Networked vehicle queue safe cruise method for coping with hybrid network attack

Abstract

The invention discloses a networked vehicle queue safe cruise method for coping with hybrid network attacks, the vehicle queue comprises a head vehicle and a following vehicle, the following vehicle collects the running state data of the vehicle, obtains the running state data of other vehicles through a communication network, continuously detects the state of the communication network, and sends the detected state to the vehicle queue. Detecting whether a network attack is currently suffered or not and the type of the suffered attack, and calculating the minimum characteristic value of the topological matrix in a normal communication state, a spoofing attack state or a replay attack state according to the identified attack type; and then solving the constructed linear matrix inequality which enables the vehicle closed-loop system to be stable to obtain a feedback gain, and generating a control input signal according to the feedback gain obtained by calculation to realize accurate control of the vehicle. The feedback gain is dynamically adjusted according to the real-time state of the vehicle, so that the safety cruise control of the vehicle queue system is more effectively realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent driving, and particularly relates to a network-connected vehicle queue safety cruise method for coping with mixed network attacks. BACKGROUND

[0002] Under the intelligent network connection environment, vehicles can share and obtain the speed, acceleration, position and other information of other vehicles in real time by means of vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication networks and corresponding vehicle-mounted sensors. Through real-time interaction of information, the following vehicles can adjust their own state according to the data of the leading vehicle or other vehicles, realize speed coordination among vehicle queues, and maintain a certain inter-vehicle distance, thereby significantly improving road traffic efficiency, reducing fuel consumption, and effectively reducing the incidence of rear-end accidents.

[0003] However, due to the openness and sharing of communication networks, vehicle queues relying on communication networks for information interaction will also face potential threats from various network attacks. Vehicle-mounted sensors and inter-vehicle communication networks are vulnerable to malicious, hidden and unpredictable information attacks such as denial of service (DoS), deception attacks, replay attacks, etc. Network attacks can cause delays, losses, tampering and other problems in the process of information reception. These inaccurate information will cause vehicles to make wrong decisions, resulting in fluctuations in vehicle speed, causing vehicles to be unable to maintain a stable vehicle distance, and even possibly causing collisions, which may lead to serious traffic accidents in severe cases.

[0004] Therefore, the research on vehicle safety under network attacks is not only a necessary requirement for technological development, but also crucial for building a safe and reliable intelligent transportation system. With the increasing diversification and complexity of network attacks, it is increasingly urgent to study the safety control of vehicles under multiple network attacks.

[0005] To improve the protection ability of the connected vehicle queue cooperative cruise system in complex network environment, relevant scholars take safety control as one of the key technologies to ensure the safe operation of vehicle queue. Considering the safety and following of vehicles, the safety controller or control protocol is studied to enhance the flexibility of the connected vehicle cooperative cruise control system in response to various attack threats, and to minimize the damage and impact of network attacks on vehicle queue system. However, the existing safety control method focuses on a single type of network attack, and fails to fully consider the case where multiple attack types exist at the same time, lacking safety enhancement control method and analysis results of connected vehicle cooperative cruise system under mixed random network attacks. Therefore, it is necessary and meaningful to study vehicle safety control under mixed network attacks, so that the control algorithm can meet more complex network environment and more accurate vehicle control requirements. SUMMARY

[0006] The purpose of the present application is to provide a connected vehicle queue safety cruise method against mixed network attacks, which overcomes the problem that the prior art cannot maintain stability when subjected to multiple network attacks at the same time.

[0007] To achieve the above purpose, the technical scheme of the present application is as follows:

[0008] A connected vehicle queue safety cruise method against mixed network attacks, the vehicle queue comprising a head vehicle and a following vehicle, the connected vehicle queue safety cruise method against mixed network attacks comprising:

[0009] The following vehicle collects the running state data of the vehicle, obtains the running state data of other vehicles through the communication network, and continuously detects the communication network state to detect whether the current vehicle is subjected to network attack and the attack type;

[0010] According to the identified attack type, the minimum eigenvalue of the topological matrix under normal communication state, deception attack state or replay attack state is calculated;

[0011] The linear matrix inequality that makes the vehicle closed-loop system stable is solved, and the feedback gain is obtained;

[0012] The control input signal is generated according to the calculated feedback gain, and the precise control of the vehicle is realized.

[0013] As a preferred, the linear matrix inequality is as follows:

[0014]

[0015] Wherein, , , , the maximum communication interval, , and respectively represent the minimum eigenvalue of the topology matrix in the normal communication state, the spoofing attack state and the replay attack state;

[0016] wherein, , represents the engine inertia time constant of the vehicle i;

[0017] , is a positive definite matrix, satisfying , and , represents the feedback gain.

[0018] As preferred, in the normal communication state, if the vehicle i can receive the information of the vehicle j, the corresponding topology matrix element is equal to 1, otherwise it is equal to 0.

[0019] As preferred, in the spoofing attack state, if the vehicle i does not receive the information of the vehicle j, the corresponding topology matrix element is equal to 0; if the vehicle i receives the information of the vehicle j and does not detect the spoofing attack, the corresponding topology matrix element is equal to 1; if the vehicle i receives the information of the vehicle j and detects the spoofing attack, the corresponding topology matrix element is equal to the multiple of the detected false data relative to the original true data.

[0020] As preferred, in the replay attack state, the topology matrix is represented by , wherein represents that the vehicle i receives the information of the vehicle j at the time t, and the corresponding topology matrix element is equal to 1, otherwise it is equal to 0.

[0021] As preferred, the process of constructing the linear matrix inequality includes:

[0022] linearizing the nonlinear dynamics model of the vehicle into a linear model, and then constructing the state space equation of the vehicle according to the linear model;

[0023] taking the speed of the vehicle following the speed of the leading vehicle and maintaining the ideal distance from the front vehicle as the control target, constructing the control input signal representation about the feedback gain, the state error of the vehicle and other communication vehicles;

[0024] obtaining the closed-loop system equation of the vehicle according to the state space equation of the vehicle and the control input signal representation;

[0025] combining various attack types to construct a mixed attack closed-loop system equation;

[0026] A Lyapunov function is constructed with a queue state error as a variable, and the Lyapunov function is derived, so that a linear matrix inequality is obtained under the condition that the Lyapunov function is greater than 0 and its derivative is less than 0.

[0027] The application provides a connected vehicle queue safety cruise method for coping with a hybrid network attack. The connected vehicle queue safety cruise control method is used for maintaining the safety and following performance of vehicles in the queue under the hybrid network attack. Considering the randomness of the network attack, the vehicle queue may be subjected to multiple types of attacks at the same time. A unified system model is established by combining the characteristics of different network attacks to describe the attack situation of each vehicle in the vehicle queue. A state feedback controller based on a linear matrix inequality is designed to maintain the stability of the queue, and the feedback gain is dynamically adjusted to adapt to the control requirements of the vehicle under different attack states. The feedback control law is solved by using the linear matrix inequality, and the calculation efficiency and real-time performance are high. In addition, the proposed scheme supports online calculation, and the feedback gain can be dynamically adjusted according to the real-time state of the vehicle, so that the safety cruise control of the vehicle queue system is more effectively realized. BRIEF DESCRIPTION OF DRAWINGS

[0028] Figure 1 A flowchart of the connected vehicle queue safety cruise method for coping with a hybrid network attack is shown.

[0029] Figure 2 A vehicle attack situation diagram under three hybrid attacks of an experimental vehicle queue is shown. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical scheme and advantages of the application clearer, the application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the application and not to limit the application.

[0031] Example 1, as shown in the drawings, a connected vehicle queue safety cruise method for coping with a hybrid network attack is proposed, the vehicle queue includes a leading vehicle and a following vehicle, the connected vehicle queue safety cruise method for coping with a hybrid network attack includes: Figure 1 Step S1, the following vehicle collects the running state data of the vehicle, obtains the running state data of other vehicles through a communication network, and continuously detects the communication network state, detects whether the current vehicle is subjected to a network attack and the type of attack.

[0032]

[0033] ​Each vehicle in the vehicle platoon is provided with an on-board sensor, which can collect the speed, acceleration, position, and inter-vehicle distance of the vehicle, and the vehicles can share and obtain the collected information of other vehicles in real time through a communication network. For example, the data can be collected by laser radar, global positioning system (GPS), inertial measurement unit (IMU), and other on-board sensors, and the data can be transmitted to other vehicles through the communication network.

[0034] In this embodiment, all the vehicles in the vehicle platoon, except the head vehicle, are follower vehicles. Each follower vehicle collects the running state data of the vehicle through an on-board sensor, and obtains the running state data of other vehicles through a communication network, including vehicle position, speed, acceleration, and inter-vehicle distance.

[0035] The on-board device of each follower vehicle receives the collected or obtained data through the communication network, and performs subsequent processing to control the running of the vehicle. The on-board device also continuously detects the state of the communication network, analyzes the data transmission based on a pre-established unified system model, and identifies whether the current network is under attack and the type of attack.

[0036] It should be noted that how to analyze the data transmission to identify whether the current network is under attack and the type of attack has a large number of technical solutions in the technical field, which will not be described here.

[0037] Step S2, according to the identified attack type, calculate the minimum eigenvalue of the topology matrix in the normal communication state, the deception attack state or the replay attack state.

[0038] Specifically, in the vehicle platoon, each vehicle has a communication relationship with other vehicles, and a topology matrix is constructed accordingly. In order to calculate the feedback gain in the subsequent steps, the minimum eigenvalue of the topology matrix in three different states (normal communication state, deception attack state or replay attack state) in the subsequent linear matrix inequality needs to be obtained.

[0039] In the normal communication state, if vehicle i can receive the information of vehicle j, the corresponding topology matrix element is equal to 1, otherwise it is equal to 0. The topology matrix includes the head vehicle matrix , which describes the communication relationship between the follower vehicle and the head vehicle, is an element in the head vehicle matrix , if vehicle i can receive the head vehicle information , otherwise . The topology matrix also includes an adjacency matrix , is an element in the adjacency matrix, and the adjacency matrix Let n be an n-order square matrix, where n represents the total number of vehicles following in the vehicle queue. The adjacency matrix is ​​used to describe the communication relationships between following vehicles. If following vehicle i can receive information from following vehicle j, then... ,on the contrary .

[0040] In step S1 above, the status of the communication network is continuously monitored to identify whether a network attack is currently occurring and the type of attack. If a spoofing attack is identified, the minimum eigenvalue of the topology matrix under the spoofing attack state needs to be calculated. If a replay attack is identified, the minimum eigenvalue of the topology matrix under the replay attack state also needs to be calculated.

[0041] The topology matrix under spoofing attack conditions describes the additional false data generated by vehicles due to the spoofing attack. If vehicle i does not receive information from vehicle j, the corresponding topology matrix element is equal to 0; if vehicle i receives information from vehicle j and does not detect the spoofing attack, the corresponding topology matrix element is equal to 1; if vehicle i receives information from vehicle j and detects the spoofing attack, the corresponding topology matrix element is equal to the multiple of the detected false data relative to the original real data.

[0042] For the topology matrix under replay attack conditions, use This indicates that when a replay attack occurs, a certain part is replayed. Historical data at any given moment; the topology matrix during a replay attack state represents the data at that moment. The vehicle's communication status at any given time, If vehicle i receives information from vehicle j at time 1, the corresponding element of the topology matrix is ​​equal to 1; otherwise, it is equal to 0.

[0043] After obtaining the topological matrix, the minimum eigenvalue of the topological matrix can be calculated, which will not be elaborated here.

[0044] Step S3: Solve the linear matrix inequalities that make the vehicle closed-loop system stable to obtain the feedback gain.

[0045] The linear matrix inequality (LMI) constructed in this embodiment is as follows:

[0046] (1)

[0047] in, , , This is the maximum communication interval (i.e., the maximum acceptable communication delay). , It is a positive definite matrix. , This indicates the feedback gain. , and denote the minimum eigenvalue of the topological matrix in three different states, respectively.

[0048] wherein, , denotes the engine inertia time constant of the vehicle i.

[0049] If there is a suitable positive definite matrix satisfying and the matrix makes the above linear matrix inequality hold, then the closed-loop system is asymptotically stable.

[0050] The linear matrix inequality is derived as follows:

[0051] Step 3.1, linearize the nonlinear dynamics model of the vehicle into a linear model, and then construct the state space equation of the vehicle according to the linear model.

[0052] Specifically, the nonlinear dynamics model of the vehicle is converted into a linear model by feedback linearization, and the simplified linear model is shown in formula (1):

[0053] (2)

[0054] wherein, denote the position and speed and acceleration of the i-th vehicle in the platoon at time t, respectively, denotes the derivative of the position, denotes the derivative of the speed, denotes the derivative of the acceleration, denotes the engine inertia time constant of the vehicle i, is the vehicle control input signal to be designed.

[0055] Let , then the state space equation of the vehicle i obtained from formula (1) can be expressed as formula (3):

[0056] (3)

[0057] wherein, is the derivative of ;

[0058] wherein, .

[0059] And the state space equation of the head vehicle can be expressed as: .

[0060] Step 3.2, the speed of the vehicle following the speed of the leading vehicle and maintaining the ideal distance from the front vehicle as the control target, the control input signal representation is constructed with respect to the feedback gain, the state error of the vehicle and other communication vehicles.

[0061] The vehicle platoon control target is to enable the following vehicle i to track the speed of the leading vehicle , while maintaining the ideal distance from the front vehicle , the control target is represented as follows:

[0062] (4)

[0063] In order to achieve the control target (4), to ensure that the following vehicle can realize speed tracking and distance keeping, based on the communication network between vehicles and the received data information, combined with the undetermined feedback gain, the following control input is created:

[0064] (5)

[0065] wherein, is the element in the adjacency matrix, the adjacency matrix is an n-order matrix, n represents the total number of following vehicles in the platoon, which is used to describe the communication relationship between the following vehicles, if the following vehicle i can receive the information of vehicle j , otherwise . is the element in the diagonal matrix of the leading vehicle, which is used to describe the communication relationship between the following vehicles and the leading vehicle, if the vehicle i can receive the information of the leading vehicle , otherwise .

[0066] and , and are the local feedback gains corresponding to the position, speed and acceleration, which need to be solved by subsequent calculation, in represents the distance between vehicle i and vehicle .

[0067] In this embodiment, each following vehicle calculates the control input signal at the corresponding time according to the vehicle state information it can obtain, through a dynamically adjustable feedback gain (which can be adaptively adjusted according to the real-time attack state of the vehicle), so as to maintain the stable tracking and safe distance of the platoon in complex network environment.

[0068] ​The control input signal representation constructed in this embodiment is based on the communication network between vehicles, uses the deviation feedback (position, speed, acceleration deviation) of the corresponding vehicles (including the front vehicle and the head vehicle), and is multiplied by the feedback gain to adjust the weight and calculate the control input.

[0069] Specifically, according to the determined tracking target (4) and the communication relationship between vehicles, the deviation (position difference, speed difference, acceleration difference) of the own state and the communication vehicle is calculated, and when , it indicates that the position, speed and acceleration of vehicle i relative to vehicle j are in an ideal state, the control target is reached, which means that no control adjustment is needed for vehicle i in this part. On the contrary, if the deviation is not zero, a control amount is needed to eliminate the deviation, which is obtained by multiplying the calculated error and the feedback gain. The feedback gain includes position deviation feedback gain , speed deviation feedback gain and acceleration deviation feedback gain . These three gain parameters determine the reaction strength of the system to different types of deviations, and their values directly affect the stability and convergence speed of the platoon control.

[0070] Step 3.3, according to the state space equation of the vehicle and the control input signal representation, the closed-loop system equation of the vehicle is obtained.

[0071] In this embodiment, the state error between vehicle i and the head vehicle is:

[0072] (6)

[0073] where .

[0074] Taking the derivative of equation (6) and substituting the state space equation in equation (3) and the control input signal representation in equation (5), the closed-loop system equation of vehicle i can be obtained:

[0075] (7)

[0076] where represents the derivative of the state error , and this equation describes how the state error of vehicle i changes over time under the influence of its own dynamics, error information of other vehicles, etc. Through the closed-loop system equation, the stability of the platoon (whether the error will gradually decrease to 0) and other key performance can be analyzed.

[0077] Step 3.4, combine various attack types to construct a mixed attack closed-loop system equation.

[0078] The goal of a DoS attack is to prevent legitimate users from accessing the network. This attack can cause the target's normal communication connections to fail by initiating a large number of fake service requests. When a vehicle is under a DoS attack, its communication links with other vehicles will be broken. At this time, the vehicle will not receive relevant information from other vehicles, and the vehicle's closed-loop system equations will become:

[0079] (8)

[0080] Spoofing attacks are a major form of cyberattack. They tamper with transmitted data packets, causing the attacked target to receive incorrect speed or location data, thus generating a certain amount of erroneous control data and further destabilizing the queue. When a vehicle is subjected to a spoofing attack, it will receive false data and correspondingly generate erroneous control data. At this point, equation (6) will become:

[0081] (9)

[0082] in, and This indicates false data originating from other vehicles.

[0083] Under a replay attack, the vehicle receives historically recorded data instead of current real-time data. In this attack, the attacker aims to mask the vehicle's current real-time data with historical data to disrupt queue performance. When a replay attack occurs, the closed-loop system equation for vehicle i is:

[0084] (10)

[0085] Therefore, based on the characteristics of different network attacks, the closed-loop system equations for vehicles under different network attacks are determined and integrated into a unified hybrid attack closed-loop equation:

[0086]

[0087]

[0088] (11)

[0089] pass Different values ​​of can be used to derive any of the formulas 7, 8, 9, and 10 from formula 11.

[0090] in, and This indicates that the data originated from other vehicles. It indicates a specific moment in history during transmission; is a time-varying parameter indicator used to represent whether vehicle i is attacked at a certain moment, the three time-varying parameters are independent of each other, when , it means that the vehicle is affected by the DoS attack, and and respectively represent that the vehicle is affected by the spoofing attack and the replay attack.

[0091] Although the forms of attacks are various, the effects of these attacks on vehicle systems ultimately manifest as abnormality of received data, therefore, a series of independent time-varying parameter indicators , which take values of 0 or 1, can be introduced to uniformly describe the attack situation of each vehicle in the vehicle platoon at a certain moment in the time and space dimensions according to the different values and combinations of the elements in . This unified modeling method can design a feedback controller that can adapt to various attack scenarios based on a single model, simplifying the design and analysis process of the safety controller.

[0092] Step 3.5, construct a Lyapunov function with the platoon state error as the variable, and derive the Lyapunov function, under the condition that the Lyapunov function is greater than 0 and its derivative is less than 0, obtain the linear matrix inequality.

[0093] In order to construct the linear matrix inequality, all vehicle closed-loop system equations are aggregated to obtain the platoon closed-loop system equation as shown in (12):

[0094] (12)

[0095]

[0096] where , , n is the number of following vehicles, , and are the topological matrices in three different states, is an n-order unit matrix.

[0097] In order to obtain the condition under which the vehicle platoon can operate stably, consider the closed-loop system (12) under the condition that the vehicle platoon is subjected to mixed attacks and the platoon control objective (4), first construct a Lyapunov function with the platoon state error as the variable:

[0098] ;

[0099] The first term in the function is the conventional term, reflecting the energy measurement of the state error at the current moment, and the second term contains the influence of time delay.

[0100] The Lyapunov function is derived, so that the function satisfies: and its derivative Finally, a linear matrix inequality (LMI) that satisfies the internal stability of the system is derived.

[0101] Then, the feedback gain that satisfies the stability requirement under the corresponding state can be obtained by solving the inequality through the LMI toolbox in MATLAB. .

[0102] Step S4, generate the control input signal according to the calculated feedback gain, and realize accurate control of the vehicle.

[0103] By real-time monitoring of the state and attack of the vehicle platoon system, the LMI is solved online according to the current operating conditions, so as to update the feedback gain in real time, and calculate the control input signal that satisfies the stability condition at the current moment, so as to control the acceleration or deceleration of the vehicle, reduce the influence of the vehicle under network attack, and enable the vehicle to recover to the original stable state faster after the attack ends. Whenever the attack characteristics or system operating point change significantly, the LMI is re-solved, the feedback gain is re-acquired, and the dynamic adjustment of the controller parameters is realized, so as to achieve the purpose of safe cruise control of the connected vehicle platoon.

[0104] In actual application, the calculation process can be simplified by setting the topological matrix, that is, setting the communication relationship between vehicles, for example, only part of the following vehicles can communicate. In particular, the following vehicles only need to communicate with the front vehicle, that is, each following vehicle communicates with the front vehicle and the head vehicle.

[0105] Taking a five-vehicle platoon as an example, which includes one head vehicle and four following vehicles, if there is a communication relationship between all the following vehicles, the adjacency matrix at this time is: and the head vehicle matrix is However, in actual situations, considering communication resources and calculation amount, it is not common to communicate with all vehicles, but to communicate with part of the vehicles.

[0106] In a specific embodiment, the vehicle communicates with the front vehicle and the head vehicle, and the adjacency matrix at this time will change to: and the head vehicle matrix remains unchanged,

[0107] Correspondingly, the control input expression is:

[0108] .

[0109] The technical scheme of the application, in actual application, firstly carries out parameter setting and state monitoring, that is, setting the related parameters of the vehicle including vehicle model parameters and controller parameters, setting the ideal safety distance of the vehicle , the dynamics constant of different vehicles , sampling period T s and maximum communication interval . After setting the parameters, the data is saved to the vehicle automatic driving control computer. Then online calculation and gain adjustment are carried out, that is, based on the identified vehicle state, the vehicle computer solves the feedback gain online through the linear matrix inequality method, and through solving the linear matrix inequality under the corresponding state, the optimal feedback gain adaptive to the current attacked state is dynamically calculated . Finally, the feedback gain obtained by calculation is applied to the state feedback controller to generate the final control input signal, which acts on the throttle, brake and other actuators of the vehicle, so as to realize accurate control of the vehicle acceleration. At the same time, the system continuously collects the latest state information of the vehicle to form a closed loop feedback, which provides real-time data support for state monitoring and feedback gain adjustment in the next control period.

[0110] The following will be described by the queue of three vehicles. First, taking the head vehicle as the coordinate origin, the initial positions of the three vehicles are set as 0 m, -1.6 m and -3.2 m respectively, the initial speed of the vehicle is set as 0 m / s, the ideal safety distance is set as 1.6 m, the dynamics constant of the following vehicle , the sampling interval is 0.1 s, the maximum communication interval . The default communication topology is PLF form, that is, the following vehicle can receive information from the front vehicle and the head vehicle. The experimental running time is 40 s, and within 40 s, the speed change of the head vehicle is set as variable speed form, which goes through acceleration, constant speed, deceleration and other stages. At the same time, three mixed network attacks are simulated within 40 s of running, and the attacked situation of the vehicle is as shown in Figure 2 . In mixed attack ①, the following vehicle 1 will receive additional false data due to deception attack, and the following vehicle 2 will receive historical recorded data due to replay attack; in mixed attack ②, the following vehicle 1 will be attacked by DoS and deception attack at the same time, resulting in data interruption and false information injection, while the following vehicle 2 is not attacked and maintains normal communication; in mixed attack ③, the following vehicle 1 is not attacked, and the following vehicle 2 is attacked by deception attack. Through the combination of the above three different types of attacks, the complex running environment of the vehicle queue under mixed random network attack is simulated, and the effectiveness of the method under mixed attack is verified.

[0111] According to the initial state of the vehicle and the related parameters, the LMI is solved at the vehicle computer end to obtain the feedback gain of the vehicle under no network attack In normal driving, the feedback gain is used. When network attacks are detected, the LMI is triggered to be resolved again, and the feedback gain in the corresponding state is obtained according to the different attack conditions of the vehicles in the queue. The feedback gains in this embodiment under three mixed attacks are respectively: , , .

[0112] During vehicle operation, the position, speed, acceleration and other data of the vehicle or the preceding vehicle can be obtained through vehicle-mounted sensors such as laser radar, global positioning system (GPS), inertial measurement unit (IMU), etc. These data can be transmitted to other vehicles through a communication network, combined with the determined feedback gain and the data information of the corresponding vehicle, and the vehicle-mounted computer can further calculate the required control input at the current time, and then control the vehicle to realize acceleration, constant speed or deceleration, and finally achieve the tracking effect of the current vehicle on the motion state of the preceding vehicle. In the next control cycle, the real-time vehicle data information of the current vehicle and the preceding vehicle is obtained again and the new control input is calculated, and this process is repeated to realize the safe cruise control of the connected vehicle queue.

[0113] The technical solution of the present application has strong ability to cope with mixed random network attacks. By establishing a unified system model, the influence of multiple network attacks on the vehicle queue can be described at the same time, and the robustness of the system in a complex attack environment is enhanced. The feedback gain is dynamically adjustable and has strong adaptability. The state feedback controller designed based on linear matrix inequality can dynamically adjust the feedback gain according to the real-time attack state of the vehicle, thereby better maintaining the stability of the queue. The calculation efficiency is high and the real-time performance is good. The control method proposed uses linear matrix inequality to solve the control law, and the calculation process is efficient, supporting online real-time calculation, which is suitable for scenarios with high real-time requirements in vehicle queue control. At the same time, safety and control performance are considered, the tracking performance and safety of the vehicle queue are effectively guaranteed while resisting network attacks, and the safe cruise control of the connected vehicle queue is realized.

[0114] The above-described embodiments only express several embodiments of the present application, and the description is more specific and detailed, but it cannot be understood as limiting the scope of the patent. It should be noted that for ordinary skilled persons in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A method for connected vehicle platoon safety cruise against mixed network attacks, characterized in that, The vehicle convoy includes a lead vehicle and following vehicles. The method for safe patrol of connected vehicle convoys in response to hybrid network attacks includes: It follows the vehicle to collect its own vehicle's operating status data, obtains the operating status data of other vehicles through the communication network, and continuously monitors the communication network status to detect whether it is currently under network attack and the type of attack. Based on the identified attack type, calculate the minimum eigenvalue of the topology matrix under normal communication state, spoofing attack state, or replay attack state. Solve the linear matrix inequalities that make the vehicle closed-loop system stable to obtain the feedback gain; The calculated feedback gain is used to generate a control input signal, thereby achieving precise control of the vehicle.

2. The cyber-physical attack handling method of claim 1, wherein, The linear matrix inequality is as follows: wherein, , , , is the maximum communication interval, , and respectively represent the minimum eigenvalue of the topology matrix in the normal communication state, the spoofing attack state and the replay attack state. wherein , represents the engine inertia time constant of the vehicle i; , is a positive definite matrix satisfying and , denotes the feedback gain.

3. The cyber-physical attack handling method of claim 1, wherein, In the normal communication state, if vehicle i can receive information from vehicle j, the corresponding element in the topology matrix is ​​equal to 1; otherwise, it is equal to 0.

4. The cyber-physical attack-resilient platoon safety cruise control method for connected vehicles according to claim 1, wherein, In the topology matrix under spoofing attack conditions, if vehicle i does not receive information from vehicle j, the corresponding topology matrix element is equal to 0; if vehicle i receives information from vehicle j and does not detect a spoofing attack, the corresponding topology matrix element is equal to 1. If vehicle i receives information from vehicle j and detects a spoofing attack, then the corresponding topology matrix element is equal to the multiple of the detected fake data relative to the original real data.

5. The cyber-physical attack-resilient platoon safety cruise control method for connected vehicles according to claim 1, wherein, In the replay attack state, in the topology matrix, use to represent the replay attack occurs, at time, vehicle i receives vehicle j information, then the corresponding topology matrix element is equal to 1, otherwise it is equal to 0.

6. The cyber-physical attack-resilient platoon safety cruise control method for connected vehicles according to claim 1, wherein, The construction process of the linear matrix inequality includes: The nonlinear dynamics model of the vehicle is linearized into a linear model, and then the state-space equation of the vehicle is constructed based on the linear model. With the control objective of following the speed of the lead vehicle at the vehicle speed and maintaining an ideal distance from the lead vehicle, a control input signal representation is constructed regarding the feedback gain and the state error between the vehicle and other communicating vehicles; Based on the vehicle's state-space equations and the control input signal representation, the vehicle's closed-loop system equations are obtained. By combining various attack types, a hybrid attack closed-loop system equation is constructed. A Lyapunov function is constructed using the queue state error as a variable. The derivative of the Lyapunov function is calculated, and a linear matrix inequality is obtained if the Lyapunov function is greater than 0 and its derivative is less than 0.

Images