Seamless handover authentication system and method based on distributed learning in intelligent internet of vehicles

Abstract

The application discloses a new edge intelligent-based switching authentication system and method in intelligent vehicle networking. The authentication system comprises at least one access point, each access point matches a corresponding area, a plurality of edge nodes are configured in the area, and more than two cooperative devices for authenticating user vehicles are included in the plurality of edge nodes; the access point comprises a cooperative identity verification module, which is used for sending user information to the cooperative devices, the cooperative devices measure the characteristics of the user vehicles, and the cooperative devices are cooperatively verified based on distributed learning, a cooperative device selection and update module is used for the access point to select a new cooperative device by predicting the trajectory of the user vehicle and evaluating the reputation of the edge node during the movement of the user vehicle, and the current cooperative device shares its authentication information with the new cooperative device to perform seamless switching authentication between different access points or cooperative devices. The seamless switching authentication system realizes fast switching through reliable edge cooperation.

Description

Technical Field

[0001] This application relates to the field of cybersecurity technology, specifically to a seamless switching authentication system and method based on distributed learning in intelligent vehicle networking under a zero-trust environment. Background Technology

[0002] The description in this section provides only background information relevant to the disclosure of this application and does not constitute prior art.

[0003] Intelligent vehicle-to-everything (V2X) networks are next-generation connected vehicle networks that combine vehicular ad-hoc networks and the Internet of Things (IoT). Secure and reliable communication in V2X networks is challenging due to the dramatic increase in the number of vehicles connected, the highly complex wireless communication environment, and the highly dynamic network topology, especially in zero-trust environments. Authentication is an indispensable part of verifying access / connection of all vehicles and their messages and uploaded data in V2X networks, protecting the network from eavesdropping, spoofing, and denial-of-service attacks. Furthermore, the proliferation of smart devices and vehicles with multiple radio interfaces, such as 4G / 5G, Wi-Fi, and WiMAX, has ushered in a new era for user connectivity to V2X networks.

[0004] User vehicles expect to be constantly connected to the internet and seamlessly switch between different access networks / access points. However, due to the shrinking coverage of access points and the high mobility of vehicles, users and different access points in intelligent vehicle networks need to perform more frequent handover authentication to prevent security threats and risks. This necessitates the design of a faster, more efficient, and more reliable handover authentication system.

[0005] Typical handover authentication methods include certificate / token-based systems, systems based on scalable authentication protocols, systems based on authentication and key negotiation, systems based on secret sharing, and systems based on bilinear pairing functions. A significant portion of the time required during handover authentication is spent on information exchange and negotiation; for example, methods based on scalable authentication protocols require an eight-way handshake, which is time-consuming and expensive. Meanwhile, the high complexity and heterogeneity of intelligent vehicle networks, along with the high mobility of vehicles, present significant challenges to the design of security systems, especially in handover authentication. Centralized security methods typically rely on trusted third parties or administrators, making them highly susceptible to single points of failure when attacked. Furthermore, when a large number of vehicles simultaneously request security services, they face problems of long latency and high communication and computational overhead. In addition, cryptographic methods require complex key generation, distribution, updating, and destruction processes. Protecting vehicle location privacy is also a critical issue in intelligent vehicle networks.

[0006] The above description of the technical background is provided solely for the purpose of clearly and completely explaining the technical solutions of this application and facilitating understanding by those skilled in the art. It should not be assumed that these technical solutions are known to those skilled in the art simply because they have been described in the background section of this application. Summary of the Invention

[0007] To overcome the aforementioned shortcomings, this application discloses a seamless switching authentication system and method based on distributed learning in intelligent vehicle networks. This method can be used in zero-trust environments, reducing response time and improving authentication accuracy.

[0008] To achieve the above objectives, this application adopts the following technical solution:

[0009] A seamless switching authentication system based on distributed learning in intelligent vehicle networks includes:

[0010] At least one access point, each access point is matched with a corresponding area, and multiple edge nodes are configured in the area, including two or more edge nodes used for authenticating user vehicles (edge ​​nodes involved in authentication are also called collaborative devices).

[0011] The access point includes a collaborative authentication module, which sends user information to a collaborative device. The collaborative device measures the characteristics of the user's vehicle and performs collaborative authentication based on distributed learning. The user information includes the user's identity and location information.

[0012] The collaborative device selection and update module is used by the access point to select a new collaborative device during the user vehicle's movement by predicting the user vehicle's trajectory and evaluating the reputation of edge nodes. Simultaneously, the current collaborative device shares its authentication information with the new collaborative device to perform seamless handover authentication between different access points. This authentication system, leveraging blockchain in a zero-trust environment, achieves seamless handover authentication with location privacy protection by moving security management from the network center to the edge and coordinating multiple collaborative devices to collect, process, and fuse user vehicle feature information.

[0013] Preferably, the current collaborative device uses blockchain to send the user's identity and location to the new collaborative device.

[0014] Preferably, the user vehicle features include at least one of Received Signal Strength Indicator (RSSI), Time of Arrival (TOA), and Time Difference of Arrival (TDOA).

[0015] Preferably, all of these collaborative devices are equipped with ECS.

[0016] This application provides a seamless handover authentication method based on distributed learning in intelligent vehicle networks, which includes the following steps:

[0017] The collaborative authentication module sends information to a selected collaborative device, which measures the characteristics of the user's vehicle and verifies it based on distributed learning.

[0018] The collaborative device selection and update module selects new collaborative devices by predicting the user vehicle's trajectory and evaluating the reputation of edge nodes. The current collaborative device transmits information to the new collaborative device before the user vehicle arrives, enabling seamless handover authentication when the user vehicle switches between different access point coverage areas or between collaborative devices. This method achieves seamless handover, reduces response time, improves authentication accuracy, and protects vehicle location privacy.

[0019] Preferably, in this seamless switching authentication method, the behavior of each collaborating device is recorded in the blockchain, and its reputation value is updated synchronously.

[0020] Preferably, the blockchain comprises a main chain and at least one sub-chain.

[0021] Preferably, the main chain runs in a cloud / service provider and is used to manage and control the sub-chains.

[0022] Preferably, the authentication switching module performs authentication based on the following model:

[0023]

[0024] in, This indicates that the certified vehicle belongs to a legitimate user. This indicates that the authenticated vehicle is an attacker. It switches the authentication threshold.

[0025] Preferably, the collaborative device is estimated based on the following model:

[0026]

[0027] If the authenticated vehicle's estimated value is close to the user's continuous estimated value, it will be identified as a legitimate user; otherwise, it is an attacker.

[0028] in, It is the estimate of the certified vehicle by the collaborative equipment at time t. It is a real-time predictive authentication feature estimation of users.

[0029] Beneficial effects

[0030] Compared with existing technologies, the seamless handover authentication system of this application exhibits superior performance in terms of handover latency and can be widely applied to intelligent vehicle networks. This authentication system utilizes collaborative devices and, with the aid of blockchain, provides fast, efficient, and reliable handover authentication services while offering location privacy protection. By moving authentication from the network center to the edge and leveraging multiple collaborative devices, this system reduces response time, improves authentication accuracy, and protects vehicle location privacy. The system proposed in this application outperforms existing technologies in terms of the time cost of handover authentication. This method enables seamless handover, reduces response time, improves authentication accuracy, and protects vehicle location privacy. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the switching authentication system according to an embodiment of this application;

[0032] Figure 2 This is a flowchart illustrating the switching authentication method according to an embodiment of this application;

[0033] Figure 3 This is a schematic diagram of the timing flow of the interactive implementation of switching authentication in an embodiment of this application;

[0034] Figure 4 This is a graph showing the relationship between authentication time and detection error in an embodiment of this application, where the signal-to-noise ratios are 20, 25, and 30 dB, respectively.

[0035] Figure 5 For the embodiments of this application, authentication is switched relative to a threshold. The accuracy of the experimental simulation curve. Detailed Implementation

[0036] The above-described solution will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of this application. The implementation conditions used in the embodiments may be further adjusted according to the conditions of specific manufacturers, and the implementation conditions not specified are generally those in routine experiments.

[0037] This application discloses a seamless handover authentication system (hereinafter referred to as the authentication system) based on distributed learning in an intelligent vehicle (V2X) network. By moving security provision from the network center to the edge and coordinating multiple collaborating devices, seamless handover authentication with location privacy protection can be achieved. During vehicle movement, the access point (AP) selects a new collaborating device, and then the current collaborating device transmits security information to the selected new collaborating device before the vehicle arrives, thus achieving seamless handover authentication. This authentication system can leverage blockchain in a zero-trust environment, moving security management from the network center to the edge and coordinating multiple collaborating devices to collect, process, and fuse user vehicle feature information, thereby achieving seamless handover authentication with location privacy protection.

[0038] This application proposes an authentication system having at least one access point, each access point matching and covering a corresponding area, in which multiple edge nodes are configured, the edge nodes being electrically connected to the access point, and the edge nodes being wirelessly or wiredly connected to each other, including two or more edge nodes (also called collaborative devices) used for authenticating user vehicles.

[0039] The access point includes a collaborative authentication module, which sends user information to a collaborative device. This collaborative device measures the characteristics of the user's vehicle and performs collaborative authentication based on distributed learning. The user information includes the user's identity and location information.

[0040] The collaborative device selection and update module is used by the access point to select a new collaborative device by predicting the trajectory of the user vehicle and evaluating the reputation of the edge node during the user vehicle's movement. At the same time, the current collaborative device shares its authentication information with the new collaborative device to perform seamless handover authentication between different access points.

[0041] The following describes in detail the authentication system proposed in this application, which is used to collect the characteristic information of user vehicles and achieve seamless handover authentication with location privacy protection, such as... Figure 1 As shown, taking three areas (A, B, and C) as an example, the authentication system includes:

[0042] Access points AP-A / AP-B / AP-C have corresponding coverage areas A / B / C (CellA / B / C).

[0043] Each area is equipped with collaborative devices. For example, area A has collaborative devices AC1 / AC2 / AC3, area B has collaborative devices AC4 / AC5 / AC6 / AC7, and area C has collaborative devices AC7 / AC8 / AC9. Collaborative device AC4 is located in both area A and area B (i.e., area A and area B share collaborative device AC4), and collaborative device AC7 is located in both area B and area C (i.e., area B and area C share collaborative device AC7).

[0044] Collaborating devices can exchange information to achieve information sharing, and each collaborative device provides an edge computing service (ECS). Roadside units (RSUs) include access points, access nodes, base stations, intelligent traffic lights, and camera devices. It should be noted that there are multiple edge nodes within a corresponding area; the edge nodes (also called edge devices) involved in authenticating user vehicles are called collaborative devices.

[0045] The system is configured with service providers, which include the main chain and subchains (such as Subchain1 / 2) managed by the main chain.

[0046] The main chain is used for management, while subchains are controlled by edge nodes. For example, subchain 1 is controlled by an edge node in region A, and subchain 2 is controlled by an edge node in region C. In other implementations, a subchain can correspond to multiple regions (e.g., subchain 1 covers regions A and B).

[0047] When a user vehicle (authenticated vehicle) moves to region A (which can be moved to region A in a zero-trust situation) and then moves to the next region, such as when moving from region A to region B and region C, the user vehicle needs to be authenticated separately. This authentication process is based on distributed learning.

[0048] During this process, the user vehicle first registers its information (including identity and location information) with the collaborative authentication module of access point AP-A in area A.

[0049] The collaborative authentication module transmits information to Subchain1 within region A and then to the main chain.

[0050] Access point AP-A selects collaborating devices to monitor the authenticated vehicle based on the received registration information. The collaborating devices estimate the characteristics of the authenticated vehicle based on a distributed learning framework and collaboratively verify it.

[0051] Collaborative device updates occur when vehicles are in motion due to changes in connectivity and network topology. For example, collaborative devices may be updated based on trajectory predictions and reputation values ​​prior to user arrival, and then collaborative authentication is performed by the new collaborative devices.

[0052] Information transmission refers to the current collaborating device sending the user's identity and location to a newly selected collaborating device with the help of a layered blockchain. In other words, the current collaborating device transmits security information to the newly selected collaborating device before the vehicle arrives.

[0053] For seamless authentication switching, a collaborative device in two regions (e.g., region A and region B) estimates the characteristics of the vehicle to be authenticated and performs the authentication switching. This authentication switching is based on an adaptive distributed learning framework. This embodiment uses three regions (regions A / B / C) as an example; in other embodiments, the number of regions is not limited. When a user's vehicle first enters a region, its information is registered with the corresponding service provider. The service provider selects a collaborative device and estimates the characteristics of the vehicle to be authenticated based on the distributed learning framework, then performs collaborative verification. Based on the user's trajectory prediction and reputation value before arrival, and with the help of a hierarchical blockchain, the user's identity and location are sent to the newly selected collaborative device. The updated collaborative device then performs collaborative authentication until the process ends.

[0054] The above-mentioned seamless authentication switching method for the authentication system includes the following steps:

[0055] The collaborative authentication module sends information to a selected collaborative device, which measures the characteristics of the user's vehicle and verifies it based on distributed learning.

[0056] The collaborative device selection and update module selects new collaborative devices by predicting the user's vehicle trajectory and evaluating the reputation of edge nodes. The current collaborative device transmits information to the new collaborative device before the user's vehicle arrives, enabling seamless handover authentication when the user's vehicle switches between different access point coverage areas or between collaborative devices. Specifically, the service provider sends the user's identity and location information to the collaborative devices. These collaborative devices then estimate the characteristics of the authenticated vehicle and perform collaborative verification based on distributed learning. The access point predicts the user's trajectory and updates the collaborative devices based on the user's location and reputation value during movement. With the help of a layered blockchain, the access point sends the user's identity and location to the new collaborative device. Simultaneously, the collaborative device shares its authentication information with the updated collaborative device. Collaborative devices in different cells continuously estimate user characteristics and perform handover authentication between access points. Furthermore, handover authentication between collaborative devices is also performed based on updates using a distributed learning framework.

[0057] Next, we will combine the appendix Figure 2 and Figure 3 And refer to Figure 1 This application will describe in detail the authentication method of the authentication system proposed in this application.

[0058] like Figure 2 The diagram shows the authentication method flow, which includes the following steps:

[0059] 1. Collaborative authentication based on distributed learning, applicable to zero-trust environments. The steps include:

[0060] 1-1 is based on a distributed learning framework, in which each collaborative device estimates the features of the vehicle to be authenticated and learns the authentication model locally. In this step, the collaborative authentication module sends the information to be authenticated to the collaborative devices, and if the vehicle to be authenticated is a user / user vehicle, a consensus is reached on the authentication result.

[0061] Its distributed learning process is represented as follows:

[0062]

[0063]

[0064] ,

[0065] in, and It is the first Parameters of the distributed learning process in steps, For the purpose of punishment The augmented Lagrangian function is expressed as:

[0066] in, Let be the objective function, expressed as

[0067]

[0068] As can be seen from the above objective function,

[0069] , for parameters and Fixed, for The process of solving the convex optimization problem for the parameters.

[0070] therefore, Updates can be made via By differentiating the derivative, we obtain the minimum value of the function, i.e. .

[0071] 1-2 Reputation Value Assessment: This step is used to assess the reputation value of collaborative devices. Since collaborative devices are not completely trustworthy, the collaborative authentication process based on distributed learning relies on the reputation value assessment of the devices. The feature values ​​observed by devices with high reputation will greatly improve the collaborative authentication results.

[0072] 1-3 If the proposed distributed learning process converges to the user's true location, the collaborative authentication result is that the authenticated vehicle is the user / user vehicle; otherwise, it is an attacker. This step is based on collaborative authentication using the following distributed learning process:

[0073]

[0074] in, This represents the convergence result of the distributed learning process. Indicates the user's real location. This indicates that the certified vehicle belongs to the user. This indicates that the authenticated vehicle is an attacker.

[0075] 2) Selection and Upgrading of Collaboration Devices. The selection of collaboration devices depends on the location and reputation value of the edge nodes, and specifically includes the following steps:

[0076] 2-1 Selection and updating of collaborative devices based on user trajectory prediction

[0077] User trajectory prediction is as follows:

[0078]

[0079] in, Indicates the red light at the user's arrival time. End time, , and These represent the user's speed values ​​when turning left, going straight, and turning right, respectively. Indicates that the user is The trajectory of time.

[0080] The selection of collaborative equipment should meet the following requirements:

[0081]

[0082] in, Represents the moment , users and The predicted distance between edge nodes This is the distance threshold.

[0083] 2-2 Selection and Updating of Collaborative Devices Based on Blockchain-Assisted Adaptive Reputation Management. To achieve better security in a zero-trust environment, edge nodes with low reputation values ​​will not be selected to join decentralized edge collaboration. The selection of collaborative devices should meet the following requirements:

[0084]

[0085] in, The threshold for reputation is updated as follows:

[0086]

[0087] in, Indicates when the first A node's reputation score is lowered for misconduct at any given moment. This penalty is applied if a node causes a high level of damage to the intelligent vehicle network system. In the collection The setting is set to high. Inappropriate behavior by edge nodes will be verified by the nodes running the blockchain and recorded in the blockchain.

[0088] 2-3 Selection and Updating of Optimal Collaboration Devices. Edge nodes located near users and with high reputation values ​​will be selected as collaboration devices for authentication. Due to the device and location-related functions used for authentication, the optimal number of collaboration devices meeting the following criteria should be determined. For accurate authentication.

[0089] First, predict the user's location based on 2-1.

[0090] Next, sort the edge nodes by their reputation values ​​from highest to lowest. If the number of optimal collaborating devices is 3 or more, the process ends. If it is less than 3, select the nearest node with a reputation value greater than the threshold, and continue this process until the number of optimal collaborating devices is 3 or more.

[0091] 3) Information transmission and reputation assessment based on hierarchical blockchain: To provide efficient information transmission services and reliable reputation management services,

[0092] 3-1 All transmitted information and improper behavior by edge nodes will be recorded as transactions on the blockchain and cannot be altered. In this way, edge nodes can be calculated and updated based on their behavior via a 2-2 mechanism. The reputation value of an edge node will affect its authorization.

[0093] A 3-2 layered blockchain comprises a main chain and multiple sub-chains. The main chain runs in the cloud / service provider, while the sub-chains are run by edge nodes from different regions. More specifically, the main chain is designed to manage and control multiple sub-chains, including sub-chain information synchronization and management, as well as sub-chain information sharing and fusion. The main chain's smart contracts include sub-chain management functions and information sharing and fusion functions to respectively implement the management, information sharing, and fusion of multiple sub-chains. Therefore, all collaborating nodes are responsible for their actions; any illegal behavior will be recorded in the blockchain, and their reputation score will be reduced.

[0094] 3-3 After the 2-2 update, authorized nodes whose reputation value falls below the threshold will be considered attackers and their actions will be recorded on the main chain. Layered blockchain-assisted reputation management does not require execution before each round of authentication switching, thus avoiding increased time latency.

[0095] 4) Seamless handover authentication: This handover authentication includes handover authentication between APs (access points) and between collaborating devices.

[0096] 4-1 Handover between Collaborating Devices: During vehicle movement, due to changes in connectivity and network topology, the collaborating devices should be updated. The AP updates the collaborating devices based on trajectory prediction prior to user arrival, according to 2-2. Then, based on the distributed learning-based collaborative authentication proposed in 1-1, collaborative authentication is performed by the new collaborating device.

[0097] 4-2 Switching between APs:

[0098] During the initialization phase, the user is first registered by the service provider. Then, the service provider sends the user's information to all collaborating devices (ACs) in Cell A. Finally, the collaborating devices in Cell A estimate the user's characteristics.

[0099] During the authentication switchover phase, firstly, the AP in region A predicts the user's trajectory using the algorithm proposed in sections 2-3 and updates the collaborating devices. Secondly, the collaborating devices send the user's authentication information to the new collaborating devices using the algorithm proposed in section 3. Then, the new collaborating devices estimate the characteristics of the authenticated vehicle. When the number of collaborating devices is greater than or equal to 3, the new collaborating devices update their learning parameters. If consensus is reached and the following conditions are met:

[0100]

[0101] Vehicles that are certified will be considered legitimate users, while those that do not will be considered attackers. This represents the convergence result of distributed learning. Represents the user's real-time location. This is the switching verification threshold. It applies when the number of collaborating devices is less than 3, and the following conditions are met:

[0102]

[0103] If the vehicle is authenticated, it is considered a legitimate user; otherwise, it will be considered an attacker. At any moment Estimates of certified vehicles by collaborative equipment. It is a real-time predictive authentication feature estimation of users.

[0104] Figure 3This approach implements timing for user equipment (intelligent vehicles) moving from region A to region B. During the initialization phase, the user is registered with a service provider, which then sends the user's information to all collaborating devices (ACs) within region A (Cell A). These collaborating devices estimate the user's characteristics. Then, steps such as distributed learning-based collaborative authentication, hierarchical blockchain-based information transmission, and reputation assessment are performed. This method moves security from the network center to the edge and coordinates multiple edge nodes (access points) to achieve seamless handover authentication with location privacy protection.

[0105] Next, we will use experimental simulations and numerical studies to evaluate the certification system proposed in this application.

[0106] Experimental conditions:

[0107] The location of the collaborating devices is randomized within the range [0, 2000]. The location is randomly generated within the range [-20, 20]m, and the user autonomously selects the collaborative device for authentication during movement. The detection error is the distance between the convergence result and the user's actual location. The user's initial location is [0, 0]m, the designed user speed is [60, 0]km / h, the location of AP-A is [0, 200]m, the location of AP-B is [1800, 200]m, and the vehicle density is 0.02m². 2 The reputation threshold is set to 0.5, the node's transmission power is 29dBm, and the vehicle's transmission power is 26dBm.

[0108] Experimental steps:

[0109] 1) With signal-to-noise ratios of 20, 25, and 30 dB, users autonomously select the AC for collaborative authentication during their movement, and the change in detection error value is observed as authentication time increases.

[0110] 2) A shared AC cooperative authentication vehicle was designed at area A and cell B to achieve seamless handover authentication between AP-A and AP-B. This experiment considers scenarios where the attacker is located 1, 1.5, and 2 meters away from the user. With the assistance of the shared AC, the experiment evaluates the handover authentication relative to a threshold. The accuracy.

[0111] Experimental results:

[0112] from Figure 4 The results show that the detection error values ​​stabilized after several certification periods, specifically at SNRs of 30, 25, and 20 dB. Furthermore, the highest detection accuracy was observed at an SNR of 30 dB. Figure 2The results validated the reliability and robustness of the switching authentication in the proposed system.

[0113] Figure 5 Threshold The relationship with switching authentication accuracy, including false alarm rate and false negative rate. From Figure 3 The results show that there is a balance between the false alarm rate and the false negative rate. As the authentication threshold increases, the false alarm rate decreases while the false negative rate increases. Considering the normal vehicle size, the scenario where the attacker is located 1 meter away from the user can be considered the worst-case scenario in the simulation, which proves the feasibility of the disclosed system in real-world intelligent vehicle network communication scenarios.

[0114] In some possible implementations, the model provided in this application can also be implemented as a program product, comprising a computer program that, when run on an electronic device, causes the electronic device to perform the method steps described above. For example, the electronic device can perform actions such as... Figure 2 or Figure 3 The steps are shown in the diagram. The program product may take the form of any combination of one or more readable media. The readable media may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CDROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0115] The above embodiments are only for illustrating the technical concept and features of this application, and are intended to enable those skilled in the art to understand the content of this application and implement it accordingly. They should not be used to limit the scope of protection of this application. All equivalent changes or modifications made in accordance with the spirit and essence of this application should be included within the scope of protection of this application.

Claims

1. A seamless handover authentication system based on distributed learning in intelligent Internet of Vehicles, characterized in that, Comprise: At least one access point, each of said access point matches a corresponding area, said area is configured with a plurality of edge nodes, at least two of said edge nodes are used as cooperative devices for authenticating user vehicles; Said access point comprises a cooperative identity authentication module for sending user information to cooperative devices, said cooperative devices measure the characteristics of user vehicles and cooperatively verify them based on distributed learning, said user information includes the identity and location information of the user, A cooperative device selection and update module is used for said access point to select a new cooperative device by predicting the trajectory of the user vehicle and evaluating the reputation of the edge node during the movement of the user vehicle, and the current cooperative device shares its authentication information with said new cooperative device to perform seamless handover authentication between different access points; Said selection of a new cooperative device by predicting the trajectory of the user vehicle and evaluating the reputation of the edge node specifically comprises the following steps: 2-1, Selection and update of cooperative devices based on user trajectory prediction: The user trajectory prediction is as follows: ， wherein, represents the time of the red light at the user arrival time , , and represent the speed values for the user when turning left, going straight and turning right, respectively; represents the trajectory of the user at the moment; The selection of cooperative devices satisfies: ， in, Represents the moment , users and The predicted distance between edge nodes Distance threshold; 2-2, Selection and update of cooperative devices based on adaptive reputation management assisted by blockchains: Edge nodes with low reputation values will not be selected to join decentralized edge cooperation, and the selection of cooperative devices should satisfy: ， wherein, is a threshold value of reputation, the threshold value being updated as follows: ， wherein, represents the penalty for misbehaving by the jth node at a certain time, which reduces its reputation value; if it causes a high level of damage in the intelligent Internet of vehicles system, will be set to a higher value in the set The misbehaving of the edge node will be verified by the node running the blockchain and recorded in the blockchain. 2-3, Selection and update of optimal cooperative devices: Those edge nodes that are located in the vicinity of the user and have a high reputation value are selected as the collaborative devices for authentication; due to the device and location related functions for authentication, the number of optimal collaborative devices that satisfy the following conditions is determined for accurate authentication; First, predict the user's location according to 2-1; Second, arrange the reputation values of the edge nodes in descending order; if the number of optimal cooperative devices is greater than or equal to 3, end the process, if it is less than 3, select the node closest to the user and with a reputation value greater than a threshold value, and repeat the process until the number of optimal cooperative devices is greater than or equal to 3.

2. The seamless handover authentication system based on distributed learning in the intelligent Internet of Vehicles according to claim 1, wherein The current cooperative device sends the identity and location information of the user to said new cooperative device using a blockchain.

3. The seamless handover authentication system based on distributed learning in the intelligent Internet of Vehicles according to claim 1, wherein Said characteristics of the user vehicle include at least one of received signal strength indicator (RSSI), time of arrival (TOA), and time difference of arrival (TDOA).

4. The seamless handover authentication system based on distributed learning in the intelligent Internet of Vehicles according to claim 1, wherein Said cooperative devices are all configured with ECS. 5.A seamless handover authentication method based on distributed learning in intelligent Internet of Vehicles, characterized in that, Comprise the following steps: Based on the cooperative identity authentication module, send information to the selected cooperative devices, said cooperative devices measure the characteristics of user vehicles and verify them based on distributed learning; Based on the cooperative device selection and update module, select a new cooperative device by predicting the trajectory of the user vehicle and evaluating the reputation of the edge node, and the current cooperative device transmits information to said new cooperative device before the user vehicle arrives, to achieve seamless handover authentication when the user vehicle switches between different access point coverage areas or between cooperative devices; Said selection of a new cooperative device by predicting the trajectory of the user vehicle and evaluating the reputation of the edge node specifically comprises the following steps: 2-1, Selection and update of cooperative devices based on user trajectory prediction: The user trajectory prediction is as follows: ， wherein, represents the time of the red light at the user arrival time, the end of the red light, , and represent the speed values for the user when turning left, going straight and turning right, respectively; represents the trajectory of the user at the moment . The selection of the collaboration device satisfies: ， wherein, represents a predicted distance between the user and the first edge node at the time instant , the distance threshold value, is a distance threshold value;​ 2-2, Selection and update of collaboration devices based on blockchain-assisted adaptive reputation management: Edge nodes with low reputation values will not be selected to join the decentralized edge collaboration, and the selection of the collaboration device should satisfy: ， wherein, is a threshold value of reputation, the threshold value being updated as follows: ， wherein, represents the penalty for misbehaving by the jth node at a certain time, which reduces its reputation value; if it causes a high level of damage in the intelligent Internet of vehicles system, will be set to be higher in the set The misbehaving of the edge node will be verified by the node running the blockchain and recorded in the blockchain. 2-3, Selection and update of optimal collaboration devices: Those edge nodes that are located in the vicinity of the user and have a high reputation value are selected as the collaborative devices for authentication; due to the device and location related functions for authentication, the number of optimal collaborative devices that satisfy the following conditions is determined for accurate authentication; First, predict the user's location according to 2-1; Second, arrange the reputation values of the edge nodes in descending order; if the number of optimal collaboration devices is greater than or equal to 3, end the process; if it is less than 3, select the node closest to the user and with a reputation value greater than the threshold, and repeat the process until the number of optimal collaboration devices is greater than or equal to 3.

6. The seamless handover authentication method of claim 5, wherein The behavior of each collaboration device is recorded in the blockchain, and the reputation value is updated synchronously.

7. The seamless handover authentication method of claim 6, wherein The blockchain includes a main chain and at least one sub-chain.

8. The seamless handover authentication method of claim 7, wherein The main chain runs in a cloud service provider and is used to manage and control the sub-chains.

9. The seamless handover authentication method of claim 5, wherein The handover authentication is based on the following model: ， wherein, represents that the vehicle is authenticated as a legitimate user, represents that the vehicle is authenticated as an attacker, is a handover authentication threshold.

10. The seamless handover authentication method of claim 9, wherein The collaboration devices are estimated based on the following model: ， If the estimate of the authenticating vehicle is close to the user's successive estimate, it will be identified as a legitimate user, otherwise, it is an attacker, wherein is the estimate of the authenticating vehicle by the cooperating device at time t, is the real-time predicted authentication feature estimate of the user.

Images