Certificate revocation method based on signal game theory in Internet of Vehicles

Abstract

The invention belongs to the field of certificate management in the Internet of Vehicles, and particularly relates to a certificate revocation method based on a signal game theory in the Internet of Vehicles. Comprising the following steps: constructing an Internet of Vehicles system, and performing initial identity authentication on each entity in the system, including a trusted authority TA, a certificate management server CMS, a roadside unit RSU and a vehicle; establishing a static certificate revocation game between the CHV and the member vehicles, and solving Bayesian Nash equilibrium; expanding a static certificate revocation game into a multi-stage dynamic certificate revocation game, and solving refined Bayesian equilibrium; the CHV makes a decision of revoking the vehicle certificate according to a certificate revocation algorithm; the CMS realizes real revocation of the malicious vehicle certificate in the whole Internet of Vehicles system by distributing CRL and revocation information; according to the invention, malicious vehicles in the Internet of Vehicles system can be rapidly and accurately detected, and the global revocation of the certificates of the malicious vehicles can be realized, so that the safety and reliability of the Internet of Vehicles system are improved.

Description

Technical Field

[0001] The present invention belongs to the field of certificate management in an Internet of Vehicles, and in particular relates to a certificate revocation method based on signaling game theory in an Internet of Vehicles. Background Art

[0002] The Internet of Vehicles (IoV) is an open mobile self-organizing network consisting of vehicles, roadside infrastructure and pedestrians. With the increase in the number of intelligent connected vehicles, it is foreseeable that the scale of the IoV will continue to expand in the future, and the corresponding IoV security issues will become more prominent. The vehicle public key infrastructure is an effective solution to ensure the security of the IoV system. In this system, the certificate management agency issues certificates to vehicles to realize identity authentication in the IoV. Before establishing communication between vehicles, the certificate needs to be used to verify the legitimacy of the identity. However, although the above security strategy based on traditional cryptographic technology can effectively defend against external attacks, it cannot defend against internal attacks, that is, legitimate vehicles with certificates in the IoV may still be invaded or destroyed, and then attack the network, such as sending false information or conducting DoS attacks.

[0003] To solve this problem, a vehicle certificate revocation method is needed that can detect the attack behavior of the vehicle and revoke the certificate so that it cannot communicate with other vehicles. The current certificate revocation methods are divided into single-vehicle accusation-based methods and multi-vehicle voting-based methods.

[0004] In the single-vehicle accusation-based approach, all vehicles in the network monitor neighboring vehicles and accuse them when they observe attacking behavior. In most cases, only one accusation is needed to initiate the certificate revocation process, which reduces the time required to revoke the certificate and the communication overhead generated by the revocation process. However, this method is vulnerable to false accusations because the accusing vehicle may also be malicious. At the same time, this method is only effective when the benefit of revoking the certificate of a normal vehicle is less than the benefit of adding malicious vehicles to the network. However, this situation is not common in the Internet of Vehicles scenario.

[0005] In the multi-vehicle voting-based method, when the number of accusations against abnormal vehicles exceeds a threshold, the vehicle's certificate will be revoked. This method avoids the problem of false accusations and improves the accuracy and reliability of the revocation process. However, it will expose the Internet of Vehicles to alliance attacks, that is, a group of vehicles mistakenly accuses a normal vehicle, causing its certificate to be revoked. At the same time, since multiple vehicles need to cooperate to complete the revocation task, this method will also generate additional communication overhead.

[0006] In summary, the current certificate revocation methods in the Internet of Vehicles have problems with low reliability and accuracy, are vulnerable to false accusation attacks and alliance attacks, and do not fully consider the security and low latency requirements of distributed Internet of Vehicles systems. Summary of the invention

[0007] To solve the above problems, the present invention proposes a certificate revocation method based on signaling game theory in an Internet of Vehicles, comprising the following steps:

[0008] S1. Build the Internet of Vehicles system. Each entity in the system performs initial identity authentication, including the trusted authority TA, the certificate management server CMS, the roadside unit RSU and the vehicle. The vehicles are clustered according to their geographical locations. The CMS completes the registration and issuance of vehicle certificates and selects the cluster head vehicle CHV in each cluster.

[0009] S2. The CHV detects an attack and interacts with the member vehicle that has made the attack. According to the signaling game theory, the present invention models the interaction process between two participants in a single time slot, namely, the CHV and the member vehicle, as a static certificate revocation game and solves the Bayesian Nash equilibrium.

[0010] S3. Expand the static certificate revocation game into a multi-stage dynamic certificate revocation game, establish a multi-stage dynamic certificate revocation game model, and solve the refined Bayesian equilibrium;

[0011] S4. Based on the mixed strategy refined Bayesian equilibrium of multi-stage dynamic certificate revocation game, a certificate revocation algorithm is proposed. CHV makes a decision to revoke the vehicle certificate according to the algorithm and sends a request message to CMS to revoke the malicious vehicle certificate;

[0012] S5. After receiving the request message from CHV, CMS distributes CRL and revocation messages to revoke the malicious vehicle certificate in the entire Internet of Vehicles system.

[0013] Beneficial effects of the present invention: The present invention proposes a certificate revocation method based on signal game theory in the Internet of Vehicles. Vehicles are clustered according to their geographic locations, and the CMS authorizes the cluster head vehicle CHV to monitor member vehicles in the cluster. At the same time, by modeling the interaction between CHV and member vehicles as a multi-stage dynamic certificate revocation game, it can flexibly respond to complex attacks in the Internet of Vehicles and improve the accuracy and real-time performance of the certificate revocation process. After receiving the revocation request, the CMS can distribute CRL and revocation messages to transmit the certificate revocation information of the malicious vehicle to all nodes in the system, thereby ensuring the overall security of the Internet of Vehicles system. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A model diagram of a distributed vehicle networking system based on vehicle clustering used in the present invention;

[0015] Figure 2 This is a flow chart of the implementation of the certificate revocation method based on signal game theory in the Internet of Vehicles proposed by the present invention;

[0016] Figure 3 This is a diagram of the expanded form of the static certificate revocation game proposed in the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] The present invention is based on Figure 1 The distributed vehicle networking system model based on vehicle clustering shown in the figure is studied, and the entities involved include vehicles, road side units (RSU), certificate management servers (CMS) and trusted authorities (TA). Among them:

[0019] TA is a fully trusted entity responsible for authorizing CMS to participate in the blockchain network.

[0020] CMS is responsible for the registration and issuance of vehicle certificates, storage and maintenance of the Certificate Revocation List (CRL), and is also a consensus node that constitutes the blockchain. It is deployed at the edge of the network and has strong computing and storage capabilities. In the Internet of Vehicles, CMSs communicate with each other through a wired network.

[0021] The RSU is located between the CMS and the vehicle, acting as a communication relay, assisting the CMS and the vehicle in implementing the certificate revocation function.

[0022] As a communication entity, the vehicle is the service object of certificate revocation. It is equipped with an on-board unit (OBU) with certain computing and storage capabilities. The vehicle can communicate with other vehicles around it through vehicle-to-vehicle communication (V2V), and can also access the RSU through vehicle-to-infrastructure communication (V2I).

[0023] Specifically, the entire Internet of Vehicles is divided into several sub-areas in geographical space, and CMS is deployed at the edge of the network. Each CMS is responsible for the registration, issuance and revocation of certificates in its own sub-area. In each sub-area, vehicles communicate using a clustered network structure, and the cluster head vehicle (CHV) is given the function of monitoring cluster members and executing revocation requests by the CMS. In the Internet of Vehicles, legal nodes include CMS, RSU and CHV. The types of cluster member vehicles are divided into ordinary vehicles and abnormal vehicles. When a vehicle is invaded or damaged, it will take direct or indirect attack behaviors (such as spreading a large number of false messages, conducting DoS attacks or man-in-the-middle attacks) to destroy the network. CHV can monitor these attack behaviors and start the revocation procedure. By applying signal game theory, the interaction between CHV and abnormal vehicles is modeled and analyzed. When it is determined that the abnormal vehicle is invaded or damaged, CHV defines it as a malicious vehicle, and then sends a revocation message to CMS to execute the decision to revoke the malicious vehicle certificate.

[0024] Based on the above analysis, if Figure 2 As shown, the present invention provides an embodiment of a certificate revocation method based on signal game theory in an Internet of Vehicles, comprising the following steps:

[0025] S1. Build a connected vehicle system. Each entity in the system performs initial identity authentication, including the trusted authority TA, the certificate management server CMS, the roadside unit RSU and the vehicle. The vehicles are clustered according to their geographical locations. The CMS completes the registration and issuance of vehicle certificates and selects the cluster head vehicle CHV in each cluster.

[0026] S2. CHV detects the attack behavior and interacts with the member vehicle that performs the attack behavior. According to the signal game theory, the present invention models the interaction process between two participants (CHV and member vehicle) in a single time slot as a static certificate revocation game and solves the Bayesian Nash equilibrium.

[0027] Specifically, the process of establishing a static certificate revocation game model includes:

[0028] S21. In a single time slot when CHV and member vehicles interact, the 5-tuple (PL, Θ, A, P, U) of the static certificate revocation game is defined as:

[0029] ①PL = {CHV (signal receiver), member vehicle (signal sender)} is the set of participants in the static certificate revocation game;

[0030] ②Θ=Θ C ×Θ V . Where Θ V ={θ V =0,θ V=1} is the type space set of member vehicles, θ V =0 represents ordinary vehicles, θ V =1 represents an abnormal vehicle, Θ C ={θ C} is the type space set of CHV. Since there is only one type of CHV in the game model, θ C It stands for CHV;

[0031] ③A=A C ×A V is a set of possible strategies that game participants may adopt. For member vehicles: A V ={A V (θ V =0),A V (θ V =1)}={{a V (θ V =0)|Normal},{a V (θ V =1)|Attack,Normal}}, where Normal represents normal behavior and Attack represents attack behavior. For CHV: A C ={a C (θ C )|Mon,Rev}, where Mon represents that CHV continues to monitor the member vehicle, and Rev represents that CHV makes a decision to revoke the vehicle certificate, i.e., sends a request message to CMS to revoke the member vehicle certificate;

[0032] ④P=(p,1-p) is Θ V Common prior probability distributions of type, Where p represents the probability that the member vehicle type is an abnormal vehicle;

[0033] ⑤U=(u C ,u V ), U is the set of utility functions, u C ,u V are the utility functions of CHV and member vehicles, respectively, which are used to quantify their preferences for different decisions or outcomes.

[0034] S22. According to the characteristics of the Internet of Vehicles scenario, the utility function parameters of the static certificate revocation game are defined as follows, and the utility function parameters are all positive numbers. A and c A They represent the benefits and costs of the abnormal vehicle after the attack, g N and c N They represent the benefits and costs of the member vehicles after normal behavior. CHV will gain a gain g after revoking the member vehicle certificate. R, while paying the cost c R In the static certificate revocation game, in order to ensure the consistency of the present invention, it is assumed that all CHVs have the same detection rate α and false alarm rate β, where the detection rate α refers to the proportion of attack behaviors recognized by CHV, and the false alarm rate β refers to the proportion of normal behaviors mistakenly recognized as attack behaviors by CHV, which will cause losses to CHV. F .

[0035] S23. Specifically, the utility function u of CHV and member vehicles C ,u V The calculation method is shown in Table 1

[0036] Table 1 Utility function calculation method

[0037]

[0038] The utility function calculation method is public knowledge in the Internet of Vehicles system, and both CHV and member vehicles, the two participants in the game, have this knowledge. A basic assumption of game theory is that the participants are rational, so CHV and member vehicles are considered rational, that is, their goal is to maximize their benefits. In order to conform to the actual situation of the Internet of Vehicles system, the present invention makes the following assumptions without loss of generality:

[0039] Assumption 1: If CHV chooses the revocation strategy Rev, the benefit of the member vehicle choosing the attack behavior should be less than the benefit of choosing the normal behavior, that is, (1-α)g A -αg R -c A <g N -c N Otherwise, the certificate revocation procedure has no meaning.

[0040] Hypothesis 2: If CHV chooses to continue monitoring strategy Mon, the benefit of member vehicles choosing attack behavior should be greater than the benefit of choosing normal behavior, that is, g A -c A > N -c N , otherwise it would be meaningless for the member vehicles to carry out offensive behavior.

[0041] S24. As an incomplete information game, the static certificate revocation game can reach a Bayesian Nash equilibrium. Since CHV does not know the type of member vehicles at the beginning, through Harsanyi transformation, the present invention introduces a virtual game participant "Nature", which first determines the type of member vehicles, thereby transforming the static certificate revocation game into a complete but imperfect information game, and establishing an extended form of the game such as Figure 3 shown.

[0042] Furthermore, the process of solving the Bayesian Nash equilibrium of the static certificate revocation game includes:

[0043] S25. Solve the pure strategy Bayesian Nash equilibrium of the static certificate revocation game, including:

[0044] S251. Member vehicle selection pure strategy {a V (θ V =0) = Normal, a V (θ V =1) = Attack}, that is, ordinary vehicles always behave normally, while abnormal vehicles always behave aggressively. At this time, the expected benefit of CHV choosing to revoke the member vehicle certificate is:

[0045] Eu C (Rev) = p·(αg R -(1-α)g A -c R )+(1-p)·(-βl F -c R ),

[0046] The expected benefit of CHV choosing to continue monitoring the vehicle is:

[0047] Eu C (Mon) = p·(-g A )+(1-p)·0,

[0048] When Eu C (Rev)≥Eu C (Mon), i.e. When {a C (θ C )=Rev} is the optimal strategy of CHV. At this time, according to Assumption 1, {a V (θ V =1) = Attack} will no longer be the optimal strategy for abnormal vehicles, so {{a V (θ V =0) = Normal, a V (θ V =1)=Attack},a C (θ C )=Rev} is not a pure strategy Bayesian Nash equilibrium.

[0049] When Eu c (Rev) <Eu c (Mon), i.e. When {a C (θ C )=Mon} is the optimal strategy of CHV. According to Assumption 2, {a V (θV =1) = Attack} is also the optimal strategy for abnormal vehicles, so {{a V (θ V =0) = Normal, a V (θ V =1)=Attack},a C (θ C )=Mon} is a pure strategy Bayesian Nash equilibrium of the static certificate revocation game.

[0050] S252. Member vehicle selection pure strategy {a V (θ V =0) = Normal, a V (θ V =1) = Normal}, that is, the member vehicles always behave normally. At this time, the optimal strategy of CHV should be {a C (θ C )=Mon}, then according to Assumption 2, {a V (θ V =1) = Attack} is the optimal strategy for the abnormal vehicle, which contradicts the pure strategy of the member vehicle, so {{a V (θ V =0) = Normal, a V (θ V =1)=Normal},a C (θ C )=Mon} is not a pure strategy Bayesian Nash equilibrium.

[0051] Furthermore, according to the above analysis, When , the static certificate revocation game has a pure strategy Bayesian Nash equilibrium:

[0052] {{a V (θ V =0) = Normal, a V (θ V =1)=Attack},a C (θ C )=Mon},

[0053] However, this strategy is not practical because under this strategy, CHV will only choose to continuously monitor member vehicles without making a decision to revoke abnormal vehicle certificates, which makes the Internet of Vehicles system vulnerable to attacks.

[0054] S26. According to S25, when When , the static certificate revocation game has a pure strategy Bayesian Nash equilibrium, which is obviously only true when When , the static certificate revocation game proposed by the present invention will have a mixed strategy Bayesian Nash equilibrium. Therefore, solving the mixed strategy Bayesian Nash equilibrium of the static certificate revocation game includes the following steps:

[0055] S261. Define abnormal vehicle (θ V =1) is a mixed strategy σ V =(ρ,1-ρ), this strategy indicates that the abnormal vehicle makes an attack with probability ρ. At this time, the expected benefit of CHV choosing to revoke the member vehicle certificate is: Eu C (Rev) = p·ρ·(αg R -(1-α)g A -c R )+p·(1-ρ)·(-βl F -c R )+(1-p)(-βl F -c R ), the expected benefit of CHV choosing to continue monitoring the vehicle is:

[0056] Eu C (Mon) = p·ρ·(-g A )+p·(1-ρ)·0+(1-p)·0=p·ρ·(-g A ),

[0057] When the expected benefits of CHV choosing to withdraw and monitor are the same, the abnormal vehicle has an optimal attack probability ρ * :

[0058]

[0059] S262. Define CHV(θ C ) is a mixed strategy σ C =(δ,1-δ), this strategy means that CHV makes a decision to revoke the vehicle certificate with probability δ. At this time, for the member vehicles, the expected benefit of choosing the attack behavior is:

[0060] Eu V (Attack) = δ·p·((1-α)g A -αg R -c A )+(1-δ)·p·(g A -c A ),

[0061] The expected payoff of choosing normal behavior is:

[0062] Eu V (Normal) = g N -c N ,

[0063] When the expected benefits of member vehicles choosing attack and normal behavior are the same, CHV has an optimal revocation probability δ * :

[0064]

[0065] S263. When , the static certificate revocation game has a mixed strategy Bayesian Nash equilibrium: Specifically, the hybrid strategy means that normal vehicles always behave normally, and abnormal vehicles behave normally with probability ρ * Make an aggressive behavior, and CHV with probability δ * Make a decision to revoke the vehicle certificate.

[0066] Further, according to the analysis of S25 and S26, in the interaction process between CHV and member vehicles in a single time slot, that is, in the static certificate revocation game, due to the different prior probabilities p of the member vehicle type being an abnormal vehicle, CHV and member vehicles will choose different strategies to make the game reach the Bayesian Nash equilibrium. Specifically, p is also called the belief of CHV that the member vehicle type is an abnormal vehicle. The increase of p value will make the abnormal vehicle abandon the pure strategy and choose the mixed strategy to disguise itself, make fewer attack behaviors, reduce the risk of its own certificate being revoked, and pursue higher benefits.

[0067] S3. Since it is difficult to determine a reasonable belief p in each individual time slot, CHV is required to dynamically update the belief p according to the actual behavior of the member vehicles. To achieve this goal, the present invention expands the static certificate revocation game into a multi-stage dynamic certificate revocation game and solves the refined Bayesian equilibrium.

[0068] Specifically, the process of establishing a multi-stage dynamic certificate revocation game includes:

[0069] S31. As the CHV interacts with its member vehicles, the static certificate revocation game will be played in each consecutive time slot t k Repeat, where k = 1, 2, ..., n (n ∈ Z + ), assuming that CHV and member vehicles are k+1 The utility function in the stage game and the k The utility functions in the stage games are the same, and participants can observe the results of each previous stage game before the next stage game begins, which enables CHV to predict the types of member vehicles and adjust future operations based on the results of the previous stage game.

[0070] The five-tuple (PL, Θ, A, P(R), U) of the multi-stage dynamic certificate revocation game is defined as:

[0071] ①The definition of PL,Θ,A,U is the same as that of the static certificate revocation game;

[0072] ②P(R)={p(θ V =1|h(t k )),1-p(θ V =1|h(t k ))}, where p(θ V =1|h(t k )) is the prior probability, indicating that at t k In the stage game, based on the historical behavior h(t k ), is the probability that the vehicle type is abnormal, also known as the CHV's prior belief that the member vehicle type is an abnormal vehicle.

[0073] S32.p(θ V =1|a V (t k ),h(t k )) is the posterior probability, also known as the posterior belief of CHV that the member vehicle type is an abnormal vehicle, indicating the tth k The probability that the vehicle type is abnormal at the end of the stage game, where a V (t k ) is the vehicle at t k Behavior during the stage game. Based on the Bayesian criterion, the posterior belief can be calculated by the following formula:

[0074]

[0075] Where p(a V (t k )|θ V ,h(t k )) indicates that the vehicle is at t k In stage game, based on historical behavior h(t k ) Make a V (t k ) behavior. Since CHV has a detection rate α and a false alarm rate β, the behavior monitored by CHV may not always objectively reflect the actual situation. Therefore, considering the detection rate and false alarm rate, we can get p(a V (t k )|θ V ,h(t k )) is calculated as follows:

[0076] p(Attack|θ V =1,h(t k ))=α·ρ A +β·(1-ρ A )

[0077] p(Normal|θ V =1,h(t k ))=(1-α)·ρ A +(1-β)·(1-ρ A )

[0078] p(Attack|θ V =0,h(t k ))=β

[0079] p(Normal|θ V =0,h(t k ))=1-β,

[0080] where ρ A =1-p1p2p3,ρ A The probability of member vehicles performing aggressive behaviors is evaluated for the IoV system, p1 is the probability that member vehicles cooperate in the IoV, p2 is the probability that member vehicles are honest, and p3 is the probability that member vehicles have a good reputation in the IoV.

[0081] Furthermore, in the multi-stage dynamic certificate revocation game, participants do not always adopt the same strategy in each stage, but adjust their future behavior according to previous results and current beliefs to obtain the maximum benefit, and choose the best strategy as the game develops. Therefore, the multi-stage dynamic certificate revocation game proposed in the present invention can be described by solving the mixed strategy refined Bayesian equilibrium. Based on the mixed strategy refined Bayesian equilibrium, CHV can establish a complete belief system about the types of member vehicles and adopt the best response strategy.

[0082] S33. Solving the mixed strategy refined Bayesian equilibrium of the multi-stage dynamic certificate revocation game.

[0083] In t k In stage game, define and They are abnormal vehicles (θ V =1) and a mixed strategy of CHV, where ρ k The probability of an abnormal vehicle performing an attack, δ k The probability of making a decision to revoke the vehicle certificate for a CHV.

[0084] For CHV, the expected benefit of choosing to revoke member vehicle certificates is:

[0085]

[0086] The expected benefit of choosing to continue monitoring the vehicle is:

[0087]

[0088] According to the member vehicle mixing strategy The indifference between revocation and continued monitoring can be used to obtain the optimal attack probability of abnormal vehicles.

[0089]

[0090] Similarly, according to the CHV mixed strategy The indifference between attack behavior and normal behavior can be obtained by

[0091]

[0092] According to S33's analysis, there is a mixed strategy refined Bayesian equilibrium in the multi-stage dynamic certificate revocation game and with α, β and belief p(θ V =1|h(t k ))related.

[0093] S4. Based on the mixed strategy refined Bayesian equilibrium of the above-mentioned multi-stage dynamic certificate revocation game, the present invention proposes a certificate revocation algorithm, the specific contents of which are as follows.

[0094]

[0095] First, CHV monitors the behavior of member vehicles in the cluster and estimates the probability ρ of the monitored member vehicles in the cluster to perform an attack A If an attack is detected, CHV performs a certificate revocation procedure and establishes a game with the member vehicles. As the game progresses, CHV's belief about the type of member vehicles also changes. Based on this belief, CHV can The best response strategy is selected in the k-th stage game with the vehicle. At the end of each stage game, CHV calculates the posterior belief of the member vehicle type to update the prior belief of the next stage game and store it.

[0096] S5. After receiving the request message for revoking the vehicle certificate from the CHV, the CMS first adds the certificate of the malicious vehicle to the CRL, and then distributes the CRL to the vehicles in the sub-area, thereby effectively revoking the certificate of the malicious vehicle in the sub-area.

[0097] Furthermore, CMS generates a revocation message consisting of five fields:

[0098] Originator: Indicates the CMS that created the revocation message. The CMS uses its certificate for identification.

[0099] Timestamp: The creation time of the revocation message, in Unix time format.

[0100] Vehicle Certificate: The certificate of the malicious vehicle.

[0101] Game history information: historical parameters recorded when the CHV that sent this revocation request played games with the malicious vehicle.

[0102] Endorsement: CMS's endorsement of this revocation message.

[0103] The revocation message can be formally expressed as:

[0104] T=(T s ,T t ,T r ,T h ,T e ),

[0105] Where T s Indicates the initiator field, T t Indicates the timestamp field, T r Represents the vehicle certificate field, T h Indicates the game history information field, T e Represents the endorsement field.

[0106] The CMS sends the revocation information to other CMSs in the Internet of Vehicles system, thereby effectively revoking the malicious vehicle certificate in the entire Internet of Vehicles system, making it impossible for it to continue to participate in the Internet of Vehicles and protecting the safety of other ordinary vehicles.

[0107] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A certificate revocation method based on signaling game theory in an Internet of Vehicles, characterized in that: The method includes: S1: Build the Internet of Vehicles system. Each entity in the system performs initial identity authentication, including the trusted authority TA, the certificate management server CMS, the roadside unit RSU and the vehicle. The vehicles are clustered according to their geographical locations. The CMS completes the registration and issuance of vehicle certificates and selects the cluster head vehicle CHV in each cluster. S2: CHV detects the attack behavior and interacts with the member vehicle that has made the attack behavior. According to the signal game theory, the present invention models the interaction process between the two participants in a single time slot, i.e., CHV and the member vehicle, as a static certificate revocation game and solves the Bayesian Nash equilibrium; S3: Expand the static certificate revocation game into a multi-stage dynamic certificate revocation game, establish a multi-stage dynamic certificate revocation game model, and solve the refined Bayesian equilibrium; S4: Based on the mixed strategy refined Bayesian equilibrium of the multi-stage dynamic certificate revocation game, a certificate revocation algorithm is proposed. CHV makes a decision to revoke the vehicle certificate according to the algorithm and sends a request message to CMS to revoke the malicious vehicle certificate; S5: After receiving the request message from CHV, CMS distributes CRL and revocation messages to revoke the malicious vehicle certificate in the entire Internet of Vehicles system.

2. The certificate revocation method based on signaling game theory in a vehicle network according to claim 1 is characterized in that: The process of establishing a static certificate revocation game model in S2 includes: S21: In a single time slot when CHV and member vehicles interact, the 5-tuple (PL, Θ, A, P, U) of the static certificate revocation game is defined as: ① PL = {CHV (signal receiver), member vehicle (signal sender)} is the set of game participants; ② Θ = Θ C ×Θ V ; where Θ V ={θ V =0,θ V =1} is the type space set of member vehicles, θ V =0 represents ordinary vehicles, θ V =1 represents an abnormal vehicle, Θ C ={θ C } is the type space set of CHV, θ C Represents CHV; S22: According to the characteristics of the Internet of Vehicles scenario, the utility function parameters of the static certificate revocation game are defined as follows. The utility function parameters are all positive numbers, g A and c A They represent the benefits and costs of the abnormal vehicle after the attack, g N and c N They represent the benefits and costs of the member vehicles after normal behavior. After CHV revokes the member vehicle certificate, it will gain a gain g R , while paying the cost c R ; S23: Specifically, the utility functions u of CHV and member vehicles C ,u V The calculation method is as follows: S24: As an incomplete information game, the static certificate revocation game can reach a Bayesian Nash equilibrium; S25: Solve the pure strategy Bayesian Nash equilibrium of the static certificate revocation game.

3. The certificate revocation method based on signaling game theory in a vehicle network according to claim 1 is characterized in that: in, S25 includes: S251: Member vehicle selection pure strategy {a V (θ V =0) = Normal, a V (θ V =1)=Attack}, that is, ordinary vehicles always behave normally, while abnormal vehicles always behave aggressively. At this time, the expected benefit of CHV choosing to revoke the member vehicle certificate is: Eu C (Rev)=p·(αg R -(1-α)g A -c R )+(1-p)·(-βl F -c R ); The expected benefit of CHV choosing to continue monitoring the vehicle is: I C (My)=p·(-g A )+(1-p)·0; When Eu C (Rev)≥Eu C (Mon), i.e. When {a C (θ C )=Rev} is the optimal strategy of CHV. At this time, according to Assumption 1, {a V (θ V =1) = Attack} will no longer be the optimal strategy for abnormal vehicles, so {{a V (θ V =0) = Normal, a V (θ V =1)=Attack},a C (θ C )=Rev} is not a pure strategy Bayesian Nash equilibrium; When Eu c (Rev) <Eu c (Mon), i.e. When {a C (θ C )=Mon} is the optimal strategy of CHV. According to Assumption 2, {a V (θ V =1) = Attack} is also the optimal strategy for abnormal vehicles, so {{a V (θ V =0) = Normal, a V (θ V =1)=Attack},a C (θ C )=Mon} is a pure strategy Bayesian Nash equilibrium of the static certificate revocation game; S252: Member vehicle selection pure strategy {a V (θ V =0) = Normal, a V (θ V =1) = Normal}, that is, the member vehicles always behave normally. At this time, the optimal strategy of CHV should be {a C (θ C )=Mon}, then according to Assumption 2, {a V (θ V =1) = Attack} is the optimal strategy for abnormal vehicles, which contradicts the pure strategy selected by member vehicles, so {{a V (θ V =0) = Normal, a V (θ V =1)=Normal},a C (θ C )=Mon} is not a pure strategy Bayesian Nash equilibrium; Define abnormal vehicles (θ V =1) is a mixed strategy σ V =(ρ,1-ρ), this strategy indicates that the abnormal vehicle makes an attack with probability ρ. At this time, the expected benefit of CHV choosing to revoke the member vehicle certificate is: EU C (Rev)=p·ρ·(αg R -(1-α)g A -c R )+p·(1-ρ)·(-βl F -c R )+(1-p)(-βl F -c R ); The expected benefit of CHV choosing to continue monitoring the vehicle is: EU C (Mon)=p·ρ·(-g A )+p·(1-p);0+(1-p);0=p·p·(-g A ); When the expected benefits of CHV choosing to withdraw and monitor are the same, the abnormal vehicle has an optimal attack probability ρ * : Define CHV(θ C ) is a mixed strategy σ C =(δ,1-δ), this strategy means that CHV makes a decision to revoke the vehicle certificate with probability δ. At this time, for the member vehicles, the expected benefit of choosing the attack behavior is: EU V (Attack)=δ·p·((1-α)g A -ag R -c A )+(1-δ)·p·(g A -c A ); The expected payoff of choosing normal behavior is: I V (Normal)=g N -w N ; When the expected benefits of member vehicles choosing attack and normal behavior are the same, CHV has an optimal revocation probability δ * : Therefore When , the static certificate revocation game has a mixed strategy Bayesian Nash equilibrium: Specifically, the hybrid strategy means that normal vehicles always behave normally, and abnormal vehicles behave normally with probability ρ * Make an aggressive behavior, and CHV with probability δ * Make a decision to revoke the vehicle certificate.

4. The certificate revocation method based on signaling game theory in a connected vehicle network according to claim 1, characterized in that: The process of establishing a multi-stage dynamic certificate revocation game includes: S31: As the CHV interacts with its member vehicles, the static certificate revocation game will be played in each consecutive time slot t k Repeat, where k = 1, 2, ..., n (n∈Z + ), assuming that CHV and member vehicles are k+1 The utility function in the stage game and the k The utility functions in the stage games are the same; participants can observe the results of each stage game before the next stage game begins, which enables CHV to predict the types of member vehicles and adjust future operations based on the results of the previous stage game; The five-tuple (PL, θ, A, P(R), U) of the multi-stage dynamic certificate revocation game is defined as follows: ① The definitions of PL, θ, A, U are the same as those of the static certificate revocation game; ② P(R) = {p(θ V =1|h(t k )),1-p(θ V =1|h(t k ))}, where p(θ V =1|h(t k )) is the prior probability, indicating that at t k In the stage game, based on the historical behavior h(t k ), the probability that the vehicle type is abnormal, also known as the CHV's prior belief that the member vehicle type is an abnormal vehicle; S32: p(θ V =1|a V (t k ),h(t k )) is the posterior probability, also known as the posterior belief of CHV that the member vehicle type is an abnormal vehicle, indicating the tth k The probability that the vehicle type is abnormal at the end of the stage game, where a V (t k ) is the vehicle at t k Based on the Bayesian criterion, the posterior belief of CHV during the stage game can be calculated by the following formula: Where p(a V (t k )|θ V ,h(t k )) indicates that the vehicle is at t k In stage game, based on historical behavior h(t k ) Make a V (t k ) behavior. Since CHV has detection rate and false alarm rate, the behavior monitored by CHV may not always objectively reflect the actual situation. Therefore, considering the detection rate and false alarm rate, we can get p(a V (t k )|θ V ,h(t k )) is calculated as follows: p(Attack|θ V =1,h(t k ))=a·r A +β·(1-ρ A ) p(Normal|θ V =1,h(t k ))=(1-a)·r A +(1-β)·(1-ρ A ) p(Attack|θ V =0,h(t k ))=b p(Normal|θ V =0,h(t k ))=1-β; where ρ A =1-p1p2p3,ρ A The probability of member vehicles performing aggressive behaviors is evaluated for the IoV system, p1 is the probability that member vehicles cooperate in the IoV, p2 is the probability that member vehicles are honest, and p3 is the probability that member vehicles have a good reputation in the IoV; Furthermore, in the multi-stage dynamic certificate revocation game, participants do not always adopt the same strategy in each stage, but adjust their future behavior according to previous results and current beliefs to obtain the maximum benefit, and choose the best strategy as the game develops; therefore, the multi-stage dynamic certificate revocation game proposed in the present invention can be described by solving the mixed strategy refined Bayesian equilibrium; based on the mixed strategy refined Bayesian equilibrium, CHV can establish a complete belief system about the types of member vehicles and adopt the best response strategy; S33: Solving the mixed strategy refined Bayesian equilibrium of the multi-stage dynamic certificate revocation game: In t k In the stage game, define σ Vk =(ρ k ,1-ρ k ) and σ Ck =(δ k ,1-δ k ) are abnormal vehicles (θ V =1) and a mixed strategy of CHV, where ρ k The probability of an abnormal vehicle performing an attack, δ k The probability of making a decision to revoke the vehicle certificate for the CHV; For CHV, the expected benefit of choosing to revoke member vehicle certificates is: The expected benefit of choosing to continue monitoring the vehicle is: According to the mixed strategy of member vehicles σ Vk The indifference between revocation and continued monitoring can be used to obtain the optimal attack probability of abnormal vehicles. Similarly, according to the CHV mixed strategy The indifference between attack behavior and normal behavior can be obtained by According to S33's analysis, there exists a mixed strategy refined Bayesian equilibrium in the multi-stage dynamic certificate revocation game and with α, β and belief p(θ V =1|h(t k ))related.

Images