according to a network topology in the wireless sensor network.

[0052] The relay analysis subsystem is used for acquiring a history set through real-time monitoring of the sensor group, and seeking multiple groups of virtual neighbors for each wireless sensor node according to the history set, analyzing a virtual proximity coefficient Xx, and preliminarily setting a to-be-replaced relay node if the virtual proximity coefficient Xx exceeds a set threshold.

[0053] The mechanism setting subsystem is used for setting an overtime duration T, starting a maintenance mechanism, updating the online state of each wireless sensor node in the task path, and starting an updating operation of a routing table in the wireless sensor network based on the online state.

[0054] The screening subsystem is used for analyzing the replacement effect of each to-be-replaced relay node according to the to-be-replaced relay node in the process of updating the routing table in the wireless sensor network, fitting a path transmission distortion index Zs in combination with a trained network security prediction model, and selecting a final relay node to reset the task path and complete the security communication between each wireless sensor node.

[0055] The present application provides a wireless sensor network node security state evaluation method and system, which has the following beneficial effects:

[0056] (1) The method determines the task path by real-time monitoring information based on network topology and sensor groups, which can automatically adjust the task path according to the changes of the network environment, so as to avoid the limitations of manually setting the path in the traditional path selection method, and flexibly select the relatively optimal path according to the real-time monitored node state and network load condition, thereby improving the data transmission efficiency. The application obtains multiple groups of virtual neighbors for each wireless sensor node, and judges based on the virtual proximity coefficient Xx, so that the network can flexibly cope with the problems of node loss or link failure. When the node fails, the lost node can be replaced in time through the virtual neighbor, avoiding the problem of network interruption caused by node loss, and improving the fault tolerance and reliability of the network. In the node online state updating process, the application sets a timeout period T and starts the maintenance mechanism to dynamically update the online state of each node in the task path, ensuring the stability and efficiency of the network. This avoids the situation of path update lag caused by network topology changes, making the network path more flexible and timely responding to node failure or fault changes. The application analyzes the replacement effect of the replacement relay node, combines the trained network security prediction model, and fits to obtain the path transmission distortion index Zs, which can effectively evaluate the quality of each candidate relay node, so as to accurately select the final relay node. This process enables the network to automatically select a relatively better replacement path when the node fails, ensuring the quality and efficiency of data transmission, and avoiding the problems of data loss or excessive transmission delay.

[0057] (2) The application can timely discover and handle potential security risks by comprehensively considering factors such as node security state, network topology, transmission delay, path quality, etc. For example, it can automatically adjust the network task path in the case of relay node failure, unstable link or path distortion, etc., to ensure safe and efficient data transmission. In addition, by dynamically adjusting the task path, the risk of data leakage or loss caused by incorrect path selection is reduced, thereby greatly improving the security of the wireless sensor network. In summary, the method of the application effectively improves the reliability, robustness and security of the network in the dynamic environment of the wireless sensor network by combining real-time monitoring, virtual neighbor mechanism, path optimization and security prediction model, and solves the problems of path update lag and network interruption caused by node loss in traditional wireless sensor networks.

[0058] (3) By introducing virtual neighbors into the wireless sensor network and analyzing the virtual adjacency relationships of each node based on historical data and prediction models, the network can identify potential nodes that can become effective connections. This predictive connection mechanism improves the fault tolerance of the network when nodes fail or the network topology changes. If some nodes fail or links are interrupted, the system can quickly switch to the pre-identified virtual neighbors, avoiding communication interruptions caused by node loss or network disconnection, ensuring the continuity and stability of network communication. By analyzing the physical connections between wireless sensor nodes and their neighbor nodes and the virtual adjacency relationships, combining historical data, node mobility, and prediction models, the network can more intelligently evaluate and select the best path. Based on factors such as virtual neighbor probability, similarity value, distance prediction difference, and path quality value, the system can dynamically evaluate the number of hops, delay, bandwidth, and packet loss rate in the task path, thereby optimizing the selection of the task path. This intelligent path selection mechanism greatly reduces path delay and improves data transmission efficiency, especially in unstable topologies or environments with large changes in node location.

[0059] (4) By analyzing the virtual adjacency relationships between each wireless sensor node in each group of physical adjacency relationships and calculating the virtual proximity coefficient Xx, the system can identify potential relay nodes in real time and replace faulty or poorly performing nodes as needed. This intelligent node replacement mechanism enables the network to quickly adjust the task path when facing node failures, link interruptions, or other unforeseen abnormal situations, ensuring data transmission continuity and stability and enhancing the network's fault tolerance. The present invention establishes a list of relay nodes to be replaced to ensure that the network can adjust and optimize data transmission paths in a changing environment. Based on the evaluation of multiple groups of virtual neighbor nodes, the network can flexibly select the best relay node, reducing path delay and packet loss rate and improving overall network resource utilization. In large wireless sensor networks, this approach not only improves communication efficiency between nodes but also supports more device access and expansion, improving the overall scalability of the network. By accurately predicting and evaluating the virtual adjacency relationships of nodes, the system can avoid redundant relay node selection and reduce unnecessary resource waste.

[0060] (5) The system adjusts the timeout length T to adapt to the characteristics of each node, ensuring that the node can respond more flexibly to changes when the network environment fluctuates, which can reduce the instability caused by frequent disconnection of node connections and improve the access reliability of the node, so that the entire wireless sensor network maintains high stability in long-term operation. Each node broadcasts information sharing tasks to physically adjacent nodes according to the dynamically set timeout length T. When the node does not complete the broadcast task within the predetermined timeout length, it will be automatically marked as an offline physical node, triggering the maintenance mechanism to update the network state and perform route table updates. This adaptive maintenance mechanism ensures real-time monitoring of each node in the wireless sensor network, timely detection of failed nodes and replacement, thereby avoiding the impact of failed nodes on the entire network and improving the self-healing ability of the network. BRIEF DESCRIPTION OF DRAWINGS

[0061] Figure 1 A wireless sensor network node safety state evaluation method step schematic diagram is provided for the application.

[0062] Figure 2 A task path display diagram in a wireless sensor network node safety state evaluation method is provided for the application.

[0063] Figure 3 A node relationship diagram in a wireless sensor network node safety state evaluation method is provided for the application.

[0064] Figure 4 A flowchart of a wireless sensor network node safety state evaluation method is provided for the application.

[0065] Figure 5 A wireless sensor network node safety state evaluation system block diagram is provided for the application. DETAILED DESCRIPTION

[0066] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0067] Embodiment 1

[0068] Please refer to Figures 1-4 The application provides a wireless sensor network node safety state evaluation method, comprising the following steps,

[0069] S1, real-time monitoring network running information in the wireless sensor network based on the sensor group, and determining a task path according to the network topology in the wireless sensor network; the task path is a process of transmitting a data packet from a source node to a target node;

[0070] S2, real-time monitoring the task based on the sensor group in S1, obtaining a historical set, and seeking a plurality of groups of virtual neighbors for each wireless sensor node according to the historical set, analyzing and obtaining a virtual proximity coefficient Xx, if the virtual proximity coefficient Xx exceeds a set threshold value, then preliminarily setting a to-be-replaced relay node;

[0071] S3, setting a timeout duration T, starting a maintenance mechanism, updating the online state of each wireless sensor node in the task path, based on the online state, starting an updating routing table task in the wireless sensor network;

[0072] S4, in the process of updating the routing table in the wireless sensor network, analyzing the replacement effect of each to-be-replaced relay node according to the to-be-replaced relay node, and combining a trained network security prediction model to fit and obtain a path transmission distortion index Zs;

[0073] S5, according to the numerical value of the path transmission distortion index Zs, selecting the final relay node to reset the task path, and completing the secure communication between each wireless sensor node.

[0074] In this embodiment, the method can accurately determine the task path by monitoring the network operation information in the wireless sensor network in real time, including the online state of each node, signal quality data, and combining the dynamic changes of network topology. By real-time acquisition of the running data of each node and historical connection information, the appropriate virtual neighbor is selected for the node, thereby enhancing the adaptive ability of the network when the topology changes. This adaptive ability effectively avoids network interruption or data loss caused by node failure or topology change. In the traditional network state evaluation method, after the node loss or connection interruption, the network needs a long time to recover. Through this method, the timeout length and virtual neighbor selection mechanism can quickly start the replacement mechanism when the node fails, and optimize the task path, reduce the path transmission delay, and ensure the efficiency and reliability of data transmission. In addition, through the virtual neighbor and path selection scheme based on the historical connection stability, the risk of data loss can be effectively reduced. Enhance the security of the network: by combining the trained network security prediction model, the method can dynamically evaluate the transmission distortion index Zs of each path in the network. Through the evaluation of the path transmission distortion index, the system can automatically identify and replace the relay node with potential risks, avoiding the threat of malicious nodes or unstable nodes to network security. The introduction of the path transmission distortion index not only helps to enhance the robustness of the network, but also improves the reliability of data transmission, preventing security risks such as network attacks or data tampering. Dynamic optimization of path selection and replacement of relay nodes: in S4 step, the method fits the path transmission distortion index Zs, analyzes the to-be-replaced relay node in combination with the actual operation status of the network, and ensures the selection of a relatively better relay node for path update. This dynamic optimization of path selection mechanism can adapt to changes in the network environment, ensure the efficiency and security of the communication path, and thus improve the overall performance of the wireless sensor network. Reduce the manual intervention of network maintenance: since the method can monitor and evaluate the state of the node, topology changes and path quality in real time, the need for manual intervention is reduced, greatly improving the automation level of the network. The automatic maintenance mechanism of the network enables the system to respond quickly when an exception occurs, automatically repair network faults, reduce operating costs and improve network availability. In summary, the present application optimizes the path selection and node replacement mechanism of the wireless sensor network by considering factors such as node state, historical data, topology structure and network security, improves the stability, security and transmission efficiency of the network, and reduces the impact of network faults on the overall system performance, which has important technical significance and practical application value.

[0075] Embodiment 2

[0076] Please refer to Figure 1 , specifically: S1 specific steps include:

[0077] S11, collecting network running information in the wireless sensor network in real time by using a sensor group in advance, wherein the network running information includes self running state data of each wireless sensor node in the wireless sensor network and running state data between each wireless sensor node, the self running state data of each wireless sensor node includes total number of connections Ls of the wireless sensor node i with other wireless sensor nodes in the past time i , independent neighbor group Z of the wireless sensor node i i , delay duration YC of the kth wireless sensor node in the independent neighbor group k , bandwidth DK of the kth wireless sensor node in the independent neighbor group k , packet loss rate DZ of the kth wireless sensor node in the independent neighbor group k , average connection time LS of the wireless sensor node in the historical period avg , and loss frequency SP of each wireless sensor node; the running state data between each wireless sensor node includes connection times Ls of each wireless sensor node with other wireless sensor nodes and independent neighbor group Z corresponding to the kth wireless sensor node in the independent neighbor group of the wireless sensor node i ik ;

[0078] The network running information is processed by a method based on Z-score standardization, which is used to convert physical quantities with different units into quantities without units. The purpose of this is to simplify the calculation, compare the relative size of different quantities, or to adapt to the needs of some algorithms (such as machine learning, numerical optimization) for data.

[0079] S12, each wireless sensor node in the wireless sensor network collects network running information of wireless sensor nodes directly communicating with it by listening to the wireless channel to establish a physical adjacency relationship, and forms a topology of the wireless sensor network through multiple groups of physical adjacency relationships. Meanwhile, each wireless sensor node in the physical adjacency relationship shares itself and the wireless sensor nodes directly communicating with it with the physical neighbor nodes through broadcasting (such as Hello message in the routing protocol) to complete the physical update of the task path in the wireless sensor network.

[0080] The physical adjacency relationship mentioned above refers to the relationship that each wireless sensor node can directly connect and communicate in the task path. Unlike the virtual neighbor and the replacement relay node, the virtual neighbor is a node that may become an effective connection in the future in the wireless sensor network through prediction or historical data, and the replacement relay node refers to a node that is determined through the virtual neighbor of the corresponding node previously mastered when the node with physical adjacency relationship in the task path fails, loses or malfunctions. In other words, the replacement relay node is a kind of replacement node.

[0081] In this embodiment, by using the Z-score standardization method to normalize the energy consumption, data transmission volume, node position, communication frequency and other data of each wireless sensor node, the differences in data units can be effectively removed, so that different physical quantities can be compared under the same standard. This not only simplifies the subsequent calculation, but also ensures the consistency of network state data, especially when facing different data sources and different calculation algorithms, providing higher calculation efficiency and accuracy. By monitoring the wireless channel and collecting real-time network operation information of the direct communication node, the physical adjacency relationship of each node can be dynamically updated, and a real-time network topology structure can be formed. This real-time updating method avoids the limitations of static routing table, so that the wireless sensor network can quickly adapt to the changes of network topology, especially in the case of node failure, joining or leaving, ensuring the rapid adjustment of data transmission path and efficient operation of the network. In the process of sharing physical adjacency information through broadcast (such as Hello message), each wireless sensor node can timely understand the status of the neighbor node directly communicating with it, helping the node to quickly select a backup path when a fault or connection loss occurs. This further enhances the fault tolerance and robustness of the network, avoids the impact of single point failure on the network, and improves the stability and reliability of the network. By establishing physical adjacency relationship and broadcasting, nodes can quickly share information, and then promote the physical update of wireless sensor network task path. This mechanism makes path update more rapid and accurate, can respond to network topology changes in real time, optimize task path, reduce delay, and ensure the real-time and reliability of data transmission. Through real-time monitoring of the running state of each node and dynamic updating of the physical adjacency relationship, accurate network state data is provided, which provides strong support for subsequent intelligent decision-making based on algorithms. For example, based on machine learning algorithms or numerical optimization methods, more scientific network routing decisions and node safety state evaluations can be made according to real-time data, thereby effectively reducing redundant nodes in the network and improving the utilization rate of network resources. By collecting information such as energy consumption and data transmission volume of nodes, combined with standardized operation methods, the present application can dynamically adjust the energy allocation and data transmission strategy of nodes while ensuring network stability, thereby effectively reducing energy consumption, prolonging the life cycle of network nodes, and improving the overall energy efficiency of the network. In summary, the method of the present application, through real-time network operation information collection, dynamic adjacency relationship updating and accurate standardization processing, not only effectively improves the stability, reliability and energy efficiency of the wireless sensor network, but also provides strong data support for subsequent intelligent routing decisions and node safety state evaluations, ensuring efficient update and optimization of task path.

[0082] Embodiment 3

[0083] Please refer to Figure 3, specifically: S2 specific steps include:

[0084] S21, by storing the network operation information acquired in real time in S11 periodically, to generate a history set, the history set including the self-operation state data of each wireless sensor node in the wireless sensor network and the operation state data between each wireless sensor node in a historical period;

[0085] S22, based on the history set acquired in S21, determining the wireless sensor nodes directly communicated by each wireless sensor node in the historical period, to establish an independent neighbor group for each wireless sensor node, and analyzing the degree of virtual adjacency relationship between each wireless sensor node in each group of physical adjacency relationship in the wireless sensor network according to the establishment of the independent neighbor group for each wireless sensor node, to respectively calculate and acquire the virtual neighbor probability P(p i →p j ), distance prediction difference Jj, similarity value Js(i, k) and path quality value Lc;

[0086] S221, based on time series analysis method and combined with the history set, predicting the probability of becoming virtual adjacency relationship between each wireless sensor node, to acquire the virtual neighbor probability P(p i →p j ), specifically acquired according to the following formula:

[0087]

[0088] In the formula, P(p i →p j ) is the virtual neighbor probability, representing the probability of wireless sensor node i becoming virtual adjacency relationship with wireless sensor node j after time t; i and j are both the number of wireless sensor nodes in the wireless sensor network; Ls ij is the connection times of wireless sensor node i and wireless sensor node j in the past time window; Ls i is the total number of connections of wireless sensor node i with other wireless sensor nodes in the past time; f(Δt) is a function, representing the influence of time difference on virtual adjacency relationship, the longer the time interval, the smaller the prediction probability of virtual adjacency relationship will be;

[0089] The above-mentioned connection times Ls ij of wireless sensor node i and wireless sensor node j in the past time window and the total number of connections Ls i of wireless sensor node i with other wireless sensor nodes in the past timeAll can be monitored and obtained by communication quality monitoring sensors (for example: signal strength sensor, link quality monitoring sensor), working principle: wireless sensor nodes will establish connection with other nodes through periodic broadcast information, the number of connections can be counted by recording each successful connection event. The sensor monitors the signal strength between nodes, the stability of connection establishment and the quality of data transmission, so as to calculate the number of connections.

[0090] Virtual neighbor refers to the node that may become an effective connection in the future through prediction or historical data. The concept of virtual neighbor usually appears in mobile network or unstable network topology, which is used to enhance the fault tolerance and flexibility of the network.

[0091] S222, for the mobility of wireless sensor nodes, predicting the distance between each wireless sensor node and the wireless sensor nodes in its corresponding independent neighbor group, and obtaining the distance prediction difference Jj based on the predicted distance between each wireless sensor node and the wireless sensor nodes in its corresponding independent neighbor group;

[0092] S223, according to each wireless sensor node and its corresponding independent neighbor group, determining the similarity between the independent neighbor group of each wireless sensor node and the independent neighbor group (because each wireless sensor node has its independent neighbor group, even the nodes in the independent neighbor group also have their own independent neighbor group, so through this process, the similarity of the connection between the nodes in the face of history can be further considered), to obtain the similarity value Js(i, k), which is obtained according to the following formula:

[0093]

[0094] In the formula, Js(i, k) is the similarity value, which represents the similarity between nodes by comparing the intersection and union of the independent neighbor group of wireless sensor node i and the independent neighbor group corresponding to the kth wireless sensor node in it;Z i Z ik Z i ik |Z i ik |Z i ik |Z i ik |Z i ik Z

[0095] ​​​​​S224, based on the multi-hop path existing in the task path, evaluate the path quality state between each wireless sensor node and the wireless sensor nodes in its independent neighbor group to obtain a path quality value Lc, which is obtained according to the following formula:

[0096]

[0097] In the formula, G is the selected number (the number of nodes in the independent neighbor group of the wireless sensor node), k is the node number in the independent neighbor group, YC k is the delay duration of the kth wireless sensor node in the selected independent neighbor group, DK k is the bandwidth of the kth wireless sensor node in the selected independent neighbor group, DZ k is the packet loss rate of the kth wireless sensor node in the selected independent neighbor group; and k is the number of the wireless sensor node in the independent neighbor group.

[0098] The delay duration YC k of the kth wireless sensor node in the selected independent neighbor group can be monitored and obtained by a time delay measurement sensor;

[0099] The bandwidth DK k of the kth wireless sensor node in the selected independent neighbor group can be monitored and obtained by a bandwidth monitoring sensor;

[0100] The packet loss rate DZ k of the kth wireless sensor node in the selected independent neighbor group can be monitored and obtained by a data packet loss monitoring sensor (for example, a packet loss detection sensor or a network performance evaluation sensor);

[0101] In this embodiment, when the physical neighbor nodes fail or the network topology changes, the system can quickly predict the possible virtual neighbors in the future using historical data to ensure the stability of the network and the continuity of the communication path, reducing the communication interruption caused by node changes or unstable connection. The present application predicts the probability of each node forming a virtual adjacency relationship with other nodes by real-time monitoring of the network running state of each node and combining time series analysis method. This method not only considers the historical connection data between nodes, but also adjusts the virtual neighbor probability according to the influence of time interval, thereby improving the dynamic adaptability of the network topology. When the topology changes, the network can more flexibly adapt to the new inter-node connection relationship, avoiding the problems of path blockage or high delay caused by the traditional static topology unable to respond to node changes in time. Based on the similarity measurement between each node and the nodes in its independent neighbor group and the path quality evaluation, the present application can evaluate the communication path quality between nodes, including delay, bandwidth, packet loss rate and other factors. By integrating these evaluation results, the system can select the optimal virtual neighbor node as the communication relay for each node, thereby improving the communication efficiency and stability of the network. Especially in the multi-hop path environment, this method can ensure the quality of data transmission by dynamically selecting the best relay node, reducing transmission delay and packet loss, and improving the overall network performance. The present application predicts the formation probability of virtual neighbors, calculates the path quality value and evaluates the similarity between nodes, effectively avoiding unnecessary resource waste in the network. Based on the historical data and prediction information between nodes, the network can intelligently select appropriate nodes as relays, avoiding redundant calculation and unreasonable resource allocation, which not only improves the resource utilization rate of each node, but also reduces unnecessary communication and calculation overhead, thereby improving the efficiency of the entire wireless sensor network. Compared with the traditional fixed path selection method, the present application provides a more flexible and efficient path selection strategy, which helps to improve the transmission rate and reduce the delay of the network. Since the present application can dynamically update the topology of the network according to the real-time monitoring running data and historical data, and timely adjust the task path according to the virtual neighbor prediction mechanism, the network can quickly respond when the topology changes, enhancing its scalability and self-healing ability.

[0102] Embodiment 4

[0103] Please refer to Figure 1 , specifically: S2 specific steps also include:

[0104] S23, based on the virtual neighbor probability P(p i →p j), distance prediction difference value Jj, similarity value Js(i, k) and path quality value Lc, analyzing the degree of virtual adjacent relationship between each wireless sensor node in each group of physical adjacent relationship in the wireless sensor network, to construct a virtual adjacent coefficient Xx, which is obtained by the following formula:

[0105]

[0106] In the formula, α1, α2, α3 and α4 are weights, wherein 0 < α1 < 1, 0 < α2 < 1, 0 < α3 < 1, 0 < α4 < 1, and the specific values are set by the user according to the situation;

[0107] S24, if the virtual adjacent coefficient Xx exceeds the set threshold value, the corresponding wireless sensor node in the independent neighbor group is preliminarily set as a to-be-replaced relay node, and after statistics, a to-be-replaced list is established, which includes a plurality of groups of to-be-replaced relay nodes corresponding to the corresponding wireless sensor node.

[0108] By comprehensively analyzing the degree of virtual adjacency relationship within each physical adjacency relationship in the wireless sensor network, a virtual proximity coefficient Xx is constructed, and the relative adaptability between nodes is dynamically evaluated according to this coefficient, which can accurately predict which nodes may become effective relay nodes in the future period. Compared with traditional static routing methods, this method introduces more dynamic factors, making the node selection more flexible and accurate, and avoiding the problems of network communication interruption or delay caused by topology changes, load fluctuations or node performance degradation. Through the prediction of virtual adjacency relationship, those nodes that may lose effective connection in the future can be identified and marked in advance, so as to replace them with more stable relay nodes in advance. This preventive replacement strategy improves the fault tolerance of the network, especially in the case of node energy depletion or connection interruption due to mobility changes, the system can intelligently replace nodes to ensure seamless data flow transmission and improve the robustness of the network. By analyzing the virtual adjacency relationship between each wireless sensor node and its neighbor nodes, and based on the calculation of virtual proximity coefficient Xx, the task path in the network can be further optimized. According to different node replacement schemes, the selection of relatively optimal relay nodes can effectively reduce the path transmission delay, reduce the packet loss rate and improve the bandwidth utilization, thereby improving the efficiency and quality of data transmission. Through the automatic node replacement and dynamic path selection mechanism, the invention can significantly reduce the complexity of manual intervention and network maintenance. Network managers do not need to frequently monitor node status or manually adjust routing, but through intelligent algorithms and adaptive mechanisms, node replacement and path optimization are automatically completed, effectively reducing operating costs and manpower investment, and improving the self-organizing ability and management efficiency of the network. As the network size expands or the topology structure changes, traditional node selection and routing algorithms may face performance bottlenecks, while the invention introduces virtual adjacency relationship analysis and virtual proximity coefficient construction, which can dynamically adjust node connection and routing path according to the changes of network state. In this way, the network not only runs efficiently under the current scale, but also maintains high adaptability and scalability in the process of continuous expansion.

[0109] In the traditional network, the failure of nodes and the change of network topology often bring about sudden performance problems, especially in wireless sensor networks. By constructing virtual adjacency relationship and calculating virtual proximity coefficient in advance, the system can predict the possible failure or connection interruption of nodes in advance according to historical data and prediction model, so as to make maintenance and replacement preparations in advance, avoid network service interruption and ensure long-term stable operation of the system. In summary, the invention can greatly improve the intelligence, fault tolerance and adaptability of wireless sensor networks through dynamic analysis based on virtual adjacency relationship, optimization of node replacement mechanism and path quality evaluation, providing an efficient, reliable and flexible solution for large-scale, dynamically changing wireless sensor networks in practical applications.

[0110] Embodiment 5

[0111] Please refer to Figures 1-4 , specifically: S3 specific steps include:

[0112] S31, based on the history set, dynamically analyze the stability of each wireless sensor node connection in the historical period, to extract the loss frequency SP of each wireless sensor node, and dynamically set an independent timeout period T for each wireless sensor node, wherein the independent timeout period T corresponding to each wireless sensor node is dynamically set as follows:

[0113] T = LS avg + (μ * LS σ ) * K;

[0114] In the formula, LS avg is the average connection time of the wireless sensor node in the historical period, indicating how long the corresponding wireless sensor node can usually maintain connection; LS σ is the standard deviation of the historical connection time of the wireless sensor node, μ is an adjustment factor, which is a constant, used to adjust the sensitivity of the timeout period to historical fluctuations; K is a coefficient, indicating the tolerance of the system to fluctuations;

[0115] The average connection time LS avg of the wireless sensor node in the historical period described above can be monitored and obtained by a connection duration monitoring sensor (such as a connection duration sensor, a connection state tracking module);

[0116] S311, the value of the coefficient K is determined by the value of the loss frequency SP of each wireless sensor node, when the value of the loss frequency SP of the corresponding wireless sensor node exceeds the average loss frequency SP avg , indicating that the connection fluctuation of the corresponding wireless sensor node is frequent, at this time the value of the coefficient K is set to 2; when the value of the loss frequency SP of the corresponding wireless sensor node does not exceed the average loss frequency SP avg , indicating that the connection fluctuation of the corresponding wireless sensor node is within the normal range, at this time the value of the coefficient K is set to 1.

[0117] S3 specific steps also include:

[0118] S32, according to S31, dynamically set an independent timeout period T for each wireless sensor node, start the maintenance mechanism for each wireless sensor node in the task path, and each wireless sensor node shares information with the directly communicating wireless sensor nodes having a physical adjacent relationship with it through broadcasting according to the independent timeout period T dynamically set by itself, when the corresponding wireless sensor node does not share information through broadcasting within the independent timeout period T dynamically set by itself, it indicates that the current wireless sensor node has a failure phenomenon, updates the current wireless sensor node to be in an offline state, and is marked as an offline physical node, which can reduce the potential impact of node failure on network transmission and ensure that the network can adjust the path in time when the node fails to avoid data transmission interruption;

[0119] When the corresponding wireless sensor node shares information through broadcasting within the independent timeout period T dynamically set by itself, the current wireless sensor node is updated to be in an online state and is not marked for processing, at this time the wireless sensor node not marked for processing will continue to transmit information according to the task path; the unmarked online node continues to execute according to the task path, maintaining the stability of the network.

[0120] When the corresponding wireless sensor node shares information through broadcasting within the independent timeout period T dynamically set by itself, the current wireless sensor node is updated to be in an online state and is not marked for processing, at this time the wireless sensor node not marked for processing will continue to transmit information according to the task path; the unmarked online node continues to execute according to the task path, maintaining the stability of the network.

[0121] S33, according to the replacement list established in S24, extract the replacement list of the offline physical node in S32, and based on the replacement list of the offline physical node, start the routing table updating task of each wireless sensor node in the task path.

[0122] Each wireless sensor node has an independent timeout period T in the network, which is dynamically adjusted according to the node loss frequency SP and its historical connection data, and the timeout period T is the maximum time for the node to maintain connection in the network. If feedback information from its adjacent nodes cannot be received within this time, it is considered that the node may have a failure or failure.

[0123] The node will periodically broadcast information to the directly communicating nodes having a physical adjacent relationship with it according to its timeout period T, and share information. Broadcast information: the node will send data or control information to the neighbor nodes directly connected to it to keep the real-time update of the network topology. Timeout without response: if the node does not complete the broadcasting task within its independently set timeout period T and does not receive the expected feedback information, it is considered that the node has a failure phenomenon.

[0124] Node failure marking: If a certain node does not update its status through broadcast information sharing job within the set timeout period, the node will be marked as an offline physical node, indicating that the node no longer participates in data transmission of the task path. Avoid data transmission interruption: Once it is found that a certain node fails, the system will adjust in time to avoid data transmission interruption caused by node failure.

[0125] In this embodiment, by dynamically analyzing the connection stability of each wireless sensor node and extracting its loss frequency SP, the application can assess the fluctuation of node connection quality in real time, and adjust the timeout duration T according to the stability of the node. This approach can more accurately identify and adapt to changes in the connection state of the node. For frequently disconnected nodes, the system will give them a shorter timeout duration, so that the network can quickly find the faulty node and avoid long waiting times, ensuring real-time updating of the task path and uninterrupted data transmission. The system dynamically sets the timeout duration T of the node according to the change of the loss frequency SP, and adjusts the value of the coefficient K to adapt to the fluctuations in node connection. For nodes with frequent connection fluctuations, the system quickly identifies disconnected nodes by setting a shorter timeout duration; for stable connections, the system reduces false positives by extending the timeout duration. This flexible timeout duration adjustment mechanism allows the network to dynamically adjust the timeout strategy based on the connection state of each node, avoiding the waste of resources and improving the accuracy and adaptability of network management. At the same time, the system extracts offline physical nodes based on the replacement list and updates the routing table in the task path, ensuring that the routing path can be automatically adjusted when a node fails. This automated path adjustment mechanism greatly reduces manual intervention, improves the response speed and efficiency of the network, and especially in complex wireless sensor networks, it can quickly adapt to node failures or changes in network topology, optimizing overall resource utilization. By establishing a dynamic timeout detection and node state update mechanism, the application improves the fault tolerance of the wireless sensor network. When a node fails in the network, the system can quickly find and replace the faulty node, avoiding the risk of the entire network being paralyzed due to a single node failure. This self-recovery capability is crucial for maintaining the stability and reliability of the wireless sensor network, especially in real-time data transmission applications, which can effectively improve the reliability and timeliness of data transmission. By accurately setting the timeout duration of each node and determining the online status of the node based on real-time data, the system can reduce unnecessary node waiting and information sharing, thereby saving network bandwidth and energy consumption. In wireless sensor networks, especially in energy-constrained devices, reducing redundant operations can significantly improve network energy efficiency and sustainability. By dynamically setting the timeout duration and marking real-time disconnected nodes, the system can effectively schedule resources, avoid over-reliance on any single node, and enhance the scalability and load balancing capabilities of the system. In summary, by introducing dynamic timeout duration and node state update mechanism based on historical data analysis, the application effectively improves the stability, reliability, and adaptability of the wireless sensor network, reduces redundant information sharing and resource waste, and quickly recovers the network path when a node fails, ensuring efficient and stable operation of the network, meeting the high requirements for real-time performance and fault tolerance in complex wireless sensor environments.

[0126] Embodiment 6

[0127] Please refer to Figures 1-4 Specifically, the S4 specific steps include:

[0128] S41, in the process of updating the routing table in the wireless sensor network, according to the offline physical node, the bandwidth condition of each to-be-replaced relay node in the to-be-replaced list of the offline physical node is determined, and the bandwidth difference ΔDc is obtained, specifically:

[0129] ΔDc g = DK g - DK;

[0130] In the formula, ΔDc g is the bandwidth difference between the gth to-be-replaced relay node in the to-be-replaced list of the offline physical node and the bandwidth of the offline physical node, DK g is the bandwidth of the gth to-be-replaced relay node in the to-be-replaced list of the offline physical node, and DK is the bandwidth of the offline physical node.

[0131] The S4 specific steps further include:

[0132] S42, analyze the replacement effect of the to-be-replaced relay node in the to-be-replaced list of the offline physical node, and construct a network security prediction model according to a deep learning technology, and after dimensionless processing, fit the path transmission distortion index Zs from the output end of the network security prediction model, the path transmission distortion index Zs is obtained by the following formula:

[0133]

[0134] In the formula, ΔCZ is the increase value of transmission delay, ΔTZ is the increase value of hop count, ΔDc is the bandwidth difference, λ is the penalty factor, indicating the influence degree of the failed node; SP is the loss frequency, ρ is the exponential of the penalty factor, indicating the nonlinear growth of the influence of node failure; η is the attenuation coefficient, LS max is the historical longest time connected by the corresponding to-be-replaced relay node, is the attenuation factor, indicating the influence of connection instability on path distortion, exp(*) is the exponential function with e as the base, I(ΔDc<0) is the indicator function, indicating that when ΔDc<0, I(ΔDc<0)=1, indicating that when ΔDc≥0, I(ΔDc≥0)=0; the indicator function specifically considers the concept of bandwidth availability.

[0135] The increase value ΔCZ of transmission delay refers to the difference between the transmission delay generated by selecting the corresponding to-be-replaced relay node and the transmission delay generated by the corresponding wireless sensor node, which can be monitored and obtained by a time delay monitoring sensor (such as a time delay increase value sensor and a network delay analysis module);

[0136] The increment of hop count \(\Delta Tz\) refers to the difference between the hop count generated by the corresponding to-be-replaced relay node and the hop count generated by the corresponding wireless sensor node, and the value thereof can be monitored and acquired by a hop count tracking sensor (for example, a routing tracking sensor or a hop count statistical module);

[0137] The bandwidth difference \(\Delta Dc\) refers to the difference between the bandwidth generated by the corresponding to-be-replaced relay node and the bandwidth generated by the corresponding wireless sensor node, and the value thereof can be monitored and acquired by a bandwidth monitoring sensor (for example, a bandwidth comparison sensor or a channel quality analysis module);

[0138] An initial model is constructed using a deep learning technology, and the initial model is trained and tested by a historical set, and the trained initial model is taken as a state recognition model, feature information in the state recognition model is acquired respectively, the state recognition model is trained and tested by the acquired feature information, and the trained state recognition model is taken as a network security prediction model in combination with the acquired to-be-replaced list.

[0139] The specific steps of S5 include:

[0140] S51, based on the manner of acquiring the path transmission distortion index Zs in S42, the path transmission distortion index Zs of each to-be-replaced relay node in the to-be-replaced list of the offline physical node relative to the offline physical node is acquired respectively, and a to-be-replaced relay node with the minimum path transmission distortion index Zs value is updated by a statistical algorithm, which is recorded as a replacement relay node, the replacement relay node replaces the corresponding offline physical node, and the task path resetting operation is realized.

[0141] In this embodiment, by analyzing the bandwidth conditions of each to-be-replaced relay node in the to-be-replaced list of the offline physical node, the system can accurately calculate the bandwidth difference and evaluate the network performance of each candidate node, avoiding direct replacement of nodes with mismatched or insufficient bandwidth, thereby ensuring the quality of data transmission and the stability of the network. The network security prediction model constructed using deep learning technology can combine dimensionless processing to comprehensively evaluate factors such as network transmission delay, hop count, and bandwidth difference, and then fit the path transmission distortion index Zs. This index can accurately quantify the impact of node replacement on network security and transmission quality, avoiding performance degradation caused by node replacement, thereby improving the security and data transmission quality of the entire network. By considering factors such as loss frequency, penalty factor, and attenuation coefficient, the path transmission distortion index Zs can comprehensively reflect the overall impact of node failure on the network, especially the nonlinear impact of failed nodes on the network. This approach effectively controls the transmission distortion caused by node failure or fluctuations in the network, ensuring efficient and stable operation of the network. The present application accurately calculates the path transmission distortion index Zs of the to-be-replaced relay node of the offline physical node through statistical algorithms, selects the replacement relay node with the smallest distortion index, and thus realizes an efficient and intelligent node replacement mechanism. By replacing the node with the optimal performance, the system can dynamically update the task path in the network, avoiding the influence of inefficient or failed nodes on task execution, and improving the response speed and processing efficiency of the entire system. By carefully evaluating each to-be-replaced node, the system can avoid frequent node replacement and reduce unnecessary network overhead. Each node replacement is based on the minimum path transmission distortion index Zs, ensuring that the selected node best meets the network requirements, thereby optimizing the allocation efficiency of network resources and improving the task execution efficiency of the wireless sensor network in complex environments.

[0142] Through this dynamic and intelligent node replacement mechanism, the wireless sensor network has strong self-healing ability and fault tolerance, and can quickly recover normal operation in the case of node failure or environmental fluctuations, ensuring the continuity and stability of data transmission. The node replacement mechanism in the application combined with the reset operation of the task path makes the network topology have stronger flexibility and adaptability in the dynamic environment. The wireless sensor network can flexibly adjust the routing and task path according to the actual situation to cope with different network states, enhancing the dynamic adaptability of the network. Through the combination of deep learning model, statistical algorithm and dimensionless processing method, the application greatly improves the intelligent degree of wireless sensor network management. The system can automatically analyze and evaluate the state of each node, and optimize node selection and task path according to the actual network condition without human intervention, thereby improving the network management efficiency. In summary, by accurately calculating the path transmission distortion index and combining factors such as bandwidth, delay and hop count, the application can realize efficient and intelligent node replacement and path optimization in the wireless sensor network, not only improving the stability, security and fault tolerance of the network, but also optimizing resource allocation and task execution efficiency, ensuring the stable operation and efficient management of the wireless sensor network in complex and dynamic environments.

[0143] Embodiment 7

[0144] Please refer to Figure 5 , specifically: a wireless sensor network node security state evaluation system, comprising a present situation determination subsystem, a relay analysis subsystem, a mechanism setting subsystem and a screening subsystem;

[0145] The present situation determination subsystem is used to monitor the network running information in the wireless sensor network in real time according to the sensor group, and to determine the task path according to the network topology in the wireless sensor network;

[0146] The relay analysis subsystem is used to obtain a historical set through real-time monitoring of the sensor group, and to seek multiple groups of virtual neighbors for each wireless sensor node according to the historical set, analyze and obtain a virtual proximity coefficient Xx, and if the virtual proximity coefficient Xx exceeds a set threshold value, a relay node to be replaced is preliminarily set;

[0147] The mechanism setting subsystem is used to set a timeout duration T, start a maintenance mechanism, update the online state of each wireless sensor node in the task path, and based on the online state, start updating the routing table operation in the wireless sensor network;

[0148] The screening subsystem is used to analyze the replacement effect of each relay node to be replaced according to the relay node to be replaced in the process of updating the routing table in the wireless sensor network, and to fit and obtain a path transmission distortion index Zs in combination with a trained network security prediction model, so as to select the final relay node, reset the task path and complete the secure communication between each wireless sensor node.

[0149] While embodiments of the application have been shown and described, it is to be understood that the application is not limited to the details of the embodiments described, since numerous changes, modifications, substitutions and variations can be made thereto without departing from the spirit and scope of the application as defined by the appended claims and their equivalents.

Claims

1. A method for security state evaluation of a wireless sensor network node, characterized in that: The method comprises the following steps: S1, real-time monitoring network running information in the wireless sensor network based on the sensor group, and determining a task path according to the network topology in the wireless sensor network; S2, real-time monitoring the task based on the sensor group in S1, obtaining a history set, and seeking multiple groups of virtual neighbors for each wireless sensor node according to the history set, analyzing and obtaining a virtual proximity coefficient Xx, if the virtual proximity coefficient Xx exceeds a set threshold, then preliminarily setting a to-be-replaced relay node; S3, setting a timeout duration T, starting a maintenance mechanism, updating the online state of each wireless sensor node in the task path, and starting to update the routing table in the wireless sensor network based on the online state; S4, in the process of updating the routing table in the wireless sensor network, analyzing the replacement effect of each to-be-replaced relay node according to the to-be-replaced relay node, and combining a trained network security prediction model to fit and obtain a path transmission distortion index Zs; S5, selecting a final relay node according to the numerical value of the path transmission distortion index Zs, resetting the task path, and completing the secure communication between each wireless sensor node; The specific steps of S2 include: S21, periodically storing the network running information obtained in S11 to generate a history set, the history set including the self-running state data of each wireless sensor node in the wireless sensor network and the running state data between each wireless sensor node in a historical period; S22, based on the history set acquired in S21, determining the wireless sensor nodes directly communicated by each wireless sensor node in the history period, to establish an independent neighbor group for each wireless sensor node, and analyzing the degree of virtual adjacency relationship between each wireless sensor node in each group of physical adjacency relationship in the wireless sensor network according to the establishment of the independent neighbor group by each wireless sensor node, to respectively calculate and acquire the virtual neighbor probability , distance prediction difference Jj, similarity value and path quality value Lc; S221、based on time series analysis method, and in combination with the history set, predict the probability of each wireless sensor node becoming a virtual adjacent relationship to obtain a virtual neighbor probability , specifically obtained according to the following formula: wherein, is the virtual neighbor probability, representing the probability that wireless sensor node i will become a virtual neighbor of wireless sensor node j after time t; i and j are both the number of wireless sensor nodes in the wireless sensor network; is the number of connections between wireless sensor node i and wireless sensor node j in the past time window; is the total number of connections of wireless sensor node i with other wireless sensor nodes in the past time; is a function representing the influence of the time difference on the virtual neighbor relationship; S222, predicting the distance between each wireless sensor node and the wireless sensor nodes in its corresponding independent neighbor group, and obtaining a distance prediction difference Jj based on the predicted distance between each wireless sensor node and the wireless sensor nodes in its corresponding independent neighbor group; S223、According to each wireless sensor node and its corresponding independent neighbor group, determine the similarity between each wireless sensor node in the independent neighbor group and the independent neighbor group to obtain a similarity value , Specifically, the following formula is obtained: wherein, is a similarity value, representing the similarity between wireless sensor nodes i by comparing the intersection and union of the independent neighbor set of wireless sensor node i and the independent neighbor set corresponding to the kth wireless sensor node within the independent neighbor set of wireless sensor node i; is the independent neighbor set of wireless sensor node i, is the independent neighbor set corresponding to the kth wireless sensor node within the independent neighbor set of wireless sensor node i, is and the number of common physical adjacent nodes, is and the total number of all physical adjacent nodes; S224, evaluating the path quality state between each wireless sensor node and the wireless sensor nodes in its independent neighbor group based on the multi-hop path existing in the task path, to obtain a path quality value Lc, specifically according to the following formula: wherein G is a selection number, is a delay duration of the kth wireless sensor node in the selected independent neighbor group, is a bandwidth of the kth wireless sensor node in the selected independent neighbor group, is a packet loss rate of the kth wireless sensor node in the selected independent neighbor group; k is a number of the wireless sensor node in the independent neighbor group. The specific steps of S2 further include: S23, based on the virtual neighbor probability acquired in S22 , distance prediction difference value Jj, similarity value and path quality value Lc, analyzing the degree of virtual adjacency relationship between each wireless sensor node in each group of physical adjacency relationship in the wireless sensor network to construct a virtual proximity coefficient Xx, which is acquired by the following formula: wherein, , , and are weights, the specific values of which are set by the user according to the situation; S24, if the virtual proximity coefficient Xx exceeds the set threshold, then preliminarily setting the corresponding wireless sensor node in the independent neighbor group as a to-be-replaced relay node, and establishing a to-be-replaced list through statistics, the to-be-replaced list including multiple groups of to-be-replaced relay nodes corresponding to the corresponding wireless sensor node.

2. The wireless sensor network node security state evaluation method according to claim 1, wherein: The specific steps of S1 include: S11, collecting network operation information in the wireless sensor network in real time in advance by using a sensor group, wherein the network operation information comprises self-operation state data of each wireless sensor node in the wireless sensor network and operation state data between each wireless sensor node, the self-operation state data of each wireless sensor node comprises total connection number of the wireless sensor node i with other wireless sensor nodes in the past time , independent neighbor group of the wireless sensor node i , delay length of the kth wireless sensor node in the independent neighbor group , bandwidth of the kth wireless sensor node in the independent neighbor group , packet loss rate of the kth wireless sensor node in the independent neighbor group , average connection time of the wireless sensor node in the historical period , and loss frequency SP of each wireless sensor node; the operation state data between each wireless sensor node comprises connection times Ls of each wireless sensor node with other wireless sensor nodes and independent neighbor group corresponding to the kth wireless sensor node in the independent neighbor group of the wireless sensor node i ; the network operation information is processed by a dimensionless method based on Z-score standardization; S12, each wireless sensor node in the wireless sensor network collects the network running information of the wireless sensor nodes directly communicating with it by listening to the wireless channel, to establish a physical adjacency relationship, and forms the topology structure of the wireless sensor network through multiple groups of physical adjacency relationships.

3. The wireless sensor network node security state evaluation method according to claim 2, wherein: The specific steps of S3 include: S31, dynamically analyze the stability of the connection of each wireless sensor node in the historical period based on the history set, to extract the loss frequency SP of each wireless sensor node, and dynamically set an independent timeout period T for each wireless sensor node, wherein the independent timeout period T corresponding to each wireless sensor node is dynamically set in the following manner: wherein, is the average connection time of the wireless sensor node over the historical period; is the standard deviation of the historical connection time of the wireless sensor node, is an adjustment factor; is a coefficient; S311, the coefficient K value is determined by the loss frequency SP value of each wireless sensor node, when the loss frequency SP value of the corresponding wireless sensor node exceeds the average loss frequency , it indicates that the connection fluctuation of the corresponding wireless sensor node is frequent, at this time the coefficient K value is set to 2; when the loss frequency SP value of the corresponding wireless sensor node does not exceed the average loss frequency , it indicates that the connection fluctuation of the corresponding wireless sensor node is in the normal range, at this time the coefficient K value is set to 1.

4. The wireless sensor network node security state evaluation method according to claim 3, characterized in that: S3 further includes the following specific steps: S32, according to the independent timeout period T dynamically set for each wireless sensor node in S31, start the maintenance mechanism for each wireless sensor node in the task path, and each wireless sensor node shares information with the directly communicating wireless sensor nodes having physical adjacency relationship with it through broadcast according to the independent timeout period T dynamically set for itself, when the corresponding wireless sensor node does not share information through broadcast within the independent timeout period T dynamically set for itself, it indicates that the current wireless sensor node has a failure phenomenon, the current wireless sensor node is updated to be in an offline state, and is marked as an offline physical node; when the corresponding wireless sensor node shares information through broadcast within the independent timeout period T dynamically set for itself, the current wireless sensor node is updated to be in an online state, and no marking processing is performed, at this time, the wireless sensor node not marked will continue to transmit information according to the task path; S33, according to the replacement list of the offline physical node extracted in S32, update the routing table of each wireless sensor node in the task path based on the replacement list of the offline physical node.

5. The wireless sensor network node security state evaluation method according to claim 4, characterized in that: S4 includes the following specific steps: S41, in the process of updating the routing table in the wireless sensor network, according to the offline physical node, determining the bandwidth condition of each to-be-replaced relay node in the to-be-replaced list of the offline physical node, and obtaining the bandwidth difference , specifically: wherein, is the difference between the bandwidth of the gth to-be-replaced relay node in the to-be-replaced list of the downline physical node and the bandwidth of the downline physical node, is the bandwidth of the gth to-be-replaced relay node in the to-be-replaced list of the downline physical node, is the bandwidth of the downline physical node.

6. The wireless sensor network node security state evaluation method according to claim 5, characterized in that: S4 further includes the following specific steps: S42, analyze the replacement effect of the replacement relay node in the replacement list of the offline physical node, and construct a network security prediction model according to the deep learning technology, fit the path transmission distortion index Zs from the output end of the network security prediction model after dimensionless processing, and the path transmission distortion index Zs is obtained by the following formula: wherein is an increase in transmission delay, is an increase in hop count, is a bandwidth difference, is a penalty factor; is a loss of frequency, is an exponent of the penalty factor; is a decay coefficient, is the historically longest time of the respective relay node connection to be replaced, is a decay factor, is an exponential function with base e, is an indicator function.

7. The wireless sensor network node security state evaluation method according to claim 6, characterized in that: S5 includes the following specific steps: S51, based on the method of obtaining the path transmission distortion index Zs in S42, obtain the path transmission distortion index Zs of each replacement relay node relative to the offline physical node in the replacement list of the offline physical node, and update the replacement relay node with the minimum path transmission distortion index Zs value through statistical algorithm, mark it as a replacement relay node, replace the corresponding offline physical node with the replacement relay node, and reset the task path.

8. A system for evaluating security state of a wireless sensor network node, which is used to implement the method for evaluating security state of a wireless sensor network node according to any one of claims 1 to 7, characterized in that: The system includes a current state determination subsystem, a relay analysis subsystem, a mechanism setting subsystem, and a screening subsystem. The status determining subsystem is configured to monitor network operation information in the wireless sensor network in real time according to the sensor groups, and determine a task path according to a network topology in the wireless sensor network; The relay analyzing subsystem is configured to acquire a history set through real-time monitoring of the sensor groups, and seek multiple sets of virtual neighbors for each wireless sensor node according to the history set, analyze a virtual proximity coefficient Xx, and preliminarily set a relay node to be replaced if the virtual proximity coefficient Xx exceeds a set threshold value; The mechanism setting subsystem is configured to set a timeout duration T, start a maintenance mechanism, update online states of the wireless sensor nodes in the task path, and start an operation of updating a routing table in the wireless sensor network based on the online states; The screening subsystem is configured to analyze replacement effects of the relay nodes to be replaced according to the relay nodes to be replaced in the operation of updating the routing table in the wireless sensor network, fit and acquire a path transmission distortion index Zs in combination with a trained network security prediction model, select a final relay node, reset the task path, and complete secure communication between the wireless sensor nodes.

Images