CN116546498B - Underwater wireless sensor network trust evaluation method based on variable membership function - Google Patents

Underwater wireless sensor network trust evaluation method based on variable membership function Download PDF

Info

Publication number
CN116546498B
CN116546498B CN202310627635.3A CN202310627635A CN116546498B CN 116546498 B CN116546498 B CN 116546498B CN 202310627635 A CN202310627635 A CN 202310627635A CN 116546498 B CN116546498 B CN 116546498B
Authority
CN
China
Prior art keywords
node
trust
data
energy
trust value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202310627635.3A
Other languages
Chinese (zh)
Other versions
CN116546498A (en
Inventor
叶方
王浩然
李一兵
孙骞
白能
张直
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Harbin Engineering University
Original Assignee
Harbin Engineering University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Harbin Engineering University filed Critical Harbin Engineering University
Priority to CN202310627635.3A priority Critical patent/CN116546498B/en
Publication of CN116546498A publication Critical patent/CN116546498A/en
Application granted granted Critical
Publication of CN116546498B publication Critical patent/CN116546498B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
    • H04W12/08—Access security
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
    • H04W12/30—Security of mobile devices; Security of mobile applications
    • H04W12/37—Managing security policies for mobile devices or for controlling mobile applications
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W84/00—Network topologies
    • H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Abstract

A trust evaluation method of an underwater wireless sensor network based on a variable membership function relates to the technical field of trust management of the underwater wireless sensor network, and aims at solving the problems that an existing trust model is difficult to give a differentiated trust evaluation strategy for specific attacks and scenes when facing a dynamic network topology, and nodes in an important area cannot be effectively protected. The method and the system define three types of safety factors in consideration of attack and environmental characteristics, and design a variable membership function fuzzy system to complete calculation of comprehensive trust values, wherein membership functions of a trust model can be regulated and controlled by the safety factors, and the membership functions are adaptively changed for specific attack modes and environmental scenes instead of being fixed. The model considers uncertainty among nodes, provides differentiated trust evaluation for different nodes, can improve the sensitivity degree to attack behaviors from important areas, shortens the malicious attack detection time, reduces the influence of attacks on a network, and can obtain higher detection rate and lower false detection rate.

Description

Underwater wireless sensor network trust evaluation method based on variable membership function
Technical Field
The invention relates to the technical field of underwater wireless sensor network trust management, in particular to an underwater wireless sensor network trust evaluation method based on a variable membership function.
Background
In recent years, underwater wireless sensor networks have been applied to the fields of pollution monitoring, oil and gas exploration, navigation control, tsunami warning, military and security, etc. The characteristics of disposable and low cost of the sensor nodes determine that most nodes are not subjected to tamper-resistant technology, and therefore the risk of being physically invaded and redeployed into compromised nodes and launching internal attacks. Internal attacks include selective forwarding attacks, data tampering attacks, DOS attacks, worm hole attacks, witches attacks, etc. The node for launching the selective forwarding attack can selectively discard the data packet, the node for launching the data tampering attack tries to send tampered false sensor information to the destination node, and the node for launching the DOS attack tries to occupy a channel through high-frequency packet sending and consume node energy. Attacks can reduce network packet delivery rates, network life, and negatively impact network performance. Because the compromised node already has the encryption information necessary for the legitimate node, the traditional cryptographic technique cannot prevent the compromised node from launching an internal attack if it exists in the network, so an effective mechanism is needed to identify the compromised node, and reduce the impact of the internal attack on the network.
Trust management mechanisms are considered to be an effective complement to traditional cryptographic mechanisms, which consist of a series of trust-based security techniques for protecting networks from attacks by compromised nodes. Since Marsh introduced trust research into the computer field, numerous trust models have been proposed in terms of distributed networks, pervasive computing, peer-to-peer computing, ad hoc networks, etc., to improve the security, reliability, and fairness of the system.
And calculating a trust value by using the fuzzy system based on a trust model of the fuzzy system, and quantifying the trust level of the evaluation node on the evaluated node. Fuzzy set theory in fuzzy systems has the characteristics of being suitable for describing and processing fuzzy concepts and objects, and by means of concepts of membership and linguistic variables, an effective method for quantitatively researching subjective trust can be provided. Although fuzzy set theory is able to model the uncertainty of a single user's understanding of semantic concepts, i.e., individual uncertainty, different users will have different understandings of the same semantic concept due to inter-individual uncertainty. Considering that the influence degree of node attack on the network is related to the node position and topology, in a real scene, the influence degree of different node launching attacks on the network is different, so that the nodes in an important area cannot be protected by using a fixed membership function in the blurring process.
Disclosure of Invention
The purpose of the invention is that: aiming at the problem that the existing trust model is difficult to give differentiated trust evaluation strategies aiming at specific attacks and scenes when facing dynamic network topology, and nodes in important areas cannot be effectively protected, the trust evaluation method of the underwater wireless sensor network based on the variable membership function is provided.
The technical scheme adopted by the invention for solving the technical problems is as follows:
the trust evaluation method of the underwater wireless sensor network based on the variable membership function comprises the following steps:
step 1: under a sliding time window mechanism, collecting communication trust evidence, data trust evidence and energy trust evidence;
step 2: based on the communication trust evidence, the data trust evidence and the energy trust evidence collected in the step 1, calculating a communication trust value of the evaluated node based on beta distribution, calculating a data trust value of the evaluated node based on cloud model, and calculating an energy trust value of the evaluated node based on Gaussian distribution algorithm;
step 3: acquiring communication safety factors, data safety factors and energy safety factors of the evaluated nodes, constructing a fuzzy system by taking the communication safety factors, the data safety factors and the energy safety factors of the evaluated nodes as membership function parameters, and then taking the communication trust values, the data trust values and the energy trust values of the evaluated nodes obtained in the step 2 as input quantities, and outputting comprehensive trust values through the fuzzy system;
step 4: acquiring a historical trust value of the evaluated node, adding a node lower than a threshold value in the final trust value into a blacklist according to the historical trust value of the evaluated node and the comprehensive trust value obtained in the step 3, and broadcasting a notification whole network.
Further, the specific steps of the step 1 are as follows:
first, a sliding time window mechanism [ delta t ] is established 1 ,Δt 2 ,...,Δt G ]Wherein G represents the total time window number during the running of the trust model, deltat represents the length of the time window, and in each time window, the evaluation node collects trust evidence of the evaluated node through monitoring and packet header information exchange, wherein the trust evidence comprises communication trust evidence, data trust evidence and energy trust evidence.
Further, the specific steps of the evaluating node collecting trust evidence of the evaluated node by monitoring and packet header information exchange are as follows:
collecting communication trust evidence: selecting the statistics times of success and failure of forwarding the data packet by the evaluated node in a time window as communication trust evidence, recording the monitored success times of forwarding the data packet by the evaluated node in the time window as s, and recording the failure times of forwarding the data packet as f;
collecting data trust evidence: in a time window, the evaluation node acquires data trust evidence of each node in the node set including the evaluated node by monitoring, and records the data trust evidence as { u } 1 ,u 2 ,...,u N N represents the number of collected data, N is an integer, and u represents the data collected by the sensor;
collecting energy trust evidence: and in a time window, the evaluation node acquires the energy consumption of each node in the node set including the evaluated node in a period of time by a monitoring mode as the energy trust evidence of the corresponding node, and the energy trust evidence is marked as e.
Further, in the step 2, the communication trust value of the node to be evaluated is calculated based on beta distribution, the data trust value of the node to be evaluated is calculated based on cloud model, and the energy trust value of the node to be evaluated is calculated based on Gaussian distribution algorithm; the specific steps of (a) are as follows:
calculating a communication trust value: selecting a method based on beta distribution to calculate a communication trust value F com Expressed as:
wherein Beta represents a Beta function;
calculating a data trust value: the evaluation node trusts the data of each node in the node set for evidence { u } 1 ,u 2 ,...,u N Using an inverse cloud algorithm, the cloud vector is converted to a cloud vector, which includes three parameters: expected ex, entropy en and super entropy he, and obtain estimated values of three parameters, the estimated values of three parameters are expressed as:
wherein the average valueSum of variances S 2 Expressed as:
wherein x is q Data trust evidence representing node q in the set of nodes, q=1, 2,..m-1;
and then, quantifying and evaluating the similarity degree between the nodes and the cloud vectors of any node in the node set by using a cloud similarity algorithm, and expressing the cloud similarity by using 2 cloud vector included angle cosine values, wherein the similarity of the two cloud vectors is as follows:
wherein,and->Representing two cloud vectors C j And C k And j and k represent cloud vectors C j And C k Converting a corresponding node before using a reverse cloud algorithm, wherein C j ＝(ex j ,en j ,he j )，C k ＝(ex k ,en k ,he k ) Relative trust RT of node j and node k data jk Expressed as:
data trust value F for node j data The average value of the relative trust of all nodes except node j including the evaluation node to node j is the data trust value F data Expressed as:
wherein S is the number of nodes in the node set, and node v is the nodes in the node set except for node j, v=1, 2,..s, v+.j;
calculating an energy trust value: energy trust value F energy Expressed as:
wherein z represents the node energy consumption value in a time window, d represents the integral sign, sigma represents the standard deviation of normal distribution, e local Representing the energy consumption of the evaluation node itself within a time window e evaluated Representing the energy consumption value of the node under evaluation within a time window.
Further, the specific steps of obtaining the communication security factor, the data security factor and the energy security factor of the evaluated node in the step 3 are as follows:
acquisition quiltEvaluating node communication security factors: let the communication range of node i be a sphere with r as radius, and the corresponding node set be denoted as C i ＝{c 1 ,c 2 ,…,c M-1 Defining a direct link in a node set of node i as an inter-node link reachable by one hop, defining an indirect link as an inter-node link unreachable by one hop but reachable by multiple hops through the node i, and representing the set of the direct links asl mn Representing the links between node m and node n, the total number of links is +.>The calculation results, wherein the node p is a node except the node i in the node set, p=1, 2..m-1, p+.i, and the communication security factor is expressed as:
wherein L represents the cardinality of the set L, M represents the number of nodes in the node set including the node i, and M is more than or equal to 1;
acquiring the data security factor of the evaluated node: defining a data security factor to quantify a threat level to a network when a node launches a data tampering attack, the data security factor being expressed as:
wherein d total Represents the maximum depth in the deployment range, d represents the current depth of the node under evaluation and 0.ltoreq.d.ltoreq.d total The depth is a vertical distance from a reference to the water bottom by taking the water surface as the reference;
acquiring an energy safety factor of the evaluated node: defining an energy security factor to quantify a threat level to a network when a node launches a DOS attack, the energy security factor being expressed as:
wherein, the integer N th1 And N th2 Is a threshold value determined according to specific scenes, satisfies N which is more than or equal to 0 th1 ＜N th2 。
Further, in the step 3, the communication security factor, the data security factor and the energy security factor of the evaluated node are used as membership function parameters to construct a fuzzy system, and then the communication trust value, the data trust value and the energy trust value of the evaluated node obtained in the step 2 are used as input quantities, and the specific steps of outputting the comprehensive trust value through the fuzzy system are as follows:
first, the communication trust value F com Data trust value F data And an energy trust value F energy As input variable and to communicate trust value F com Data trust value F data And an energy trust value F energy Respectively dividing into 3 fuzzy sets { Low (Low), medium (High) }, dividing the comprehensive trust value as an output variable into 6 fuzzy sets
{ extremely Low (VL), low (Low), lower (LL), higher (LH), high (High), extremely High (VH) } membership function selects a triangular membership function with a belief value argument of [0,1 ]]，U f ＝{u f }，u f Representing U f Defining the low in the fuzzy set as A l Middle is A m The height is A h Int represents the size of a membership function adjustment interval, int is more than or equal to 0 and less than or equal to 1, and membership functions of the fuzzy set are mu respectively Al (u f |SF,Int)、μ Am (u f |SF,Int)、μ Ah (u f SF, int) defined as follows:
SF is a communication trust value F com Data trust value F data Or energy trust value F energy ；
Then fuzzy reasoning is carried out by using a Mandarin fuzzy rule based on membership function of the fuzzy set, and finally fuzzy reasoning results are defuzzified by adopting a gravity center method to obtain an accurate output quantity T of the comprehensive trust value synthesized 。
Further, in the Mandarin fuzzy rule, an AND operator is filled with a min method AND an OR operator is filled with a max method in logic operation, the weight of all rules is 1, fuzzy sets under all rules are obtained by using min implication rules, AND output fuzzy sets of all rules are combined into a comprehensive trust value fuzzy set by using a max aggregation rule.
Further, the specific steps of the step 4 are as follows:
obtaining historical trust value { T ] h1 ,T h2 ,T h3 ,T h4 Then combine the integrated trust value T synthesized Obtaining the final trust value T final Final trust value T final Expressed as:
T final ＝w 0 *T synthesized +w 1 *T h1 +w 2 *T h2 +w 3 *T h3 +w 4 *T h4
wherein w is 0 、w 1 、w 2 、w 3 And w 4 To be the attenuation factor weight, w 0 ∈[0,1]，w 1 ∈[0,1]，w 2 ∈[0,1]，w 3 ∈[0,1]，w 4 ∈[0,1]，w 0 +w 1 +w 2 +w 3 +w 4 ＝1；
Updating the historical trust value to { T ] final ,T h1 ,T h2 -as a historical trust value;
setting a trust value threshold T th ，0≤T th Not more than 1, when the evaluation node detects T final ≤T th And adding the addresses of the nodes to be evaluated into a blacklist of the nodes and broadcasting the nodes in the whole network, and then taking measures to isolate the nodes from the network.
The beneficial effects of the invention are as follows:
the method and the system define three types of safety factors in consideration of attack and environmental characteristics, and design a variable membership function fuzzy system to complete calculation of comprehensive trust values, wherein membership functions of a trust model can be regulated and controlled by the safety factors, and the membership functions are adaptively changed for specific attack modes and environmental scenes instead of being fixed. The model considers uncertainty among nodes, provides differentiated trust evaluation for different nodes, can improve the sensitivity degree to attack behaviors from important areas, shortens the malicious attack detection time, reduces the influence of attacks on a network, and can obtain higher detection rate and lower false detection rate. Meanwhile, the trust model is more sensitive to the attack behavior of the nodes in the important area of the network, the recognition time is shorter, and the protection degree of the nodes in the important area is enhanced. In addition, the method and the device also solve the problems that the parameters of the trust model of the existing underwater wireless sensor network are fixed, uncertainty among nodes is ignored, and the adaptability and accuracy of the model environment are limited.
Drawings
FIG. 1 is a schematic diagram of a trust model process based on a variable membership function fuzzy system;
FIG. 2 is a schematic diagram of the operation of the trust model within a time window;
FIG. 3 is a schematic diagram of a process for computing a comprehensive trust value for a variable membership function fuzzy system.
Detailed Description
It should be noted in particular that, without conflict, the various embodiments disclosed herein may be combined with each other.
The first embodiment is as follows: referring to fig. 1, a specific description is given of a trust evaluation method for an underwater wireless sensor network based on a variable membership function according to the present embodiment, including the following steps:
step 1: under a sliding time window mechanism, collecting communication trust evidence, data trust evidence and energy trust evidence;
step 2: based on the communication trust evidence, the data trust evidence and the energy trust evidence collected in the step 1, calculating a communication trust value of the evaluated node based on beta distribution, calculating a data trust value of the evaluated node based on cloud model, and calculating an energy trust value of the evaluated node based on Gaussian distribution algorithm;
step 3: acquiring communication safety factors, data safety factors and energy safety factors of the evaluated nodes, constructing a fuzzy system by taking the communication safety factors, the data safety factors and the energy safety factors of the evaluated nodes as membership function parameters, and then taking the communication trust values, the data trust values and the energy trust values of the evaluated nodes obtained in the step 2 as input quantities, and outputting comprehensive trust values through the fuzzy system;
step 4: acquiring a historical trust value of the evaluated node, adding a node lower than a threshold value in the final trust value into a blacklist according to the historical trust value of the evaluated node and the comprehensive trust value obtained in the step 3, and broadcasting a notification whole network.
The present application also includes such structural features:
the step 3 comprises the following steps: definition of communication Security factor SF com Quantifying the threat level to the network when a node initiates a selective forwarding attack. Let the communication range of node i be a sphere with r as radius, and the corresponding neighbor set be denoted as C i ＝{c 1 ,c 2 ,…,c M-1 Defining a direct link in a neighbor node set of the node i as an inter-node link reachable by one hop, defining an indirect link as an inter-node link unreachable by one hop but reachable by multiple hops of the node i, and expressing the set of the direct links asThe total number of links can be defined by +.>Calculated, M is node i including itselfThe number M of neighbor nodes in the network is more than or equal to 1. The calculation formula of the communication security factor is as follows:
where |L| represents the cardinality of set L.
Further, a data security factor is calculated. Definition of data Security factor SF data The threat level of the node to the network when launching a data tamper attack is quantified. The calculation formula of the data security factor is as follows:
wherein d is total For the maximum depth in the node deployment range, d is the current depth of the evaluated node and d is more than or equal to 0 and less than or equal to d total The depth is the vertical distance from the reference to the water bottom based on the water surface.
Further, an energy safety factor is calculated. Definition of energy safety factor SF energy The threat level to the network when a node initiates a DOS attack is quantified. The energy safety factor is calculated as follows:
wherein the integer N th1 And N th2 Is a threshold value determined according to specific scenes, satisfies N which is more than or equal to 0 th1 ＜N th2 。
Further, a fuzzy logic method is adopted, and a variable membership function fuzzy system is used for calculating the comprehensive trust value. First three communication trust values F as input variables com Data trust value F data And an energy trust value F energy Is divided into 3 fuzzy sets { Low (Low), medium (High), high) }, and the integrated trust value as an output variable is divided into 6 fuzzy sets { extremely Low (VL), low (Low), low (LL), high (LH), high (High), extremely High (VH) },the membership function selects a triangular membership function. Taking a communication trust value as an example, a communication trust value domain U cf ＝[0,1]，u cf Representing U cf The element in (a) is denoted as U cf ＝{u cf }. Define the fuzzy set "low" asThe fuzzy set "middle" is->Fuzzy set "high" is->Int c The size of the interval is adjusted for the membership function, and Int is not less than 0 c As being regulated by communication safety factors, the membership functions of the fuzzy sets are respectively +.>The definition is as follows:
the membership function formula of fuzzy sets in the data trust value and energy trust value domains is similar to the membership function formula of fuzzy sets in the communication trust value domains, except that the communication security factor SF com Respectively by SF data And SF (sulfur hexafluoride) energy Instead, the membership function adjusts the interval Int c From Int d 、Int e Instead, 0.ltoreq.Int d ≤1、0≤Int e ≤1。
Further, the Mandarin fuzzy rule is used for fuzzy reasoning. A total of 27 fuzzy rules are expressed in the form of IF CT is CT1 and DT is DT1 and ET is ET1THEN OT is ST1, where CT1, DT1, ET1 and ST1 represent fuzzy linguistic variables over three inputs CT, DT, ET and one output ST, respectively. AND filling min for an AND operator AND max for an OR operator in logic operation, wherein the weights of all rules are 1, fuzzy sets under all rules are obtained by using min implication rules, AND output fuzzy sets of all rules are combined into a comprehensive trust value fuzzy set by using max aggregation rules. Finally, resolving the fuzzy by using a gravity center method to obtain an accurate output quantity T of the comprehensive trust value synthesized 。
The purpose of the application is to design an underwater wireless sensor network trust model, which mainly aims to solve the problems that the existing underwater wireless sensor network trust model is fixed in parameters, so that the adaptability and accuracy of the model are limited. According to the method, the acoustic communication openness characteristic is periodically utilized under a time window mechanism, trust evidence is collected, trust values are calculated, attack and environmental characteristics are considered on the basis, three types of safety factors are defined, a variable membership function fuzzy system is designed to complete calculation of the comprehensive trust values, and the historical trust values and attenuation factors are combined, so that final trust values are calculated.
Example 1
FIG. 1 is a schematic diagram of a trust model process based on a variable membership function fuzzy system.
Step 1: under the sliding time window mechanism, nodes gather communication, data, and energy trust evidence distributively.
Step 1.1: a sliding time window mechanism is established. FIG. 2 is a schematic diagram of the trust model of the present application during a time window. First, a sliding time window mechanism [ delta t ] is established 1 ,Δt 2 ,...,Δt G ]And G is the total time window number during the running of the trust model, and in each time window, the node utilizes the openness characteristic of acoustic communication to collect communication, data and energy trust evidence of the neighbor nodes in a distributed manner by monitoring and packet head information exchange modes. Trust certificateAnd according to the real-time updating in each time window, settling is carried out at the end of the time window, and the node trust value is calculated. The trust value comprises a communication trust value F com Data trust value F data And an energy trust value F energy From [0,1]The value in between means that a higher trust value means that the node under evaluation has a higher degree of trust at that observation angle. And calculating the comprehensive trust value through the fuzzy system, then combining the historical trust value to calculate the final trust value, finally clearing the trust evidence, and repeatedly executing the operation again in the next time window.
Step 1.2: communication trust evidence is collected. And aiming at the attack mode of selectively discarding the data packet by the selective forwarding attack node, collecting communication trust evidence related to the forwarding of the data packet and calculating a communication trust value. And selecting the success and failure statistics times of forwarding the data packet by the node in a period of time as communication trust evidence, recording the monitored success times of forwarding the data packet by the evaluated node in a time window as s, and recording the failure times of forwarding the data packet as f.
Step 1.3: data trust evidence is collected. And aiming at an attack mode of the data tampering attack node tampering sensor data, collecting data trust evidence related to the sensor data and calculating a data trust value. The data collected by the sensor node is selected as the data trust evidence of the node itself and recorded as { x } 1 ,x 2 ,...,x N And, where the integer N represents the number of recently collected data.
Step 1.4: collecting energy trust evidence. And aiming at the attack mode of the DOS attack node for transmitting the data packet at high frequency, collecting the energy trust evidence related to the energy consumption and calculating the energy trust value. And selecting the energy consumption of the sensor node in a period of time as energy trust evidence, and marking as e.
Step 2: based on the communication, data and energy trust evidence collected in the step 1, communication, data and energy trust values of the evaluated nodes are calculated by using a beta distribution-based algorithm, a cloud model-based algorithm and a Gaussian distribution-based algorithm.
Step 2.1: a communication trust value is calculated. Selecting a method based on beta distribution to calculate a communication trust value F com 。
Beta(x,y)＝(Γ(x+y)/Γ(x)Γ(y))ε x-1 (1-ε) y-1 (7)
Wherein 0 < ε < 1, s > 0, f > 0, and x represents the number of honest behaviors, y represents the number of malicious behaviors, and the two variables are recorded as s and f in the trust model. In the algorithm, s represents the number of times that the evaluated node successfully forwards the data packet, and is defined as the number of times that the evaluating node monitors the action of forwarding the data packet within a period of time after monitoring the ACK message sent by the evaluated node. f represents the number of times that the evaluated node fails to forward the data packet, and is defined as the number of times that the evaluating node does not monitor the action of forwarding the data packet within a period of time after monitoring the ACK message sent by the evaluated node. The probability distribution can be represented by a beta distribution taking s and F as parameters, and the communication trust value F com Then it is the statistical expectation of the probability density function.
Step 2.2: a data trust value is calculated. Selecting a group trust algorithm based on a cloud model to calculate a data trust value F data . First, the conversion from quantitative values to qualitative concepts is achieved using the inverse cloud algorithm. Each node uses N data stored locally as cloud droplets { x } 1 ,x 2 ,...,x N Local trust Yun Xiangliang is computed distributively using the inverse cloud algorithm, the cloud vector comprising three parameters: ex, entropy en and super entropy he are desired.
Calculating estimated values of three parameters:
wherein the mean valueSum of variances S 2 The calculation is as follows
Second, the degree of similarity between 2 clouds was quantified using a cloud similarity algorithm. Using cosine values of 2 cloud vector included angles to represent cloud similarity: if two clouds C j And C k Digital feature component vector of (a)And->The similarity of the two clouds is:
the relative trust is then used to describe the relative degree of trust of the two nodes in the same period of data based on cloud similarity. Given a cloud model of the same period of two adjacent nodes j and k as C j And C k Wherein C j ＝(ex j ,en j ,he j )，C k ＝(ex k ,en k ,he k ) RT of data of node j and node k in the period jk The relative trust is shown in formulas (2-6):
and finally, calculating the data trust value. J data trust value F for node data Refers to the average of the relative trust of all neighbor nodes, except node j, that the evaluating node will include itself, to node j.
Step 2.3: an energy trust value is calculated. Selecting an algorithm based on a gaussian distribution to calculate an energy trust value F energy . UsingThe energy trust value is calculated by a normal distribution probability density function, and the probability density function of energy consumption can be calculated by the following formula:
wherein z is a node energy consumption value in a time window, sigma is a standard deviation of normal distribution, and the value can be specifically set according to specific scenes, e local The energy consumption of the evaluation node in a time window is defined as the energy trust value:
wherein e evaluated Is the energy consumption value of the evaluated node within a time window.
Step 3: and (3) calculating communication, data and energy safety factors of the evaluated nodes as membership function parameters, taking the trust value obtained in the step (2) as input quantity, and outputting the comprehensive trust value through a fuzzy system. Fig. 3 is a schematic diagram of a process for calculating an integrated trust value by using the variable membership function fuzzy system.
Step 3.1: a communication security factor is calculated. A communication security factor is defined to quantify the level of threat to the network when a node initiates a selective forwarding attack. Let the communication range of node i be a sphere with r as radius, and the corresponding neighbor set be denoted as C i ＝{c 1 ,c 2 ,…,c M-1 Defining a direct link in a neighbor node set of the node i as an inter-node link reachable by one hop, defining an indirect link as an inter-node link unreachable by one hop but reachable by multiple hops of the node i, and expressing the set of the direct links asThe total number of links can be defined by +.>And (5) calculating to obtain the product. Calculation formula of communication safety factorThe following are provided:
where |L| represents the cardinality of set L.
Step 3.2: a data security factor is calculated. A data security factor is defined to quantify the level of threat to the network when a node is launching a data tamper attack. The calculation formula of the data security factor is as follows:
wherein d is total For the maximum depth in the deployment range, d is the current depth of the evaluated node and d is more than or equal to 0 and less than or equal to d total 。
Step 3.3: an energy safety factor is calculated. An energy security factor is defined to quantify the level of threat to the network when a node initiates a DOS attack. The energy safety factor is calculated as follows:
wherein the integer N th1 And N th2 Is a threshold value determined according to specific scenes, satisfies N which is more than or equal to 0 th1 ＜N th2 。
Step 3.4: and calculating the comprehensive trust value by using a fuzzy logic method and a variable membership function fuzzy system. First three communication trust values F as input variables com Data trust value F data And an energy trust value F energy Is divided into 3 fuzzy sets { Low (Low), medium (High) }, and the comprehensive trust value as an output variable is divided into 6 fuzzy sets { extremely Low (VL), low (Low), low (LL), high (LH), high (High), extremely High (VH) }, and the membership function selects a triangular membership function. Taking a communication trust value as an example, a communication trust value domain U cf ＝[0,1]，u cf Representing U cf The element in (a) is denoted as U cf ＝{u cf }. Define the fuzzy set "low" asThe fuzzy set "middle" is->Fuzzy set "high" is->Int c The size of the interval is adjusted for the membership function, and Int is not less than 0 c As being regulated by communication safety factors, the membership functions of the fuzzy sets are respectively +.>The definition is as follows:
/>
the membership function formula of fuzzy sets in the data trust value and energy trust value domains is similar to the membership function formula of fuzzy sets in the communication trust value domains, except that the communication security factor SF com Respectively by SF data And SF (sulfur hexafluoride) energy Instead, the membership function adjusts the interval Int c From Int d 、Int e Instead, 0.ltoreq.Int d ≤1、0≤Int e ≤1。
And then performing fuzzy reasoning by using a Mandarin fuzzy rule. A total of 27 fuzzy rules are expressed in the form of IF CTisCT1 and DTisDT1 and ETisET1THENOTIS ST1, wherein CT1, DT1, ET1 and ST1 are substituted respectivelyFuzzy linguistic variables over three inputs CT, DT, ET and one output ST are shown. AND filling min for an AND operator AND max for an OR operator in logic operation, wherein the weights of all rules are 1, fuzzy sets under all rules are obtained by using min implication rules, AND output fuzzy sets of all rules are combined into a comprehensive trust value fuzzy set by using max aggregation rules. Finally, resolving the fuzzy by using a gravity center method to obtain an accurate output quantity T of the comprehensive trust value synthesized 。
Step 4: and (3) calculating a final trust value by combining the historical trust value and the comprehensive trust value obtained in the step (3), adding a blacklist to the node with the final trust value lower than the threshold value, and broadcasting a notification to the whole network.
Step 4.1: and calculating a final trust value. Taking into account the integrated trust value T at the same time when calculating the final trust value synthesized And 4 historical trust values { T h1 ,T h2 ,T h3 ,T h4 Final trust value T for settlement of current time window final The method comprises the following steps:
T final ＝w 0 *T syn thesized+w 1 *T h1 +w 2 *T h2 +w 3 *T h3 +w 4 *T h4 (22)
wherein w is 0 、w 1 、w 2 、w 3 And w 4 To be the attenuation factor weight, w 0 ∈[0,1]，w 1 ∈[0,1]，w 2 ∈[0,1]，w3∈[0,1]，w 4 ∈[0,1]，w 0 +w 1 +w 2 +w 3 +w 4 ＝1。
Calculating to obtain a final trust value, and updating the historical trust value to { T } final ,T h1 ,T h2 And (3) reserved for use in the next final trust value calculation.
Step 4.2: setting a trust value threshold T th ，0≤T th Not more than 1, when the evaluation node detects T final ≤T th And adding the addresses of the nodes to be evaluated into a blacklist of the nodes and broadcasting the nodes in the whole network, and then taking measures to isolate the nodes from the network.
Example two
Table 1 is simulation parameter settings including deployment range, node communication range, mobile model, simulation time, etc.; table 2 shows the statistical malicious node detection rate and false detection rate after 100 random seed experiments. The malicious node detection rate is defined as the ratio of the number of detected malicious nodes to the total number of malicious nodes, and the false detection rate is defined as the ratio of the number of legal nodes detected as malicious nodes to the total number of legal nodes.
TABLE 1
TABLE 2
Malicious node detection rate 97.43％
False detection rate 0.45％
In a second embodiment of the present application, using an NS-3 network emulation simulator, using an Aqua-Sim-NG module, the transmitted packet event for each node follows a poisson distribution. By adopting the trust model in the first embodiment of the application, the random seed simulation is modified for 100 times, and the malicious node detection rate 97.43% and the false detection rate 0.45% are obtained, which shows that the method provided by the invention has higher reliability.
The application relates to the technical field of underwater wireless sensor network trust management, and provides a trust model based on a variable membership function fuzzy system. The membership function of the trust model may be governed by a security factor, adaptively changing for a particular attack pattern and environmental scenario, rather than being fixed. The model considers uncertainty among nodes, provides differentiated trust evaluation for different nodes, can improve the sensitivity degree to attack behaviors from important areas, shortens malicious attack detection time, reduces the influence of attacks on a network, and can obtain higher detection rate and lower false detection rate. The trust model of the first embodiment of the application has strong applicability, high robustness and better interpretability, overcomes the defects of the prior method, can adaptively adjust model parameters, makes differentiated trust evaluation, and improves the protection degree of important regional nodes.
It should be noted that the detailed description is merely for explaining and describing the technical solution of the present invention, and the scope of protection of the claims should not be limited thereto. All changes which come within the meaning and range of equivalency of the claims and the specification are to be embraced within their scope.

Claims (3)

1. The underwater wireless sensor network trust evaluation method based on the variable membership function is characterized by comprising the following steps of:
step 1: under a sliding time window mechanism, collecting communication trust evidence, data trust evidence and energy trust evidence;
step 2: based on the communication trust evidence, the data trust evidence and the energy trust evidence collected in the step 1, calculating a communication trust value of the evaluated node based on beta distribution, calculating a data trust value of the evaluated node based on a cloud model, and calculating an energy trust value of the evaluated node by using a Gaussian distribution-based algorithm;
step 3: acquiring communication safety factors, data safety factors and energy safety factors of the evaluated nodes, constructing a fuzzy system by taking the communication safety factors, the data safety factors and the energy safety factors of the evaluated nodes as membership function parameters, and then taking the communication trust values, the data trust values and the energy trust values of the evaluated nodes obtained in the step 2 as input quantities, and outputting comprehensive trust values through the fuzzy system;
step 4: acquiring a historical trust value of the evaluated node, adding a node lower than a threshold value in the final trust value into a blacklist according to the historical trust value of the evaluated node and the comprehensive trust value obtained in the step 3, and broadcasting a notification whole network;
the specific steps of the step 1 are as follows:
first, a sliding time window mechanism [ delta t ] is established 1 ,Δt 2 ,...,Δt G ]Wherein G represents the total time window number during the running of the trust model, deltat represents the length of the time window, and in each time window, the evaluation node collects trust evidence of the evaluated node in a monitoring and packet header information exchange mode, wherein the trust evidence comprises communication trust evidence, data trust evidence and energy trust evidence;
the specific steps of the evaluating node collecting trust evidence of the evaluated node by monitoring and packet header information exchange are as follows:
collecting communication trust evidence: selecting the statistics times of success and failure of forwarding the data packet by the evaluated node in a time window as communication trust evidence, recording the monitored success times of forwarding the data packet by the evaluated node in the time window as s, and recording the failure times of forwarding the data packet as f;
collecting data trust evidence: in a time window, the evaluation node acquires data trust evidence of each node in the node set including the evaluated node by monitoring, and records the data trust evidence as { u } 1 ,u 2 ,...,u N N represents the number of collected data, N is an integer, and u represents the data collected by the sensor;
collecting energy trust evidence: in a time window, the evaluation node acquires the energy consumption of each node in a node set including the evaluated node in a period of time in a monitoring mode as energy trust evidence of the corresponding node, and the energy trust evidence is marked as e;
in the step 2, the specific steps of calculating the communication trust value of the node to be evaluated based on the beta distribution, calculating the data trust value of the node to be evaluated based on the cloud model, and calculating the energy trust value of the node to be evaluated by using the algorithm based on the Gaussian distribution are as follows:
calculating a communication trust value: selecting a method based on beta distribution to calculate a communication trust value F com Expressed as:
wherein Beta represents a Beta function;
calculating a data trust value: the evaluation node trusts the data of each node in the node set for evidence { u } 1 ,u 2 ,...,u N Using an inverse cloud algorithm, the cloud vector is converted to a cloud vector, which includes three parameters: expected ex, entropy en and super entropy he, and obtain estimated values of three parameters, the estimated values of three parameters are expressed as:
wherein the average valueSum of variances S 2 Expressed as:
wherein x is q Data trust evidence representing node q in the set of nodes, q=1, 2,..m-1;
and then, quantifying and evaluating the similarity degree between the nodes and the cloud vectors of any node in the node set by using a cloud similarity algorithm, and expressing the cloud similarity by using 2 cloud vector included angle cosine values, wherein the similarity of the two cloud vectors is as follows:
wherein,and->Representing two cloud vectors C j And C k And j and k represent cloud vectors C j And C k Converting a corresponding node before using a reverse cloud algorithm, wherein C j ＝(ex j ,en j ,he j )，C k ＝(ex k ,en k ,he k ) Relative trust RT of node j and node k data jk Expressed as:
data trust value F for node j data The average value of the relative trust of all nodes except node j including the evaluation node to node j is the data trust value F data Expressed as:
wherein S is the number of nodes in the node set, and node v is the nodes in the node set except for node j, v=1, 2,..s, v+.j;
calculating an energy trust value: energy trust value F energy Expressed as:
wherein z represents one timeNode energy consumption value in interval window, d represents integral sign, sigma represents standard deviation of normal distribution, e local Representing the energy consumption of the evaluation node itself within a time window e evaluated Representing the energy consumption value of the evaluated node in a time window;
the specific steps of acquiring the communication safety factor, the data safety factor and the energy safety factor of the evaluated node in the step 3 are as follows:
acquiring the communication security factor of the evaluated node: let the communication range of node i be a sphere with r as radius, and the corresponding node set be denoted as C i ＝{c 1 ,c 2 ,…,c M-1 Defining a direct link in a node set of node i as an inter-node link reachable by one hop, defining an indirect link as an inter-node link unreachable by one hop but reachable by multiple hops through the node i, and representing the set of the direct links asl mn Representing the links between node m and node n, the total number of links is +.>The calculation results, wherein the node p is a node except the node i in the node set, p=1, 2..m-1, p+.i, and the communication security factor is expressed as:
wherein L represents the cardinality of the set L, M represents the number of nodes in the node set including the node i, and M is more than or equal to 1;
acquiring the data security factor of the evaluated node: defining a data security factor to quantify a threat level to a network when a node launches a data tampering attack, the data security factor being expressed as:
wherein d total Represents the maximum depth in the deployment range, d represents the current depth of the node under evaluation and 0.ltoreq.d.ltoreq.d total The depth is a vertical distance from a reference to a node by taking the water surface as the reference;
acquiring an energy safety factor of the evaluated node: defining an energy security factor to quantify a threat level to a network when a node launches a DOS attack, the energy security factor being expressed as:
wherein, the integer N th1 And N th2 Is a threshold value determined according to specific scenes, satisfies N which is more than or equal to 0 th1 ＜N th2 ；
In the step 3, the communication security factor, the data security factor and the energy security factor of the evaluated node are used as membership function parameters to construct a fuzzy system, and then the communication trust value, the data trust value and the energy trust value of the evaluated node obtained in the step 2 are used as input quantities, and the specific steps of outputting the comprehensive trust value through the fuzzy system are as follows:
first, the communication trust value F com Data trust value F data And an energy trust value F energy As input variable and to communicate trust value F com Data trust value F data And an energy trust value F energy Respectively dividing into 3 fuzzy sets { Low (Low), medium (High) }, dividing the comprehensive trust value as an output variable into 6 fuzzy sets
{ extremely Low (VL), low (Low), lower (LL), higher (LH), high (High), extremely High (VH) } membership function selects a triangular membership function with a belief value argument of [0,1 ]]，U f ＝{u f }，u f Representing U f Defining the low in the fuzzy set as A l Middle is A m The height is A h Int represents the size of a membership function adjustment interval, int is more than or equal to 0 and less than or equal to 1, and membership functions of the fuzzy set are mu respectively Al (u f |SF,Int)、μ Am (u f |SF,Int)、μ Ah (u f SF, int) defined as follows:
SF is a communication trust value F com Data trust value F data Or energy trust value F energy ；
Then fuzzy reasoning is carried out by using a Mandarin fuzzy rule based on membership function of the fuzzy set, and finally fuzzy reasoning results are defuzzified by adopting a gravity center method to obtain an accurate output quantity T of the comprehensive trust value synthesized 。
2. The underwater wireless sensor network trust evaluation method based on the variable membership function according to claim 1, wherein in the Mandarin fuzzy rules, an AND operator is filled with a min method AND an OR operator is filled with a max method in logic operation, weights of all rules are 1, fuzzy sets under all rules are obtained by using min implication rules, AND output fuzzy sets of all rules are combined into a comprehensive trust value fuzzy set by using a max aggregation rule.
3. The method for evaluating the trust of the underwater wireless sensor network based on the variable membership function according to claim 2, wherein the specific steps of the step 4 are as follows:
obtaining historical trust value { T ] h1 ,T h2 ,T h3 ,T h4 Then combine the integrated trust value T synthesized Obtaining the final trust value T final Final trust value T final Expressed as:
T final ＝w 0 *T synthesized +w 1 *T h1 +w 2 *T h2 +w 3 *T h3 +w 4 *T h4
wherein w is 0 、w 1 、w2 、 w 3 And w 4 To be the attenuation factor weight, w 0 ∈[0,1]，w 1 ∈[0,1]，w 2 ∈[0,1]，w 3 ∈[0,1]，w 4 ∈[0,1]，w 0 +w 1 +w 2 +w 3 +w 4 ＝1；
Updating the historical trust value to { T ] final ,T h1 ,T h2 -as a historical trust value;
setting a trust value threshold T th ，0≤T th Not more than 1, when the evaluation node detects T final ≤T th And adding the addresses of the nodes to be evaluated into a blacklist of the nodes and broadcasting the nodes in the whole network, and then taking measures to isolate the nodes from the network.
CN202310627635.3A 2023-05-30 2023-05-30 Underwater wireless sensor network trust evaluation method based on variable membership function Active CN116546498B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310627635.3A CN116546498B (en) 2023-05-30 2023-05-30 Underwater wireless sensor network trust evaluation method based on variable membership function

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310627635.3A CN116546498B (en) 2023-05-30 2023-05-30 Underwater wireless sensor network trust evaluation method based on variable membership function

Publications (2)

Publication Number Publication Date
CN116546498A CN116546498A (en) 2023-08-04
CN116546498B true CN116546498B (en) 2024-01-26

Family

ID=87447027

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310627635.3A Active CN116546498B (en) 2023-05-30 2023-05-30 Underwater wireless sensor network trust evaluation method based on variable membership function

Country Status (1)

Country Link
CN (1) CN116546498B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101765231A (en) * 2009-12-30 2010-06-30 北京航空航天大学 Wireless sensor network trust evaluating method based on fuzzy logic
CN101835158A (en) * 2010-04-12 2010-09-15 北京航空航天大学 Sensor network trust evaluation method based on node behaviors and D-S evidence theory
CN106251047A (en) * 2016-07-20 2016-12-21 国网浙江省电力公司电力科学研究院 Consider the relay protection change power fuzzy synthesis method for evaluating state that degree of membership is variable
CN109246155A (en) * 2018-12-07 2019-01-18 重庆邮电大学 A method of attack is trusted in the wireless sensor network defence based on trust management
CN116015769A (en) * 2022-12-12 2023-04-25 上海交通大学 Zero trust system main body trust degree dynamic evaluation system based on fuzzy reasoning
CN116017463A (en) * 2022-12-26 2023-04-25 通化师范学院 Wireless sensor network malicious node identification method based on dynamic trust mechanism

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101765231A (en) * 2009-12-30 2010-06-30 北京航空航天大学 Wireless sensor network trust evaluating method based on fuzzy logic
CN101835158A (en) * 2010-04-12 2010-09-15 北京航空航天大学 Sensor network trust evaluation method based on node behaviors and D-S evidence theory
CN106251047A (en) * 2016-07-20 2016-12-21 国网浙江省电力公司电力科学研究院 Consider the relay protection change power fuzzy synthesis method for evaluating state that degree of membership is variable
CN109246155A (en) * 2018-12-07 2019-01-18 重庆邮电大学 A method of attack is trusted in the wireless sensor network defence based on trust management
CN116015769A (en) * 2022-12-12 2023-04-25 上海交通大学 Zero trust system main body trust degree dynamic evaluation system based on fuzzy reasoning
CN116017463A (en) * 2022-12-26 2023-04-25 通化师范学院 Wireless sensor network malicious node identification method based on dynamic trust mechanism

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
The Method of Fuzzy Comprehensive Evaluation on Radar Jamming Effect Based on Variable Trapezoidal Membership Function;Bing Xiao 等;《 2020 9th Asia-Pacific Conference on Antennas and Propagation (APCAP)》;正文第1-2页 *
基于模糊隶属度函数的传感器失效管理方法;倪建军 等;《计算机工程》;第33卷(第20期);正文第55-57页 *

Also Published As

Publication number Publication date
CN116546498A (en) 2023-08-04

Similar Documents

Publication Publication Date Title
Eziama et al. Malicious node detection in vehicular ad-hoc network using machine learning and deep learning
Devi et al. Implementation of intrusion detection system using adaptive neuro-fuzzy inference system for 5G wireless communication network
Alrehan et al. Machine learning techniques to detect DDoS attacks on VANET system: A survey
Zhang et al. AATMS: An anti-attack trust management scheme in VANET
Sherazi et al. DDoS attack detection: A key enabler for sustainable communication in internet of vehicles
Alheeti et al. On the detection of grey hole and rushing attacks in self-driving vehicular networks
Narayanadoss et al. Crossfire attack detection using deep learning in software defined its networks
Rani et al. Supervised machine learning based network intrusion detection system for Internet of Things
Gu et al. k-Nearest Neighbours classification based Sybil attack detection in Vehicular networks
Wei et al. Virus-traffic coupled dynamic model for virus propagation in vehicle-to-vehicle communication networks
Cabelin et al. SVM-based detection of false data injection in intelligent transportation system
Laptiev et al. Dynamic Model of Cyber Defense Diagnostics of Information Systems With The Use of Fuzzy Technologies
Liang et al. GaDQN-IDS: a novel self-adaptive IDS for VANETs based on bayesian game theory and deep reinforcement learning
Wang et al. Game-theoretic actor–critic-based intrusion response scheme (gtac-irs) for wireless sdn-based iot networks
Sedar et al. Reinforcement learning based misbehavior detection in vehicular networks
CN116546498B (en) Underwater wireless sensor network trust evaluation method based on variable membership function
Xie et al. UWPEE: Using UAV and wavelet packet energy entropy to predict traffic-based attacks under limited communication, computing and caching for 6G wireless systems
Wei et al. Classification of channel access attacks in wireless networks: A deep learning approach
Garg et al. An attack tree based comprehensive framework for the risk and security assessment of VANET using the concepts of game theory and fuzzy logic
Shams et al. Flow-based intrusion detection system in vehicular ad hoc network using context-aware feature extraction
Guo et al. A flow based detection mechanism against flooding attacks in mobile ad hoc networks
CN115802358A (en) Multi-step DDoS prediction poisoning attack based on reinforcement learning and defense method thereof
CN116232707A (en) APT attack defense method based on deep reinforcement learning
Gebremariam et al. Secure Intrusion Detection System for Hierarchically Distributed Wireless Sensor Networks
Almalki et al. Prospectus: An online polymorphic attack detection model for intelligent transportation systems

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant