A trust-driven vehicle-to-everything (V2X) security authentication and key negotiation system and method
By employing a trust-driven vehicle authentication and key negotiation method, utilizing a trusted center, a trust coordination server, and roadside units, and combining physically unclonable functions and Chebyshev mapping, this method solves the problem in existing vehicle authentication and key negotiation methods that cannot continuously correlate the legitimacy of a single session's identity with its long-term trust state. This achieves vehicle identity privacy protection and reliable traffic message authentication.
Patent Information
- Application Number
- CN202611451060.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-18
- Publication Date
- 2026-11-17
AI Technical Summary
Existing vehicle authentication and key negotiation methods cannot effectively and continuously link the legitimacy of a single session's identity with its long-term trust status. Traffic messages lack reliable evidence after authentication, and trust management is susceptible to low-credibility or conflicting evidence, thus failing to guarantee vehicle identity privacy and communication security.
A trust-driven authentication and key negotiation method is adopted, which is jointly executed by a trusted center, a trust coordination server, roadside units, and vehicle nodes. Secure session keys between vehicles and roadside units and between vehicles are established through physical non-clonable functions, fuzz extractors, and Chebyshev mappings. One-time dynamic pseudonyms are generated by combining anonymous trust indexes and Hash Chains to achieve closed-loop management of message discrimination and trust updates.
It achieves the protection of vehicle identity privacy while continuously associating the legitimacy of a single session's identity with its long-term trust status, improving the reliability of authentication results and the basis for security strategies, reducing the risk of stored secret leakage, enhancing device binding capabilities, and mitigating the impact of evidence sparsity and historical information expiration through multi-dimensional consistency scoring and trust updates.
Smart Images

Figure CN123122825A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of vehicle network communication security and trusted computing technology, specifically to a trust-driven vehicle network security authentication and key negotiation system and method. Background Technology
[0002] Vehicle-to-everything (V2X) communication provides real-time information for accident warning, road condition awareness, collaborative driving, and traffic scheduling through collaborative communication between vehicle nodes, roadside units, and traffic management infrastructure. However, because V2X uses open wireless channels, vehicles move at high speeds, and network topology changes frequently, the communication process is vulnerable to attacks such as identity impersonation, message tampering, replay attacks, man-in-the-middle attacks, privacy tracking, and injection of false traffic information.
[0003] Existing vehicle authentication and key negotiation methods typically verify the registration status of entities in the current session through dynamic pseudonyms, hash authentication, elliptic curve cryptography, or chaotic mappings, and establish session keys for both communicating parties. While these methods can prevent unregistered entities from directly accessing the system, they do not guarantee that authenticated vehicles will continue to send trusted messages in subsequent communications. Legitimate vehicles may still continuously report false information after gaining access, switch between normal and malicious behavior, or collude with other vehicles to provide misleading recommendations.
[0004] On the other hand, existing vehicle trust management methods mostly rely on historical interactions or third-party recommendations to calculate trust values, but there is a lack of stable correlation between authenticated identity, dynamic pseudonyms, traffic message quality, and long-term trust accounts. Repeated reporting may lead to the same message being accumulated multiple times, dynamic pseudonym updates may disrupt historical trust states, and low-credibility recommendations and highly conflicting evidence may manipulate recommended trust. Furthermore, trust inflation is prone to occur when the amount of evidence is limited, and expired evidence may also have a long-term impact on current judgments.
[0005] Therefore, there is a need for a technical solution that can protect vehicle identity privacy while unifying identity authentication, vehicle-roadside unit and vehicle-vehicle key negotiation, authentication protection message discrimination, long-term trust update and anomaly tracking, so that the authentication result can not only be used for current session access, but also provide a reliable basis for post-authentication behavior evaluation and subsequent security strategies. Summary of the Invention
[0006] The purpose of this invention is to provide a trust-driven vehicle network security authentication and key negotiation system and method to solve the problems of existing methods that only verify the legitimacy of a single session, the difficulty in continuously associating dynamic pseudonyms with long-term trust states, the lack of reliable evidence conversion of traffic messages after authentication, and the susceptibility of recommended trust to low-credibility or conflicting evidence.
[0007] To achieve the above objectives, this invention discloses a trust-driven vehicle-to-everything (V2X) security authentication and key negotiation method. The technical solution employed involves joint execution by a Trusted Center (TA), a Trust Coordination Server (TCS), Roadside Units (RSUs), and vehicle nodes, and includes the following steps: Step 1: The Trust Center completes system initialization and entity registration, and configures the vehicle with Physically Unclonable Function (PUF) challenge and fuzz extractor auxiliary data for secret recovery, long-term Chebyshev parameters for session key negotiation, one-time dynamic pseudonyms for identity anonymity, and anonymous trust indexes for long-term trust tracing. Step 2: The vehicle recovers long-term Chebyshev private parameters using PUF and a fuzz extractor to verify the validity of the local login; Step 3: The vehicle, the trusted center, and the target roadside unit perform three-party authentication to establish a secure session key between the vehicle and the target roadside unit. At the same time, the target roadside unit obtains the vehicle's neighbor authentication parameters. Step 4: The target roadside unit synchronizes the neighbor authentication parameters with other vehicles in the coverage area. Any two vehicles perform two-way authentication and negotiate the inter-vehicle session key based on the neighbor authentication parameters and a one-time dynamic pseudonym. Step 5: The sending vehicle generates and sends an authenticated and protected traffic message using the inter-vehicle session key. After receiving the vehicle verification message, it reports the reception record using the session key between itself and the roadside unit. Step 6: The roadside unit performs multi-dimensional consistency quality judgment on the traffic messages corresponding to the reported received records, and classifies the traffic messages into multiple quality categories, including correct, incorrect or seriously abnormal, based on the judgment results. Step 7: The roadside unit converts the quality category into local trust evidence and uploads it to the trust coordination server. The trust coordination server deduplicates and merges the evidence, and updates the vehicle's global trust status. Step 8: The trust coordination server synchronizes the global trust status to the relevant roadside units. The roadside units execute a tiered response strategy based on the trust status and request the Trust Center to restore the true identity of the vehicle involved when the abnormal tracking conditions are met. By linking single-session identity verification with long-term behavior evaluation after authentication, a closed loop of authentication, message discrimination, trust update, and policy feedback is formed.
[0008] As a preferred embodiment of the present invention, in step one, the center of confidence is selected from a large prime number p and an order field. Common initial value x, one-way hash function H(·), Chebyshev mapping The fuzz extractor generates the Gen(·) algorithm, the recovery algorithm Rep(·) algorithm, and the effective time window ΔT. The main private parameter s is selected and calculated. ; vehicle Select long-term private parameters And calculate Fuzzy extractor output and long-term Chebyshev public values Construct protection value and local check value ; Trust Center establishes vehicle's true identity Registering fake identities Dynamic kana set With the anonymous trust index Protected association mapping between them; Vehicle long-term private parameters are not stored directly, but are restored on demand through PUF, fuzz extractor and protection value, thereby enhancing device binding capabilities and reducing the risk of leakage of stored secrets.
[0009] As a preferred embodiment of the present invention, in step two, the vehicle will perform a PUF challenge. Inputting PUF will get a response Using auxiliary data Perform fuzz extractor recovery Restore long-term private parameters And calculate the check value ; Only when If the local login is confirmed to be valid, the user is allowed to enter the online authentication phase using the restored long-term private parameters; otherwise, the authentication process is terminated.
[0010] As a preferred embodiment of the present invention, the vehicle-roadside unit tripartite authentication in step three includes: The vehicle generates a first authentication message M1 containing the next unused dynamic pseudonym, a vehicle random number protection field, a target roadside unit identity protection field, a vehicle temporary Chebyshev public value protection field, a request authentication tag, and a timestamp, and sends it to the Trusted Center. The Trusted Center locates the vehicle registration record based on the dynamic pseudonym, verifies M1, generates a second authentication message M2 containing a Trusted Center random number protection field, a dynamic pseudonym protection field, a vehicle temporary public value protection field, neighbor authentication auxiliary parameters, an integrity label, and a timestamp, and sends it to the target roadside unit. After verifying M2, the target roadside unit generates its own temporary Chebyshev public value and vehicle-roadside unit session key, and forms a third authentication message M3 to return to the Trust Center; After the trusted center verifies M3, it generates a fourth authentication message M4 and forwards it to the vehicle. After verifying the vehicle M4 and its key confirmation value, the vehicle-roadside unit accepts the vehicle-roadside unit session key. By using Hash Chain to generate one-time dynamic pseudonyms, the correlation between multiple communications is reduced, while the trusted center retains the ability to trace conditions.
[0011] As a preferred embodiment of the present invention, in step three, the vehicle uses temporary private parameters. Calculate temporary public values Roadside units use temporary private parameters Calculate temporary public values ; Roadside elements are calculated based on the semigroup property of the Chebyshev map. Vehicle calculation , making Then, the same vehicle-roadside unit session key is generated respectively.
[0012] As a preferred embodiment of the present invention, in step four, the roadside unit maintains a set of neighbor authentication parameters. And by connecting with the current vehicle The established session key is synchronized with the vehicle to the set of neighbor authentication parameters; Current vehicle Obtain the target vehicle from the neighbor authentication parameter set. Neighbor authentication parameters The neighbor authentication parameters are encapsulated by the Trusted Center during the vehicle-roadside unit authentication process in step three, and added to the neighbor authentication parameter set after being verified and recovered by the roadside unit.
[0013] As a preferred embodiment of the present invention, in step four, any two vehicles and According to respectively and Obtain the same long-term authentication shared value; vehicle Generate a first vehicle-to-vehicle authentication message containing its own dynamic pseudonym, the target dynamic pseudonym, its own temporary Chebyshev public value, the HashChain index, and the authentication tag, and send it to the vehicle. ,vehicle After verification, a second vehicle authentication message is generated, containing the vehicle's own dynamic pseudonym, its own temporary Chebyshev public value, and authentication tag, and returned to the vehicle. ; Both parties calculate a temporary shared value based on their own temporary private parameters and the other party's temporary public value, and generate the same vehicle-to-vehicle session key based on the long-term authentication shared value and the temporary shared value.
[0014] As a preferred embodiment of the present invention, in step six, the roadside unit adopts a unified consistency scoring function. When the deviation value Take 1 at time. Time to take ,when Take 0 at the time; The scoring function is used to obtain the speed score, step speed score, prediction position error score, event space deviation score, event numerical deviation score, semantic consistency score, and time freshness score, respectively, and the message quality comprehensive score is obtained by weighting and summing them according to the preset weights. If any of the following hard failure scenarios occur: speed exceeding the physical limit, position jump exceeding the maximum reachable distance, timestamp severely expired, or event position severely inconsistent, then regardless of the overall score, it will be judged as a serious abnormal message. Otherwise, when the overall score is greater than or equal to the correct threshold, it is judged as a correct message; When the overall score is greater than or equal to the severe anomaly threshold but less than the correct threshold, it is judged as an error message; When the overall score is less than the severe anomaly threshold, it is judged as a severe anomaly message.
[0015] As a preferred embodiment of the present invention, in step seven, the trust coordination server follows... Correct evidence regarding the vehicle erroneous evidence and serious abnormalities Perform time decay, and respectively with , and Accumulate new evidence for the corresponding category, among which ; Basic direct trust is based on prior smoothing parameters and three types of cumulative evidence , , Calculated as ; Evidence Maturation Factor With the total amount of accumulated evidence The value increases monotonically and approaches 1, eventually leading to direct trust. ; For recommendation trust, the recommendation reliability is calculated based on the positional and motion similarity between the recommended vehicle and the evaluation vehicle, combined with the amount of historical evidence of the recommended vehicle. Recommendation evidence with reliability below the threshold is discarded, and the retained evidence is converted into a quality allocation after reliability discount. The Dempster combination rule is used to fuse the evidence in order of reliability from high to low, and evidence with conflict factor greater than the threshold is skipped. Recommendation trust is obtained based on the correct quality and uncertain quality after fusion. When valid recommendation evidence exists, direct trust weights and recommendation trust weights are dynamically allocated based on the cumulative amount of direct evidence. The weighted average is then used to obtain the comprehensive trust value. The vehicle's trustworthiness, medium trustworthiness, or untrustworthiness level is determined based on the trustworthiness threshold and the untrustworthiness threshold. Finally, the vehicle's node label is determined based on whether the proportion of observers whose relationship-level comprehensive trust is below the pairwise threshold reaches the node threshold.
[0016] A trust-driven vehicle-to-everything (V2X) security authentication and key negotiation system, comprising: The Trust Center is used to perform system initialization, entity registration of vehicles and roadside units, dynamic pseudonym lifecycle management via Hash Chain, auxiliary verification during the authentication process, and recovery of the vehicle's true identity based on dynamic pseudonym mapping when responding to anomaly tracking requests. The trust coordination server is used to maintain anonymous trust accounts that are separate from the vehicle's real identity. It performs global deduplication and relation-level deduplication on the received local trust evidence, and performs time decay, direct trust calculation, evidence maturity adjustment, recommendation evidence screening based on reliability discount and Dempster-Shafer fusion to generate the vehicle's comprehensive trust value, trust level and node label. The roadside unit is used to participate in vehicle-roadside unit three-party authentication, synchronize neighbor authentication parameters to vehicles within the coverage area, receive authentication protection traffic messages reported by vehicles, perform five-dimensional consistency quality judgment, convert the judgment results into local evidence for uploading, and receive and execute forwarding, rate limiting or revocation policies fed back by the trust coordination server. Vehicle nodes are used to recover long-term private parameters through PUF and fuzz extractor, participate in vehicle-roadside unit and vehicle-to-vehicle two-way authentication, negotiate session keys, use vehicle-to-vehicle session keys to authenticate and protect traffic messages, and use vehicle-roadside unit session keys to securely report received records. The trusted center, trust coordination server, roadside unit, and vehicle node cooperate with each other to execute the above method.
[0017] Compared with the prior art, the beneficial effects of the present invention are: the present invention connects the single-session identity legitimacy verification with the long-term behavior evaluation after authentication, forming a closed loop of authentication, message discrimination, trust update and policy feedback.
[0018] Vehicle long-term private parameters are not stored directly, but are restored on demand through PUF, fuzz extractor and protection value, thereby enhancing device binding capabilities and reducing the risk of leakage of stored secrets.
[0019] By using Hash Chain to generate one-time dynamic pseudonyms, the correlation between multiple communications is reduced, while the trusted center retains the ability to trace conditions.
[0020] By utilizing the semigroup property of the Chebyshev chaotic map, V2R and V2V session keys are established respectively to realize third-party authentication and two-way authentication between vehicles.
[0021] By employing a five-dimensional physical and contextual consistency score combined with hard failure rules, the authenticated traffic messages are transformed into graded behavioral evidence.
[0022] By employing prior smoothing, time decay, evidence maturity factors, recommendation reliability discounting, and conflict skipping, the impact of evidence sparsity, outdated historical information, low-reliability recommendations, and collusion evaluations can be reduced.
[0023] The anonymous trust index maintains long-term trust continuity before and after dynamic pseudonym updates and supports tiered strategies such as normal forwarding, increased inspection frequency, access restriction, revocation, and real identity tracking. Attached Figure Description
[0024] Figure 1 This is a diagram of the vehicle network authentication and dynamic trust feedback system architecture of the present invention; Figure 2 This is a flowchart of the vehicle-roadside unit three-party authentication and key negotiation process of the present invention; Figure 3 This is a flowchart of the vehicle-to-vehicle two-way authentication and key negotiation process of the present invention; Figure 4 This is a flowchart illustrating the quality assessment, trust update, and policy feedback processes for authentication protection messages in this invention. Figure 5 This is a schematic diagram illustrating the vehicle-to-everything (V2X) deployment scenario and entity interaction of the present invention; Figure 6 This is the vehicle trust level state transition diagram of the present invention; Figure 7 This is a comparison chart of the communication and computational overhead of the present invention and existing solutions. Detailed Implementation
[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0026] Example 1 like Figures 1 to 7 As shown, this invention discloses a trust-driven vehicle network security authentication and key negotiation method, the technical solution of which is: 1. The method, jointly executed by the Trusted Center (TA), Trust Coordination Server (TCS), Roadside Unit (RSU), and vehicle nodes, includes: Step 1: The Trust Center generates common parameters for the Chebyshev chaotic mapping and completes the registration of roadside units and vehicles; it also configures vehicles with physical non-cloning function challenges, fuzz extractor auxiliary data, long-term Chebyshev parameters, HashChain state, one-time dynamic pseudonyms, and anonymous trust indexes that are separated from real identities. Step 2: The vehicle uses the physical unclonable function and fuzz extractor to recover the stable secret, thereby recovering the long-term Chebyshev private parameters and verifying them locally. Step 3: The vehicle, the trusted center, and the target roadside unit perform vehicle-roadside unit three-way authentication. The vehicle and the target roadside unit calculate the same vehicle-roadside unit session key. The target roadside unit simultaneously obtains the vehicle's neighbor authentication parameters and dynamic pseudonym-anonymity trust index binding item. Step 4: The target roadside unit synchronizes the neighbor authentication parameters (excluding itself) to vehicles within the coverage area via the vehicle-roadside unit session key. Any two vehicles perform bidirectional authentication based on the neighbor authentication parameters, one-time dynamic pseudonym, and temporary Chebyshev public value and calculate the same vehicle-to-vehicle session key. Step 5: The sending vehicle uses the vehicle-to-vehicle session key to generate a traffic message authentication tag, and the receiving vehicle verifies the tag and then uses the session key between itself and the roadside unit to report the reception record. Step 6: The roadside unit calculates kinematic consistency, spatial consistency, numerical consistency, semantic consistency, and temporal consistency scores for traffic messages that have passed the basic legality check, and classifies the traffic messages into correct messages, error messages, or seriously abnormal messages. Step 7: The roadside unit converts the message quality category along with the sending relationship into local trust evidence and uploads it to the trust coordination server. The trust coordination server performs time decay, direct trust update, evidence maturity adjustment, recommended reliability discount and evidence fusion after deduplication of the evidence to obtain the comprehensive trust value, trust level and node label. Step 8: The trust coordination server synchronizes the trust status to the relevant roadside units, which then perform message forwarding, frequency adjustment checks, access restriction or revocation policies, and request the Trust Center to track the vehicle's true identity based on dynamic pseudonym mapping when abnormal conditions are met.
[0027] A trust-driven vehicle-to-everything (V2X) security authentication and key negotiation system employs the following technical solution: (1) System composition and symbol conventions like Figure 1 As shown, the system includes a Trusted Center (TA), a Trust Coordination Server (TCS), and roadside units. And vehicle nodes. TA is a trusted registration and authentication auxiliary entity, responsible for public parameter generation, entity registration, dynamic pseudonym mapping maintenance, registration status checking, and abnormal vehicle tracking; TCS maintains anonymous trust accounts separated from real identities, receives local evidence uploaded by RSU, and updates the global trust state; RSU is responsible for access authentication within the coverage domain, neighbor authentication parameter synchronization, traffic message reception, local quality judgment, and evidence uploading; vehicle nodes are equipped with onboard units (OBU) and PUFs for local secret recovery, V2R authentication, V2V authentication, and authentication protection message transmission.
[0028] (2) System initialization and entity registration TA based on safety parameters Choose a large prime number p and the order field. Common initial value x, one-way hash function H(·), Chebyshev mapping The fuzz extractor Gen(·), Rep(·), and the effective timestamp window ΔT; randomly select the master private parameters. Its long-term common parameters are: Identity is submitted to TA via a secure channel before deployment. TA is Assign long-term private parameters ,calculate save and via a secure channel Write to the RSU secure storage area.
[0029] vehicle Submit your real identity through the registration secure channel TA assigns registration of fake identities. PUF Challenge Vehicles randomly select long-term private parameters. And calculate: Vehicles submit to TA via the registration security channel Save the registration session after TA verification. and And generate a Hash Chain of length L. The k-th chain value As a current one-time dynamic kana TA saves vehicle registration records: The vehicle saves the state of the next available chain. as well as Vehicles are not directly saved. .
[0030] TA generates an anonymous trust index for vehicles. ,maintain The protected affiliation will only include Synchronize to TCS. TCS with Create long-term trusted accounts to store three types of historical evidence, overall trust value, trust level, and node tags.
[0031] (3) Local vehicle login and secure recovery Before initiating V2R or V2V certification, vehicles will face challenges. Input PUF, get and using auxiliary data implement Then, the long-term private parameters are restored and the local checksum is calculated: like If the two are equal, then the login will be terminated; if they are equal, then... As a long-term private parameter currently available for authentication, and from Read the next unused Hash Chain value And index k+1. The vehicle only enters the online authentication phase when the local login is valid and the chain value is not used.
[0032] (4) Vehicle-Roadside Unit Tripartite Authentication and Key Negotiation vehicle , TA and according to Figure 2 Perform the exchange of four core messages M1 to M4. The current session terminates immediately if any timestamp, registration status, hash chain status, or authentication tag verification fails.
[0033] (4.1) Vehicle request generation (M1).
[0034] Vehicle Selection , And generate a timestamp T1, let ,calculate: The vehicle sends to TA .
[0035] (4.2) TA verification and forwarding (M2).
[0036] TA according to position ,calculate By the Chebyshev semigroup property, we have And restore according to the following formula , and And recalculate the authentication tag for the request: Recalculate authentication tag: TA checks timestamps, vehicle status, hash chain status, and After verification, select... And generate T2, calculate And construct: TA to send .
[0037] (4.3) RSU verification and response (M3).
[0038] calculate By the properties of semigroups And restore according to the following formula , and And recalculate the authentication label: Recalculation and authentication tags: RSU checks timestamps, registration status, and After successful verification, restore the V2V auxiliary authentication parameters: RSU is formed accordingly. And add the neighbor authentication parameter set . choose Generate T3 and calculate: Send to TA .
[0039] (4.4) TA response verification and forwarding (M4).
[0040] TA restores the temporary common value of RSU from M3 according to the following formula. and key confirmation value And recalculate the response authentication tags: TA checks timestamps, RSU registration status, and After successful verification, T4 is generated and constructed: TA sends to the vehicle .
[0041] (4.5) Vehicle key verification.
[0042] The vehicle is restored from M4 according to the following formula. and And recalculate the forwarding authentication tag: Vehicle inspection timestamp and key confirmation value and calculate By the Chebyshev semigroup property: therefore After successful authentication, the vehicle updates its local hash chain state; this session key is used for the vehicle to report security information to the RSU.
[0043] (5) Synchronization of neighbor authentication parameters and anonymous trust mapping RSU maintains the set of neighbor authentication parameters for the coverage area: For vehicles that have already established a safety reporting channel RSU passed synchronous The vehicle obtains the target vehicle from it. of This parameter is encapsulated by the TA and verified by the RSU to prevent the communicating peer from declaring its own authentication parameters.
[0044] TA also Derivate the background state synchronization key, for Confidentiality and integrity protection are implemented. A temporary mapping is established after RSU verification. When dynamic pseudonyms are updated, the old mapping is replaced with the new binding; when a vehicle leaves the coverage area, its registration expires, or it is revoked, the corresponding parameters and mappings are deleted.
[0045] (6) Vehicle-to-vehicle two-way authentication and key negotiation vehicle and according to Figure 3 Perform V2V certification. Already held , Already held Both parties completed their local logins and established connections with the RSU. and .
[0046] (6.1) Initiate a request choose Generate T1 and calculate Towards send .
[0047] (6.2) from Obtain and ,calculate And recalculate the authentication tag: examine , Unused as well as Select after verification. Generate T2 and calculate: Towards send .
[0048] (6.3) Verification and confirmation Verify timestamp, Usage status and ,calculate: when When accept By the Chebyshev semigroup property, we have ,and Therefore .
[0049] (7) Authentication protection messages and local evidence generation After successful V2V certification, the vehicle Send traffic messages Generate timestamp And certification labels: vehicle Only verify the message source and integrity. Once verification is successful, construct the reception record. and through generate Report to RSU .
[0050] RSU verifies the security reporting channel, timestamp, and session identifier, based on... Obtain and If the mapping does not exist, has expired, or the dynamic pseudonym has become invalid, the record is rejected; otherwise, message quality assessment is completed and the following is generated: TCS uses global evidence Deduplication ensures that the same message is updated only once; this is used on relation-level evidence. Deduplication allows different evaluated vehicles to form separate evidence of their relationship, but the same evaluated vehicle cannot be reported repeatedly.
[0051] (8) Message quality judgment RSU first checks the message format, timestamp freshness, and authentication status. For messages that pass the basic checks, they are evaluated across five dimensions: kinematics, spatiality, numericality, semantics, and time. The unified consistency scoring function is: Let the vehicle's position, speed, and event at time t be... , and .
[0052] Speed Score Step speed is obtained from continuous position Step speed score Predicted location Prediction error Predicted position score The kinematic consistency score is: Event space bias The deviation between the event value and the vehicle status is , Semantic score Message generation time and reception time deviation The overall quality score is: in in This is a hard failure indicator variable; it takes the value 1 if any hard failure occurs, and 0 otherwise. Message quality category is determined by the following formula: If a hard failure occurs, such as speed exceeding the physical limit, position jump exceeding the maximum reachable distance, timestamp severely expired, or event location being severely inconsistent with vehicle location, it will be directly judged as Faulty regardless of the overall score.
[0053] In an urban road implementation example, vehicles use a 1-second beacon cycle, with a positioning error standard deviation not exceeding 1.5 m, and a normal maximum speed configured at 120 km / h. Normal threshold. This represents the upper bound of the tolerance that requires no penalty, and the warning threshold. This represents the upper bound where the score of this dimension drops to zero. During deployment, it can be initialized with the 95th and 99th percentiles of the normal sample error, respectively, and then calibrated using the normal message false positive rate and abnormal message recall rate on the validation set.
[0054] The default parameters for five-dimensional message scoring are shown in the table below: The sum of the five weights is 1. Correct message threshold. Severe abnormal threshold : Output Correct; Output: Incorrect; Output Faulty. If the speed is not less than 50 m / s, the adjacent position exceeds the maximum reachable distance, the time deviation is not less than 5s, or the event position deviation is not less than 150m, then a hard failure is triggered and Faulty is output directly.
[0055] During redeployment, two total thresholds can be searched in 0.01 increments, and the set with the highest anomaly recall rate should be selected from the candidate values that meet the upper limit of the normal message false positive rate. When road speed limits, positioning accuracy, or beacon cycles change, the normal thresholds and warning thresholds for each dimension should be recalibrated synchronously.
[0056] (9) Direct Trust Update and Evidence Maturity Adjustment TCS for vehicles Maintaining three types of evidence: correct, incorrect, and seriously abnormal. , and Within adjacent update intervals Δt, the following is adopted: Attenuate historical evidence, and in accordance with Accumulated new evidence: Parameters satisfy This makes error messages and critical exception messages have a stronger negative impact.
[0057] Introducing a symmetric prior parameter α and a Faulty penalty coefficient The basic direct trust is: Let the cumulative amount of evidence, the evidence maturity factor, and the final direct trust be as follows: When the amount of evidence is small, Suppress the expansion of direct trust; as evidence accumulates, the maturation factor gradually releases basic direct trust.
[0058] The default values for trust update and decision parameters are shown in the table below: (10) Recommended Trust Integration When evaluating vehicles Regarding the target vehicle When direct evidence is insufficient, TCS introduces other observers. Relationship-level evidence. Position, motion, and context similarity are: Recommended vehicles The amount of historical evidence, uncertainty and its impact on vehicle evaluation The recommended reliability is as follows: when At that time, TCS discarded the recommendation evidence.
[0059] In the identification framework Order Convert recommendation evidence into quality assignment: TCS Press Valid recommendation evidence is ranked from highest to lowest and then fused sequentially using the standard Dempster combination rule. If the evidence to be fused conflicts with the current result... If so, skip that piece of evidence. After fusion is complete, about The definition of recommendation trust is: Calculate global recommendation trust At that time, TCS performs the same discounting and fusion process on all valid relation-level evidence; if there is no valid recommendation evidence, the recommendation trust is not included in the comprehensive trust calculation.
[0060] (11) Integrating trust, node determination and strategy feedback The overall trust strategy prioritizes direct trust and supplements it with recommended trust. When valid recommended evidence exists, weights are dynamically allocated based on the cumulative amount of direct evidence. If no valid evidence of recommendation exists, then let The new vehicle is initialized to... , .
[0061] Trust level is determined as follows: Regarding the evaluation relationship TCS replaces global evidence in the direct trust formula with relation-level evidence. , and ,get ; and then utilize other effective evaluation methods regarding vehicles Relationship-level evidence obtained .make: If no valid evidence of recommendation exists, then let The relation-level suspicious indicator variable is defined as follows: set up For the statistics window on vehicles To form a valid set of observers, The node-level malicious reporting ratio and the final label are as follows: when The previous time-lapse tag is retained. TCS feeds back the comprehensive trust, trust level, and node tag to the relevant RSU. The RSU forwards messages normally to trusted vehicles; increases message inspection frequency for vehicles with medium trust; restricts forwarding or access for untrusted vehicles; and revoks the status of vehicles that are consistently judged as malicious, and TCS submits a tracking request to TA. TA then restores the real identity based on the protected mapping between PID, RID, and VID. This forms... Figure 4 The closed loop is shown.
[0062] (12) Data structure definition This embodiment uses network byte order. All fields are unsigned fixed-endian encoded according to the length specified by the data structure, i.e., the most significant byte first. Identity identifier, dynamic pseudonym, Chebyshev mapping output, hash digest, authentication tag, session identifier, and shared and temporary public values in the key derivation process are uniformly encoded as 160 bits; timestamp, HashChain index, and expiration date are encoded as 32 bits; and status enumeration is encoded as 8 bits. The score and trust value use 16-bit Q15 fixed-point numbers during transmission and IEEE 754 double-precision floating-point numbers during internal TCS calculations.
[0063] definition This is an unsigned 160-bit fixed-endian encoding of the integer z. When the actual length of the integer is less than 160 bits, zeros are padded to the high bits.
[0064] When multiple fields are involved in a hash operation, the fixed-length binary codes of each field are directly concatenated according to the field order specified in the protocol. The hash function is defined as follows: That is, the SHA-256 algorithm is used and the leftmost 160 bits are truncated.
[0065] Used to generate a 352-bit mask of the same length as the auxiliary parameter packet; its output length is not affected by ordinary hash functions. The 160-bit truncation rule limit.
[0066] The vehicle registration record has a fixed 1160-bit portion, and the RSU registration record has a fixed 360-bit portion. The mapping between the real identity and TrustID is protected and stored by TA, and cannot be directly read by ordinary vehicles or ordinary RSUs.
[0067] The vehicle registration record data structure is as follows: A single neighbor parameter item, including expiration time and status, is 552 bits. Set N A hash table with PID as the key is used, and the average lookup complexity is O(1). Entries are deleted when a dynamic pseudonym is used, a vehicle leaves the domain, or its status is revoked.
[0068] Single neighbor parameter Ω m The data structure is as follows: The relation-level evidence payload is fixed at 712 bits, with an additional 160-bit authentication tag in the outer RSU-TCS secure channel; the trusted account header is fixed at 656 bits. The global deduplication key is {TrustID}. i The relation-level deduplication key is {TrustID}, msgHash, messageTime}. j TrustID i , msgHash, sessionID}.
[0069] (13) Key negotiation: verify that the SK of both parties are equal To facilitate manual recalculation, smaller prime numbers are used. , The temporary private parameters for vehicles and RSUs are as follows: , The result is obtained from the Chebyshev polynomial recursion: Temporary sharing values are calculated separately for the RSU side and the vehicle side: Both are equal to ,therefore .
[0070] Let the example integer values for the dynamic pseudonym, RSU identity, and two temporary public values be 101, 77, 3, and 17, respectively. The shared value, dynamic pseudonym, RSU identity, and temporary public value are all encoded using unsigned 160-bit fixed-endian encoding. The encodings for each field are as follows: The five fields are concatenated directly in the order specified by the session key formula to form the hash input of the session key: The total length of the input is 5 × 160 = 800 bits, or 100 bytes. Both parties calculate: The complete SHA-256 digest before truncation is: RSU further calculates the key confirmation value. At this point, SK participates in the concatenation as its original 20-byte binary value, and must not be converted to 40-character hexadecimal text. The total length of the input is 20 + 3 × 20 = 80 bytes, and the key confirmation value is: Since both parties receive the same temporary shared value and use the same field order, fixed-endian encoding, and truncation rules, therefore And the key confirmation value recalculated by the vehicle is the same as The two must be equal. Any difference in dynamic pseudonyms, identities, temporary public values, field encodings, or concatenation order will result in key verification failure and session rejection.
[0071] Security boundary: p=23 is only used to demonstrate the human computability of equations and hash inputs, and does not possess practical cryptographic security. It must not be used for production deployment. Actual deployment should use large prime numbers that meet the target security strength and extended Chebyshev parameters that have undergone security assessment.
[0072] (14) Message rating: Incorrect vs. Faulty The first exception message did not trigger a hard failure. The scores for velocity, step speed, and predicted position error are 1.00, 0.90, and 0.825, respectively; therefore, the kinematic consistency score is: The event location deviation is 70 m, corresponding to The numerical deviation for velocity-related events is 6.85 m / s, corresponding to... Semantic consistency The message is delayed by 1.4 seconds, corresponding to... .
[0073] because The message was determined to be Incorrect.
[0074] The five-dimensional scores for the second abnormal message are 0.45, 0.35, 0.40, 0, and 0.80, with a total score of [missing information]. Therefore, it is judged as Faulty. If its reporting speed reaches 56 m / s, even if the weighted total score is higher than 0.60, it will trigger a hard failure and be judged as Faulty because it exceeds the physical limit of 50 m / s.
[0075] (15) Ten consecutive rounds of credit changes Assuming a 1-second interval between rounds, initially... The ten message categories are C, C, C, C, C, I, I, F, F, C; where C, I, and F represent Correct, Incorrect, and Faulty, respectively. To demonstrate the process by which direct evidence gradually takes over decision-making, a discounted neutral recommendation trust is assumed. Always valid. Press first in each round. We reduce historical evidence and then add evidence from the current round.
[0076] The changes in overall trust over ten rounds are shown in the table below: The first two rounds Slightly lower than the initial value, this represents the intentional suppression of the evidence sparsity phase by the maturation factor; at this point, the total amount of evidence is less than... Maintain the Medium and EvidenceInsufficient states; do not perform an undo. Fifth round. It rose to 0.5191; consecutive Incorrect and Faulty errors caused it to drop to 0.4524 in the ninth round; in the tenth round, a Correct error only restored it to 0.4642, which could not immediately offset the two Faulty errors. In the absence of valid evidence of recommendation, the original rules apply. However, the minimum evidence threshold is still used to prevent high-risk actions from being triggered during the insufficient evidence stage.
[0077] (16) Comparison of safety and performance (16.1) Attack protection and boundary The following analysis is based on the Dolev-Yao network attacker model: attackers can listen to, replay, insert, and modify messages in public channels, but cannot directly crack hash functions, PUFs, or Chebyshev hard problems. This analysis is a mechanism-level security statement and does not replace formal proofs such as BAN Logic, ROR models, ProVerif, or AVISPA.
[0078] The attack types, resistance mechanisms, and applicable boundaries are shown in the table below: (16.2) Online computing and communication overhead The computational overhead is measured by the number of cryptographic operations performed by each entity during a single online authentication and session key negotiation process, excluding operations generated during system initialization, entity registration, business message processing, and background trust updates. Following the evaluation method used in Zhang et al.'s CSAP-IoD scheme (J. Zhang, Q. Cheng, X. Chen, and X. Luo, “CSAP-IoD: Achaotic map-based secure authentication protocol for Internet of Drones,” IEEE Transactions on Information Forensics and Security, vol. 20, pp. 8848-8862, 2025), the number of basic operations such as hashing, Chebyshev mapping, PUF, fuzz extractor, symmetric encryption / decryption, and ECC are first counted based on the authentication process of each scheme, and then the same execution time is uniformly applied. The time complexity of operations such as fuzzy extractor, AEAD, hashing, Chebyshev mapping, ECC, symmetric encryption / decryption, and bilinear pairing is taken from the aforementioned CSAP-IoD work, while the PUF response generation time is taken from the PUF-Chebyshev Power-IoT scheme proposed by Jin et al. (X. Jin, N. Lin, Z. Li, W. Jiang, Y. Jia, and Q. Li, “A lightweight authentication scheme for Power IoT based on PUF and Chebyshev chaotic map,” IEEE Access, vol. 12, pp. 83692-83706, 2024). The PUF response generation and fuzzy extractor recovery performed before vehicle authentication are necessary operations and are included in the online computation overhead.
[0079] Computational overhead: Comparative schemes for vehicle-RSU three-entity authentication include CSAP-IoD by Zhang et al., the scheme by Tomar and Tripathi (A. Tomar and S. Tripathi, “A Chebyshev polynomial-based authentication scheme using blockchain technology for fog-based vehicular network,” IEEE Transactions on Mobile Computing, vol. 23, no. 10, pp. 9075-9089, 2024), PSAS-MTS proposed by Dwivedi et al. (SK Dwivedi et al., “Design of a provable secure ECC and HMAC-based robust and efficient authentication scheme for maritime transportation system,” IEEE Transactions on Intelligent Transportation Systems, vol. 26, no. 12, pp. 23125-23138, 2025), and the scheme by Wang and Han (L. Wang and C. Han, “Multi-factor authentication and key agreement scheme based on PUF and Chebyshev chaotic map for wireless sensor)). (networks), Scientific Reports, vol. 16, art. 3311, 2026, and the proposal by Akram et al. (MW Akram et al., “A secure and lightweight drones-access protocol for smart city surveillance,” IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 10, pp.19634-19643, 2022).The total computation time for a single V2R certification using this technical solution is 7.204ms, which is lower than CSAP-IoD's 8.168ms, Tomar-Tripathi's 15.752ms, PSAS-MTS's 22.162ms, and Wang-Han's 10.636ms, but higher than Akram's 2.588ms. Figure 7 As shown in (e), as the scale of authentication sessions increases, the cumulative execution time of each scheme increases approximately linearly. This technical solution maintains a lower computational overhead compared to the aforementioned four schemes.
[0080] Comparative schemes for V2V two-entity authentication include Cross-domain UAV proposed by Liang et al. (S. Liang, J. Ning, B. Duo, and R. Zhou, “A lightweight authentication and key agreements scheme for cross-domain UAV networks based on extended Chebyshev chaotic map and hash chain,” in Proc. 4th International Conference on Cloud Computing, BigData Application and Software Engineering (CBASE), 2025, pp. 621-624), PUF-Chebyshev Power-IoT proposed by Jin et al. (ibid.), the scheme proposed by Li et al. (Z. Li, Z. Ju, H. Zhao, Z. Wei, and G. Lan, “A lightweight certificateless authenticated key agreement scheme based on Chebyshev polynomials for the Internet of Drones,” Sensors, vol. 25, no. 14, art. 4286, 2025), and the scheme proposed by Rajasekaran et al. (AS Rajasekaran et al., “Lightweight Chebyshev polynomial-based authentication”). and a signature framework with antenna array for IoT-enabled V2V and V2X communications, IEEE Internet of Things Journal, vol. 12, no. 24, pp. 55227-55240, 2025) and the scheme of Nyangaresi et al. (VO Nyangaresi, M. Ahmad, LA Maghrabi, and T.Althaqafi, “Cost-effective PUF and ECC-based authentication protocol for secure Internet of Drones communication,” IEEE Internet of Things Journal, vol. 12, no. 16, pp. 33844-33857, 2025). The total computation time for a single V2V authentication attempt using this technical solution is 4.655ms, lower than solutions by Li, Rajasekaran, and Nyangaresi, but higher than Cross-domain UAV and PUF-Chebyshev Power-IoT. Overall, it falls within the lower-middle range of the compared solutions. Figure 7 As shown in (b), the computational overhead of this technical solution mainly comes from Chebyshev mapping, PUF response generation, and fuzz extractor recovery; the solution avoids the use of bilinear pairing and a large number of ECC scalar multiplications, and binds the device physical characteristics, long-term authentication parameters, and temporary session information together.
[0081] Communication overhead: This technical solution only considers the fixed communication overhead generated by a single online authentication and session key negotiation. V2V authentication is compared only with the two-entity scheme, and V2R authentication is compared only with the three-entity scheme. Referring to the communication overhead evaluation method in Zhang et al.'s CSAP-IoD scheme (as above), identity, pseudo-identity, random number, hash output, authentication tag, and Chebyshev mapping output are all taken as 160 bits, timestamp and hash chain index are taken as 32 bits, and compressed ECC point is taken as 320 bits; for encrypted fields whose length varies with plaintext, the payload length specified in the original protocol is retained.
[0082] This V2V authentication solution exchanges only two online messages, with message lengths of 736 bits and 544 bits respectively, resulting in a total communication overhead of 1280 bits. Compared to Jin et al.'s PUF-Chebyshev Power-IoT, Li et al.'s solution, Rajasekaran et al.'s solution, and Nyangaresi et al.'s solution, the communication overhead is reduced by approximately 14.9%, 23.1%, 31.0%, and 47.4%, respectively. Liang et al.'s Cross-domain UAV solution has a communication overhead of 832 bits, lower than this solution, mainly due to its more compact message structure and fewer transmitted identity and status fields. Figure 7As shown in (a), as the size of the authentication session increases, this technical solution consistently maintains a lower cumulative communication overhead compared to the four solutions other than Cross-domain UAV.
[0083] This V2R authentication solution exchanges four core messages, with the vehicle, RSU, and TA handling 832 bits, 512 bits, and 1184 bits of communication respectively, for a total communication overhead of 2528 bits. This result reduces communication overhead by approximately 33.6% and 45.9% compared to the Tomar-Tripathi and Dwivedi et al.'s PSAS-MTS solutions, respectively, but is higher than Zhang et al.'s CSAP-IoD, Wang-Han, and Akram solutions. Figure 7 As shown in (d), the cumulative communication overhead of each scheme increases approximately linearly with the size of the authentication session, and this technical scheme is generally at the middle level of the three-entity scheme. If a V2V auxiliary parameter packet with a length of 352 bits is further transmitted, the complete communication volume is 2880 bits; since this parameter packet does not participate in the V2R core authentication, the main comparison result still uses 2528 bits.
[0084] In summary, this technical solution eliminates the need to repeatedly transmit long-term common vehicle parameters in messages for V2V authentication, thus maintaining low communication overhead even with dynamic pseudonyms, hash chain indexes, and two-way authentication tags. V2R authentication, requiring the participation of the vehicle, RSU, and TA, has relatively high communication volume, but it can simultaneously complete identity authentication, session key negotiation, condition tracking, and the secure distribution of subsequent V2V auxiliary authentication parameters. Figure 7 (c) and Figure 7 (f) further provides a joint comparison of single-time computation overhead and communication overhead.
[0085] (16.3) Trust detection effect With a 40% malicious node ratio, each attack scenario uses 20 independent random seeds. The following are the original simulation averages. In a fixed-pattern continuous attack, the MessageAcc, BadRecall, and Good→Bad rates of the proposed scheme are 93.17%, 86.16%, and 1.78%, respectively, while the baseline rates are 90.14%, 80.43%, and 2.61%.
[0086] The node-level results with a 40% malicious node ratio are shown in the table below: (16.4) Theoretical convergence of message correct reception / discrimination rate make This represents the proportion of messages that have been authenticated and correctly classified out of the total number of received messages, using a first-order convergent approximation: Take cold start value time constant (Consistent with node-level statistics window), and using the final MessageAcc under the condition of continuous attack with a fixed pattern and 40% malicious nodes as... The proposed solution is 0.9317, with a baseline of 0.9014.
[0087] The theoretical convergence estimate of the message correct reception / discrimination rate is shown in the table below: (17) Implementation effect and scope of application The above implementation forms a complete closed loop: TA registers entities and establishes a protected mapping between real identity, dynamic pseudonym, and anonymous trust index; vehicles complete V2R and V2V key negotiation after recovering secrets via PUF; V2V session keys protect traffic messages, and vehicle-RSU session keys protect reception records; RSU generates three types of evidence based on five-dimensional scoring; TCS deduplicates and updates direct trust, recommended trust, and comprehensive trust; the trust level is ultimately fed back to RSU for adjusting inspection frequency, forwarding permissions, and revocation policies.
[0088] The default parameters are suitable for single RSU urban roads, 1-second beacon period, and given positioning error conditions. For multi-RSU cross-domain scenarios, highways, longer message periods, or scenarios with different sensor accuracies, the physical thresholds, time decay, and minimum evidence quantity should be recalibrated. Minor and physically consistent internal forgery, majority-controllable observers, and wireless jamming remain the main limitations of this implementation.
[0089] Components not described in detail in this article are existing technologies.
[0090] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A trust-driven method for vehicle-to-everything (V2X) security authentication and key negotiation, characterized in that, This process is jointly executed by the Trusted Center (TA), Trust Coordination Server (TCS), Roadside Unit (RSU), and vehicle nodes, and includes the following steps: Step 1: The Trust Center completes system initialization and entity registration, and configures the vehicle with Physically Unclonable Function (PUF) challenge and fuzz extractor auxiliary data for secret recovery, long-term Chebyshev parameters for session key negotiation, one-time dynamic pseudonyms for identity anonymity, and anonymous trust indexes for long-term trust tracing. Step 2: The vehicle recovers long-term Chebyshev private parameters using PUF and a fuzz extractor to verify the validity of the local login; Step 3: The vehicle, the trusted center, and the target roadside unit perform three-party authentication to establish a secure session key between the vehicle and the target roadside unit. At the same time, the target roadside unit obtains the vehicle's neighbor authentication parameters. Step 4: The target roadside unit synchronizes the neighbor authentication parameters with other vehicles in the coverage area. Any two vehicles perform two-way authentication and negotiate the inter-vehicle session key based on the neighbor authentication parameters and a one-time dynamic pseudonym. Step 5: The sending vehicle generates and sends an authenticated and protected traffic message using the inter-vehicle session key. After receiving the vehicle verification message, it reports the reception record using the session key between itself and the roadside unit. Step 6: The roadside unit performs multi-dimensional consistency quality judgment on the traffic messages corresponding to the reported received records, and classifies the traffic messages into multiple quality categories, including correct, incorrect or seriously abnormal, based on the judgment results. Step 7: The roadside unit converts the quality category into local trust evidence and uploads it to the trust coordination server. The trust coordination server deduplicates and merges the evidence, and updates the vehicle's global trust status. Step 8: The trust coordination server synchronizes the global trust status to the relevant roadside units. The roadside units execute a tiered response strategy based on the trust status and request the Trust Center to restore the true identity of the vehicle involved when the abnormal tracking conditions are met.
2. The trust-driven vehicle network security authentication and key negotiation method according to claim 1, characterized in that: In step one, the center of trust is selected from a large prime number p and an order field. Common initial value x, one-way hash function H(·), Chebyshev mapping The fuzz extractor generates the Gen(·) algorithm, the recovery algorithm Rep(·) algorithm, and the effective time window ΔT. The main private parameter s is selected and calculated. ; vehicle Select long-term private parameters And calculate Fuzzy extractor output and long-term Chebyshev public values Construct protection value and local check value ; Trust Center establishes vehicle's true identity Registering fake identities Dynamic kana set With the anonymous trust index Protected association mapping between them.
3. The trust-driven vehicle network security authentication and key negotiation method according to claim 1, characterized in that: In step two, the vehicle will undergo a PUF challenge. Inputting PUF will get a response Using auxiliary data Perform fuzz extractor recovery Restore long-term private parameters And calculate the check value ; Only when If the local login is confirmed to be valid, the user is allowed to enter the online authentication phase using the restored long-term private parameters; otherwise, the authentication process is terminated.
4. The trust-driven vehicle network security authentication and key negotiation method according to claim 1, characterized in that, The vehicle-roadside unit tripartite authentication in step three includes: The vehicle generates a first authentication message M1 containing the next unused dynamic pseudonym, a vehicle random number protection field, a target roadside unit identity protection field, a vehicle temporary Chebyshev public value protection field, a request authentication tag, and a timestamp, and sends it to the Trusted Center. The Trusted Center locates the vehicle registration record based on the dynamic pseudonym, verifies M1, generates a second authentication message M2 containing a Trusted Center random number protection field, a dynamic pseudonym protection field, a vehicle temporary public value protection field, neighbor authentication auxiliary parameters, an integrity label, and a timestamp, and sends it to the target roadside unit. After verifying M2, the target roadside unit generates its own temporary Chebyshev public value and vehicle-roadside unit session key, and forms a third authentication message M3 to return to the Trust Center; After the trusted center verifies M3, it generates a fourth authentication message M4 and forwards it to the vehicle. After verifying the vehicle M4 and its key confirmation value, the vehicle-roadside unit accepts the vehicle-roadside unit session key.
5. The trust-driven vehicle network security authentication and key negotiation method according to claim 1, characterized in that: In step three, the vehicle uses temporary private parameters. Calculate temporary public values Roadside units use temporary private parameters Calculate temporary public values ; Roadside elements are calculated based on the semigroup property of the Chebyshev map. Vehicle calculation , making Then, the same vehicle-roadside unit session key is generated respectively.
6. The trust-driven vehicle network security authentication and key negotiation method according to claim 1, characterized in that: In step four, the roadside unit maintains the neighbor authentication parameter set. And by connecting with the current vehicle The established session key is synchronized with the vehicle to the set of neighbor authentication parameters; Current vehicle Obtain the target vehicle from the neighbor authentication parameter set. Neighbor authentication parameters The neighbor authentication parameters are encapsulated by the Trusted Center during the vehicle-roadside unit authentication process in step three, and added to the neighbor authentication parameter set after being verified and recovered by the roadside unit.
7. The trust-driven vehicle network security authentication and key negotiation method according to claim 1, characterized in that: In step four, any two vehicles and According to respectively and Obtain the same long-term authentication shared value; vehicle Generate a first vehicle-to-vehicle authentication message containing its own dynamic pseudonym, the target dynamic pseudonym, its own temporary Chebyshev public value, a hash chain index, and an authentication tag, and send it to the vehicle. ,vehicle After verification, a second vehicle authentication message is generated, containing the vehicle's own dynamic pseudonym, its own temporary Chebyshev public value, and authentication tag, and returned to the vehicle. ; Both parties calculate a temporary shared value based on their own temporary private parameters and the other party's temporary public value, and generate the same vehicle-to-vehicle session key based on the long-term authentication shared value and the temporary shared value.
8. The trust-driven vehicle network security authentication and key negotiation method according to claim 1, characterized in that: In step six, the roadside unit adopts a unified consistency scoring function. When the deviation value Take 1 at time. Time to take ,when Take 0 at the time; The scoring function is used to obtain the speed score, step speed score, prediction position error score, event space deviation score, event numerical deviation score, semantic consistency score, and time freshness score, respectively, and the message quality comprehensive score is obtained by weighting and summing them according to the preset weights. If any of the following hard failure scenarios occur: speed exceeding the physical limit, position jump exceeding the maximum reachable distance, timestamp severely expired, or event position severely inconsistent, then regardless of the overall score, it will be judged as a serious abnormal message. Otherwise, if the overall score is greater than or equal to the correct threshold, it is judged as a correct message; When the overall score is greater than or equal to the severe anomaly threshold but less than the correct threshold, it is judged as an error message; When the overall score is less than the severe anomaly threshold, it is judged as a severe anomaly message.
9. The trust-driven vehicle network security authentication and key negotiation method according to claim 1, characterized in that: In step seven, the trust coordination server follows... Correct evidence regarding the vehicle erroneous evidence and serious abnormalities Perform time decay, and respectively with , and Accumulate new evidence for the corresponding category, among which ; Basic direct trust is based on prior smoothing parameters and three types of cumulative evidence , , Calculated as ; Evidence Maturation Factor With the total amount of accumulated evidence The value increases monotonically and approaches 1, eventually leading to direct trust. ; For recommendation trust, the recommendation reliability is calculated based on the positional and motion similarity between the recommended vehicle and the evaluation vehicle, combined with the amount of historical evidence of the recommended vehicle. Recommendation evidence with reliability below the threshold is discarded, and the retained evidence is converted into a quality allocation after reliability discount. The Dempster combination rule is used to fuse the evidence in order of reliability from high to low, and evidence with conflict factor greater than the threshold is skipped. Recommendation trust is obtained based on the correct quality and uncertain quality after fusion. When valid recommendation evidence exists, direct trust weights and recommendation trust weights are dynamically allocated based on the cumulative amount of direct evidence. The weighted average is then used to obtain the comprehensive trust value. The vehicle's trustworthiness, medium trustworthiness, or untrustworthiness level is determined based on the trustworthiness threshold and the untrustworthiness threshold. Finally, the vehicle's node label is determined based on whether the proportion of observers whose relationship-level comprehensive trust is below the pairwise threshold reaches the node threshold.
10. A trust-driven vehicle-to-everything (V2X) security authentication and key negotiation system, characterized in that, include: The Trust Center is used to perform system initialization, entity registration of vehicles and roadside units, dynamic pseudonym lifecycle management of Hash Chain, auxiliary verification during the authentication process, and to restore the true identity of vehicles based on dynamic pseudonym mapping when responding to anomaly tracking requests. The trust coordination server is used to maintain anonymous trust accounts that are separate from the vehicle's real identity. It performs global deduplication and relation-level deduplication on the received local trust evidence, and performs time decay, direct trust calculation, evidence maturity adjustment, recommendation evidence screening based on reliability discount, and Dempster-Shafer fusion to generate the vehicle's comprehensive trust value, trust level, and node label. The roadside unit is used to participate in vehicle-roadside unit three-party authentication, synchronize neighbor authentication parameters to vehicles within the coverage area, receive authentication protection traffic messages reported by vehicles, perform five-dimensional consistency quality judgment, convert the judgment results into local evidence for uploading, and receive and execute forwarding, rate limiting or revocation policies fed back by the trust coordination server. Vehicle nodes are used to recover long-term private parameters through PUF and fuzz extractor, participate in vehicle-roadside unit and vehicle-to-vehicle two-way authentication, negotiate session keys, use vehicle-to-vehicle session keys to authenticate and protect traffic messages, and use vehicle-roadside unit session keys to securely report received records. The trusted center, trust coordination server, roadside unit, and vehicle node cooperate with each other to perform the method as described in any one of claims 1 to 9.
