Active fault-tolerant control method for intelligent networked automobile queue under communication time delay

Abstract

The invention discloses an active fault-tolerant control method for an intelligent networked automobile queue under communication time delay, and relates to intelligent safety and automatic driving of automobiles. A vehicle-mounted sensor and a V2X wireless communication system are used for achieving information interaction between the vehicle and other vehicles, the inverse model compensation and feedback linearization technology is adopted, a vehicle queue closed-loop control model with sensor faults and actuator faults is established, and the active fault-tolerant control method of the distributed observer is designed under the influence of communication time delay. Based on a dynamic proportional integral observer, sensor faults and actuator faults are estimated, communication time delay is considered in a controller part, the faults are compensated, the stability of a queue system is achieved, and the expected vehicle-to-vehicle distance and the expected running speed of each vehicle in the queue are ensured. A distributed controller is designed on the basis of fault estimation value design and delayed output information of other vehicles, so that the influence caused by faults of sensors and actuators is restrained, and the aim of queue control is achieved.

Description

Technical Field

[0001] The present invention belongs to the field of automobile intelligent safety and automatic driving, and in particular to an active fault-tolerant control method for an intelligent networked vehicle queue under the combined influence of communication delay, actuator failure, and sensor failure. Background Art

[0002] As the number of private cars continues to rise, the existing road traffic infrastructure can no longer effectively carry the growing traffic demand. This situation directly leads to frequent traffic congestion, and vehicles have to start and stop frequently, which not only greatly increases energy consumption, but also indirectly leads to an increase in road safety accidents, thereby significantly increasing the public's travel costs and living burden. Intelligent connected car platoon driving can adjust the spacing to improve road utilization and alleviate energy waste through speed coordination, which has great application potential in solving the above pain points.

[0003] Intelligent connected vehicle platoon control technology uses V2X (Vehicle to Everything) technology to share vehicle information. Each vehicle in the platoon adjusts its own motion state based on the information shared by the vehicles to achieve the desired spacing with adjacent vehicles and the same speed as the lead vehicle. Reference 1 (Yu G, Wong PK, Huang W, et al. Distributed adaptive consensus protocol for connected vehicle platoon with heterogeneous time-varying delays and switching topologies [J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23 (10): 17620-17631.) proposes an adaptive control method for vehicle platoons with heterogeneous time-varying delays. Reference 2 (Liu C, Zhao J, Patton RJ. Distributed antittack fault-tolerant tracking control for vehicle platoon systems under cyber-physical threats[J]. IEEE Transactions on Industrial Informatics, 2023, 19(6): 7825-7834.) proposes a control method based on unknown input observer to solve actuator faults and switching topologies.

[0004] Intelligent connected vehicle platoons use V2X vehicle networking technology to improve the perception capabilities of vehicles. However, the introduction of wireless communications will inevitably bring about the problem of information transmission delay, that is, communication delay, which may lead to information synchronization problems between vehicles, increase the risk of rear-end collisions, and reduce the efficiency of platoon coordination. In addition, in practical applications, the sensors and actuators of intelligent connected vehicles may fail due to various factors, resulting in the inability of vehicles to accurately perceive environmental information or to execute control instructions, resulting in a decrease in control performance or even failure. In view of the impact of communication delays, sensor failures, and actuator failures on intelligent connected vehicle platoons, the present invention proposes an active fault-tolerant control method for intelligent connected vehicle platoons under communication delays. Summary of the invention

[0005] The purpose of the present invention is to solve the problem of car platoon system formation in the prior art with communication delay, sensor failure and actuator failure, and to provide an active fault-tolerant control method for intelligent connected car platoon under communication delay, which can maintain the stable operation of the platoon even in the case of sensor or actuator failure, and ensure the safe distance and driving speed between vehicles. The present invention first designs a distributed dynamic proportional integral observer to estimate the sensor and actuator failure of each car in the platoon, and then designs a distributed controller based on the fault estimation value and the delayed output information of other vehicles, thereby suppressing the impact of sensor and actuator failures and achieving the goal of platoon control.

[0006] The active fault-tolerant control method for a queue of intelligent connected vehicles under communication delay of the present invention comprises the following steps:

[0007] Step 1: Composition and information collection of the vehicle platoon: The vehicle platoon consists of N+1 vehicles, including a pilot vehicle and N following vehicles. The vehicle-mounted sensors are used to obtain the vehicle's own status information, including vehicle position, driving speed and acceleration. The output information of adjacent vehicles and the pilot vehicle is collected in real time through the V2X wireless communication network, the output information of the vehicle with which the communication connection is established is received in real time, and the output information of the vehicle itself is broadcast.

[0008] Step 2: Based on the information obtained from the on-board sensors and the V2X wireless communication network, a longitudinal dynamic state space model of a single vehicle with actuator failure and sensor failure is established; specifically: the nonlinear longitudinal dynamic expression of a single vehicle is derived based on Newton's second law, feedback linearization is performed based on the inverse model compensation technology, and a linearized longitudinal dynamic model of a single vehicle is obtained; the position, velocity and acceleration information of the ego vehicle is used as the state vector, actuator failure is considered, and a vehicle longitudinal motion model including actuator failure is established; sensor failure is considered, and sensor output information is constructed so that these failures can be estimated and compensated during the control process;

[0009] Step 3: Based on graph theory, the information interaction between intelligent connected vehicles is described to establish the objective function of the intelligent connected vehicle platoon control. Specifically: the communication topology of the vehicle platoon is defined based on graph theory, and the definitions of the adjacency matrix and the Laplace matrix are given; the speed difference between the ego vehicle and the pilot vehicle, and the distance difference between the ego vehicle and the front vehicle are set as independent variables, and the objective function of the vehicle platoon control is constructed to keep the platoon in a constant desired formation.

[0010] Step 4: For the car platoon system affected by sensor failure and actuator failure, establish a distributed error and fault observer, design a distributed active fault-tolerant controller without delay, design a distributed active fault-tolerant controller under the influence of communication delay, and solve the wheel motor driving torque required for platoon control in real time. Specifically:

[0011] 4.1) Construct an augmented system including sensor failure, design a distributed dynamic proportional integral observer, and establish a closed-loop error system for the vehicle platoon observer;

[0012] 4.2) Design a distributed active fault-tolerant controller without communication delay and substitute it into the formation model to establish a closed-loop error control system for the vehicle platoon;

[0013] 4.3) Based on Lyapunov stability theory and linear matrix inequality method, the conditions for the active fault-tolerant controller to achieve the stability of the queue closed-loop error system without time-varying delay are given, and the design method of the controller gain matrix and the decentralized observer gain is obtained;

[0014] 4.4) Design a distributed active fault-tolerant controller under the influence of communication delay, substitute it into the formation model, and establish a closed-loop error control system for the vehicle queue under the influence of time-varying communication delay;

[0015] 4.5) Based on Lyapunov stability theory, linear matrix inequality method and Halanay inequality, the conditions for the active fault-tolerant controller to achieve the stability of the queue closed-loop error system under time-varying delay are given, that is, the upper bound of the delay that can be tolerated by the controller and observer gains obtained in step 4.3) is obtained.

[0016] 4.6) Substitute the active fault-tolerant controller into the feedback linearization model in step 2 to calculate the expected driving torque of the vehicle wheels in real time to realize the platoon control of intelligent connected vehicles.

[0017] Compared with the prior art, the present invention has the following outstanding technical effects and advantages:

[0018] The present invention uses vehicle-mounted sensors and V2X wireless communication systems to realize information interaction between itself and other vehicles, adopts inverse model compensation and feedback linearization technology, establishes a closed-loop control model of a car queue with sensor failure and actuator failure, and designs an active fault-tolerant control method of a distributed observer under the influence of the communication industry. Through feedback linearization, the present invention can transform the originally complex nonlinear control system into a relatively simple linear control system, making the control design easier to implement. Based on a dynamic proportional integral observer, the present invention estimates sensor failure and actuator failure, considers communication delay and compensates for failure in the controller part, realizes the stability of the queue system, and ensures the expected vehicle spacing and expected driving speed of each car in the queue. By considering actuator failure and sensor failure, the designed controller can better cope with these uncertain factors and improve the robustness and reliability of the system. The present invention uses vehicle-mounted sensors and V2X wireless communication networks to collect and exchange information in real time, can quickly respond to changes in vehicle status, and improve the real-time performance of the system. The method of the present invention is not only applicable to the control of a single vehicle, but can also be extended to the coordinated control of the entire fleet, which helps to realize the automation and intelligent management of the fleet. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Schematic diagram of an active fault-tolerant control system for an intelligent connected vehicle platoon under communication delay according to an embodiment of the present invention.

[0020] Figure 2 The figure is a flowchart of a method according to an embodiment of the present invention.

[0021] Figure 3 This is a diagram showing the fault signal tracking result of the first following vehicle under the influence of communication delay and fault in an embodiment of the present invention. Figure 3 In the figure, (a) is the dynamic tracking curve of the first following vehicle actuator fault (sinusoidal signal) compared with the estimated and true value; (b) is the dynamic tracking curve of the first following vehicle sensor fault (sinusoidal signal) compared with the estimated and true value.

[0022] Figure 4 This is a diagram showing the fault signal tracking result of the third following vehicle under the influence of communication delay and fault according to an embodiment of the present invention. Figure 4 In the figure, (a) is the comparison between the estimated and true value of the dynamic tracking curve of the actuator failure (combined signal) of the third following vehicle; (b) is the comparison between the estimated and true value of the dynamic tracking curve of the sensor failure (combined signal) of the third following vehicle.

[0023] Figure 5 This is a simulation result diagram of active fault-tolerant control of a queue system under the influence of communication delay in an embodiment of the present invention. Figure 5In the figure, (a) is the simulation result curve of the distance tracking error of the queue system; (b) is the simulation result curve of the speed tracking error of the queue system; (c) is the simulation result curve of the acceleration of the queue system. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the following embodiments will further illustrate the present invention in conjunction with the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0025] like Figure 1 As shown, the formation control system of the embodiment of the present invention includes a distributed observer i, a distributed active fault-tolerant controller i, sensors, actuators and other modules to form a typical hierarchical structure. The lowest layer is the physical layer, and the sensor and actuator modules may be affected by unknown faults. Next is the distributed fault estimation layer, which independently estimates sensor faults and actuator faults through distributed proportional integral observers deployed on each vehicle. Next is the distributed active fault-tolerant control layer, and the active fault-tolerant control layer and the distributed fault estimation layer are bidirectionally interactive. The distributed fault-tolerant control layer receives information from automobiles through the V2X wireless communication network, and at the same time transmits the vehicle information and the distributed observer information to other vehicles in the queue through the V2X wireless communication network, thereby completing the collaborative control of the automobile queue.

[0026] like Figure 2 As shown, the present invention proposes an active fault-tolerant control method for an intelligent connected vehicle queue under communication delay, comprising the following steps:

[0027] Step 1: The car platoon consists of N+1 cars, numbered 0,...N, where car 0 is the lead car and cars 1,...,N are the following cars. The vehicle's driving state information and the output information of the adjacent vehicles and the lead car are collected in real time through the on-board sensors and the V2X wireless communication network.

[0028] Step 1.1: Use advanced vehicle-mounted sensor technology to integrate GPS positioning function to capture and analyze the vehicle's own status information, that is, determine the vehicle's position, measure the driving speed and acceleration, and calculate the output information of the vehicle;

[0029] Step 1.2: The ego vehicle exchanges information with the pilot vehicle and other following vehicles in the queue through the V2X wireless communication network, receives the output information of the vehicles with established communication connection in real time, and broadcasts the output information of the ego vehicle;

[0030] Step 2: Based on the information obtained from the on-board sensors and the V2X wireless communication network, a longitudinal dynamic state space model of a single vehicle with actuator failure and sensor failure is established.

[0031] Step 2.1: Perform a dynamic analysis on the longitudinal motion of the vehicle. Based on Newton's second law, the nonlinear dynamic model of the i-th vehicle in the queue can be obtained.

[0032]

[0033]

[0034] Among them, m i is the mass of the vehicle, F i is the driving force of the vehicle, v i is the vehicle speed, μ i is the mechanical efficiency of the transmission system, T i is the actual driving / braking torque of the vehicle, R i is the wheel radius, c fi is the lumped air resistance coefficient, g is the gravity acceleration constant, f i is the rolling resistance coefficient, ι i is the vehicle dynamic time lag constant, T ie is the expected torque.

[0035] Design the feedback linearization module:

[0036]

[0037] Combining equations (10), (11), (12) and (13), and assuming that the dynamics of the vehicles in the platoon are isomorphic, we can obtain ι i =ι>0, the linear dynamic model of the i-th electric car can be obtained:

[0038]

[0039] Among them, a i is the vehicle acceleration, u i is the control input, i is the vehicle dynamic lag constant.

[0040] Step 2.2, take the position, velocity and acceleration of the vehicle as the state vector: Considering the impact of actuator failure and sensor failure on the system, the feedback linearization model of the i-th intelligent connected vehicle is established:

[0041]

[0042] in, is the output of the system; is an output matrix with full column rank and (A, C) observable; represents the unknown time-varying actuator fault (in a car platoon, generally E = B), where q represents the number of types of vehicle actuator faults. Without loss of generality, we assume that the actuator fault f ai (t) The matching condition is satisfied, and the constraint is rank(B,F) = rank(B) = 1. In addition, the actuator fault f ai (t) belongs to That is, the energy accumulated by the fault over time is controllable, and f ai (t) is a continuous smooth function whose first-order derivative is bounded, that is, represents unknown time-varying sensor faults, where r represents the number of types of vehicle sensor faults. Without losing generality, it is assumed that the sensor fault matrix E satisfies and with a time-invariant system matrix of compatible dimensions.

[0043] For the sensor fault matrix, when r = 3 and This means that the vehicle's position sensor, velocity sensor, and acceleration sensor channels are affected by three different, unknown fault inputs.

[0044] Define the feedback linearization model of the pilot vehicle:

[0045]

[0046] Among them, the definitions of A and B are the same as those of the following vehicles, x0 is the state of the leading vehicle, and y0(t) is the input of the leading vehicle. In our model, the leading vehicle is a virtual and complete leader that will not be affected by sensor-actuator failures.

[0047] Step 3: Based on graph theory, describe the information interaction between intelligent connected vehicles and establish the objective function of intelligent connected vehicle formation control.

[0048] Step 3.1: Based on graph theory, describe the communication connection topology of the N following vehicles in the car platoon system as a directed graph in Represents the set of following vehicles in the queue, Represents the set of edges with direct connectivity between vehicles; edge (i, j)∈ε indicates that vehicle j can receive information from vehicle i; based on graph theory, the communication relationship between N following vehicles can be expressed using the adjacency matrix The communication connection Laplacian matrix of N following vehicles can be defined as in is the in-degree matrix.

[0049] Based on graph theory, the traction matrix of the following vehicle and the leading vehicle can be expressed as diag is used to construct diagonal elements into a diagonal matrix, m i0 =1 indicates that there is a communication connection, otherwise there is no communication connection. Therefore, the Laplace matrix of the pilot vehicle and the N following vehicles can be defined as:

[0050] When the above directed graph Contains a directed spanning tree with the leader as the root node. Then, is reversible, and there exists a matrix G = diag{g1,…,g N}>0, so that:

[0051]

[0052] in, As defined above, it represents the Laplacian matrix of the lead vehicle and the N following vehicles;

[0053] For the convenience of description, For {g1,…,g N}. for The maximum and minimum eigenvalues ​​of .

[0054] Step 3.2: Take the speed difference between the ego vehicle and the pilot vehicle, and the distance difference between the ego vehicle and the leading vehicle as independent variables to establish the objective function of vehicle formation control, that is, to satisfy:

[0055]

[0056] Among them, p i With v i denote the position and speed of vehicle i, d i,i-1 represents the expected distance between the i-th vehicle and the i-1-th vehicle. When the objective function is satisfied, the queue can achieve the desired formation. i,i-1 =d0 is a constant value, which means that the distance between the vehicle and the front vehicle is a constant value.

[0057] Step 4: For the car platoon system affected by sensor failure and actuator failure, first build an augmented system and establish a distributed error and fault observer. Then design a distributed active fault-tolerant controller without delay, and then design a distributed active fault-tolerant controller under the influence of communication delay. Finally, solve the wheel motor driving torque required for formation control in real time.

[0058] Step 4.1: In order to simultaneously observe the state and sensor failure of vehicle i in the queue, construct the following augmented system:

[0059]

[0060] in, It can be concluded that

[0061] Then construct an unknown full row rank matrix To make the following conditions hold;

[0062]

[0063] Unknown full row rank matrix Can be designed as:

[0064]

[0065] in:

[0066]

[0067] Then, the augmented system can be written as:

[0068]

[0069] For the above augmented system, a dynamic proportional-integral observer is constructed to estimate the actuator fault and augmented state:

[0070]

[0071] in, and They are f ai (t) and y i (t) observation, η i (t) is the auxiliary variable; M1 is the proportional gain, is the integral gain; for the derivative of the actuator fault, the observation effect is enhanced by introducing the dynamic quantity v, M2 and is the estimated gain of actuator fault, where ω is a variable selected empirically.

[0072] The estimation error of augmented state and actuator fault is defined as:

[0073]

[0074] and

[0075]

[0076] Among them, the augmented state estimation error is The estimated error of actuator failure is

[0077] Therefore, the actuator fault error dynamic system can be derived as:

[0078]

[0079] Among them, the estimated error of actuator failure is

[0080] Combined with actuator fault estimation, we further construct an augmented vector containing actuator faults and augmented state estimation including actuator faults The augmented state estimation error including actuator faults can be defined as:

[0081]

[0082] Among them, e fi (t) is the augmented state estimation error.

[0083] Its kinetic equation is:

[0084]

[0085] in, Obviously J0 is reversible, we have:

[0086]

[0087] in,

[0088] Then the i-th decoupled observation error dynamics in the decentralized proportional-integral fault observer can be rewritten as:

[0089]

[0090] in,

[0091] Its compact form is:

[0092]

[0093] in, represents the Kronecker product,

[0094] Step 4.2: Define I a =[0 q×3 ,0 q×r ,I q×q ]、I s =[0 r×3 ,I r×r ,0 r×q ],So The distributed active fault-tolerant controller of the i-th following vehicle is designed as:

[0095]

[0096] Among them, actuator fault compensation gain K is a constant gain matrix, which is designed later, and δ>0 is the coupling strength; To achieve compensation for actuator failure, The sensor failure is compensated by the above compensation, and active fault-tolerant control is achieved.

[0097] Define the queue tracking error variable:

[0098] e i (t) = x i (t)-x0(t)D i,0 , (34)

[0099] Among them, D i,0 =[d i,0 ,0,0], constant spacing d i,0 = -i × d0, d0 is the desired constant spacing, define d i,j =-(ij)×d0, then D ij =[d i,j ,0,0].

[0100] Then, according to the above state tracking error e i (t), the state tracking error dynamics can be obtained as:

[0101]

[0102] Among them, e i (t) is the state tracking error.

[0103] Written in compact form:

[0104]

[0105] in, Represents the Kronecker product operator. N is the N×N identity matrix. e(t) is the compact form of the state tracking error. f (t) is a compact form of the state estimation error.

[0106] Step 4.3: Consider the Lyapunov function V(e(t),e f (t))=V1(e(t))+V2(e f (t)), where the Lyapunov function V1(e(t)) with symmetric positive definite matrices P and G is defined as:

[0107]

[0108] Among them, P -1 is the inverse of the positive definite matrix P, and G is a diagonal matrix defined in step 3.1.

[0109] Another Lyapunov function V2(e f (t)) is selected as:

[0110]

[0111] Among them, ρ is a positive scalar, reflecting the convergence speed of the queue error system, and Q is a positive definite matrix.

[0112] Then the car platoon system (36) can achieve consistent ultimate boundedness condition: for a given positive scalar ρ, if there exists a symmetric positive definite matrix P, Q and positive scalars c, ω that satisfies the following inequality conditions:

[0113]

[0114] Among them, λ max (·) represents the maximum eigenvalue of the matrix, c is a positive scalar, and the positive coupling strength is limited to The constant gain matrix is ​​designed as Observer Gain ρ is a positive scalar, and the remaining parameters have been defined previously.

[0115] Proof: Definition The derivative of the first Lyapunov function V1(e(t)) is:

[0116]

[0117] For the second Lyapunov function V2(e f (t)) is derived as follows:

[0118]

[0119] comprehensive and We can get V(e(t),e f The time derivative of (t) is:

[0120]

[0121] Among them, θ is the upper bound of the first-order derivative of the actuator fault, and no information about the upper bound of the sensor is needed here.

[0122] When equation (39) holds, the queue error system is uniformly ultimately bounded.

[0123] Step 4.4: Consider the impact of communication delay, that is, only the information y of the neighboring vehicle at time t-τ(t) can be obtained j (t-τ(t)), construct the following controller:

[0124]

[0125] Among them, actuator fault compensation gain K is the constant gain matrix, δ>0 is the coupling strength, To achieve compensation for actuator failure, since the output information is delayed, the output delay information is used to obtain the sensor failure estimate, so The two compensations are used to compensate for sensor failures, and active fault-tolerant control is achieved through the above two compensations.

[0126] Then, according to the state tracking error e in (34) i (t), we can get the state tracking error dynamic equation as follows:

[0127]

[0128] Among them, e i (t-τ(t)) is the state tracking error at time t-τ(t), e fi (t-τ(t)) is the augmented state estimation error at time t-τ(t).

[0129] Its compact form is:

[0130]

[0131] where e(t-τ(t)) is the compact form of the state tracking error at time t-τ(t), e f (t-τ(t)) is a compact form of the augmented state estimate error at time t-τ(t).

[0132] According to the Newton-Leibniz formula, we have:

[0133]

[0134] The delay term in (45) can be converted into the following form:

[0135]

[0136] where e(t-τ(t)) is a compact form of the state tracking error at time t-τ(t).

[0137] Step 4.5: Design a Lyapunov function V3(e(t)) with symmetric positive definite matrices P and G,

[0138]

[0139] Among them, P -1 is the inverse of the positive definite matrix P, and G is a diagonal matrix defined in step 3.1.

[0140] The car platoon system can achieve consistent ultimate boundedness condition: for a given positive scalar ρ, if there exists a symmetric positive definite matrix P, Q, a positive scalar c, ω and an upper bound on the communication delay When the following inequality conditions are met

[0141]

[0142] The positive coupling strength is limited by The positive coupling strength is limited by The constant gain matrix is ​​designed as Observer Gain ρ is a positive scalar,

[0143] Proof: The derivative of the Lyapunov function V3(e(t)) is:

[0144]

[0145] in,

[0146]

[0147] According to the Halanay inequality, we have:

[0148]

[0149] Among them, the specific forms of β1, β2, β3, β4, and β5 are as described above.

[0150] Solving the above inequality, we have:

[0151]

[0152] When the upper bound of the delay satisfies the above conditions, we have:

[0153]

[0154] in, It means taking the maximum value of the interval function. ∈ is a positive constant, and f is a positive constant related to ∈;

[0155] Further:

[0156]

[0157] Here, ||·|| represents the bi-norm of the matrix.

[0158] The car platoon error system is exponentially convergent and uniformly ultimately bounded, while the stability of the estimation error system of the decentralized observer is guaranteed by (39).

[0159] Step 4.6: Substitute the obtained controller into the feedback linearization strategy (13) to obtain the real-time desired control torque and realize the corresponding vehicle control.

[0160] Consider a platoon consisting of 1 leader vehicle and 6 following vehicles, whose communication topology is a typical predecessor-leader following (PLF). The dynamic model of the vehicles in the platoon is given by Equation (15), and its initial state value is randomly selected. At the same time, the acceleration input a0 of the leader vehicle is set to:

[0161]

[0162] In the first vehicle, the sensor fault input f s1 and the actuator with the fault input f a1 is set to:

[0163]

[0164]

[0165] The actuator of the third vehicle is connected to the fault input f a3 and sensor fault input f s3 Set to:

[0166]

[0167] The controller gains are set to:

[0168] K=[0.7039 1.3828 0.3542], δ=5.9538, (59)

[0169] The observer gain is set as:

[0170]

[0171] In addition, the upper bound of the time-varying delay obtained by the above gain is Therefore, the time-varying communication delay is set to τ(t)=||0.0015sin(t)||+0.00001, and the above delay meets the upper bound requirement.

[0172] Figure 3The demonstration shows the tracking effect of the distributed proportional integral observer for the actuator fault and sensor fault of the first following vehicle. Figure 3 In the figure, (a) is the dynamic tracking curve of the first following vehicle actuator fault (sinusoidal signal) and the comparison between the estimated and true value; (b) is the dynamic tracking curve of the first following vehicle sensor fault (sinusoidal signal) and the comparison between the estimated and true value. Figure 3 It can be seen that the designed distributed proportional-integral observer can quickly track large-amplitude, high-frequency sinusoidal fault signals, converge quickly when the fault signal switches, has good real-time tracking performance and high accuracy, and is less affected by time-varying communication delays.

[0173] Figure 4 The demonstration shows the tracking effect of the distributed proportional integral observer for the actuator fault and sensor fault of the third following vehicle. Figure 4 In the figure, (a) is the comparison between the estimated and true value of the dynamic tracking curve of the actuator failure (combined signal) of the third following vehicle; (b) is the comparison between the estimated and true value of the dynamic tracking curve of the sensor failure (combined signal) of the third following vehicle. Figure 4 It can be seen that the designed distributed proportional-integral observer can quickly track large-value steady-state fault signals, exponential signals, linear changes and other combined fault signals, converge quickly when the fault signal switches, and is less affected by time-varying communication delays.

[0174] Figure 5 The following diagram shows the trajectory curves of the distance error, speed error, speed and acceleration of the following vehicle under the distributed controller designed by the present invention. Figure 5 In the figure, (a) is the simulation result curve of the distance tracking error of the queue system; (b) is the simulation result curve of the speed tracking error of the queue system; (c) is the simulation result curve of the acceleration of the queue system. Figure 5 It can be seen that with each adjustment of the speed of the lead vehicle, the distance error and speed error of the queue can converge to zero within about 10 seconds. This fact proves that when the car queue is in a driving environment with sensor failure, actuator failure and communication delay, the stability of the system is guaranteed under the action of the controller, that is, each following vehicle can track the lead vehicle and maintain the desired vehicle-to-vehicle distance.

[0175] The above embodiments are only preferred embodiments of the present invention and cannot be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. Active fault-tolerant control method for intelligent connected vehicle platoon under communication delay, characterized by The following steps are involved: Step 1: Composition and information collection of vehicle platoons: A vehicle platoon consists of N+1 vehicles, including a leading vehicle and N following vehicles; Use on-board sensors to obtain the vehicle's own status information, including vehicle position, driving speed and acceleration; and collect output information from adjacent vehicles and pilot vehicles in real time through the V2X wireless communication network, receive output information from vehicles with established communication connections in real time, and broadcast the output information of the vehicle itself; Step 2: Based on the information obtained from the on-board sensors and the V2X wireless communication network, a longitudinal dynamic state space model of a single vehicle with actuator failure and sensor failure is established; specifically: the nonlinear longitudinal dynamic expression of a single vehicle is derived based on Newton's second law, feedback linearization is performed based on the inverse model compensation technology, and a linearized longitudinal dynamic model of a single vehicle is obtained; the position, velocity and acceleration information of the ego vehicle is used as the state vector, actuator failure is considered, and a vehicle longitudinal motion model including actuator failure is established; sensor failure is considered, and sensor output information is constructed so that these failures can be estimated and compensated during the control process; Step 3: Based on graph theory, the information interaction between intelligent connected vehicles is described to establish the objective function of the intelligent connected vehicle formation control. Specifically: the communication topology of the vehicle queue is defined based on graph theory, and the definitions of the adjacency matrix and the Laplace matrix are given; the speed difference between the ego vehicle and the pilot vehicle, and the distance difference between the ego vehicle and the front vehicle are set as independent variables, and the objective function of the vehicle queue control is constructed to keep the formation in a constant desired formation; Step 4: For the vehicle platoon system affected by sensor failure and actuator failure, a distributed error and fault observer is established, a distributed active fault-tolerant controller is designed without delay, and a distributed active fault-tolerant controller is designed under the influence of communication delay, and the wheel driving torque required for formation control is solved in real time.

2. The active fault-tolerant control method for a platoon of intelligent connected vehicles under communication delay as claimed in claim 1, characterized in that In step 2, the longitudinal dynamic state space model of a single vehicle with actuator failure and sensor failure is established, and the specific steps include: Step 2.1: Perform dynamic analysis on the longitudinal motion of the vehicle and obtain the nonlinear dynamic model of the i-th vehicle in the queue based on Newton's second law; Among them, m i is the mass of the vehicle, F i is the driving force of the vehicle, v i is the vehicle speed, μ i is the mechanical efficiency of the transmission system, T i is the actual driving / braking torque of the vehicle, R i is the wheel radius, c fi is the lumped air resistance coefficient, g is the gravity acceleration constant, f i is the rolling resistance coefficient, ι i is the vehicle dynamic time lag constant, T ie is the expected torque; Design the feedback linearization module: Combining equations (10), (11), (12) and (13), and assuming that the dynamics of the vehicles in the platoon are isomorphic, we can obtain ι i =ι>0, the linear dynamic model of the i-th electric car is obtained: Among them, a i is the vehicle acceleration, u i is the control input, i is the vehicle dynamic time lag constant; Step 2.2, take the position, velocity and acceleration of the vehicle as the state vector: Considering the impact of actuator failure and sensor failure on the system, the feedback linearization model of the i-th intelligent connected vehicle is established: in, is the output of the system; is an output matrix with full column rank and (A, C) observable; represents the unknown time-varying actuator fault, where q represents the number of types of vehicle actuator faults. Without loss of generality, we assume that the actuator fault f ai (t) The matching condition is satisfied, and the constraint is rank(B,F) = rank(B) = 1. In addition, the actuator fault f ai (t) belongs to That is, the energy accumulated by the fault over time is controllable, and f ai (t) is a continuous smooth function whose first-order derivative is bounded, that is, represents unknown time-varying sensor faults, where r represents the number of types of vehicle sensor faults. Without loss of generality, it is assumed that the sensor fault matrix E satisfies and with a time-invariant system matrix of compatible dimensions; Define the feedback linearization model of the pilot vehicle: Among them, the definitions of A and B are the same as those of the following vehicles, x0 is the state of the leading vehicle, and y0(t) is the input of the leading vehicle. In our model, the leading vehicle is a virtual and complete leader that will not be affected by sensor-actuator failures.

3. The active fault-tolerant control method for a platoon of intelligent connected vehicles under communication delay as claimed in claim 1, characterized in that In step 3, the objective function of intelligent networked vehicle formation control is established based on the graph theory description of the information interaction between intelligent networked vehicles, including the following steps: Step 3.1: Based on graph theory, describe the communication connection topology of the N following vehicles in the car platoon system as a directed graph in Represents the set of following vehicles in the queue, Represents the set of edges with direct connectivity between vehicles; edge (i, j)∈ε indicates that vehicle j can receive information from vehicle i; based on graph theory, the communication relationship between N following vehicles is represented by the adjacency matrix The communication connection Laplacian matrix of N following vehicles is defined as in is the in-degree matrix; Based on graph theory, the traction matrix of the following vehicle and the leading vehicle is used Among them, diag means constructing a diagonal matrix from diagonal elements, m i0 =1 indicates that there is a communication connection, otherwise there is no communication connection; the Laplace matrix of the pilot vehicle and the N following vehicles is defined as: When the above directed graph contains a directed spanning tree with the leader as the root node; then, is reversible, and there exists a matrix G = diag{g1,…,g N }>0, so that: in, As defined above, it represents the Laplacian matrix of the lead vehicle and the N following vehicles; remember For {g1,…,g N }'s largest and smallest elements; for The maximum and minimum eigenvalues ​​of Step 3.2: Take the speed difference between the ego vehicle and the pilot vehicle, and the distance difference between the ego vehicle and the leading vehicle as independent variables to establish the objective function of vehicle formation control, that is, to satisfy: Among them, p i With v i denote the position and speed of vehicle i, d i,i-1 It represents the expected distance between the i-th vehicle and the i-1-th vehicle. When the objective function is satisfied, the queue can achieve the expected formation.

4. The active fault-tolerant control method for a platoon of intelligent connected vehicles under communication delay as claimed in claim 1, characterized in that In step 4, the specific steps of solving the wheel motor driving torque required for formation control in real time include: 4.1) Construct an augmented system including sensor failure, design a distributed dynamic proportional integral observer, and establish a closed-loop error system for the vehicle platoon observer; 4.2) Design a distributed active fault-tolerant controller without communication delay and substitute it into the formation model to establish a closed-loop error control system for the vehicle platoon; 4.3) Based on Lyapunov stability theory and linear matrix inequality method, the conditions for the active fault-tolerant controller to achieve the stability of the queue closed-loop error system without time-varying delay are given, and the design method of the controller gain matrix and the decentralized observer gain is obtained; 4.4) Design a distributed active fault-tolerant controller under the influence of communication delay, substitute it into the formation model, and establish a closed-loop error control system for the vehicle queue under the influence of time-varying communication delay; 4.5) Based on Lyapunov stability theory, linear matrix inequality method and Halanay inequality, the conditions for the active fault-tolerant controller to achieve the stability of the queue closed-loop error system under time-varying delay are given, that is, the upper bound of the delay that the controller and observer gain obtained in step 4.3) can tolerate is obtained; 4.6) Substitute the active fault-tolerant controller into the feedback linearization model in step 2 to calculate the expected driving torque of the vehicle wheels in real time to realize the platoon control of intelligent connected vehicles.

Images