Dynamic reputation evaluation and PBFT hierarchical consensus optimization system and method oriented to Internet of Vehicles

Abstract

The invention discloses a dynamic reputation evaluation and PBFT hierarchical consensus optimization system and method oriented to the Internet of Vehicles. The dynamic reputation evaluation system for the Internet of Vehicles comprises a network construction module, a comprehensive reputation value calculation module, a probability calculation module, a node selection module and a malicious behavior detection and punishment module. The PBFT hierarchical consensus optimization system comprises a network construction module, a supervision framework deployment module and a double-layer supervision module. According to the method, a double-layer interaction supervision framework is introduced to construct an Internet of Vehicles alliance chain, a reputation evaluation model and a dynamic reputation balance mechanism are integrated, and the method which is low in malicious node detection rate, poor in data sharing safety and low in consensus efficiency in the Internet of Vehicles is optimized.

Description

Technical Field

[0001] The present invention belongs to the interdisciplinary field of blockchain technology and secure communication in the Internet of Vehicles (IoV), and specifically relates to a dynamic reputation evaluation and PBFT layered consensus optimization system and method for the IoV. Background Art

[0002] In recent years, the deep integration of blockchain technology and the Internet of Vehicles (IoV) has provided innovative solutions to core challenges such as data storage security, vehicle identity verification, and privacy protection. However, its technical characteristics have also posed a series of new challenges. Research focuses on three key areas: first, building efficient mechanisms for identifying and isolating malicious nodes; second, dynamically protecting user privacy within the open and transparent nature of blockchain; and third, improving the real-time and scalability of the consensus process. Notably, the Practical Byzantine Fault Tolerance (PBFT) consensus mechanism, due to its efficient consensus performance and low latency, has become a research hotspot for improving data credibility and attack resistance.

[0003] The PBFT (Practical Byzantine Fault Tolerance) algorithm is a widely used consensus algorithm in consortium blockchains. It can ensure secure and reliable consensus even with up to one-third of malicious nodes. PBFT also offers high transaction throughput and determinism, making it particularly suitable for enterprise-level blockchain applications requiring high performance and high reliability. However, PBFT has limitations on the number of participating nodes. When the number of nodes increases, communication and computational overhead increases dramatically. Furthermore, PBFT relies on a master node to coordinate the consensus process, but lacks an effective mechanism for identifying malicious nodes. If a malicious node becomes the master node, it could pose a threat to the security and stability of the system.

[0004] In the Internet of Vehicles (IoV) environment, vehicle nodes (OBUs) and roadside units (RSUs) need to efficiently share real-time traffic data. However, the traditional PBFT consensus mechanism has the following problems: insufficient node participation enthusiasm, the fixed master node leads to centralization, and ordinary nodes lack the motivation to participate; data quality is difficult to guarantee, and malicious nodes may tamper with or forge data, threatening network security; dynamic adaptability is poor, and the high mobility of IoV nodes leads to frequent topology changes, making traditional static reputation models unable to effectively evaluate node behavior. In existing technologies, some improved solutions use reputation models to screen nodes, but lack dynamic weight adjustment and malicious behavior suppression mechanisms, and the incentive mechanism is simple. Therefore, a consensus method that adapts to the dynamic characteristics of the IoV is needed to improve the reliability of data sharing and the robustness of the system. Summary of the Invention

[0005] In order to solve the problems of frequent malicious behaviors and low efficiency of traditional consensus mechanisms when sharing data in the Internet of Vehicles environment, the present invention proposes a dynamic reputation evaluation and PBFT layered consensus optimization system and method for the Internet of Vehicles.

[0006] A dynamic reputation evaluation system for the Internet of Vehicles, which achieves one of the objectives of the present invention, includes:

[0007] Network construction module: used to build a multi-node IoV alliance chain network consisting of roadside units (RSUs) and onboard units (OVUs); nodes include fixed nodes and dynamic nodes; multiple RSUs are deployed as fixed nodes of the IoV alliance chain network at main road intersections and transportation hubs as needed; when a vehicle enters the IoV alliance chain network and successfully registers with the network, the OVU becomes a dynamic node of the IoV alliance chain network;

[0008] Comprehensive reputation calculation module: used to calculate the comprehensive reputation of each node based on the performance data and behavior data of each node in the Internet of Vehicles alliance chain network; the performance data includes: response delay, network-wide average response delay, and network-wide delay standard deviation; the behavior data includes: the number of data uploads by the node and the total number of data uploads by the node;

[0009] Probability calculation module: used to calculate the node selection probability of each node based on the comprehensive reputation value of each node;

[0010] Node selection module: used to select a consensus node based on the selection probability of each node, and select a master node from the consensus nodes according to the PBFT protocol;

[0011] Supervisory node group selection module: used to select on-board units that meet the set conditions to form multiple supervisory node groups;

[0012] The set conditions include: the comprehensive reputation value is greater than or equal to the set value; there is no violation record in the recent multiple cycles; the comprehensive reputation value is calculated based on the performance data and behavior data of each node in the Internet of Vehicles Alliance Chain network; the performance reputation value is obtained based on the performance data; the behavior reputation value is obtained based on the behavior data, and the comprehensive reputation value is obtained by weighted calculation of the performance reputation value and the behavior reputation value. The calculation method of the performance reputation value includes:

[0013]

[0014] Where R p,i represents the performance reputation value of the i-th node; d i 、dV i , σ i They represent the response delay of the i-th node, the average response delay of the entire network, and the standard deviation of the entire network delay respectively;

[0015] The behavioral reputation value is calculated as follows, where the number of successful data uploads and the total number of data uploads are the number from the previous cycle to the current time;

[0016]

[0017] Malicious Behavior Detection and Punishment Module: This module monitors master nodes through multiple supervisory node groups. If a master node exhibits malicious behavior, its overall reputation value is reduced. Malicious behavior refers to any behavior by a node in the IoV consortium chain that endangers data security, consensus efficiency, or system stability, such as data forgery, denial of service, or forged control instructions.

[0018] A further technical solution is: the calculation method of the comprehensive reputation value of the node includes:

[0019] The node's performance reputation value is calculated based on the node's response delay, the average response delay of the entire network, and the standard deviation of the entire network delay;

[0020] The node's behavioral reputation value is calculated based on the number of successful node data uploads and the total number of data uploads;

[0021] The performance reputation value and the behavior reputation value are weighted and summed to obtain the comprehensive reputation value of the node.

[0022] A further technical solution is: further comprising a reputation adjustment module: for dynamically adjusting the comprehensive reputation value of each node, the adjustment method includes:

[0023] When the comprehensive reputation value of a node is greater than the set upper limit and there is no new good behavior in multiple consecutive cycles, the reputation value is adjusted according to the following formula:

[0024] R new,i =R total,i -0.005·(R total,i -70) 2

[0025] When the comprehensive reputation value of a node is less than the set lower limit and there is no violation in multiple consecutive cycles, the reputation value is adjusted according to the following formula:

[0026]

[0027] Where R new,i represents the updated comprehensive reputation value of the i-th node; R total,i represents the updated comprehensive reputation value of the i-th node before adjustment; t is the number of violation-free cycles.

[0028] Good behavior refers to node behavior that contributes to the stable, secure, and efficient operation of the system. For example: responding to system tasks or other node requests within the specified time; ensuring that the traffic information provided by the node (such as accident and congestion information) matches the data from the RSU or other nodes; and proactively detecting and preventing the spread of forged, tampered, or fraudulent data (such as denial of service attack protection). Illegal behavior refers to the behavior of a node providing false, incomplete, or unauthorized data in data sharing.

[0029] A further technical solution is: the calculation method of the node selection probability includes:

[0030] Normalizing the comprehensive reputation value of the node to obtain the normalized comprehensive reputation value of the node;

[0031] The mapping formula is used to generate the first initial parameter α of the Beta distribution for the normalized node's comprehensive reputation value. base,i and the second initial parameter β base,i ;

[0032] The first initial parameter α is adjusted based on the interval of the node's comprehensive reputation value before normalization. base,i and the second initial parameter β base,i Perform adjustment to obtain an adjusted first parameter α and a second parameter β;

[0033] The node selection probability of the node is calculated according to the adjusted first parameter α and second parameter β and the contribution weight of the node's performance reputation and the contribution weight of the behavior reputation.

[0034] A further technical solution is: the calculation method of the node selection probability includes:

[0035]

[0036] Where w 1,i and w 2,i are the contribution weights of the performance reputation and behavior reputation of the i-th node respectively; α i and β i represent the adjusted first parameter and second parameter respectively.

[0037] A further technical solution is: the method for selecting a consensus node according to the node selection probability includes:

[0038] A hash algorithm is used to randomly divide all nodes registered in the Internet of Vehicles alliance chain network into multiple groups. The nodes in the group are sorted according to the probability of node selection, and the top-ranked nodes are selected as consensus nodes.

[0039] A further technical solution is: in the malicious behavior detection and punishment module, the method of reducing the comprehensive reputation value of the master node with malicious behavior includes:

[0040]

[0041] In the formula, PunishValue represents the penalty value, and R1 is the initial value of the penalty value, which can be set according to actual needs. If you want to slowly reduce the comprehensive reputation value, you can set the value of R1 to a smaller value, such as a positive number less than 10 or 5; otherwise, you can set it to a larger value. t =P1×PunishValue t-1 ; P1 is a positive number greater than 1, and the preferred value is 1.5.

[0042] Furthermore, if the comprehensive reputation value of the punished node R new,i <0, then force setting R new,i If it is 0, it means that the node is frozen. The frozen node needs to recalculate its behavior reputation value R b,i and performance reputation value R p,i Restore comprehensive reputation value R total,i .

[0043] Furthermore, it also includes a comprehensive reputation value recovery module for restoring the reputation of nodes with a comprehensive reputation value of 0. The recovery method includes:

[0044] The performance reputation value R is calculated according to the following formula p,i :

[0045]

[0046] The behavioral reputation value R is calculated according to the following formula b,i :

[0047]

[0048] The comprehensive reputation value R is calculated according to the following formula new,i :

[0049] R total,i =w p,i ×R p,i +w b,i ×R b,i

[0050] d i 、dV i , σ i They represent the response delay of the i-th node, the average response delay of the entire network, and the standard deviation of the entire network delay respectively; w p,i and w b,i They represent the weights of the performance reputation value and behavior reputation value of the i-th node respectively.

[0051] A dynamic reputation evaluation method for the Internet of Vehicles (IoV) to achieve the second objective of the present invention includes:

[0052] Build a multi-node Internet of Vehicles alliance chain network consisting of roadside units and on-board units;

[0053] Calculate the comprehensive reputation value of each node based on the performance data and behavior data of each node in the Internet of Vehicles alliance chain network;

[0054] Calculate the node selection probability of each node based on the comprehensive reputation value of each node;

[0055] Select a consensus node based on the selection probability of each node, and select a master node from the consensus nodes according to the PBFT protocol;

[0056] Selecting on-board units that meet the set conditions to form multiple supervisory node groups;

[0057] The supervisory node group monitors the master node, and when the master node has malicious behavior, the comprehensive reputation value of the master node with malicious behavior is reduced.

[0058] A PBFT hierarchical consensus optimization system for achieving the third objective of the present invention includes:

[0059] Network construction module: used to build a multi-node Internet of Vehicles consortium chain network consisting of roadside units and on-board units;

[0060] Supervision framework deployment module: used to deploy each node in the IoV alliance chain network to obtain a two-layer interactive supervision framework; the edge layer of the two-layer interactive supervision framework is composed of all nodes that are roadside units; the center layer of the two-layer interactive supervision framework is composed of a plurality of selected nodes that are onboard units;

[0061] Two-layer supervision module: used to build multiple cross-regional supervision node groups using the central layer, and the cross-regional supervision node groups supervise the behavior of all edge layer nodes; also used to use the edge layer to monitor the node behavior of the central layer nodes in the geographical area.

[0062] Furthermore, the method for building an Internet of Vehicles alliance chain network includes:

[0063] Based on the geographical scope of administrative streets and base station coverage density, the Internet of Vehicles is divided into N geographical areas. In each geographical area, multiple roadside units are deployed as fixed core nodes, distributed at main road intersections and transportation hubs. Multiple vehicle-mounted units are used as dynamic participating nodes, and traffic data is collected in real time through vehicle-mounted sensors.

[0064] The roadside unit is authenticated through the alliance chain management contract. After the identity authentication is passed, the roadside unit becomes the core node of the Internet of Vehicles alliance chain network.

[0065] When a vehicle enters a geographical area under the jurisdiction of the Internet of Vehicles alliance chain network, the vehicle's on-board unit sends a registration request to the roadside unit in that geographical area. After the identity certificate is verified, the on-board unit is registered as a participating node of the Internet of Vehicles alliance chain network.

[0066] Furthermore, the method for deploying each node in the Internet of Vehicles alliance chain network includes:

[0067] The roadside units registered in the IoV alliance chain network are divided into the edge layer, which is used to collect the behavior data of the on-board units of vehicles in the geographical area where the roadside unit itself is located and detect abnormal behavior;

[0068] During each supervision cycle, the roadside unit extracts nodes registered as on-board units and meeting the set conditions from the Internet of Vehicles alliance chain network to form a cross-regional supervision node group to supervise the behavior of the nodes in the edge layer of the Internet of Vehicles alliance chain network; the set conditions include: the comprehensive reputation value is greater than or equal to the set value (such as 70); there is no violation record in the last multiple (such as 3) cycles (verified by the alliance chain audit traceability chain); the comprehensive reputation value is calculated based on the performance data and behavior data of each node in the Internet of Vehicles alliance chain network.

[0069] Furthermore, the method for the cross-regional supervisory node group to supervise the behavior of the nodes in the edge layer of the Internet of Vehicles alliance chain network includes:

[0070] When an edge layer roadside unit node exhibits abnormal behavior, the period of abnormal behavior is marked as a suspicious period, and a cross-regional supervision process is initiated for the behavior of the roadside unit node. The abnormal behavior includes: multiple consecutive data interaction delays exceeding the threshold (500ms); multiple consecutive failures of business data verification code matching; and receiving collaborative warnings from other edge layer roadside units. The cross-regional supervision process includes:

[0071] Multiple nodes from different groups are randomly selected from all cross-regional supervisory node groups and anonymously verified for abnormal voting behavior using ring signature technology. This ring signature technology combines the public keys of each node into a signature ring, generating an anonymous signature that cannot be traced back to a specific node. The signature includes the hash value of the voting vector (such as SHA-256 (Vote1, Vote2, Vote3), where Vote represents the voting content), ensuring that the voting content is verifiable but not tampered with.

[0072] The anonymous voting verification process includes: performing a hash check on the roadside unit's behavior data during the questionable period to determine whether the data is complete and has not been tampered with; using a machine learning anomaly detection model to compare the questionable behavior with the roadside unit's historical behavior baseline and calculate the degree of deviation (e.g., using an isolation forest algorithm to identify outliers); and checking whether the questionable behavior conforms to the malicious behavior signature library defined in the Internet of Vehicles security protocol specification (e.g., data forgery, denial of service, etc.).

[0073] The final voting value of each node is the logical OR result of multi-dimensional verification. If any dimension passes, it will be counted as 1 vote, and if all dimensions fail, it will be counted as 0 votes. For example, if the verification result of a node is [1,0,0] (only data integrity (no tampering) passes), the final voting value is equal to 1.

[0074] When more than the set percentage of nodes vote to confirm that the roadside unit has abnormal behavior, the cross-regional supervision team will initiate a secondary review. The secondary review includes:

[0075] Synchronize abnormal behavior data from edge-layer roadside unit nodes to all cross-regional monitoring groups. During the synchronization process, zero-knowledge proof (ZKP) is used to encapsulate abnormal behavior data, revealing only the minimum information required for verification to the OBU node, thus avoiding the leakage of original data content.

[0076] Each monitoring group executes the PBFT consensus algorithm to generate a group verification conclusion on whether the roadside unit has malicious behavior;

[0077] When the internal verification conclusion of at least two different supervisory groups is that the suspected abnormal behavior is confirmed to be abnormal, obtaining dual signatures of the two different supervisory groups;

[0078] The dual signature is verified, and when the dual signature verification passes, it is considered that the behavior of the roadside unit is abnormal.

[0079] Furthermore, the standard for passing the double signature verification is: the public key of the signature belongs to the valid supervision group node within the period, and the consistency of the verification conclusions of two different supervision groups is greater than or equal to a set ratio (such as 90%); the consistency of the verification conclusion is calculated based on the cosine similarity of the behavioral feature vector.

[0080] Furthermore, the method of using the edge layer to monitor the node behavior of the central layer nodes in the geographical area includes:

[0081] Collect node behaviors of central layer nodes, including data upload frequency (number of uploads per minute), response delay (data transmission time from vehicle onboard unit (OBU) to roadside unit (RSU), and effective response rate (effective data volume / total transmitted data volume);

[0082] When the response delay is detected to fluctuate, if it exceeds the network average delay multiple times (such as 2 times) for multiple times (such as 3 times) in a row, the node is considered to have suspected abnormal behavior.

[0083] A non-transitory computer-readable storage medium that implements the fourth objective of the present invention has a computer program stored thereon, and when the computer program is executed by a processor, the steps of the dynamic reputation evaluation and PBFT layered consensus optimization method for the Internet of Vehicles are implemented.

[0084] A computer program product for achieving the fifth objective of the present invention includes a computer program / instruction, which, when executed by a processor, implements the steps of the dynamic reputation evaluation and PBFT hierarchical consensus optimization method for the Internet of Vehicles.

[0085] The beneficial effects of the present invention include:

[0086] The present invention introduces a two-layer interactive supervision framework to construct an Internet of Vehicles alliance chain, a comprehensive reputation evaluation model and a dynamic reputation balance mechanism, thereby optimizing the methods with low malicious node detection rate, poor data sharing security and low consensus efficiency in the Internet of Vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 This is a flow chart of a dynamic reputation evaluation and PBFT layered consensus optimization method for the Internet of Vehicles in an embodiment of the present invention;

[0088] Figure 2 is a diagram of a two-layer interactive supervision framework in an embodiment of the present invention;

[0089] Figure 3 It is a flowchart of the blockchain consensus and supervision mechanism;

[0090] Figure 4 is a schematic diagram of a dynamic reputation evaluation system for the Internet of Vehicles according to an embodiment of the present invention;

[0091] Figure 5 Schematic diagram of the PBFT hierarchical consensus optimization system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0092] The following specific embodiments are provided to explain the technical solutions of the present invention so that those skilled in the art can understand the present invention. The scope of protection of the present invention is not limited to the specific implementation structures described below. Any implementation schemes created by those skilled in the art that include the technical solutions of the present invention but differ from the following specific implementation schemes are also within the scope of protection of the present invention.

[0093] The embodiment of the present invention provides a PBFT layered consensus optimization method for the Internet of Vehicles. Figures 1 to 3 As shown, the following steps are included:

[0094] S1. Build a car networking alliance chain network

[0095] The construction process includes area division and node registration.

[0096] The method of regional division includes: dividing the Internet of Vehicles into N geographical areas (such as urban administrative divisions) according to the geographical scope of administrative streets and the coverage density of 5G base stations, deploying multiple roadside units (RSUs) as fixed nodes in each geographical area, distributed at the intersections of main roads and transportation hubs, responsible for blockchain ledger storage, node authority management and communication coordination within the area; at the same time, incorporating multiple vehicle on-board units (OBUs) as dynamic nodes, and collecting traffic data (such as vehicle speed, location, road conditions) in real time through on-board sensors.

[0097] The node registration method includes: RSU registration: completing identity authentication through the consortium chain management contract, deploying a complete blockchain ledger, and authorizing it to serve as a core node of the Internet of Vehicles consortium chain network, responsible for maintaining ledger consistency within its region and reviewing OBU node access. OBU registration: When a vehicle enters the geographical area governed by the Internet of Vehicles consortium chain network, the OBU sends a registration request to the RSU in that geographical area. After verification using an identity certificate (such as a digital certificate issued by a CA), it becomes a lightweight, dynamic participating node of the Internet of Vehicles consortium chain network, storing only block header information and verifying the legitimacy of transactions through the Simplified Payment Verification (SPV) mechanism, reducing storage overhead.

[0098] S2. Deploy a two-layer interactive supervision framework based on the Internet of Vehicles alliance chain network

[0099] The double layer includes an edge layer and a center layer;

[0100] The edge layer construction method includes: dividing the roadside units (RSUs) registered in the Internet of Vehicles consortium chain network into the edge layer, which is used to collect node behaviors of the on-board units (OBUs) within the geographical area where the roadside units (RSUs) are located and detect abnormal behaviors, as follows:

[0101] The node behavior includes: data upload frequency (number of uploads per minute), response delay (data transmission time from OBU to RSU), and effective response rate (effective data volume / total transmitted data volume);

[0102] The method for detecting abnormal behavior includes: monitoring the response delay fluctuation of the on-board unit OBU, and if it exceeds twice the average network delay for multiple times (such as 3 times) in a row, marking it as suspected abnormal behavior.

[0103] The method for forming the central layer includes: in each cycle, the roadside unit (RSU) randomly selects nodes with a comprehensive reputation value ≥ 80 and no violation records from the nodes registered in the Internet of Vehicles Alliance Chain network and on-board units (OBUs) as supervisory nodes. The supervisory nodes constitute the central layer, and multiple supervisory nodes are randomly selected from the central layer to form k cross-regional supervision groups (k ≥ 2, such as k = 3), each group contains 5-10 OBUs. The responsibilities of the cross-regional supervision group include:

[0104] Secondary verification: Cross-check the malicious behavior of edge layer RSUs, requiring independent confirmation by at least two supervisory groups (such as a dual signature mechanism) to prevent a single RSU from being maliciously manipulated.

[0105] Privacy protection: When malicious behavior is detected, the node identity is partially revealed through threshold signature technology (only disclosed to the alliance chain management contract), and zero-knowledge proof is used to verify the authenticity of the supervision results to prevent the leakage of original data;

[0106] Dynamic rotation: The supervisory group nodes are randomly replaced every cycle to reduce the risk of collusion. The identities of the supervisory nodes are anonymized (for example, hash values replace real IDs) to prevent targeted attacks by malicious nodes. Zero-knowledge proof is used to verify the authenticity of the supervisory results without exposing the original data.

[0107] The master node is dynamically elected by high-reputation candidate consensus nodes (with a comprehensive reputation value ≥ 70 and no violation records for three consecutive cycles) through the PBFT protocol. The master node is responsible for coordinating the consensus process (such as proposal initiation and message broadcasting) and verifying the local decisions submitted by the edge layer RSU.

[0108] When the master node detects abnormal RSU behavior (such as data inconsistency or latency exceeding a threshold), it must distribute the verification request to at least three independent supervisory nodes, triggering multi-node collaborative verification. The specific verification process includes: the supervisory node uses ring signature technology to anonymize its identity to ensure that its true identity cannot be identified by the RSU or other nodes during the verification process to prevent targeted attacks. Each supervisory node independently judges the RSU behavior (such as "compliant" or "violation") and submits the voting results to the master node through an encrypted channel. The master node summarizes the voting results, and it must meet the requirements of at least 2 / 3 of the supervisory nodes (such as at least 2 out of 3) to confirm that the RSU has engaged in malicious behavior before it can be judged as a violation.

[0109] Based on the above-constructed IoV consortium chain network, an embodiment of the present invention further provides a dynamic reputation evaluation method for IoV, comprising the following steps:

[0110] S3. Build a dynamic comprehensive reputation evaluation model for nodes to comprehensively evaluate node performance and behavior;

[0111] The node dynamic comprehensive reputation evaluation model includes: a node performance model and a behavior reputation model.

[0112] The node performance model evaluates node performance by assessing behavioral data such as the node's response delay and delay fluctuation range. The response delay of a normal node should fluctuate within a reasonable range and match the overall network status and the node's own performance. Byzantine nodes may deliberately delay responses or experience abnormal response delays, such as a sudden and significant increase in response delay or irregular fluctuations. Therefore, by setting a threshold for response delay and an upper limit for the fluctuation range, when a node's response delay exceeds these ranges, it can be identified as a suspected Byzantine node. The calculation method for each node's performance reputation value includes:

[0113]

[0114] Where: R p,i represents the performance reputation value of the i-th node; d i 、dV i , σ i They represent the response delay of the i-th node, the average response delay of the entire network, and the standard deviation of the entire network delay respectively.

[0115] The behavioral reputation model focuses on various node behaviors within the network, such as the accuracy of data uploads and compliance with network rules. Each node's behavioral reputation is calculated by quantifying data such as the quality of data contribution, compliance with rules, and interaction frequency, and then applying a weighted composite approach. A node's performance reputation and behavioral reputation are then assigned different weights, and a weighted combination of these two values is used to generate a comprehensive reputation score. As a node's performance within the network changes, its scores in both sub-models are adjusted in real time, and each node's comprehensive reputation is updated accordingly.

[0116] The formula for calculating behavioral reputation value is:

[0117]

[0118] Where: R b,i Represents the behavioral reputation value of the i-th node.

[0119] The roadside unit RSU collects the performance reputation value and behavior reputation value of each node, and calculates the reputation value R of each node based on the node performance and behavior reputation value. total,i ; The calculation methods include the following:

[0120] R total,i =w p,i ×R p,i +w b,i ×R b,i

[0121] Rtotal,i represents the comprehensive reputation value of the i-th node; R p,i and R b,i Represent the performance reputation value and behavior reputation value of the i-th node respectively; w p,i and w b,i They represent the weights of the performance reputation value and behavior reputation value of the i-th node respectively.

[0122] In the Internet of Vehicles environment, different application scenarios have different requirements for the performance and behavior of nodes, so the weight ratio of the node's performance reputation value and behavior reputation value is w p,i and w b,i Need to adjust according to the scene requirements, w p,i +w b,i =1, the present invention does not limit this. For example: in a high-security scenario, the Internet of Vehicles system has extremely high requirements for data accuracy and security, and the weight of the behavior reputation value w b,i Need to be adjusted to a higher level, such as w p,i =0.3,w b,i =0.7. In high-throughput scenarios, the IoV system needs to efficiently process large amounts of data to meet the needs of real-time traffic information sharing, vehicle navigation optimization, etc. The weight of the performance reputation value w p,i Need to be adjusted to a higher level, such as w p,i =0.6,w b,i =0.4.

[0123] The response delay of normal nodes should be significantly lower than the average level of the network, with a narrow fluctuation range, high data consistency, and stable transmission, which is suitable for application in high-security scenarios; the response delay of malicious nodes is much higher than the network delay threshold, the fluctuation range is significantly abnormal, the effective response rate is much lower than the normal level, and there are data forgery characteristics; therefore, the weight distribution is dynamically optimized according to the real-time network status (such as the proportion of malicious nodes and data load), and the comprehensive credibility value is initially distributed to the core node RSU with a higher initial credibility value.

[0124] When a node accumulates malicious behavior or its behavior decays to a certain reputation threshold, it will be judged as a low-reputation node and frozen; the reputation threshold is set to 30%-50% of the initial reputation value.

[0125] S4. Establish a dynamic reputation balance mechanism to achieve segmented dynamic adjustment of the node's comprehensive reputation value;

[0126] The dynamic reputation balancing mechanism specifically includes:

[0127] If a high-reputation node (comprehensive reputation value greater than the set upper limit of 70) has no new good behavior (such as correctly verifying a traffic incident) for two consecutive cycles (such as each cycle = 30 minutes), its reputation will decay, and the comprehensive reputation value after decay will be R new,iThe calculation formula is as follows:

[0128] R new,i =R total,i -0.005·(R total,i -70) 2

[0129] If a low-reputation node (reputation score is less than the set lower limit of 30) has no violations for two consecutive cycles, the attenuated comprehensive reputation value R new,i The calculation method is as follows:

[0130]

[0131] t is the number of violation-free cycles.

[0132] S5: Mathematically couple the Beta distribution with the dynamic reputation balance mechanism to generate node selection probability;

[0133] The mathematical coupling based on Beta distribution and dynamic reputation balance mechanism specifically includes:

[0134] In order to standardize the reputation value, a high reputation value (close to 1) corresponds to a higher α value, the Beta distribution is biased towards 1, and the probability of the node being selected is significantly increased; a low reputation value (close to 0) corresponds to a higher β value, the Beta distribution is biased towards 0, and the probability of the node being selected is reduced. By controlling the concentration distribution degree through the k value, high-reputation nodes have a greater advantage in being selected. Therefore, the balanced comprehensive reputation value R new,i Mapping from [0,100] to [0,1] interval:

[0135]

[0136] According to the normalized comprehensive reputation value R norm,i , use the mapping formula to generate the first initial parameter α of the Beta distribution base,i and the second initial parameter β base,i :

[0137] α base,i =R norm,i ×k+1,β base,i =(1-R norm,i )×k+1

[0138] Where k is a parameter used to control the concentration of the distribution. The larger the k value, the higher the probability of high-reputation nodes. It is usually a fixed constant, such as 10, and is adjusted according to the needs of node selection. Increasing the k value will make the distribution more concentrated, giving priority to high-reputation nodes; decreasing the k value will make the distribution more dispersed, increasing the tolerance of low-reputation nodes.

[0139] Further adjust α according to the reputation area of the node i and βi , reflecting the dynamic impact of historical behavior, and obtaining the adjusted first parameter α and second parameter β. The adjustment method includes:

[0140] For high reputation nodes (R new,i ≥70): Reduce α i , increase β i , accelerated decay:

[0141] α i =α base,i ×γ,β i =β base,i ×(2-γ)

[0142] For low reputation nodes (R new,i <30): Improve α i , reducing β i , promote recovery:

[0143] α i =α base,i +Δ×γ,β i =β base,i -Δ×γ

[0144] Where γ = 0.9, Δ = 2 (increment factor)

[0145] Handling boundary conditions and maintaining a reasonable reputation distribution:

[0146] Credibility lower limit (R=0): mandatory setting of α i =1,β i =k+1, avoiding division by zero error.

[0147] Credit value upper limit (R=100): mandatory setting α i =k+1,β i =1, keep the distribution reasonable.

[0148] The updated α i and β i The node selection probability used to generate the node. The node selection probability is calculated as follows:

[0149]

[0150] where w 1,i and w 2,i are the contribution weights of the performance reputation and behavior reputation of the i-th node respectively; w 1,i +w 2,i =1; For high-security scenarios, behavioral reputation is more important, so w 2,i The weight is higher, for example, w 1,i =0.3,w 2,i= 0.7. For high throughput scenarios: performance reputation is more important, so w 1,i The weight of is higher. For example: 1,i =0.6,w 2,i =0.4.

[0151] S6. Dynamically select the master node through hierarchical voting

[0152] The steps for dynamically selecting a master node include:

[0153] First, a hash algorithm is used to randomly divide all nodes registered in the Internet of Vehicles Alliance Chain network into multiple groups, and the probability of selecting nodes in each group is P. i Sorting, selecting the top three nodes to become consensus nodes;

[0154] All consensus nodes vote to elect the master node through the PBFT protocol. The master node must also meet the following requirements: recent reputation value change |Δ|≤5 (stability constraint), reputation value ≥70, and no penalty record for three consecutive cycles.

[0155] S7. Set up a supervisory node group. The supervisory node monitors the behavior of the master node in real time and triggers view switching when malicious operations are detected.

[0156] If a supervisory node detects malicious behavior from a master node, it triggers a view switching mechanism and reselects a new master node. Malicious behavior refers to any behavior within the IoV consortium chain that compromises data security, consensus efficiency, or system stability. Nodes engaging in malicious behavior will be directly labeled as "low-reputation nodes" and have their permissions frozen. Unlike the aforementioned violations, violations can temporarily reduce reputation, but can be restored through positive behavior.

[0157] The supervisory node establishes a malicious behavior processing mechanism, which specifically includes: using a dynamic penalty formula to process illegal nodes, punishing them based on the comprehensive reputation score obtained from the comprehensive reputation evaluation, and reducing their comprehensive reputation value. PunishValue represents the penalty value, and its initial value is 5. t =1.5×PunishValue t-1 The calculation method of the comprehensive reputation value of the penalized node is as follows:

[0158]

[0159] For example, if the violation is the first time, the comprehensive reputation value is reduced by 5 using the above formula 1. If the violation is repeated, the penalty value is calculated using the above formula 2, and each penalty value is 1.5 times the previous penalty value. If the violation is caused by malicious code implantation, 100 is directly deducted. If R new,i <0, then force setting R new,iIf it is 0, it means that the node is frozen. The frozen node needs to return to step S3 to recalculate its behavior reputation value R b,i and performance reputation value R p,i Restore comprehensive reputation value R total,i and adjust the comprehensive reputation value through a dynamic reputation balance mechanism.

[0160] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0161] The embodiment of the present invention also provides a dynamic reputation evaluation system for the Internet of Vehicles. Figure 4 As shown, including:

[0162] Network construction module: used to build a multi-node IoV alliance chain network consisting of roadside units (RSUs) and onboard units (OVUs); nodes include fixed nodes and dynamic nodes; multiple RSUs are deployed as fixed nodes of the IoV alliance chain network at main road intersections and transportation hubs as needed; when a vehicle enters the IoV alliance chain network and successfully registers with the network, the OVU becomes a dynamic node of the IoV alliance chain network;

[0163] Comprehensive reputation value calculation module: used to calculate the comprehensive reputation value of each node based on the performance data and behavior data of each node in the Internet of Vehicles alliance chain network;

[0164] Probability calculation module: used to calculate the node selection probability of each node based on the comprehensive reputation value of each node;

[0165] Node selection module: used to select a consensus node based on the selection probability of each node, and select a master node from the consensus nodes according to the PBFT protocol;

[0166] Supervisory node group selection module: used to select on-board units that meet the set conditions to form multiple supervisory node groups;

[0167] Malicious Behavior Detection and Punishment Module: This module monitors master nodes through multiple supervisory node groups. If a master node exhibits malicious behavior, its overall reputation value is reduced. Malicious behavior refers to any behavior by a node in the IoV consortium chain that endangers data security, consensus efficiency, or system stability, such as data forgery, denial of service, or forged control instructions.

[0168] The embodiment of the present invention also provides a PBFT layered consensus optimization system, such as Figure 5 As shown, including:

[0169] Network construction module: used to build a multi-node IoV alliance chain network consisting of roadside units (RSUs) and onboard units (OVUs); nodes include fixed nodes and dynamic nodes; multiple RSUs are deployed as fixed nodes of the IoV alliance chain network at main road intersections and transportation hubs as needed; when a vehicle enters the IoV alliance chain network and successfully registers with the network, the OVU becomes a dynamic node of the IoV alliance chain network;

[0170] Supervision framework deployment module: used to deploy each node in the IoV alliance chain network to obtain a two-layer interactive supervision framework; the edge layer of the two-layer interactive supervision framework is composed of all nodes that are roadside units; the center layer of the two-layer interactive supervision framework is composed of a plurality of selected nodes that are onboard units;

[0171] Two-layer supervision module: used to build multiple cross-regional supervision node groups using the central layer, and the cross-regional supervision node groups supervise the behavior of all edge layer nodes; also used to use the edge layer to monitor the node behavior of the central layer nodes in the geographical area.

[0172] An embodiment of the present invention further provides a non-transitory computer-readable storage medium, which stores a computer program. The computer program includes program instructions, which implement the various steps of the method described in the present invention when executed by a processor, and will not be repeated here.

[0173] The computer-readable storage medium may be the data transmission device provided in any of the aforementioned embodiments or an internal storage unit of a computer device, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., provided on the computer device.

[0174] Furthermore, the computer-readable storage medium may include both an internal storage unit of the computer device and an external storage device. The computer-readable storage medium is used to store the computer program and other programs and data required by the computer device. The computer-readable storage medium may also be used to temporarily store data to be output or that has been output.

[0175] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0176] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0177] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0178] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0179] The contents not described in detail in this specification belong to the prior art known to professional and technical personnel in this field.

Claims

1. A dynamic reputation evaluation system for the Internet of Vehicles, characterized by: include: Network construction module: used to build a multi-node Internet of Vehicles consortium chain network consisting of roadside units and on-board units; Comprehensive reputation value calculation module: used to calculate the comprehensive reputation value of each node based on the performance data and behavior data of each node in the Internet of Vehicles alliance chain network; Probability calculation module: used to calculate the node selection probability of each node based on the comprehensive reputation value of each node; Node selection module: used to select a consensus node based on the selection probability of each node, and select a master node from the consensus nodes according to the PBFT protocol; Supervisory node group selection module: used to select on-board units that meet the set conditions to form multiple supervisory node groups; Malicious behavior detection and punishment module: used to monitor the master node through multiple supervisory node groups. When the master node has malicious behavior, the comprehensive reputation value of the master node with malicious behavior is reduced.

2. The dynamic reputation evaluation and PBFT layered consensus optimization system for the Internet of Vehicles according to claim 1 is characterized in that: The calculation method of the node's comprehensive reputation value includes: The node's performance reputation value is calculated based on the node's response delay, the average response delay of the entire network, and the standard deviation of the entire network delay; The node's behavioral reputation value is calculated based on the number of successful node data uploads and the total number of data uploads; The performance reputation value and the behavior reputation value are weighted and summed to obtain the comprehensive reputation value of the node.

3. The dynamic reputation evaluation and PBFT layered consensus optimization system for the Internet of Vehicles according to claim 1 is characterized in that: It also includes a reputation adjustment module: used to dynamically adjust the comprehensive reputation value of each node. The adjustment method includes: When the comprehensive reputation value of a node is greater than the set upper limit and there is no new good behavior in multiple consecutive cycles, the reputation value is adjusted according to the following formula: R new,i =R total,i -0.005·(R total,i -70) 2 When the comprehensive reputation value of a node is less than the set lower limit and there is no violation in multiple consecutive cycles, the reputation value is adjusted according to the following formula: Where R new,i represents the updated comprehensive reputation value of the i-th node; R total,i represents the updated comprehensive reputation value of the i-th node before adjustment; t is the number of violation-free cycles.

4. The dynamic reputation evaluation and PBFT layered consensus optimization system for the Internet of Vehicles according to claim 1 is characterized in that: The calculation method of node selection probability includes: Normalizing the comprehensive reputation value of the node to obtain the normalized comprehensive reputation value of the node; A mapping formula is used to generate the first initial parameter and the second initial parameter of the Beta distribution for the normalized node's comprehensive reputation value; Adjusting the first initial parameter and the second initial parameter according to the interval of the node's comprehensive reputation value before normalization to obtain adjusted first and second parameters; The node selection probability of the node is calculated according to the adjusted first parameter and second parameter and the contribution weight of the performance reputation and the contribution weight of the behavior reputation of the node.

5. The dynamic reputation evaluation and PBFT layered consensus optimization system for the Internet of Vehicles according to claim 1 or 4, characterized in that: The calculation method of node selection probability includes: Where w 1,i and w 2,i are the contribution weights of the performance reputation and behavior reputation of the i-th node respectively; α i and β i represent the adjusted first parameter and second parameter respectively.

6. A dynamic reputation evaluation method for the Internet of Vehicles, characterized in that: include: Build a multi-node Internet of Vehicles alliance chain network consisting of roadside units and on-board units; Calculate the comprehensive reputation value of each node based on the performance data and behavior data of each node in the Internet of Vehicles alliance chain network; Calculate the node selection probability of each node based on the comprehensive reputation value of each node; Select a consensus node based on the selection probability of each node, and select a master node from the consensus nodes according to the PBFT protocol; Selecting on-board units that meet the set conditions to form multiple supervisory node groups; The supervisory node group monitors the master node, and when the master node has malicious behavior, the comprehensive reputation value of the master node with malicious behavior is reduced.

7. A PBFT hierarchical consensus optimization system for the Internet of Vehicles, characterized by: include: Network construction module: used to build a multi-node Internet of Vehicles consortium chain network consisting of roadside units and on-board units; Supervision framework deployment module: used to deploy each node in the IoV alliance chain network to obtain a two-layer interactive supervision framework; the edge layer of the two-layer interactive supervision framework is composed of all nodes that are roadside units; the center layer of the two-layer interactive supervision framework is composed of a plurality of selected nodes that are onboard units; Two-layer supervision module: used to build multiple cross-regional supervision node groups using the central layer, and the cross-regional supervision node groups supervise the behavior of all edge layer nodes; also used to use the edge layer to monitor the node behavior of the central layer nodes in the geographical area.

8. The PBFT layered consensus optimization system for the Internet of Vehicles according to claim 7 is characterized in that: The method for a cross-region supervisory node group to supervise the behavior of edge layer nodes includes: During each monitoring cycle, nodes registered as on-board units (OBUs) are extracted from the IoV consortium chain network and meet the following conditions: a comprehensive reputation value greater than or equal to a set value; no violation records in the recent cycles. The comprehensive reputation value is calculated based on the performance and behavior data of each node in the IoV consortium chain network. Randomly generate multiple cross-region supervisory node groups from the extracted nodes; each group contains multiple nodes; When an edge layer roadside unit node exhibits suspected abnormal behavior, a cross-regional supervision process is initiated for the behavior of the roadside unit node. The supervision process includes: Randomly select multiple nodes from different groups from all cross-regional supervisory node groups and use ring signature technology to conduct anonymous voting verification for suspected abnormal behavior; When more than the set percentage of nodes vote to confirm that the roadside unit has abnormal behavior, the cross-regional supervision team will initiate a secondary review. The secondary review includes: Synchronize suspected abnormal behavior data of edge layer roadside unit nodes to all cross-regional supervision groups; Each supervisory group executes the PBFT consensus algorithm internally to generate an internal verification conclusion on whether the suspected abnormal behavior is confirmed; When the internal verification conclusion of at least two different supervisory groups is that the suspected abnormal behavior is confirmed to be abnormal, obtaining dual signatures of the two different supervisory groups; The dual signature is verified, and when the dual signature verification passes, it is considered that the behavior of the edge layer roadside unit is abnormal.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the dynamic reputation evaluation method for the Internet of Vehicles as claimed in claim 6 are implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the steps of the dynamic reputation evaluation method for the Internet of Vehicles as described in claim 6 are implemented.

Images