Auxiliary lightweight anonymous authentication Internet of Vehicles data sharing method based on block chain

By adopting blockchain's auxiliary lightweight anonymous authentication method in the Internet of Vehicles, two-way anonymous authentication and message signatures are performed, cluster head nodes are selected, and smart contracts and DBFT consensus algorithms are used to solve the problem of data leakage risk in the Internet of Vehicles. , realizes the security and efficiency of data sharing.

CN119031369APending Publication Date: 2024-11-26CHONGQING UNIV OF POSTS & TELECOMM
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410373373.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-03-29
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

There is a risk of data leakage in the Internet of Vehicles, especially the parameters transmitted through blockchain may be acquired by malicious nodes, resulting in privacy leakage and reasoning attacks.

Method used

Adopting an auxiliary lightweight anonymous authentication method based on blockchain, through two-way anonymous authentication and message signature, it ensures authentication and data encryption during data sharing between the vehicle and roadside units, and selects the vehicle node with the largest weight measurement as the Cluster head, which uses blockchain's smart contract verification contract and DBFT consensus algorithm to protect data privacy.

Benefits of technology

It effectively prevents the leakage of malicious nodes in the middle, ensures the privacy and security of vehicle data, reduces communication and computing overhead through anonymous authentication and message signatures, and improves the security and efficiency of data sharing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119031369A_ABST
    Figure CN119031369A_ABST
Patent Text Reader

Abstract

The invention relates to an Internet of Vehicles security technology, in particular to an auxiliary lightweight anonymous authentication Internet of Vehicles data sharing method based on a block chain, which comprises the following steps: a trusted server registers a roadside unit or a vehicle machine requesting for registration, and a vehicle performs bidirectional anonymous authentication to the roadside unit initiating the request for the first time after registration is completed; after the bidirectional anonymous authentication is completed, the road side unit distributes a training task to a vehicle, the vehicle performs local training according to a distributed model to obtain a local model, signs the local model, and sends a signature to the road side unit; the road side unit verifies the received signature, selects one of the vehicle nodes passing the verification as a cluster head node, and sends the trained model data to the cluster head node by other nodes passing the verification; the cluster head node collects all training data passing verification of the vehicle nodes and forwards the training data to the road side unit, and the road side unit generates blocks according to the received data and issues the blocks to the block chain; according to the invention, data privacy can be protected, and intermediate malicious leakage can be prevented.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to vehicle networking security technology, and in particular to a blockchain-based auxiliary lightweight anonymous authentication vehicle networking data sharing method. Background Art

[0002] With the development of 5G and edge computing technologies, the Internet of Vehicles (IoV) formed by vehicle interconnection has attracted widespread attention. In the IoV, the server collects the shared data of the car and combines it with artificial intelligence technology for collaborative analysis, which can improve the driving experience and vehicle applications. At the same time, information sharing between vehicles is carried out through on-board units, roadside units (RSUs) and mobile networks. Usually, the vehicle transmits the raw data it collects to the server or RSU, which contains sensitive information such as location and speed. Therefore, this leads to serious risks of privacy leakage.

[0003] Vehicles transmit data obtained from advanced on-board devices to the cloud through edge nodes. Data from the cloud can be used for model training and provide services to the application layer, such as map services, vehicle services, etc. However, a large amount of privacy data is gathered on the central server, and there are huge security and privacy challenges in training. Federated learning, as a new type of distributed machine learning, is not centralized and can combine multiple local devices (i.e., vehicles) to jointly train machine learning models by sharing local model updates (i.e., gradient parameters). Existing research combines blockchain and federated learning to achieve collaborative training of global models without sharing their original data. And the blockchain guarantees the secure transmission of parameters. However, the parameters stored in the blockchain global ledger as transactions are accessible to all participants, which may lead to privacy leakage. In addition, some studies have shown that malicious nodes can obtain training data information of other nodes by collecting these parameters, thereby conducting inference attacks. Summary of the invention

[0004] In view of the risk of data leakage in the existing vehicle network, the present invention proposes a blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method, which specifically includes the following steps:

[0005] The trusted server registers the roadside unit or vehicle that requested the registration. After the registration is completed, the vehicle performs two-way anonymous authentication with the roadside unit that initiated the request for the first time.

[0006] After completing the two-way anonymous authentication, the roadside unit distributes the training task to the vehicle. The vehicle performs local training based on the distributed model, obtains the local model, signs the local model, and sends the signature to the roadside unit.

[0007] The roadside unit verifies the received signature and selects one of the verified vehicle nodes as the cluster head node. The other verified nodes send the trained model data to the cluster head node.

[0008] The cluster head node will collect all the training data that have passed the verification of the vehicle nodes and forward it to the roadside unit. The roadside unit will generate blocks from the received data and publish them to the blockchain.

[0009] Furthermore, after completing the registration, the vehicle performs two-way anonymous authentication with the roadside unit that initiates the request for the first time, including the following steps:

[0010] Each roadside unit or vehicle obtains public parameters from a trusted server Among them, G is an additive cyclic group with generator P and order q, G T is a multiplicative cyclic group with generator P and order q, where q is a prime number, is a bilinear mapping: G×G→G T , P, P pub1 , P pub2 is the public key parameter, H(·) is the hash function;

[0011] The trusted server sends the private parameters A, D and ω of the roadside unit through a secure channel and satisfies Ad=ω, and sends the private parameters to the roadside unit for bidirectional anonymous verification as a shared secret between the roadside unit and the trusted server; the trusted server generates a column vector d for each legitimate vehicle node i , and let d i is the solution of Ad=ω;

[0012] The roadside unit randomly selects a parameter k, where parameter k belongs to the finite field The random number is used to calculate the authentication parameters R=kA and S=kω of the roadside unit;

[0013] The roadside unit will authenticate the information Sent to vehicle user V i , where t 1 is a timestamp related to the message transmission time, || represents a concatenation operation, It is the real ID of the roadside unit;

[0014] Vehicle User V i Receive the message from the roadside unit and calculate r = Rd i ,verify Is it true? If so, Sent to the roadside unit; where t 2 is a timestamp related to the time the message was transmitted;

[0015] After receiving the message, the roadside unit verifies Is it true? If true, the RSU selection belongs to the finite domain The random parameter y i , and generate the parameter ρ i =y i P is sent to the vehicle user V i ; Otherwise, communication will be interrupted.

[0016] Furthermore, in the message signing stage, the roadside unit and the vehicle user jointly generate a message signature for identity verification between vehicles, which specifically includes the following steps:

[0017] Vehicle User V i Randomly select parameter μ i 、r i1 、r i2 , and set the randomization parameter C i =μ i ρ i , calculate the pseudonym AID i =RID i t 2 , calculate the signature key β i and γ i , that is, β i =r il H(AID i ||C i ||M i )+r i2 P,γ i =r i1 μ i (modq); where RID i For vehicle users V i Real ID of M i For the message sent;

[0018] Vehicle User V i The signing key (β i , γ i ) is sent to the roadside unit;

[0019] The roadside unit signs the message sent to the vehicle, that is, T i =x 1 β i +x 2 yγ i P and sends it to the vehicle user; where x 1 、x 2 is the private master key of the roadside unit; y is a finite field Random parameters of

[0020] The vehicle user receives the T sent by the roadside unit i , calculate the final signature information

[0021] Vehicle User V i Get the roadside unit to message M i The signature σ i ={AID i ,M i ,S i ,C i}.

[0022] Furthermore, the process of verifying the signed message includes:

[0023] In single message verification, the real identity ID of the message receiver verifies the signature of the message senderσ i , if e^(S i ,P)=e^(C i ,P pub2 )e^(H(AID i ||C i ||M i ),P pub1 ) is established, the verification process is passed, indicating that the vehicle identity is legal, and the message M is received i ; Otherwise, the message will be rejected;

[0024] In the batch message verification process, the message receiver verifies the validity of n messages at the same time. This proves that the message signatures are valid and that the verifier received the messages.

[0025] Furthermore, the vehicle node with the largest weight metric is selected as the cluster head. The calculation of the weight metric includes:

[0026]

[0027] ω 1 +ω 2 +ω 3 =1

[0028] in, Represents the vehicle node V i The weight measure of Represents the vehicle node V i The total number of vehicle nodes covered by a specific node in a one-hop neighborhood; Represents the vehicle node V i The remaining energy Represents the vehicle node V i Credibility value; 1 ,ω 2 ,ω3 They are The weight parameter of .

[0029] Furthermore, the vehicle node V i The total number of vehicle nodes covered by a specific node in a single-hop neighborhood It is expressed as:

[0030]

[0031] in, Represents a slave node arrive The number of hops; N represents The set of directly connected nodes.

[0032] Furthermore, the vehicle node V i The remaining energy It is expressed as:

[0033]

[0034]

[0035]

[0036]

[0037] in, For Node The initial energy of For Node Energy used; E Tx (n, Distance) represents the energy required to transmit n bits of data between two nodes with a distance of Distance; E RX(n) Indicates the energy of a node receiving n bits of data; E Elec is the energy consumed by the electronic circuit; E FS is the energy required by the free-space model amplifier, E MP is the energy required by the amplifier in the multipath fading model.

[0038] Furthermore, the vehicle node V i Reputation It is expressed as:

[0039]

[0040] in, Represents the reputation of the vehicle node; express The utility parameter.

[0041] Furthermore, a verification contract Verification Contract is embedded in the smart contract on the blockchain. The contract verifies the digital signature of the vehicle through the function Authentication(), and estimates the training accuracy of each local training model update through the function LocalModelVerification(). The data can be packaged to generate a block only when the training accuracy of the local training model update is greater than the set threshold.

[0042] In order to protect data privacy, the present invention sends training tasks to vehicles requesting data sharing training at the application layer. Each vehicle iteratively trains a shared global model based on its local data and generates a local model update. In order to prevent malicious leakage in the middle, all vehicles will anonymously verify with the RSU to select qualified model updates. The RSU will upload to the blockchain through authentication, and the message signature and local model update will be stored in the blockchain as a transaction for the next round of federated learning; at the blockchain layer, the edge nodes (the present invention uses RSU as an edge node) have powerful computing power, communication capabilities, and storage capabilities. They can be used as blockchain miners. The message signature and local model update are stored in the data block after being verified by the miners through the consensus algorithm. Finally, the vehicle downloads the latest local model updates of all vehicles from the blockchain and aggregates the global model. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the framework structure of the blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method of the present invention;

[0044] Figure 2 This is an operation diagram of the anti-tampering device in the present invention;

[0045] Figure 3 Schematic diagram of the digital signature process in the present invention;

[0046] Figure 4 This is a schematic diagram of the message verification phase process in the present invention;

[0047] Figure 5 This is a schematic diagram of the consensus process of DBFT in the present invention. DETAILED DESCRIPTION

[0048] 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.

[0049] The present invention proposes a blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method, which specifically includes the following steps:

[0050] The trusted server registers the roadside unit or vehicle that requested the registration. After the registration is completed, the vehicle performs two-way anonymous authentication with the roadside unit that initiated the request for the first time.

[0051] After completing the two-way anonymous authentication, the roadside unit distributes the training task to the vehicle. The vehicle performs local training based on the distributed model, obtains the local model, signs the local model, and sends the signature to the roadside unit.

[0052] The roadside unit verifies the received signature and selects one of the verified vehicle nodes as the cluster head node. The other verified nodes send the trained model data to the cluster head node.

[0053] The cluster head node will collect all the training data that have passed the verification of the vehicle nodes and forward it to the roadside unit. The roadside unit will generate blocks from the received data and publish them to the blockchain.

[0054] The present invention provides a blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method, which includes three stages, namely:

[0055] The first stage is local training. When the data requester assigns training tasks to vehicles through roadside units, the data requester hopes that vehicles can share data for collaborative training. The vehicle downloads the initial global model, trains the model locally using its own data, and sends the local model update to the nearest roadside unit.

[0056] In the second phase, in order to prevent malicious leakage in the middle, all vehicles will conduct anonymous verification with the roadside unit to select qualified model updates. The roadside unit will upload the authenticated local model to the blockchain. The message signature and local model update will be stored in the blockchain as transactions for the next round of federated learning. This phase includes the sender's identity authentication and digital signature process, that is, the knowledge learned in the source domain is migrated to achieve real-time monitoring of whether the SDN network controller is faulty;

[0057] In the third stage, the model update is aggregated with anonymously authenticated data parameters. The message signature and local model update are verified by miners through a consensus algorithm and stored in the data block. Finally, the vehicle downloads the latest local model updates of all vehicles from the blockchain and aggregates the global model.

[0058] In the process of implementing the second stage of this embodiment, the roadside unit is used as the PSN and the vehicle node is used as the SN, where the PSN is granted the right to write data and participate in consensus. The local storage device in the PSN is responsible for collecting the sensor data uploaded by the SN and obtaining the data shared by other PSNs, and automatically collating and analyzing the data using the originally deployed smart contract; the SN can access and synchronize the copy, but does not participate in the consensus. The traffic management center or the trusted server can adjust the traffic conditions of the VANET according to the analysis results to improve the traffic efficiency and the authenticity of the safety warning. The trusted server and the traffic management center communicate with the PSN through a wired connection security channel (such as a transport layer security pool). The PSN is usually supervised and authorized by the trusted server and scheduled and managed by the traffic management center. Finally, the data in the entire network can be interconnected to form a fully automated data transmission network, thereby ensuring two-way information transmission and on-demand data sharing between nodes in the VANET.

[0059] In the third phase, DBFT consensus is used to select a leader and generate a new block. Qualified transactions for local model updates are packaged into blocks after hash calculation. During this process, eligible local model update transactions will form a transaction pool, and the leader node will take out the transaction from the transaction pool for hash calculation. The leader node will get a Merkle root and save it in the block; all transaction information will be reflected in the hash value of the new block; finally, the model update is aggregated, and the vehicle downloads a new block containing the latest local model update, and the vehicle uses these latest local model updates to aggregate the global model.

[0060] like Figure 1 The blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method proposed in this embodiment includes the following steps:

[0061] Step 1: The data requester assigns training tasks to vehicles through the roadside unit. The data requester hopes that the vehicles can share data for collaborative training. During the collaborative training process, the vehicle downloads the initial global model, trains the model locally using its own data, and sends the local model update to the nearest roadside unit, realizing data sharing through local data collaborative training.

[0062] Step 2: In order to prevent malicious leakage in the middle, all vehicles will conduct anonymous verification with the roadside unit to select qualified model updates. The roadside unit will upload the authenticated local model to the blockchain. This stage is divided into four stages, including registration stage, authentication and signature key generation stage, message signing stage, and message verification stage. After the local model is anonymously authenticated, the update verification is uploaded to the blockchain.

[0063] The specific steps are as follows:

[0064] 1. Registration stage

[0065] Before network deployment, the roadside unit nodes and vehicle nodes need to register with a trusted server. After each roadside unit and vehicle passes legal authentication, the trusted server uses the following steps to generate system parameters for them:

[0066] (1) Let \(G\) be an additive cyclic group with generator \(P\) and prime order \(q\), and \(G T be a multiplicative cyclic group with generator \(P\) and prime order \(q\), and the bilinear pair

[0067] (2) The trusted server randomly generates an \(m\times n\) - dimensional matrix \(A(2\leq m\lt n)\) and an \(m\) - dimensional column vector \(\omega\) that satisfy the linear equation system \(Ad=\omega\) and have infinite solutions. Let \(R(\cdot)\) denote the rank of matrix \(A\), and \(\widetilde{A}\) denote the augmented matrix of \(A\).

[0068] (3) The trusted server generates a unique \(n\) - dimensional column vector \(d i for each legal vehicle node, and \(d i satisfies \(Ad i =\omega\), such that \(d i is a solution of the linear equation system \(Ad = \omega\); the trusted server sends the vector \(d i to the corresponding vehicle node \(V i as its real identity information. The trusted server randomly selects an \(m\) - dimensional column vector \(D\), and then calculates the real identity \(ID\) of \(V i as follows:

[0069] RID i =D T d i

[0070] The trusted server sends \(A\), \(D\), and \(\omega\) to the roadside unit through a secure channel as the shared secret between the roadside unit and the trusted server.

[0071] (4) The roadside unit generates its own private key \(x in the multiplicative group of units of the prime modulus \(q\) and the finite field 1 and calculates another private key \(x 2 as follows:

[0072] x 2 =(D T \omega)\bmod q

[0073] The corresponding public key dimension is \(P pub1 =x 1 P\) and \(P pub2 =x 2 P\), and \(H:\{0,1\} *→G is a one-way hash function. Each roadside unit and vehicle can publicly access the system parameters

[0074] 2. Authentication and signature key generation phase

[0075] The vehicle node needs to authenticate before communicating with the roadside unit. After the vehicle node is ready to receive the vehicle's message signature, it sends a signature request signal to the roadside unit. The specific steps are as follows:

[0076] 1) The roadside unit randomly selects k, where k belongs to a finite field Calculate the authentication parameters R and S, namely:

[0077] R=kA

[0078] S=kω

[0079] Authentication information Sent to vehicle user V i , where t 1 A timestamp related to the message transmission time.

[0080] 2) Vehicle user V i The message from the roadside unit is received and saved in the tamper-proof device. The tamper-proof device on the vehicle is considered to be reliable and its information will never be leaked. First, calculate:

[0081] r=Rd i

[0082] Then, verify Is it true? If they are equal, Sent to the roadside unit, where t 2 Is a timestamp related to the time the message was transmitted.

[0083] 3) After receiving the message, the roadside unit verifies Is it true? If the verification is successful, continue with the vehicle user V i Communication, that is, random parameter y i is selected, where y i For a finite field A random number, then ρ i =y i P calculation, parameter ρ i is sent to the vehicle user; otherwise, communication will be interrupted.

[0084] If i and RID i Once all are verified, the tamper-proof device will start generating anonymous identities and signing keys. The operation steps of the tamper-proof device are as follows: Figure 2 shown.

[0085] When a vehicle communicates with a roadside unit for the first time within the coverage of the RSU, it needs to perform two-way anonymous authentication to complete the reception of subsequent message signatures. After the authentication is completed, when the vehicle communicates with other vehicles and needs the assistance of the roadside unit to generate a signature, the vehicle and the roadside unit no longer need to authenticate each other, thereby reducing the communication and computing overhead of the entire communication process.

[0086] 3. Message signing phase

[0087] In the message signing phase, the RSU and the vehicle jointly generate a message signature for identity authentication between vehicles. The specific steps are as follows:

[0088] 1) Vehicle user V i Randomly select parameter μ i 、r i1 、r i2 , and set the randomization parameter C i Calculate the vehicle user V i AID i , that is, calculate the vehicle user V i The fake ID includes:

[0089] C i =μ i ρ i

[0090] AID i =RID i t 2

[0091] Calculate the signature key β i and γ i ,Right now:

[0092] β i =r il H(AID i ||C i ||M i )+r i2 P

[0093] γ i =r i1 μ i (modq)

[0094] Final output of the tamper-proof device (β i , γ i ), vehicle users will (β i , γ i ) is sent to the signatory roadside unit.

[0095] 2) Upon receiving the (βi , γ i ) message, the message sent by the vehicle is signed, that is:

[0096] T i =x 1 β i +x 2 yγ i P

[0097] Calculate the signature T i And send it to the vehicle user, the vehicle user receives the signature T sent by the signer's roadside unit i , calculate the final signature information, namely:

[0098]

[0099] Finally, the vehicle obtains the signatory roadside unit's response to the message M i The signature σ i ={AID i ,M i ,S i ,C i}, the digital signature process is as follows Figure 3 shown.

[0100] 4. Message verification phase

[0101] In the message verification phase, the assumption is that i∈(1,n) is randomly selected, and the RSU can receive the verification parameter ver i (AID i ,S i ,M i ,C i ), and each step runs normally. AID i The pseudonym generated for the vehicle, S i is the final signature information, M i For the message, C i is a randomized parameter, then when the messages shared between nodes are batch verified, the plural number of the authentication parameter is: BatchVer n ((AID 1 ,S 1 ,M 1 ,C 1 ),…,(AID n ,S n ,M n ,C n )).

[0102] Therefore, if all n signatures are valid, the batch verification passes; if one or more of the n signatures are invalid, the batch verification fails. Batch verification can be divided into three types, namely verifying that the signer signed different messages, verifying that different signers signed the same message, and verifying that different signers signed different messages. The process of single message verification and batch message verification includes:

[0103] 1) In a single message verification, the receiving vehicle receives the signature of the sending vehicle σ i , and checks if {AID i ,S n ,M n ,C i}.

[0104]

[0105] If the above formula is satisfied, the verification process is passed, indicating that the vehicle identity is legal, and the message M is received. i Otherwise, the message will be rejected. The message verification process is as follows Figure 4 shown.

[0106] 2) During batch message verification, the roadside unit or vehicle user V i The validity of multiple messages can be verified at the same time. If the validity of n messages is verified at the same time, each message is recorded as σ 1 ={AID 1 ,S 1 ,M 1 ,C 1},σ 2 ={AID 2 ,S 2 ,M 2 ,C 2},……,σ n ={AID n ,S n ,M n ,C n}, where M 1 ,M 2 ,……M n Probably the same.

[0107]

[0108] If the above equation is satisfied, it proves that the message signature is valid and the verifier received these messages.

[0109] 5. Cluster head selection

[0110] After the verifier successfully receives the message, the present embodiment selects a cluster head (CH) among all vehicle nodes to achieve energy saving. Different parameters are considered when selecting the CH node, such as degree of connectivity (DoC), remaining energy, distance from the trusted server, and reputation of the node.

[0111] In this embodiment, the vehicle node with the largest weight metric is selected as the cluster head, and the calculation of the weight metric includes:

[0112]

[0113] ω 1 +ω 2 +ω 3 =1

[0114] in, Represents the vehicle node V i The weight measure of Represents the vehicle node V i The total number of vehicle nodes covered by a specific node in a one-hop neighborhood; Represents the vehicle node V i The remaining energy Represents the vehicle node V i Credibility value; 1 ,ω 2 ,ω 3 They are The weight parameter of .

[0115] Vehicle Node V i The total number of vehicle nodes covered by a specific node in a single-hop neighborhood It is expressed as:

[0116]

[0117] in, Represents a slave node arrive The number of hops; N represents The set of directly connected nodes.

[0118] Vehicle Node V i The remaining energy It is expressed as:

[0119]

[0120]

[0121]

[0122]

[0123] in, For Node The initial energy of For Node Energy used; E Tx (n, Distance) represents the energy required to transmit n bits of data between two nodes with a distance of Distance; E RX(n) Indicates the energy of a node receiving n bits of data; E Elec is the energy consumed by the electronic circuit; E FS is the energy required by the free-space model amplifier, E MP is the energy required by the amplifier in the multipath fading model.

[0124] Vehicle Node V i Reputation It is expressed as:

[0125]

[0126] in, Represents the reputation of the vehicle node; express The utility parameter.

[0127] Trusted servers from After collecting the information, it is broadcasted to the cloud server, which acts as a fog node. The information of the trusted server is divided into two parts, namely (1) data management and (2) certificate management. All the large-volume information or data is stored in the cloud network, which has various benefits such as reduced latency, reduced time consumption due to the availability of information in the cloud, and storage efficiency. The registration information, random keys, and revocation certificates of all vehicle nodes are managed using blockchain technology.

[0128] Step 3: All information stored in the blockchain is subject to the consensus process using the DBFT consensus algorithm. Compared with other consensus mechanisms, such as computationally intensive PoW consensus or communication-complex Byzantine fault tolerance, DBFT is developed on the basis of PBFT and can flexibly replace consensus nodes without the need for a fixed consensus group. Compared with DPoS, DBFT has better performance for the finality of block data. Therefore, this embodiment integrates local model update verification into the DBFT consensus algorithm and establishes a flexible alliance chain, which will help ensure reliable federated learning. The consensus process is as follows: Figure 5 As shown, the specific steps include:

[0129] (1) Initialization: Before executing the consensus algorithm, the identity recognition contract is called to verify the identity of the consensus node (the roadside unit is used as the consensus node in this embodiment). The authenticated roadside unit will become the delegated consensus node. The leader node generates and verifies block data as the representative of the consensus node. A leader is randomly selected from the consensus node to generate and verify the new block. This embodiment assumes that there are N representatives in the alliance chain and the maximum number of malicious representatives is m. In DBFT, N>3m+1. The leader will change in the next round of consensus; each delegated consensus node has an equal chance to be selected as the leader, and the leader will perform block generation, verification and broadcasting; before packaging the transaction into a new block, the delegated consensus node calls the verification contract to estimate the quality of the local model update. This will eliminate low-quality local model updates through this step. This embodiment embeds the verification contract Verification Contract in the smart contract on the blockchain. The contract verifies the digital signature of the vehicle through the function Authentication() and estimates the training accuracy of each local training model update through the function LocalModelVerification(). When the training accuracy of the local training model update is greater than the set threshold, the training data can be packaged into the new block.

[0130] (2) Preliminary preparation stage: In the preliminary preparation stage, the work of the consensus node is to execute blocks, generate signature packages, and broadcast new blocks to all consensus nodes.

[0131] (3) Preparation phase: After receiving the broadcast, other delegated consensus nodes generate verification results for the block data and send the results to other consensus nodes for confirmation. The consensus nodes verify the legitimacy of the signature package. As long as the delegated consensus node receives at least 2m+1 verification results, the preparation phase ends and the consensus node generates a submission package.

[0132] (4) Delivery phase: After receiving feedback from different delegated consensus nodes, the leader node will first verify the legitimacy of the commit package and check whether more than 2m+1 miners agree with the block data. If 2-3 miners reach the same conclusion, the block will be recorded in the blockchain.

[0133] Finally, the model updates are aggregated on the blockchain, and vehicles download new blocks containing the latest local model updates and use these latest local model updates to aggregate the global model.

[0134] 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 blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method, characterized in that: The specific steps include: The trusted server registers the roadside unit or vehicle that requested the registration. After the registration is completed, the vehicle performs two-way anonymous authentication with the roadside unit that initiated the request for the first time. After completing the two-way anonymous authentication, the roadside unit distributes the training task to the vehicle. The vehicle performs local training based on the distributed model, obtains the local model, signs the local model, and sends the signature to the roadside unit. The roadside unit verifies the received signature and selects one of the verified vehicle nodes as the cluster head node. The other verified nodes send the trained model data to the cluster head node. The cluster head node will collect all the training data that have passed the verification of the vehicle nodes and forward it to the roadside unit. The roadside unit will generate blocks from the received data and publish them to the blockchain.

2. According to claim 1, the blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method is characterized in that: After registration is completed, the vehicle performs two-way anonymous authentication with the roadside unit that initiates the request for the first time, including the following steps: Each roadside unit or vehicle obtains public parameters from a trusted server Among them, G is an additive cyclic group with generator P and order q, G T is a multiplicative cyclic group with generator P and order q, where q is a prime number, is a bilinear mapping: G×G→G T , P, P pub1 , P pub2 is the public key parameter, H(·) is the hash function; The trusted server sends the private parameters A, D and ω of the roadside unit through a secure channel and satisfies Ad=ω, and sends the private parameters to the roadside unit for bidirectional anonymous verification as a shared secret between the roadside unit and the trusted server; the trusted server generates a column vector d for each legitimate vehicle node i , and let d i is the solution of Ad=ω; The roadside unit randomly selects a parameter k, where parameter k belongs to the finite field The random number is used to calculate the authentication parameters R=kA and S=kω of the roadside unit; The roadside unit will authenticate the information Sent to vehicle user V i , where t1 is the timestamp related to the message transmission time, || represents the splicing operation, It is the real ID of the roadside unit; Vehicle User V i Receive the message from the roadside unit and calculate r = Rd i ,verify Is it true? If so, Sent to the roadside unit; where t2 is a timestamp related to the message transmission time; After receiving the message, the roadside unit verifies Is it true? If true, the roadside unit selection belongs to the finite domain The random parameter y i , and generate the parameter ρ i =y i P is sent to the vehicle user V i ; Otherwise, communication will be interrupted.

3. The blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method according to claim 2 is characterized in that: In the message signing phase, the roadside unit and the vehicle user jointly generate a message signature for identity authentication between vehicles, which specifically includes the following steps: Vehicle User V i Randomly select parameter μ i 、r il 、r i2 , and set the randomization parameter C i =μ i ρ i , calculate the pseudonym AID i =RID i t2, calculate the signature key β i and γ i , that is, β i =r il H(AID i ||C i ||M i )+r i2 P,γ i =r il μ i (modq); where RID i For vehicle users V i Real ID of M i For the message sent; Vehicle User V i The signing key (β i , γ i ) is sent to the roadside unit; The roadside unit signs the message sent to the vehicle, that is, T i =x1β i +x2yγ i P and sends it to the vehicle user; x1, x2 are the private master keys of the roadside unit; y is a finite field Random parameters of The vehicle user receives the T sent by the roadside unit i , calculate the final signature information Vehicle User V i Get the roadside unit to message M i The signature σ i ={AID i ,M i ,S i ,C i }.

4. The blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method according to claim 3 is characterized in that: The process of verifying a signed message includes: In single message verification, the real identity ID of the message receiver verifies the signature of the message senderσ i , if e ^ (S i ,P)=e ^ (C i ,P pub2 ) ^ (H(AID i ||C i ||M i ),P pub1 ) is established, the verification process is passed, indicating that the vehicle identity is legal, and the message M is received i ; Otherwise, the message will be rejected; In the batch message verification process, the message receiver verifies the validity of n messages at the same time. This proves that the message signatures are valid and that the verifier received the messages.

5. The blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method according to claim 1 is characterized in that: The vehicle node with the largest weight metric is selected as the cluster head. The calculation of the weight metric includes: ω1+ω2+ω3=1 in, Represents the vehicle node V i The weight measure of Represents the vehicle node V i The total number of vehicle nodes covered by a specific node in a one-hop neighborhood; Represents the vehicle node V i The remaining energy Represents the vehicle node V i The credibility value of The weight parameter of .

6. The blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method according to claim 5 is characterized in that: Vehicle Node V i The total number of vehicle nodes covered by a specific node in a single-hop neighborhood It is expressed as: in, Represents a slave node arrive The number of hops; N represents The set of directly connected nodes.

7. The blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method according to claim 5 is characterized in that: Vehicle Node V i The remaining energy It is expressed as: in, For Node The initial energy of For Node Energy used; E Tx (n, Distance) represents the energy required to transmit n bits of data between two nodes with a distance of Distance; E RX(n) Represents the energy of a node receiving n bits of data; E Elec is the energy consumed by the electronic circuit; E FS is the energy required by the free-space model amplifier, E MP is the energy required by the amplifier in the multipath fading model.

8. The blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method according to claim 5 is characterized in that: Vehicle Node V i Reputation It is expressed as: in, Represents the reputation of the vehicle node; express The utility parameter.

9. The blockchain-based auxiliary lightweight anonymous authentication vehicle network data sharing method according to claim 1 is characterized in that: A verification contract is embedded in the smart contract on the blockchain. The contract verifies the digital signature of the vehicle through the function Authentication(), and estimates the training accuracy of each local training model update through the function LocalModelVerification(). The data can be packaged to generate a block only when the training accuracy of the local training model update is greater than the set threshold.