Dynamic separation and combination control method of vehicle platoon based on trust evaluation in zero-trust environment

Abstract

The present invention discloses a method for dynamic splitting and merging control of vehicle queues based on trust evaluation in a zero-trust environment, including: in a zero-trust environment, reconstructing a vehicle dynamics model that integrates the trust of vehicle nodes, and then constructing a corresponding vehicle longitudinal motion trajectory prediction model; based on a C‑F uncertain reasoning model, combining the vehicle longitudinal motion trajectory prediction model to evaluate the trust of each vehicle node in the vehicle queue in real time; using a dynamic splitting and merging strategy for the vehicle queue, combining the real-time evaluation results of the trust of each vehicle node in the vehicle queue to split or merge the vehicle queue, so as to ensure the safe operation of the vehicle queue in a zero-trust scenario. The present invention evaluates the trust of vehicle nodes in real time based on the C‑F uncertain reasoning model combined with the trajectory prediction model, and uses a dynamic splitting and merging strategy to split or merge the vehicle queue, thereby improving the safety and control flexibility of the vehicle queue during operation control.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation systems and zero trust, and in particular to a vehicle queue dynamic separation and combination control method based on trust evaluation in a zero trust environment. Background Art

[0002] With the rapid development of intelligent transportation systems and autonomous driving technologies, vehicles can exchange information with the external environment (such as other vehicles, pedestrians, roadside units, etc.) through V2X network technology. At present, cooperation between connected vehicles has become a key factor in achieving safe, efficient and intelligent transportation, and its role in alleviating various traffic problems and improving energy utilization is becoming more and more significant.

[0003] In autonomous driving, the trust factor between connected vehicles is crucial to improving safety. Too low a trust level may cause autonomous driving to make dangerous driving behaviors. For this reason, a general framework is proposed to collect and integrate data from different sources with different capabilities and needs to determine the driver's trust in autonomous driving. In the process of coordinated control of intelligent connected vehicle platoons, the degree of trust between connected vehicles also plays a vital role in improving safety and efficiency. In actual traffic scenarios, unmanned vehicles communicate with each other through on-board mobile networks. The communication between vehicles is exposed to various network attacks at any time. Although corresponding protection measures will be taken, it is still difficult to cope with the increasingly complex network security and privacy protection challenges in the face of endless attack methods.

[0004] J. Kinderval first proposed the zero trust model for information security in 2010. In 2011, Forrester and the National Institute of Standards and Technology (NIST) of the United States jointly proposed the concept of zero trust and the definition of zero trust architecture. Subsequently, more and more research on zero trust was published, and the concept and architecture of zero trust were defined in different ways. Although there is no unified definition of the concept and architecture of zero trust, the core concept of "never trust, constantly verify" is consistent.

[0005] When the communication between vehicles in the queue fails, switching the queue communication topology is an effective solution. However, the applicant found that the existing research has certain defects. The previous method assumes that the faulty vehicle drives independently as a smart car, and the other vehicles do not consider the existence of the faulty vehicle, that is, the software communication level still maintains a queue form, which is equivalent to directly removing the faulty vehicle from the queue in the physical topology, considering only the network topology, and ignoring the physical topology of the vehicle queue. This is not applicable to the intelligent networked vehicle queue with spatial geometric properties. In this case, the entire queue ignores the influence of the motion trajectory of the faulty vehicle, which will cause the vehicles in the queue to have a collision risk: if vehicle N and vehicle N-2 become adjacent vehicles, and the desired distance d is maintained between the two vehicles, the motion state of vehicle N-1 is ignored, and vehicle N may collide with vehicle N-1. At the same time, in the existing research on vehicle collaborative control, few studies consider the impact of the degree of trust between connected vehicles on vehicle queue control, which leads to potential safety risks when the vehicle queue is running. Therefore, how to improve the safety of vehicle queue operation control is a technical problem that needs to be solved urgently. Summary of the invention

[0006] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is: how to provide a vehicle queue dynamic splitting and merging control method based on trust evaluation in a zero-trust environment, which evaluates the trust of vehicle nodes in real time based on the CF uncertain reasoning model combined with the vehicle longitudinal motion trajectory prediction model, and uses the vehicle queue dynamic splitting and merging strategy to split or merge the vehicle queue, thereby improving the safety and control flexibility of the vehicle queue operation control.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0008] The vehicle queue dynamic separation and combination control method based on trust evaluation in a zero-trust environment includes:

[0009] S1: Reconstruct the vehicle dynamics model integrating the trust of vehicle nodes in a zero-trust environment;

[0010] S2: Construct the corresponding vehicle longitudinal motion trajectory prediction model based on the vehicle dynamics model integrated with the trust of vehicle nodes;

[0011] S3: Based on the CF uncertain reasoning model, combined with the vehicle longitudinal motion trajectory prediction model, the trust of each vehicle node in the vehicle queue is evaluated in real time;

[0012] S4: Use the dynamic splitting and merging strategy of the vehicle queue and combine the real-time evaluation results of the trust of each vehicle node in the vehicle queue to control the splitting or merging of the vehicle queue to ensure the safe operation of the vehicle queue in the zero-trust scenario;

[0013] The dynamic split and merge strategy of vehicle queues allows sub-queues in a vehicle queue to take split or merge measures according to whether abnormal trust nodes appear in the queue, and dynamically splits the sub-queue into multiple sub-queues or merges multiple sub-queues into a larger sub-queue. If abnormal trust nodes continue to appear in the split sub-queues or the abnormal trust nodes disappear, each sub-queue will further take split or merge measures; after each split or merge, the newly generated sub-queue re-selects the lead vehicle and a trust-based spacing strategy is adopted between different sub-queues.

[0014] Preferably, in step S1, the vehicle dynamics model of the vehicle is expressed as:

[0015]

[0016] Where: p i (t), v i (t), a i (t) represent the position, velocity and acceleration of vehicle i at time t respectively; κ i ∈[0, 1] represents the trustworthiness of vehicle i; τ i represents the time constant of vehicle i; u i (t) represents the control input of vehicle i at time t.

[0017] Preferably, in step S2, a vehicle longitudinal motion trajectory prediction model is constructed by the following steps:

[0018] S201: Convert the vehicle dynamics model of the vehicle into the following formula:

[0019]

[0020] Where: T, w p (t), w v (t), w a (t) represents the prediction model sampling time, vehicle position interference term, vehicle speed interference term, and vehicle acceleration interference term respectively;

[0021] S202: Construct the following state vector for the vehicle:

[0022]

[0023] in

[0024] S203: Based on the vehicle's state equation, the predicted state value of the vehicle at Np moments in the future is calculated using the following formula:

[0025]

[0026] Where: Nc 、N p They represent the control period and prediction period in the prediction model respectively;

[0027] S204: Convert the formula in step S203 into a matrix form to generate the following vehicle longitudinal motion trajectory prediction model:

[0028]

[0029] in

[0030] Preferably, in step S3, the trustworthiness of the vehicle node is evaluated by the following steps:

[0031] S301: Input the communication interaction between vehicles in the vehicle queue and the state information perceived by the vehicle itself into the vehicle longitudinal motion trajectory prediction model to obtain the future N p The vehicle's own status information and adjacent vehicle status information at each moment;

[0032] S302: Adopting the CF uncertain reasoning model, using the future N of each vehicle node p The trust calculation model is established based on the vehicle's own state information and the state information of adjacent vehicles at each moment, and the trust of each vehicle node to its adjacent vehicles is calculated;

[0033] S303: For a certain vehicle node, the trust of the vehicle node to the adjacent vehicle node is taken as the direct trust, and the trust of the adjacent vehicle node to the vehicle node is taken as the indirect trust, and then the direct trust and the indirect trust are weighted to obtain the comprehensive trust of the vehicle node.

[0034] Preferably, in step S302, the trust of the vehicle node to its adjacent vehicles is calculated by the following formula:

[0035] CF(H,E)=E(g(S ovp , S ovs ));

[0036]

[0037] Where: CF(H, E) represents the trust of the vehicle node to its neighboring vehicles; E(·) represents the expectation; S ovp Represents the vehicle node’s own state information; S ovs Represents the state information shared by the vehicle node and its neighboring vehicles.

[0038] Preferably, in step S303, the comprehensive trust of the vehicle node is calculated by the following formula:

[0039]

[0040] Where: κ represents the comprehensive trust of the vehicle node; CF D (H, E) represents the direct trust of the vehicle node; CF I (H, E) represents the indirect trust of the vehicle node; γ represents the weight of the direct trust; n represents the number of neighboring vehicles of the vehicle node.

[0041] Preferably, in step S4, the processing steps of the vehicle queue dynamic splitting and merging strategy are as follows:

[0042] S401: Determine whether there is a trust abnormality node with a trust level lower than a trust level threshold in the vehicle queue: if so, execute step S402; otherwise, execute step S403;

[0043] S402: Split the vehicle queue into multiple sub-queues with the trust abnormal node as the center, and reselect the first vehicle in the queue as the pilot vehicle in each sub-queue; a constant distance interval strategy d is adopted within the sub-queue exp =d, the interval strategy d based on trust is adopted between two adjacent sub-queues exp =(1-κ)v+d;

[0044] S403: Determine whether the vehicle queue contains multiple sub-queues: If so, execute step S404; otherwise, the vehicle queue maintains the current state;

[0045] S404: Determine whether the trust of the latter leading vehicle and the former trailing vehicle in two adjacent sub-queues are both higher than or equal to the trust threshold: if so, merge the two adjacent sub-queues into a larger queue; otherwise, each sub-queue maintains the current state;

[0046] S405: Control the operation of each sub-queue or vehicle queue through a queue controller, and ensure the stable operation of the vehicle queue in coordination with the internal stability and chord stability constraints of the queue.

[0047] Preferably, in step S405, the queue controller of the vehicle queue is expressed as follows:

[0048]

[0049] Where: u i (t) represents the control input of vehicle i at time t; k p , k v , k a are the control gains of the linear feedback controller respectively; p0(t), v0(t), a0(t) represent the position, velocity and acceleration of the pilot vehicle; d exp is a desired spacing d.

[0050] Preferably, after executing the dynamic splitting and merging strategy of the vehicle queue, the state equations of all split sub-queues or merged larger-scale sub-queues are expressed as follows:

[0051]

[0052] Preferably, after executing the dynamic splitting and merging strategy of the vehicle queue, the state equations of all split sub-queues or merged larger-scale sub-queues are expressed as follows:

[0053]

[0054] Where: I N is the N×N identity matrix; represents the Kronecker product; G is the remaining constant term, including the state of the pilot vehicle and the vehicle interval distance; is the traction matrix; is the Laplace matrix; A cl It is the closed-loop matrix for vehicle platoon control.

[0055] Compared with the prior art, the vehicle queue dynamic separation and combination control method based on trust evaluation in a zero-trust environment of the present invention has the following beneficial effects:

[0056] The present invention evaluates the trust of each vehicle node in the vehicle queue in real time based on the CF uncertain reasoning model combined with the constructed vehicle longitudinal motion trajectory prediction model. First, through real-time trust evaluation, potential unsafe vehicles or abnormal behaviors can be detected in time. When the trust of a vehicle node is lower than the preset threshold, corresponding measures can be taken quickly, such as removing the vehicle from the queue or reducing its priority in the queue, avoiding potential safety risks, thereby improving the safety of vehicle queue operation control. Secondly, the CF uncertain reasoning model can process incomplete and uncertain information, can adapt to various complex and changing traffic environments, and can dynamically adjust the parameters and strategies of trust evaluation as the traffic environment changes, thereby ensuring the control efficiency and flexibility during the operation control of the vehicle queue. Finally, through the CF uncertain reasoning model combined with the vehicle longitudinal motion trajectory prediction model, the future behavior of the vehicle can be predicted, and then potential unstable factors can be discovered in advance. By timely adjusting the queue configuration and driving strategy, the stable operation of the vehicle queue can be maintained, and the queue chaos or accidents caused by the uncertainty of vehicle behavior can be reduced.

[0057] The present invention further utilizes the dynamic split-and-merge strategy of the vehicle queue to split or merge the vehicle queue based on the real-time evaluation of the trust of the vehicle node. First, according to the real-time evaluation result of the trust of the vehicle node, the vehicle queue can be flexibly split or merged. When an unsafe factor is found in a sub-queue, it can be split out in time to avoid affecting the entire queue. At the same time, when multiple safe and reliable sub-queues are merged, a more stable and safe driving environment can be formed, thereby further improving the safety of the vehicle queue operation control. Secondly, the dynamic split-and-merge strategy allows the scale and structure of the vehicle queue to be flexibly adjusted. When the traffic flow changes or the road conditions change, the queue configuration can be quickly adjusted as needed to adapt to different traffic scenarios, thereby improving the control flexibility of the vehicle queue operation control. Finally, the dynamic split-and-merge strategy can maintain the stable operation of the vehicle queue. During the splitting process, each sub-queue reselects the pilot vehicle and adopts a trust-based spacing strategy to ensure the stability and safety of the sub-queue. During the merging process, the conflict and confusion in the merging process are reduced by optimizing the driving strategy and queue configuration.

[0058] The vehicle queue dynamic splitting and merging strategy designed by the present invention can ensure the safety of the vehicle queue during the splitting or merging process. By real-time evaluation of the trust of vehicle nodes and optimization of driving strategies, it can ensure that the sub-queues after splitting and the new queues after merging can maintain a stable driving state and avoid potential safety risks. In addition, the strategy ensures the stability of the queue by reselecting the pilot vehicle and adopting a trust-based spacing strategy. During the splitting process, each sub-queue can operate independently and remain stable; during the merging process, conflicts and confusion are reduced by optimizing the driving strategy and queue configuration, thereby maintaining the stable operation of the entire vehicle queue. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to make the purpose, technical solution and advantages of the invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which:

[0060] Figure 1 The logical block diagram of the dynamic separation and combination control method of vehicle queues based on trust evaluation in a zero-trust environment;

[0061] Figure 2 It is a schematic diagram of LPF communication topology;

[0062] Figure 3 Provides the overall framework for the dynamic splitting and merging control strategy of intelligent connected vehicle platoons in a zero-trust environment;

[0063] Figure 4 Flowchart of the dynamic splitting and merging strategy for vehicle queues. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but only represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0065] It should be noted that similar numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. In the description of the present invention, it should be noted that the orientation or position relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. is based on the orientation or position relationship shown in the drawings, or the orientation or position relationship in which the invention product is usually placed when used, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance. In addition, the terms "horizontal", "vertical", etc. do not mean that the components are required to be absolutely horizontal or suspended, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted. In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0066] The following is a further detailed description through specific implementation methods:

[0067] Example:

[0068] This embodiment discloses a method for dynamically splitting and combining vehicle queues based on trust evaluation in a zero-trust environment.

[0069] like Figure 1 As shown, the vehicle queue dynamic separation and combination control method based on trust evaluation in a zero-trust environment includes:

[0070] S1: Reconstruct the vehicle dynamics model integrating the trust of vehicle nodes in a zero-trust environment;

[0071] S2: Construct the corresponding vehicle longitudinal motion trajectory prediction model based on the vehicle dynamics model integrated with the trust of vehicle nodes;

[0072] S3: Based on the CF uncertain reasoning model, combined with the vehicle longitudinal motion trajectory prediction model, the trust of each vehicle node in the vehicle queue is evaluated in real time;

[0073] S4: Use the dynamic splitting and merging strategy of the vehicle queue and combine the real-time evaluation results of the trust of each vehicle node in the vehicle queue to control the splitting or merging of the vehicle queue to ensure the safe operation of the vehicle queue in the zero-trust scenario;

[0074] The dynamic split and merge strategy of vehicle queues allows sub-queues in a vehicle queue to take split or merge measures according to whether abnormal trust nodes appear in the queue, and dynamically splits the sub-queue into multiple sub-queues or merges multiple sub-queues into a larger sub-queue. If abnormal trust nodes continue to appear in the split sub-queues or the abnormal trust nodes disappear, each sub-queue will further take split or merge measures; after each split or merge, the newly generated sub-queue re-selects the lead vehicle and a trust-based spacing strategy is adopted between different sub-queues.

[0075] The present invention evaluates the trust of each vehicle node in the vehicle queue in real time based on the CF uncertain reasoning model combined with the constructed vehicle longitudinal motion trajectory prediction model. First, through real-time trust evaluation, potential unsafe vehicles or abnormal behaviors can be detected in time. When the trust of a vehicle node is lower than the preset threshold, corresponding measures can be taken quickly, such as removing the vehicle from the queue or reducing its priority in the queue, avoiding potential safety risks, thereby improving the safety of vehicle queue operation control. Secondly, the CF uncertain reasoning model can process incomplete and uncertain information, can adapt to various complex and changing traffic environments, and can dynamically adjust the parameters and strategies of trust evaluation as the traffic environment changes, thereby ensuring the control efficiency and flexibility during the operation control of the vehicle queue. Finally, through the CF uncertain reasoning model combined with the vehicle longitudinal motion trajectory prediction model, the future behavior of the vehicle can be predicted, and then potential unstable factors can be discovered in advance. By timely adjusting the queue configuration and driving strategy, the stable operation of the vehicle queue can be maintained, and the queue chaos or accidents caused by the uncertainty of vehicle behavior can be reduced.

[0076] The present invention further utilizes the dynamic split-and-merge strategy of the vehicle queue to split or merge the vehicle queue based on the real-time evaluation of the trust of the vehicle node. First, according to the real-time evaluation result of the trust of the vehicle node, the vehicle queue can be flexibly split or merged. When an unsafe factor is found in a sub-queue, it can be split out in time to avoid affecting the entire queue. At the same time, when multiple safe and reliable sub-queues are merged, a more stable and safe driving environment can be formed, thereby further improving the safety of the vehicle queue operation control. Secondly, the dynamic split-and-merge strategy allows the scale and structure of the vehicle queue to be flexibly adjusted. When the traffic flow changes or the road conditions change, the queue configuration can be quickly adjusted as needed to adapt to different traffic scenarios, thereby improving the control flexibility of the vehicle queue operation control. Finally, the dynamic split-and-merge strategy can maintain the stable operation of the vehicle queue. During the splitting process, each sub-queue reselects the pilot vehicle and adopts a trust-based spacing strategy to ensure the stability and safety of the sub-queue. During the merging process, the conflict and confusion in the merging process are reduced by optimizing the driving strategy and queue configuration.

[0077] The vehicle queue dynamic splitting and merging strategy designed by the present invention can ensure the safety of the vehicle queue during the splitting or merging process. By real-time evaluation of the trust of vehicle nodes and optimization of driving strategies, it can ensure that the sub-queues after splitting and the new queues after merging can maintain a stable driving state and avoid potential safety risks. In addition, the strategy ensures the stability of the queue by reselecting the pilot vehicle and adopting a trust-based spacing strategy. During the splitting process, each sub-queue can operate independently and remain stable; during the merging process, conflicts and confusion are reduced by optimizing the driving strategy and queue configuration, thereby maintaining the stable operation of the entire vehicle queue.

[0078] In order to better introduce the technical solution of the present invention, this embodiment is described through the following parts.

[0079] 1. Vehicle Dynamics Model

[0080] Considering that the longitudinal force of the vehicle is affected by vehicle gravity, air resistance, powertrain transmission efficiency, etc., the vehicle longitudinal dynamics model simplified by Newton's second law is as follows:

[0081]

[0082] where p i (t) is the absolute position of vehicle i at time t, v i (t) is the absolute speed of vehicle i at time t, a i (t) is the absolute acceleration of vehicle i at time t, τ i is the time constant of vehicle i, T i(t) is the actual driving or braking torque of vehicle i, m i refers to the mass of vehicle i, η i is the maneuvering efficiency of the power transmission system of vehicle i, R i is the tire radius of vehicle i, h is the air density, A i is the cross-sectional area of ​​vehicle i, C i is the air resistance coefficient of vehicle i, g is the acceleration due to gravity, f i is the rolling friction coefficient of vehicle i, T i，des (t) are the desired driving and braking torques of vehicle i.

[0083] In order to carry out high-level control design, the inverse model compensation technique is used to eliminate the nonlinearity in the longitudinal dynamics. At the same time, considering the influence of the degree of trust between vehicles on the vehicle state equation in the zero-trust scenario, a third-order dynamic model (i.e., the vehicle dynamics model of the vehicle) is obtained, which is expressed as:

[0084]

[0085] Where: p i (t), v i (t), a i (t) represent the position, velocity and acceleration of vehicle i at time t respectively; κ i ∈[0, 1] represents the trust of vehicle i, indicating the degree of trust between connected vehicles. The magnitude of the trust reflects the aggressiveness of the vehicle's maneuvering behavior; τ i represents the time constant of vehicle i; u i (t) represents the control input of vehicle i at time t.

[0086] Let x i (t) = [p i (t) v i (t) a i (t)] T , the state equation of the vehicle can be described as follows:

[0087]

[0088] in

[0089] The change of trust degree k will affect the control performance of the vehicle, and thus the longitudinal dynamic model of the vehicle platoon can be constructed.

[0090] make:

[0091] where X(t) = [x1(t); x2(t); ...; x N (t)];

[0092]

[0093] U(t)=[u1(t);u2(t);…;u N (t)];

[0094] The vehicle platoons considered in the present invention are isomorphic, that is, the state matrix A i τ1=τ2=…=τ i , i∈N.

[0095] In a vehicle platoon, data transmission between vehicles is transmitted through the Vanet network. According to the multi-agent system (MAS) framework, the communication links between vehicles in the vehicle platoon can be transmitted by using a directed graph. To model. Using the adjacency matrix Represents the traffic connection relationship between vehicles, the in-degree matrix Indicates how much information a vehicle can receive from other vehicles, Laplace matrix In terms of communication, it represents the changing coupling relationship and traction matrix between vehicles. To represent the communication connection relationship between the following vehicle and the leading vehicle. The present invention adopts the leader-predecessor following topology (LPF), such as Figure 2 shown.

[0096] 2. Vehicle Longitudinal Motion Trajectory Prediction Model

[0097] In the present invention, a vehicle longitudinal motion trajectory prediction model is constructed through the following steps:

[0098] S201: Convert the vehicle dynamics model of the vehicle into the following formula:

[0099]

[0100] Where: T, w p (t), w v (t), w a (t) represents the prediction model sampling time, vehicle position interference term, vehicle speed interference term, and vehicle acceleration interference term respectively;

[0101] S202: Construct the following state vector for the vehicle:

[0102] In a zero-trust scenario, there may be communication delays or interruptions between vehicles, as well as network attacks. The types of communication network attacks on vehicle queues are divided into two categories: delay attacks and information tampering attacks. Both attacks will cause the information received by vehicles in the queue to be ahead or behind, and the final result is that the information received by the vehicle is inconsistent with the actual information. Therefore, the new state variable is adopted as s(t)=x i (t)+ξ i (t), where is the attack vector. Different from the noise that affects the system performance, the attacker can attack the vehicle in a targeted manner (offset injection attack), which can be simplified as follows:

[0103]

[0104] in

[0105] S203: Based on the vehicle's state equation, the predicted state value of the vehicle at Np moments in the future is calculated using the following formula:

[0106]

[0107] Where: N c 、N p They represent the control period and prediction period in the prediction model respectively;

[0108] S204: Convert the formula in step S203 into a matrix form to generate the following vehicle longitudinal motion trajectory prediction model:

[0109]

[0110] in

[0111] 3. Trust Assessment between Intelligent Connected Vehicles in Zero Trust Scenarios

[0112] The literature (from D. Huang, Y. Na, Y. Liu, Z. Zhang and B. Mi, "Overview of Cooperative Fault-Tolerant Control Driven by the Full Information Chain of Intelligent Connected Vehicle Platforms Under the Zero-Trus Framework: Opportunities and Challenges") emphasizes that in the zero-trust scenario, trust assessment between intelligent connected vehicles is a prerequisite for vehicle safety collaborative control. The literature (from S. Gilletal., "How Physicality Enables Trust: a New Era of Trust-Centered Cyberphysical Systems") also proposes a trust-centered cyber-physical system, taking trust assessment as the core point. Therefore, in the cooperative control of intelligent connected vehicles in the zero-trust scenario, trust assessment between vehicle nodes is a key issue. In order to solve the problem that traffic sensor data is distributed and stored in a zero-trust manner by different organizations or parties, which hinders the sharing of multi-party sensor data, the literature (from F. Wang et al., "Privacy-aware traffic flow prediction based on multi-party sensor data with zero trust in smart city") proposes a traffic flow prediction method based on precise locality-sensitive hashing (LSH) with the ability to protect privacy, which is conducive to the implementation of trust assessment methods based on multi-source data.

[0113] During operation, the vehicle platoon faces network malicious attacks, communication and other problems at any time, resulting in inaccurate data on information exchange between vehicles. These factors will cause changes in the degree of trust between connected vehicles. At present, trust evaluation is mainly divided into two categories, one is identity-based trust evaluation, and the other is data-based trust evaluation. Since it is in a zero-trust scenario, there is no potential trust between connected vehicles, and when facing attacks, attackers often target objects with important identities. Therefore, the present invention uses the CF uncertain reasoning model to perform trust evaluation of vehicle nodes based on vehicle status data. The trust of vehicle nodes consists of direct trust and indirect trust.

[0114] In a zero-trust environment, vehicles in a queue can establish communication and self-perception to obtain information about surrounding vehicles. To facilitate the establishment of the trust model later, the symbolic definition of vehicle trajectory is first given. The predicted state information of the ego vehicle (egovehicle) for the adjacent vehicles that it can perceive is recorded as S ovp The state information shared with neighboring vehicles is denoted as S ovs , and the information received by the self-vehicle from other vehicles is recorded as S ovn , n=1,2...n.

[0115] The CF model is a basic method of uncertainty reasoning based on trust representation. Its advantage is that it can be used to reason under different trust levels of different evidence, and is suitable for dealing with uncertainty and ambiguity in practical problems. In CF model reasoning, the reasoning process from evidence to result is usually expressed as: IF E THEN H(CF(H, E)), where E represents evidence, H represents conclusion, and CF(H, E) is the trust factor, which represents the degree of support of evidence E for conclusion H. CF(H, E)∈[-1, 1], from which the trust level of conclusion H can be obtained as:

[0116] CF(H)=CF(H,E)×max{0,CF(E)};

[0117] Among them, CF(E)∈[-1, 1] represents the trust of evidence. In the zero-trust scenario, the trust evaluation between intelligent connected vehicles mainly consists of two pieces of evidence. One is the predicted trajectory data E perceived by the vehicle itself. ego Since the data information is not forwarded, CF(E ego )=1. For communication receiving information S ovs , respectively, with the symbol E o It means that since it is external information, its trustworthiness cannot be set to 1.

[0118] For E o The trusted computing method can be defined as:

[0119]

[0120] in represents the expected reference trajectory of the neighboring vehicle after the corresponding transformation to the ego vehicle, and α∈(0,1) is the adjustment parameter for the evidence trust calculation.

[0121] At this time, the trust factor CF(H, E) of the ego vehicle to its neighboring vehicles is defined as:

[0122] CF(H,E)=E(g(S ovp , S ovs ));

[0123] Where E(·) represents the expectation, is the function g(S ovp , S ovs ) finds the probability of each element appearing when expected, indicating that data closer to the current moment is given more importance, g(S ovp , S ovs ) is defined as follows:

[0124]

[0125] The final trust of the ego vehicle to its neighboring vehicles is calculated by combining the direct trust calculated by the ego vehicle and the direct trust of other vehicles to the neighboring vehicle (for the ego vehicle, that is, indirect trust), and its calculation can be defined as:

[0126]

[0127] Where γ is the weight of direct trust, CF D (H, E) is the direct trust of the ego vehicle to its neighboring vehicles, CF I (H, E) is the direct trust of the surrounding vehicles of the adjacent vehicle. If there is a communication connection with the self vehicle, it is sent to the self vehicle, otherwise it is not taken into account. In the present invention, the designed node trust evaluation model is implemented based on the communication between vehicles. Therefore, for vehicles without direct and indirect communication connections, the trust of the self vehicle is zero.

[0128] In summary, in the present invention, the trustworthiness of the vehicle node is evaluated through the following steps:

[0129] S301: Input the communication interaction between vehicles in the vehicle queue and the state information perceived by the vehicle itself into the vehicle longitudinal motion trajectory prediction model to obtain the future N of each vehicle node p The vehicle's own status information and adjacent vehicle status information at each moment;

[0130] S302: Adopting the CF uncertain reasoning model, using the future N of each vehicle node p The trust calculation model is established based on the vehicle's own state information and the state information of adjacent vehicles at each moment, and the trust of each vehicle node to its adjacent vehicles is calculated;

[0131] The trust of a vehicle node to its neighboring vehicles is calculated using the following formula:

[0132] CF(H,E)=E(g(S ovp , S ovs ));

[0133]

[0134] Where: CF(H, E) represents the trust of the vehicle node to its neighboring vehicles; E(·) represents the expectation; S ovp Represents the vehicle node’s own state information; S ovs Represents the state information shared by the vehicle node and its neighboring vehicles.

[0135] S303: For a certain vehicle node, the trust of the vehicle node to the adjacent vehicle node is taken as the direct trust, and the trust of the adjacent vehicle node to the vehicle node is taken as the indirect trust, and then the direct trust and the indirect trust are weighted to obtain the comprehensive trust of the vehicle node.

[0136] The trust of the vehicle node is calculated by the following formula:

[0137]

[0138] Where: k represents the comprehensive trust of the vehicle node; CF D (H, E) represents the direct trust of the vehicle node; CF I (H, E) represents the indirect trust of the vehicle node; γ represents the weight of the direct trust; n represents the number of neighboring vehicles of the vehicle node.

[0139] 4. Design of Dynamic Splitting and Merging Strategies for Vehicle Queues

[0140] Most of the current research on communication topology switching of intelligent connected vehicle platoons only considers the network topology attributes of intelligent connected vehicles, but does not consider the physical topology attributes. This is not applicable to the formation control of intelligent connected vehicles with spatial geometric attributes, and collisions and other problems may occur in actual intelligent connected vehicle platoons. In order to achieve safe collaborative control of intelligent connected vehicle platoons with both network topology and physical topology attributes under a zero-trust architecture, a dynamic split-merge strategy for intelligent connected vehicle platoons that considers both network topology and physical topology is proposed based on the trust evaluation results between vehicles. Under this strategy, the intelligent connected vehicle platoon is allowed to split into multiple sub-platoons or merge into a larger intelligent connected vehicle platoon according to the trust status between vehicle nodes. The split or merged platoon reselects the lead vehicle. The vehicles within the platoon adopt a constant distance spacing strategy d exp =d, the interval strategy based on trust is adopted between different queues d exp =(1-κ)v+d to ensure the safe distance between different vehicle queues in the zero-trust scenario. Each vehicle in the intelligent connected vehicle queue is a potential pilot vehicle, that is, each vehicle has two states: a follower vehicle and a pilot vehicle. When the queue is split or merged, the corresponding vehicle switches between the two states.

[0141] In summary, the dynamic splitting and merging control strategy for the intelligent connected vehicle queue in the zero-trust scenario is based on the communication data between the connected vehicles and the vehicle's own perception data, combined with the vehicle longitudinal trajectory prediction model, and the trust of adjacent vehicles is calculated under the CF uncertain reasoning algorithm, and the vehicle node trust evaluation results are integrated into the vehicle dynamics model. On this basis, the present invention designs a dynamic splitting control strategy for vehicle queues based on vehicle node trust evaluation. The strategy adopts a distributed linear controller. There are two spacing strategies for the entire queue operation. Vehicles in the same queue adopt a constant distance spacing strategy, and different queues adopt a trust-based spacing strategy. Finally, the internal stability and chord stability constraints of the vehicle queue are used to ensure the safe and stable operation of the vehicle queue. The overall framework is as follows: Figure 3 shown.

[0142] Combination Figure 4 As shown, in the present invention, the processing steps of the vehicle queue dynamic splitting and merging strategy are as follows:

[0143] S401: Determine whether there is a trust abnormal node with a trust level lower than a trust level threshold (set to 0.7 in this embodiment) in the vehicle queue (using LPF communication topology): if yes, execute step S402; otherwise, execute step S403;

[0144] S402: Split the vehicle queue into multiple sub-queues with the trusted abnormal node as the center (still using the LPF communication topology), and each sub-queue reselects the first vehicle in the queue as the pilot vehicle; the sub-queue adopts a constant distance interval strategy d exp =d, the interval strategy d based on trust is adopted between two adjacent sub-queues exp =(1-κ)v+d, which can play the role of getting closer to high-trust vehicles and away from low-trust vehicles, thereby improving the safety of platoon operation.

[0145] S403: Determine whether the vehicle queue contains multiple sub-queues: If so, execute step S404; otherwise, the vehicle queue maintains the current state;

[0146] S404: Determine whether the trust of the latter leading vehicle and the former trailing vehicle in two adjacent sub-queues are both higher than or equal to the trust threshold: if so, merge the two adjacent sub-queues into a larger queue; otherwise, each sub-queue maintains the current state;

[0147] S405: Controlling the operation of each sub-queue or vehicle queue through a queue controller, cooperating with the internal stability and chord stability constraints of the vehicle queue, to ensure the stable operation of the vehicle queue.

[0148] 5. Distributed Controller Design and Stability Analysis

[0149] The state of a vehicle platoon during operation can be divided into a stable state and a non-stable state. The stable state refers to the control target of the vehicle platoon, that is, all vehicles in the vehicle platoon maintain the same speed and the specified spacing. The stable state is the main control target of the vehicle platoon, which is characterized by all vehicles maintaining the same speed and complying with the specified spacing. This operating state ensures that the entire vehicle platoon moves as a cohesive whole, thereby optimizing efficiency and safety. Conversely, if these conditions are not met, a non-stable state will occur, which may lead to instability and reduce the safety and efficiency of vehicle platoon control. Therefore, achieving and maintaining a stable state is crucial to vehicle platoon control, and the specific control targets are as follows:

[0150]

[0151] Among them, d exp >0, indicating the expected spacing between vehicle nodes i-1 and i. In the present invention, the spacing strategy adopts a constant distance spacing strategy, that is, d is a constant, including the length of the vehicle itself. The specific control law (i.e., queue controller or distributed controller) is as follows:

[0152]

[0153] Where: u i (t) represents the control input of vehicle i at time t; k p , k v , k a are the control gains of the linear feedback controller respectively; p0(t), v0(t), a0(t) represent the position, velocity and acceleration of the pilot vehicle; d exp is a desired spacing d, which is expressed as d exp =[d 0 0] T , including the length of the vehicle itself.

[0154] Let k = [k p k v k a ], the control law can be rewritten as

[0155] u i (t) = -k[2x i (t)-(x i-1 (t)+x0(t))+(i+1)d exp ].

[0156] Substituting it into the state space equation, the state equation under feedback control (that is, the state equation of the vehicle queue after executing the dynamic splitting and merging strategy of the vehicle queue) is:

[0157]

[0158] The communication topology adopted by the vehicle platoon, based on its Laplace matrix and traction matrix, the state equation of the entire vehicle platoon can be rewritten as follows:

[0159]

[0160] Among them, I N is the N×N identity matrix, represents the Kronecker product, and G is the remaining constant term, including the state of the pilot vehicle and the vehicle interval distance.

[0161] After executing the dynamic splitting and merging strategy of the vehicle queue, the vehicle queue is represented as follows:

[0162]

[0163] Where: I N is the N×N identity matrix; represents the Kronecker product; G is the remaining constant term, including the state of the pilot vehicle and the vehicle interval distance; is the traction matrix; is the Laplace matrix; A cl It is the closed-loop matrix for vehicle platoon control.

[0164] 6. Stability Analysis

[0165] In order to ensure the stability of the system when the length of the vehicle queue increases infinitely, it is necessary to analyze the internal stability and chord stability of the vehicle queue control system. For a linear time-invariant system, according to the Routh-Hurwitz stability criterion, the necessary and sufficient condition for the internal stability of the system is that the eigenvalue of the closed-loop matrix of the system is negative. Before analyzing the chord stability, we first give the symbolic definitions of the state errors between several vehicles. The spacing error of vehicle i is denoted by e i (t) = p i-1 -dp i , speed error Acceleration error The distance error between the leading vehicle and the following vehicle i is denoted as e i，0 (t) = p0-p i -i×d, the speed error is recorded as The acceleration error is recorded as

[0166] Definition 1: The reference trajectory x of the vehicles in the queue ref(t) After a step change occurs at time t, the control objective is to make the error state of each vehicle in the queue asymptotically stable to the origin (from WB Dunbar, DS Caveney, "Distributed Receding Horizon Control of Vehicle Platoons: Stability and String Stability,").

[0167] Definition 2: String stability requires the vehicle reference trajectory x ref (t) After a step change occurs at time t, the vehicle platoon system has 0<ρ i <1, so that the position error of the following vehicles i=1, 2, ...N in the vehicle queue satisfies The vehicle platoon is said to be string stable (from WB Dunbar, DS Caveney, "Distributed Receding Horizon Control of Vehicle Platoons: Stability and String Stability,").

[0168] Lemma 1: According to the Routh-Hurwitz stability criterion, if λ i , i = 1, 2, ... N is a matrix The eigenvalues ​​of may be the same or different. If the vehicle platoon is asymptotically stable, then and only if A-λ i Bk,i=1,2,...N satisfies asymptotic stability.

[0169] Based on Lemma 1, we only need to discuss how to analytically give the stable range of the controller gain under a fixed queue size N and LPF communication topology. This problem is equivalent to analyzing the matrix A cl If all the characteristic roots of are negative real parts, then the queue satisfies global asymptotic uniform stability.

[0170] Proposal 1: Given a linearly isomorphic vehicle platoon with LPF communication topology, its closed-loop control system is Then the queue closed loop is stable if and only if the eigenvalue of the system closed loop matrix satisfies kp>0,

[0171]

[0172] Proof: According to Lemma 1, the vehicle platoon is asymptotically stable if and only if A-λ i All eigenvalues ​​of Bk,i=1,2,...N have negative real parts, and the matrix A-λ i The characteristic polynomial of Bk is

[0173]

[0174] According to the above formula, the Routh table can be written for judgment. Under the LPF communication topology adopted in the present invention, The matrix is ​​positive definite, so λ i ≥0, i=1, 2, ...N, since τ>0, according to the Routh-Hurwitz stability criterion, the vehicle platoon control system must be stable if all characteristic roots are negative, that is, the first column of the Routh table has the same sign, which proves Proposition 1.

[0175] The string stability of the queue is closely related to the queue communication topology. The proof of string stability is based on the communication topology. The vehicle queue of the present invention adopts the LPF communication topology. The formula of the queue controller can be obtained as follows:

[0176]

[0177] Considering the error relationship between the leading vehicle and the following vehicle and between the following vehicles in the vehicle queue, the following relationship expression can be obtained:

[0178]

[0179] Will Rewriting in the S domain gives the following

[0180]

[0181] Theorem 1: Considering the vehicle platoon system When the queue controller is applied to vehicle queue control under the assumptions 1-4 and the internal stability of the queue, if the vehicle queue system is chord-stable, the controller gain satisfies the following conditions:

[0182]

[0183] Proof: Rewrite in the frequency domain, let S = jω, and simplify to get:

[0184]

[0185] If the vehicle platoon is to be stable, it is necessary to From this, we can derive the range of the feedback control gain of the controller under the condition of vehicle platoon chord stability. The model is regarded as At this time, we only need to consider the value of the control gain. After algebraic operation, we can get the following expression:

[0186]

[0187] You can get it if you want to satisfy If and only if Since τ>0, κ∈(0,1), through the transformation On the right side, we get:

[0188]

[0189] Similarly, we get the following two inequality constraints:

[0190] According to the relevant parameter range, Theorem 1 can be proved by solving the above inequality.

[0191] Corollary 1: In a zero-trust scenario, if the vehicle platoon control system is asymptotically stable and chordally stable, if the control gain is positive and satisfies the following conditions:

[0192]

[0193] Proof: After some algebraic manipulation, this corollary can be proved by the conditions given in Proposition 1 and Theorem 1.

[0194] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit the technical solution. Those skilled in the art should understand that those modifications or equivalent substitutions of the technical solution of the present invention that do not depart from the purpose and scope of the technical solution should be included in the scope of the claims of the present invention.

Claims

1. A vehicle queue dynamic separation and combination control method based on trust evaluation in a zero-trust environment, characterized in that: include: S1: Reconstruct the vehicle dynamics model integrating the trust of vehicle nodes in a zero-trust environment; In step S1, the vehicle dynamics model of the vehicle is expressed as: Where: p i (t), v i (t), a i (t) represent the position, velocity and acceleration of vehicle i at time t respectively; κ i ∈[0,1] represents the trustworthiness of vehicle i; τ i represents the time constant of vehicle i; u i (t) represents the control input of vehicle i at time t; S2: Construct the corresponding vehicle longitudinal motion trajectory prediction model based on the vehicle dynamics model integrated with the trust of vehicle nodes; In step S2, a vehicle longitudinal motion trajectory prediction model is constructed through the following steps: S201: Convert the vehicle dynamics model of the vehicle into the following formula: Where: T, w p (t), w v (t), w a (t) represents the prediction model sampling time, vehicle position interference term, vehicle speed interference term, and vehicle acceleration interference term respectively; S202: Construct the following state vector for the vehicle: in S203: Based on the state equation of the vehicle, the predicted state value of the vehicle at the next Np moments is calculated in combination with the following formula: Where: N c 、N p They represent the control period and prediction period in the prediction model respectively; S204: Convert the formula in step S203 into a matrix form to generate the following vehicle longitudinal motion trajectory prediction model: in S3: Based on the CF uncertain reasoning model, combined with the vehicle longitudinal motion trajectory prediction model, the trust of each vehicle node in the vehicle queue is evaluated in real time; In step S3, the trustworthiness of the vehicle node is evaluated by the following steps: S301: Input the communication interaction between vehicles in the vehicle queue and the state information perceived by the vehicle itself into the vehicle longitudinal motion trajectory prediction model to obtain the future N p The vehicle's own status information and adjacent vehicle status information at each moment; S302: Adopting the CF uncertain reasoning model, using the future N of each vehicle node p The trust calculation model is established based on the vehicle's own state information and the state information of adjacent vehicles at each moment, and the trust of each vehicle node to its adjacent vehicles is calculated; S303: For a certain vehicle node, the trust of the vehicle node to the adjacent vehicle node is taken as the direct trust, and the trust of the adjacent vehicle node to the vehicle node is taken as the indirect trust, and then the direct trust and the indirect trust are weighted to obtain the trust of the vehicle node; S4: Use the dynamic splitting and merging strategy of the vehicle queue and combine the real-time evaluation results of the trust of each vehicle node in the vehicle queue to control the splitting or merging of the vehicle queue to ensure the safe operation of the vehicle queue in the zero-trust scenario; The dynamic split and merge strategy of vehicle queues allows sub-queues in a vehicle queue to take split or merge measures according to whether abnormal trust nodes appear in the queue, and dynamically splits the sub-queue into multiple sub-queues or merges multiple sub-queues into a larger sub-queue. If abnormal trust nodes continue to appear in the split sub-queues or the abnormal trust nodes disappear, each sub-queue will further take split or merge measures; after each split or merge, the newly generated sub-queue re-selects the lead vehicle and a trust-based spacing strategy is adopted between different sub-queues.

2. The method for controlling dynamic separation and combination of vehicle queues based on trust evaluation in a zero-trust environment according to claim 1, characterized in that: In step S302, the trust of the vehicle node to its neighboring vehicles is calculated by the following formula: CF(H,E)=E(g(S ovp ,S ovs )); Where: CF(H,E) represents the trust of the vehicle node to its neighboring vehicles; E(·) represents the expectation; S ovp Represents the vehicle node’s own state information; S ovs Represents the state information shared by the vehicle node and its neighboring vehicles.

3. The method for controlling dynamic separation and combination of vehicle queues based on trust evaluation in a zero-trust environment according to claim 1, characterized in that: In step S303, the trustworthiness of the vehicle node is calculated by the following formula: Where: κ represents the trust of the vehicle node; CF D (H, E) represents the direct trust of the vehicle node; CF I (H, E) represents the indirect trust of the vehicle node; γ represents the weight of the direct trust; n represents the number of adjacent vehicles of the vehicle node.

4. The method for controlling dynamic separation and combination of vehicle queues based on trust evaluation in a zero-trust environment according to claim 1, characterized in that: In step S4, the processing steps of the vehicle queue dynamic splitting and merging strategy are as follows: S401: Determine whether there is a trust abnormality node in the vehicle queue whose trust is lower than the trust threshold: if so, execute step S402; Otherwise, execute step S403; S402: Split the vehicle queue into multiple sub-queues with the trust abnormal node as the center, and reselect the first vehicle in the queue as the pilot vehicle in each sub-queue; a constant distance interval strategy d is adopted within the sub-queue exp =d, the interval strategy d based on trust is adopted between two adjacent sub-queues exp =(1-κ)v+d; S403: Determine whether the vehicle queue contains multiple sub-queues: If so, execute step S404; otherwise, the vehicle queue maintains the current state; S404: Determine whether the trust of the latter leading vehicle and the former trailing vehicle in two adjacent sub-queues are both higher than or equal to the trust threshold: if so, merge the two adjacent sub-queues into a larger queue; otherwise, each sub-queue maintains the current state; S405: Control the operation of each sub-queue or vehicle queue through a queue controller, and ensure the stable operation of the vehicle queue in coordination with the internal stability and chord stability constraints of the queue.

5. The vehicle queue dynamic separation and combination control method based on trust evaluation in a zero-trust environment as claimed in claim 1, characterized in that: In step S405, the queue controller of the vehicle queue is expressed as follows: Where: u i (t) represents the control input of vehicle i at time t; k p , k v , k a are the control gains of the linear feedback controller respectively; p0(t), v0(t), a0(t) represent the position, velocity and acceleration of the pilot vehicle; d exp is a desired spacing d.

6. The method for controlling dynamic separation and combination of vehicle queues based on trust evaluation in a zero-trust environment according to claim 5, characterized in that: After executing the dynamic splitting and merging strategy of the vehicle queue, the state equations of all split sub-queues or merged larger sub-queues are expressed as follows:

7. The method for controlling dynamic separation and combination of vehicle queues based on trust evaluation in a zero-trust environment according to claim 5, characterized in that: After executing the dynamic splitting and merging strategy of the vehicle queue, the state equations of all split sub-queues or merged larger sub-queues are expressed as follows: Where: I N is the N×N identity matrix; represents the Kronecker product; G is the remaining constant term, including the state of the pilot vehicle and the vehicle interval distance; is the traction matrix; is the Laplace matrix.

Images