Data-driven vehicle queue fault-tolerant tracking performance recovery control method and system

Abstract

The invention belongs to the technical field of vehicle queue control, and discloses a data-driven vehicle queue fault-tolerant tracking performance recovery control method and system. According to the method, a preset performance function is introduced into output of a vehicle queue system, a complex time-varying inequality constraint problem is converted into a simple variable bounded problem, and then a model-free adaptive controller design of the vehicle queue system in a normal state is completed by designing a pseudo-partial derivative estimator and an observer. Then, an attack exponential function is introduced on the basis of a model-free adaptive controller in a normal state, and a model-free adaptive controller based on an attack compensation mechanism is designed. Meanwhile, aiming at the problem of fault delay replacement of the actuator, a shift function is introduced into tracking errors, and a reconfigurable fault-tolerant model-free adaptive controller based on the shift function is designed. According to the invention, after the vehicle queue is subjected to actuator fault and delayed replacement, the tracking error can be reconverged to the preset performance boundary within the specified recovery time.

Description

Technical Field

[0001] This invention belongs to the field of vehicle queuing control technology, specifically relating to a data-driven vehicle queuing fault-tolerant tracking performance recovery control method and system. Background Technology

[0002] Vehicle platooning control is an intelligent transportation method that coordinates vehicle operation, enabling vehicles to travel in platoons with smaller intervals. This effectively improves road capacity, reduces fuel consumption, and enhances driving safety, and has received widespread attention in the field of intelligent transportation systems. Existing research has proposed various vehicle platooning control methods, including minimum headway analysis for heterogeneous platoons considering communication delays and pre-constrained control schemes for nonlinear third-order vehicle platoons. These vehicle platooning control methods provide a theoretical basis for achieving precise vehicle spacing maintenance and speed tracking.

[0003] However, in practical applications, vehicle platooning systems face several key challenges. First, the introduction of communication networks brings the potential risk of network attacks. Denial-of-service (DoS) attacks are particularly destructive, intermittently blocking communication channels and hindering the timely transmission of sensor data to the controller. Second, during the operation of a vehicle platooning system, the failure of one or more vehicles can lead to serious traffic accidents. Therefore, fault-tolerant control is crucial for the smooth operation of vehicle platooning systems. To achieve fault-tolerant control, existing control strategies mostly employ backup actuators; backup actuators can be divided into two main categories: static backup and dynamic backup. In a static backup structure, the backup actuator continues to operate even when there are no faults, and long-term operation makes the backup actuator components prone to wear and tear failures. A dynamic backup structure typically consists of a monitoring module and replacement rules; when the monitoring module detects a failure in the current actuator, it replaces it with a backup actuator. In actual systems, most backup actuators are in a cold standby state before activation to save spare parts life, thus requiring preheating before use, resulting in a delay in deployment. However, during delayed replacement, the system may violate its preset performance indicators. The tracking error of the faulty vehicle may deviate significantly from the expected range during replacement, or even violate safety constraints, leading to collision risks. Furthermore, vehicle platooning systems possess complex nonlinear dynamic characteristics, making it difficult to obtain accurate mathematical models. Factors in actual systems, such as throttle characteristics, mechanical transmission friction, and aerodynamic drag, are highly nonlinear and their parameters are time-varying, making it difficult to establish accurate models using traditional mechanistic modeling methods. With the increasing complexity of industrial processes, modeling methods based on first principles or system identification face numerous challenges. Summary of the Invention

[0004] The purpose of this invention is to propose a data-driven vehicle queue fault-tolerant tracking performance recovery control method to solve the problem of rapid recovery of vehicle queue tracking error performance under the influence of actuator failure and aperiodic DOS attacks.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A data-driven vehicle queuing fault-tolerant tracking performance recovery control method includes the following steps: Step 1. Establish a discrete system model of the leader vehicle and following vehicles in the vehicle platoon system; Step 2. Based on the discrete system model in Step 1, a preset performance function is introduced into the output transformation process of the vehicle queuing system to transform the complex time-varying inequality constraint problem into a simple variable boundedness problem; Step 3. Based on the output of the vehicle queuing system in Step 2, establish a compact form dynamic linearization model using the compact form dynamic linearization method; Step 4. Design an estimator to estimate the pseudo-partial derivatives, and introduce an observer to complete the design of a model-free adaptive controller for the vehicle queuing system under normal conditions. Step 5. Based on the model-free adaptive controller under normal conditions in Step 4, an attack indication function is introduced, and a model-free adaptive controller based on an attack compensation mechanism is designed to cope with aperiodic DoS attacks. Step 6. Introduce a shift function into the tracking error and design a reconfigurable fault-tolerant model-free adaptive controller based on the shift function to solve the problem of delayed replacement after actuator failure; Step 7. Use the model-free adaptive controller based on the attack compensation mechanism designed in Step 5 to realize the tracking control of the vehicle queuing system; use the reconfigurable fault-tolerant model-free adaptive controller based on the shift function designed in Step 6 to realize the recovery of the tracking performance of the vehicle queuing system after a fault occurs.

[0006] Furthermore, based on the aforementioned data-driven vehicle platoon fault-tolerant tracking performance recovery control method, this invention also proposes a corresponding data-driven vehicle platoon fault-tolerant tracking performance recovery control system, which adopts the following technical solution: The data-driven vehicle platoon fault-tolerant tracking performance recovery control system includes the following modules: The model building module is used to build discrete system models of the leader and follower vehicles in a vehicle platoon system. The problem transformation module is used to introduce a preset performance function during the output transformation process of the vehicle queuing system, transforming complex time-varying inequality constraint problems into simple variable boundedness problems. The model linearization module is used to establish a compact form dynamic linearization model based on the output of the vehicle queuing system using the compact form dynamic linearization method. The controller design module is used to design an estimator to estimate the pseudo-partial derivatives, and at the same time introduces an observer to complete the design of a model-free adaptive controller for the vehicle queuing system under normal conditions. Based on the model-free adaptive controller under normal conditions, an attack indication function is introduced, and a model-free adaptive controller based on an attack compensation mechanism is designed to deal with non-periodic DoS attacks. A shift function is introduced into the tracking error, and a reconfigurable fault-tolerant model-free adaptive controller based on the shift function is designed to solve the problem of delayed replacement after actuator failure. The system also includes a tracking control and performance recovery module, which is used to implement tracking control of the vehicle queuing system using a model-free adaptive controller based on an attack compensation mechanism; and to restore the tracking performance of the vehicle queuing system after a failure using a reconfigurable fault-tolerant model-free adaptive controller based on a shift function.

[0007] Furthermore, based on the aforementioned data-driven vehicle queue fault-tolerant tracking performance recovery control method, this invention also proposes a computer device comprising a memory and one or more processors. Executable code is stored in the memory. When the processor executes the executable code, it implements the steps of the aforementioned data-driven vehicle queue fault-tolerant tracking performance recovery control method.

[0008] Furthermore, based on the aforementioned data-driven vehicle queue fault-tolerant tracking performance recovery control method, this invention also proposes a computer-readable storage medium storing a program thereon, which, when executed by a processor, is used to implement the steps of the aforementioned data-driven vehicle queue fault-tolerant tracking performance recovery control method.

[0009] The present invention has the following advantages: As described above, this invention discloses a data-driven fault-tolerant tracking performance recovery control method for vehicle platoons. This method combines the concept of preset performance control with a model-free adaptive control method. By performing a nonlinear transformation on the system output, it transforms the complex time-varying inequality constraint problem into a simple variable boundedness problem. Thus, within a purely data-driven framework, it simultaneously guarantees the steady-state and transient performance of the vehicle platoon system's tracking error. Existing technologies typically only guarantee steady-state convergence and lack constraints on the transient process. This invention, by introducing a preset performance function, can quantitatively design the convergence speed and maximum overshoot of the tracking error, fundamentally ensuring the driving safety and comfort of the vehicle platoon. Furthermore, this invention cleverly introduces a shift function to address the problem of delayed actuator replacement and integrates it into the design of a reconfigurable fault-tolerant model-free controller. This mechanism can actively "pull back" the deteriorated tracking error in the passive case of delayed replacement of the backup actuator, enabling it to meet the preset performance requirements again within a specified recovery time. This solves the problem that existing technologies cannot handle performance degradation during the replacement window. Furthermore, this invention constructs a unified control architecture capable of simultaneously handling actuator failures with delayed replacement times and aperiodic DoS attacks. This architecture, through the collaborative work of the controller, monitoring functions, shift functions, and a reconfigurable, fault-tolerant, model-free controller, achieves resilient operation and performance self-recovery under multiple adverse factors. Existing vehicle queuing control systems are primarily designed for sensor failures; this invention addresses actuator failures, which have a more direct impact, and considers the unavoidable delay in spare actuator replacement in practical applications, thus having a wider range of applications and being more closely aligned with engineering realities. Attached Figure Description

[0010] Figure 1 This is a flowchart of the data-driven vehicle queue fault-tolerant tracking performance recovery control method in an embodiment of the present invention; Figure 2 This is a schematic diagram of a vehicle queuing system in an embodiment of the present invention; Figure 3 This is a schematic diagram of an aperiodic DoS attack in an embodiment of the present invention; Figure 4 This is a control framework diagram of the vehicle queuing system in an embodiment of the present invention; Figure 5 This is a diagram illustrating the effect of the shift function on tracking error in an embodiment of the present invention. (a) represents the tracking error of the vehicle queuing system without the shift function; (b) represents the new error variable of the vehicle queuing system after the shift function is introduced. Figure 6 The original error trajectory of the faulty vehicle based on the reconfigurable model-free adaptive controller and the new error trajectory after introducing the shift function are shown in the embodiments of the present invention. Figure 7The error trajectory curve of a faulty vehicle in an embodiment of the present invention is shown, without the use of a reconfigurable fault-tolerant model-free adaptive controller. Detailed Implementation

[0011] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments: Example 1 To address the issue of rapid performance recovery for vehicle queuing tracking errors under the influence of actuator failures and aperiodic DoS attacks, this embodiment 1 proposes a data-driven fault-tolerant tracking performance recovery control method for vehicle queuing. This method, based on a preset performance and a compact-format dynamic linearization model, transforms the nonlinear vehicle queuing system into a linear data model. On this basis, a model-free adaptive controller with preset performance constraints under normal conditions is designed. Furthermore, a model-free adaptive controller with a compensation mechanism is proposed to address the aperiodic DoS attack problem. By introducing a shift function, a reconfigurable fault-tolerant model-free adaptive controller based on the shift function is designed to achieve performance recovery after a fault, ultimately realizing the control objective of the vehicle queuing system. Figure 1 As shown, the specific steps are as follows: Step 1. Establish the dynamic equations of the leading and following vehicles in the vehicle platoon system and obtain their discrete system model.

[0012] The dynamic equations of the leading vehicle are established as follows: (1) in and It is the position and speed of the leader's vehicle. It is an unknown nonlinear function.

[0013] Following vehicle The dynamic equation is: (2) in, and It is the first The position and speed of the following vehicle; It is the first The input following the vehicle is the control input, which represents the vehicle's traction or braking force; It is an unknown nonlinear function that represents all nonlinear dynamic effects following the vehicle; Indicates the first One following vehicle This indicates the total number of following vehicles.

[0014] Therefore, the discrete system model of the leading vehicle is obtained as follows: (3) in Represents discrete time nodes. Indicates the sampling period. and They represent Time and Always monitor the location of the vehicle. and They represent Time and Always keep track of the vehicle's speed.

[0015] The discrete system model for following the vehicle is as follows: (4) in and They represent Time and Time of the first The location of the following vehicle. and They represent Time and Time of the first The speed of the following vehicle express Time of the first The input of a following vehicle.

[0016] Step 2. Based on the discrete system model in Step 1, a preset performance function is introduced into the output transformation process of the vehicle queuing system, which transforms the complex time-varying inequality constraint problem into a simple variable boundedness problem.

[0017] To achieve simultaneous position and velocity tracking, an output adjustment factor is introduced, and the output of the vehicle queuing system is redefined as follows: (5) in express Time of the first An input following the vehicle, This represents the output adjustment factor.

[0018] Further, based on formula (4), we obtain: (6) in ; in express Output adjustment factor at time.

[0019] The output tracking error of the vehicle queuing system Defined as: (7) in Indicates the reference output trajectory. Indicates the leadership vehicle and the first The safe distance between following vehicles, such as Figure 2 As shown.

[0020] Define preset performance functions for: (8) in It is defined as hyperbolic cosecant function, and It is a positive number. It is a pre-allocated time constant.

[0021] Constraint boundaries are established by using preset performance functions. This ensures that the preset tracking error performance converges and satisfies the initial conditions. ; This indicates taking the absolute value.

[0022] Substituting the specific representation of the tracking error in formula (7) into the constraint boundary, we get: (9) Therefore, the output of the vehicle queuing system can be converted and constructed as follows: (10) in It is a transformation function. Select it as ; Therefore, the output of the transformed vehicle queuing system is obtained. for: (11) in express Time of the first The output follows the vehicle's transformation, guiding the vehicle in... Output after time transformation .

[0023] Use the output transformation function The output constraint relationship, i.e., formula (9), is embedded into the output after system transformation. Therefore, it is only necessary to ensure in the controller design. If it is bounded, the output constraint relationship can be satisfied, thereby ensuring the tracking error. It satisfies the specified performance boundary constraints; it transforms the complex time-varying inequality constraint problem into a simple variable boundedness problem, thereby avoiding excessive complexity in the subsequent controller design.

[0024] Step 3. Based on the transformed output of the vehicle queuing system in Step 2, establish a compact form dynamic linearization model using the compact form dynamic linearization method.

[0025] Assuming a nonlinear function Relative to control input If the partial derivatives exist and are continuous, then the system output... Relative to control input The partial derivatives also exist and are continuous, and according to the performance function... and Given the properties of the function, the output after transformation... Relative to control input The partial derivatives also exist and are continuous.

[0026] Assuming the system output Satisfying the generalized Lipschitz condition, that is, for any fixed... and satisfy ;in , , The output is a positive constant; derived from the transformed value. Similarly, we can conclude that there exists a positive constant. Make ;in .

[0027] Based on the above assumptions, it can be determined by algebraic methods that there exists a pseudo-partial derivative. This makes the transformed output It can be transformed into a compact-format dynamic linearized model, as shown in Equation (12): (12) in .

[0028] Therefore, it was obtained and Based on this tight-form dynamic linearization model, control methods can be designed using only input and output data to establish the relationship between the two.

[0029] Step 4. Design an estimator to estimate the pseudo-partial derivatives, and introduce an observer to complete the design of a model-free adaptive controller for the vehicle queuing system under normal conditions.

[0030] The pseudo-partial derivatives obtained from step 3 The exact expression for is difficult to obtain, so an estimator is designed to estimate it.

[0031] Design pseudo-partial derivative estimation standard function for: (13) in express The estimated value, express The estimated value, for The pseudo-partial derivative at time t, These are weighting coefficients; .

[0032] By solving equation (13) about The minimum value of is used to obtain the pseudo-partial derivative estimator: (14) in It is the step size coefficient. .

[0033] To eliminate the assumption about the sign of pseudo-partial derivatives, the observer is designed as follows: (15) in It is the observation gain. express The observed values, express The observed values.

[0034] To design a model-free adaptive controller that simultaneously achieves preset tracking performance and smooth control input, a control input criterion function is designed. for: (16) in It is a weighting factor. Indicates that the leader's vehicle is in Output after time change express Time of the first The input of a following vehicle.

[0035] By finding equation (16) regarding Find the minimum value and use an observer to obtain the model-free adaptive controller under normal conditions: (17) in It is the step size factor.

[0036] Step 5. Based on the model-free adaptive controller under normal conditions in Step 4, an attack indication function is introduced, and a model-free adaptive controller based on an attack compensation mechanism is designed to deal with aperiodic DoS attacks.

[0037] Assuming the aforementioned vehicle queuing system suffers from, for example Figure 3 The non-periodic DoS attack shown is described in detail below, where the j-th attack activity period is... The attack pause period is ;in Indicates the first The start of the attack Indicates the first The moment the attack ended. Indicates the first The moment the attack ended.

[0038] gather Total DoS attack duration Defined as: (18) Total intervals between no DoS attacks for: ,and satisfy: (19) in and These are the initial attack parameters and the attack interval coefficient, respectively. This represents the difference set.

[0039] To counter the aforementioned aperiodic DoS attacks that prevent data packets from being transmitted from the sensor to the controller, an attack indication function is introduced. as follows: (20) Therefore, for The attack compensation mechanism is given by the following formula: (twenty one) in express Time of the first The output follows the vehicle's transformation.

[0040] Furthermore, based on the model-free adaptive controller under normal conditions in step 4, by introducing an attack indicator function, a model-free adaptive controller based on an attack compensation mechanism is designed as follows: (twenty two) (twenty three) (twenty four) in For positive integers, This represents a pseudo-partial derivative estimator based on an attack compensation mechanism. This refers to an observer based on an attack compensation mechanism. This represents a model-free adaptive controller based on an attack compensation mechanism. express The observed values.

[0041] Unlike the control input in step 4 At this point, the estimator and observer included in the controller both employ corresponding attack compensation mechanisms, and This is a normal number introduced to ensure the efficiency of algorithm updates. The model-free adaptive controller based on the attack compensation mechanism designed in this step can ensure that the tracking error meets the preset performance requirements.

[0042] Step 6. While the vehicle queuing system is subjected to the aperiodic DoS attack in Step 5, to address the delayed replacement problem after actuator failure, a shift function is introduced into the tracking error, and a reconfigurable fault-tolerant model-free adaptive controller based on the shift function is designed to solve the delayed replacement problem after actuator failure.

[0043] Considering the entire vehicle queuing system is subjected to the aforementioned DoS attack, one of the following vehicles in time An actuator failure occurred. The architecture of the vehicle queuing system is as follows: Figure 4 As shown.

[0044] To handle potential actuator failures, the vehicle queuing system is equipped with two identical actuators, with only one connected to the controller at any given time; the designed monitoring function detects the failure at the moment of fault detection. If a fault is detected in the currently active actuator, the faulty actuator should be immediately deactivated; within the actually determined replacement delay time... Subsequently, the backup actuator at the time replacement moment It is activated and put into operation.

[0045] During the delayed replacement period of the actuator after a fault is detected, the system experiences a significant performance degradation; the recovery time is defined as... The resolution time is And satisfy the relation .

[0046] Assuming the required recovery time Greater than the replacement delay time During the replacement delay, the faulty vehicle maintains a safe distance from the vehicles in front and behind it, and the subsequently activated backup actuator will not fail again.

[0047] After analysis, the monitoring function was designed as follows: (25) in It is a constant that can be determined in advance and can be derived from step 5.

[0048] The fault detection time is given by the following formula: (26) in It represents the infimum, which is the smallest integer that satisfies the conditions within the set.

[0049] Due to the replacement delay, the backup actuator cannot be... It doesn't start running immediately, but rather after an uncertain replacement delay. Then it starts running. It's important to note that during this period, tracking errors... The preset performance requirements may no longer be met. Specifically, the inequality... If this condition is met, the transformation function used in step 2 will become invalid, and thus the algorithm in step 5 will also fail.

[0050] In order to achieve the required recovery time Internal error Restore to range In tracking error Introducing shift functions based on this Its expression is as follows: (27) in It is a design constant. , express Time of the first The tracking error of each following vehicle.

[0051] According to the shift function The property of the replacement moment Next, new error variables were designed for the faulty vehicles. as follows: (28) in .

[0052] Functions such as Figure 5 As shown, by introducing this shift function, the new error variable is now... It always stays within the preset performance boundaries, that is, it satisfies: ; in Indicates the faulty vehicle is in The new error variable at time 10:00 Indicates the faulty vehicle is in The shift function at time step.

[0053] This means that the transformation function in step 2 can be used. For the new error variable Transformation is performed to achieve the preset performance. Therefore, the system output can be further transformed into: (29) in This indicates that after introducing a shift function transformation, the vehicle is following... The output at time t, after introducing a shift function transformation, leads the vehicle in Output at time .

[0054] Similar to the principle in step 3, there must exist pseudo-partial derivatives. , making It can be transformed into a compact-format dynamic linearized model, as shown in equation (30): (30) in , , For positive integers, To introduce a shift function transformation, the following vehicle is in Output at any given moment.

[0055] Combining step 4, design pseudo-partial derivatives. estimator for: ; , These are design parameters.

[0056] Based on step 5, incorporating an attack compensation mechanism, a reconfigurable, fault-tolerant, model-free adaptive controller based on a shift function is designed as follows: (31) (32) (33) in This represents a reconfigurable estimator. Indicates a reconfigurable observer. This represents a reconfigurable, fault-tolerant, model-free adaptive controller based on a shift function. , , For design parameters, It is a constant.

[0057] Existing vehicle queuing control systems do not consider performance recovery after failures, especially when there is a delay in replacing spare actuators. When the primary actuator fails and the backup actuator requires warm-up or activation time, the vehicle's tracking error may increase sharply during this "uncontrolled" or "out-of-control" delay, exceeding safety boundaries. Existing methods cannot guarantee performance during and after this period, and cannot restore the system to satisfactory control accuracy within a specified time. The reconfigurable fault-tolerant model-free adaptive controller based on shift functions proposed in this invention can proactively restore system performance to a preset level within a specified time after actuator replacement, achieving "performance recovery" in fault-tolerant control rather than merely "stability maintenance." This represents a qualitative improvement over existing fault-tolerant control capabilities.

[0058] Step 7. Use the model-free adaptive controller based on the attack compensation mechanism designed in Step 5 to realize the tracking control of the vehicle queuing system; use the reconfigurable fault-tolerant model-free adaptive controller based on the shift function designed in Step 6 to realize the recovery of the tracking performance of the vehicle queuing system after a fault occurs.

[0059] Figure 4 The control framework diagram of the vehicle queuing system is shown. For vehicles that are not faulty and for faulty vehicles before actuator failure, a preset performance function is introduced through output transformation based on initial data measured by sensors. The initially defined output Transform into Due to the pseudo-partial derivatives in the compact form dynamic linearization model Since these values ​​are not readily available, a pseudo-partial derivative estimator was designed at the sensor end to obtain their estimated values. Therefore, data packets Data is transmitted from the sensor to the controller via the network. Due to the presence of aperiodic DoS attacks in the network channel, attackers can disrupt data transmission from the sensor to the controller by attacking the network channel. This invention addresses this by introducing an observer to design a model-free adaptive controller based on an attack compensation mechanism. It acts on the actuator 1, which is in good working condition, to control the vehicle queuing system, so that the tracking error between the following vehicle and the leader vehicle meets the preset performance requirements.

[0060] If an actuator malfunctions unexpectedly in a vehicle queuing system, and the delayed actuator replacement process in step 6 is used for handling, then for the malfunctioning vehicle after the delayed replacement, a shift function is introduced based on the data measured by the sensors. Define a new error variable so that a preset performance function can be introduced using the output transformation. , will output Transform into Similarly, the pseudo-partial derivatives in the compact scheme dynamic linearized model at this time Since these values ​​are not readily available, a pseudo-partial derivative estimator was designed at the sensor end to obtain their estimated values. Therefore, data packets Transmission is made to the controller via the network; since aperiodic DoS attacks still exist in the network channel, this invention designs a reconfigurable fault-tolerant model-free adaptive controller based on a shift function by introducing an observer and an attack compensation mechanism. It acts on the replacement backup actuator 2 to control the faulty vehicle, enabling the vehicle's tracking error to be restored to the preset performance requirements.

[0061] To verify the effectiveness of the proposed method under actuator failure and DoS attack scenarios, this section provides a simulation example.

[0062] Consider a configuration consisting of one lead vehicle and five follower vehicles. The following vehicle queuing system is formed: ; ; in, Indicates traction or braking force. This represents nonlinear dynamic effects. In simulations, this model is used only to generate the system's input and output data, and not for designing control schemes.

[0063] The control inputs are driven by the vehicle queuing system actuators, such as Figure 4 As shown, the system has another actuator as a backup. Initially, actuator 1 is put into operation, and the corresponding actual control input is designed as follows: Suppose that while the vehicle queuing system is under an aperiodic DoS attack, the fourth following vehicle in the queue experiences an actuator failure at 6.6 seconds, causing the control input of that vehicle's actuator drive to remain at a value of 0.1. When the monitoring function detects the failure at 13.455 seconds, it activates and puts a backup actuator 2 into use within 4 seconds. During this 4-second delay, the tracking error of the faulty vehicle no longer meets the preset performance requirements. If a model-free adaptive controller based on an attack compensation mechanism is still used... The control objective cannot be achieved, therefore, in this case, the backup actuator 2 drives the reconfigurable fault-tolerant model-free adaptive controller based on the shift function. To restore the performance of the faulty vehicle.

[0064] Figure 6 The original error of a faulty vehicle based on a reconfigurable model-free adaptive controller Trajectory and new error variables The trajectory, from Figure 6 It can be seen that the standby actuator drives a reconfigurable fault-tolerant model-free adaptive controller based on a shift function. This makes the new error variable After replacement, the preset performance function is met to ensure the tracking error of the faulty vehicle is correct. Able to restore preset performance within the specified time For ease of display, , , Use respectively , , Indicates. On the contrary, Figure 7 This illustrates a situation where the controller was not redesigned after the actuator was replaced with a delayed one, resulting in tracking errors exceeding the preset performance range. This situation is dangerous for vehicle operation and could lead to serious traffic accidents.

[0065] This invention keyly solves the problem of delayed actuator replacement, specifically the situation where a backup actuator cannot immediately connect to the vehicle queuing system after a fault is detected. By introducing a shift function, the output transformation function becomes effective at the replacement time, thus designing a reconfigurable model-free adaptive controller for the backup actuator. This controller can restore the tracking error of the faulty vehicle to within a preset performance boundary within a specified time. The unified control architecture constructed by this invention can simultaneously handle actuator failures (including delayed replacement) and aperiodic DoS attacks. It can simultaneously resist actuator failures and aperiodic DoS attacks within a data-driven framework, exhibiting stronger system resilience and robustness. This invention ensures that the tracking error of all vehicles (including faulty vehicles after replacement) always meets the preset transient and steady-state performance requirements throughout the entire operation; simultaneously, it ensures that after experiencing actuator failure and delayed replacement, the tracking error of the faulty vehicle can converge back to the preset performance boundary within a specified recovery time, achieving rapid performance recovery.

[0066] Example 2 This embodiment 2 describes a data-driven vehicle queue fault-tolerant tracking performance recovery control system, which is based on the same inventive concept as the data-driven vehicle queue fault-tolerant tracking performance recovery control method in embodiment 1 above.

[0067] The data-driven vehicle platoon fault-tolerant tracking performance recovery control system includes the following modules: The model building module is used to build discrete system models of the leader and follower vehicles in a vehicle platoon system. The problem transformation module is used to introduce a preset performance function through the output transformation of the vehicle queuing system, thereby transforming complex time-varying inequality constraint problems into simple variable boundedness problems. The model linearization module is used to establish a compact form dynamic linearization model based on the output of the vehicle queuing system using the compact form dynamic linearization method. The controller design module is used to design an estimator to estimate the pseudo-partial derivatives, and at the same time introduces an observer to complete the design of a model-free adaptive controller for the vehicle queuing system under normal conditions. Based on the model-free adaptive controller under normal conditions, an attack indication function is introduced, and a model-free adaptive controller based on an attack compensation mechanism is designed to deal with non-periodic DoS attacks. A shift function is introduced into the tracking error, and a reconfigurable fault-tolerant model-free adaptive controller based on the shift function is designed to solve the problem of delayed replacement after actuator failure. The system also includes a tracking control and performance recovery module, which is used to implement tracking control of the vehicle queuing system using a model-free adaptive controller based on an attack compensation mechanism; and to restore the tracking performance of the vehicle queuing system after a failure using a reconfigurable fault-tolerant model-free adaptive controller based on a shift function.

[0068] It should be noted that any content not mentioned in the above-described functional modules of the system described in Embodiment 2 can be referred to the step description of the corresponding method in Embodiment 1 above, and will not be repeated in detail here.

[0069] Example 3 This embodiment 3 describes a computer device including a memory and one or more processors. Executable code is stored in the memory. When the processor executes the executable code, it implements the steps of the data-driven vehicle queue fault-tolerant tracking performance recovery control method in embodiment 1 above.

[0070] Example 4 This embodiment 4 describes a computer-readable storage medium storing a program that, when executed by a processor, is used to implement the steps of the data-driven vehicle queue fault-tolerant tracking performance recovery control method in embodiment 1 above.

[0071] The computer-readable storage medium can be an internal storage unit of any device or apparatus with data processing capabilities, such as a hard disk or memory, or an external storage device of any device with data processing capabilities, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc.

[0072] Of course, the above description is only a preferred embodiment of the present invention. The present invention is not limited to the above-described embodiments. It should be noted that any equivalent substitutions or obvious modifications made by those skilled in the art under the guidance of this specification fall within the scope of this specification and should be protected by the present invention.

Claims

1. A data-driven vehicle queuing fault-tolerant tracking performance recovery control method, characterized in that, Includes the following steps: Step 1. Establish a discrete system model of the leader vehicle and following vehicles in the vehicle platoon system; Step 2. Based on the discrete system model in Step 1, a preset performance function is introduced into the output transformation process of the vehicle queuing system to transform the complex time-varying inequality constraint problem into a simple variable boundedness problem; Step 3. Based on the output of the vehicle queuing system in Step 2, establish a compact form dynamic linearization model using the compact form dynamic linearization method; Step 4. Design an estimator to estimate the pseudo-partial derivatives, and introduce an observer to complete the design of a model-free adaptive controller for the vehicle queuing system under normal conditions. Step 5. Based on the model-free adaptive controller under normal conditions in Step 4, an attack indication function is introduced, and a model-free adaptive controller based on an attack compensation mechanism is designed to cope with aperiodic DoS attacks. Step 6. Introduce a shift function into the tracking error and design a reconfigurable fault-tolerant model-free adaptive controller based on the shift function to solve the problem of delayed replacement after actuator failure; Step 7. Use the model-free adaptive controller based on the attack compensation mechanism designed in Step 5 to realize the tracking control of the vehicle queuing system; use the reconfigurable fault-tolerant model-free adaptive controller based on the shift function designed in Step 6 to realize the recovery of the tracking performance of the vehicle queuing system after a fault occurs.

2. The data-driven vehicle queuing fault-tolerant tracking performance recovery control method according to claim 1, characterized in that, Step 1 specifically involves: The dynamic equations of the leading vehicle are established as follows: (1) in and It is the position and speed of the leader's vehicle. It is an unknown nonlinear function; Following vehicle The dynamic equation is: (2) in, and It is the first The position and speed of the following vehicle; It is the first The input following the vehicle is the control input, which represents the vehicle's traction or braking force; It is an unknown nonlinear function that represents all nonlinear dynamic effects following the vehicle; Indicates the first One following vehicle Indicates the total number of following vehicles; Therefore, the discrete system model of the leading vehicle is obtained as follows: (3) in Represents discrete time nodes. Indicates the sampling period. and They represent Time and Always monitor the location of the vehicle. and They represent Time and Always keep track of the vehicle's speed; The discrete system model for following the vehicle is as follows: (4) in and They represent Time and Time of the first The location of the following vehicle. and They represent Time and Time of the first The speed of the following vehicle express Time of the first The input of a following vehicle.

3. The data-driven vehicle queuing fault-tolerant tracking performance recovery control method according to claim 2, characterized in that, Step 2 specifically involves: To achieve simultaneous position and velocity tracking, the output of the vehicle queuing system is redefined as: (5) in express Time of the first The output of a following vehicle Indicates the output adjustment factor; Further, based on formula (4), we obtain: (6) in ; in express Output adjustment factor at time; The output tracking error of the vehicle queuing system Defined as: (7) in Indicates the reference output trajectory. Indicates the leadership vehicle and the first The safe distance between following vehicles; Define preset performance functions for: (8) in It is defined as hyperbolic cosecant function, and It is a positive number. It is a pre-allocated time constant; Constraint boundaries are established by using preset performance functions. This ensures that the preset tracking error performance converges and satisfies the initial conditions. ; Indicates taking the absolute value; Substituting the specific representation of the tracking error in formula (7) into the constraint boundary, we get: (9) Therefore, the output of the vehicle queuing system is transformed and constructed as follows: (10) in It is a transformation function. , ; Therefore, the output of the transformed vehicle queuing system is obtained. for: (11) in express Time of the first The output follows the vehicle's transformation, guiding the vehicle in... Output after time transformation ; Using transformation functions The output constraint relationship, i.e., formula (9), is embedded into the transformed output of the vehicle queuing system. Therefore, it is only necessary to ensure in the controller design. If it is bounded, the output constraint relationship can be satisfied, thereby ensuring the tracking error. It meets the specified performance boundary constraints.

4. The data-driven vehicle queuing fault-tolerant tracking performance recovery control method according to claim 3, characterized in that, Step 3 specifically involves: Assuming a nonlinear function Relative to control input If the partial derivatives exist and are continuous, then the system output... Relative to control input The partial derivatives also exist and are continuous, and according to the performance function... and Given the properties of the function, the output after transformation... Relative to control input The partial derivatives also exist and are continuous; Assuming the system output Satisfying the generalized Lipschitz condition, that is, for any fixed... and satisfy ;in , , The output is a positive constant; derived from the transformed value. Similarly, we can conclude that there exists a positive constant. Make ;in ; Based on the above assumptions, algebraic methods reveal the existence of pseudo-partial derivatives. This makes the transformed output It can be transformed into a compact-format dynamic linearized model, as shown in Equation (12): (12) in .

5. The data-driven vehicle queuing fault-tolerant tracking performance recovery control method according to claim 4, characterized in that, Step 4 specifically involves: Design pseudo-partial derivative estimation standard function for: (13) in express The estimated value, express The estimated value, for The pseudo-partial derivative at time t, These are weighting coefficients; ; By solving equation (13) about The minimum value of is used to obtain the pseudo-partial derivative estimator: (14) in It is the step size coefficient. ; To eliminate the assumption about the sign of pseudo-partial derivatives, the observer is designed as follows: (15) in It is the observation gain. express The observed values, express Observed values; Design control input criterion function for: (16) in It is a weighting factor. Indicates that the leader's vehicle is in Output after time change express Time of the first Input for a following vehicle; By finding equation (16) regarding Find the minimum value and use an observer to obtain the model-free adaptive controller under normal conditions: (17) in It is the step size factor.

6. The data-driven vehicle queuing fault-tolerant tracking performance recovery control method according to claim 5, characterized in that, Step 5 specifically involves: Assuming the above vehicle queuing system is subjected to a non-periodic DoS attack, the j-th attack activity period is... The attack pause period is ;in Indicates the first The start of the attack Indicates the first The moment the attack ended. Indicates the first The moment the attack ends; gather Total DoS attack duration Defined as: (18) Total intervals between no DoS attacks for: And the length of the time interval satisfy: (19) in and These are the initial attack parameters and the attack interval coefficient, respectively. Represents the difference set; To counter the aforementioned aperiodic DoS attacks, an attack indicator function is introduced. as follows: (20) Therefore, for The attack compensation mechanism is given by the following formula: (21) in express Time of the first The output follows the vehicle's transformation; Furthermore, based on the model-free adaptive controller under normal conditions in step 4, by introducing an attack indicator function, a model-free adaptive controller based on an attack compensation mechanism is designed as follows: (22) (23) (24) in For positive integers, This represents a pseudo-partial derivative estimator based on an attack compensation mechanism. This refers to an observer based on an attack compensation mechanism. This represents a model-free adaptive controller based on an attack compensation mechanism. express The observed values.

7. The data-driven vehicle queuing fault-tolerant tracking performance recovery control method according to claim 6, characterized in that, Step 6 specifically involves: To handle potential actuator failures, the vehicle queuing system is equipped with two identical actuators, with only one connected to the controller at any given time; the designed monitoring function detects the failure at the moment of fault detection. If a fault is detected in the currently active actuator, the faulty actuator should be immediately deactivated; within the actually determined replacement delay time... Subsequently, the backup actuator at the time replacement moment Activated and put into operation; During the delayed replacement period of the actuator after a fault is detected, the system experiences a significant performance degradation; the recovery time is defined as... The resolution time is And satisfy the relation ; Assuming the required recovery time Greater than the replacement delay time During the replacement delay, the faulty vehicle maintains a safe distance from the vehicles in front and behind it, and the subsequently activated backup actuator will not fail again; After analysis, the monitoring function was designed as follows: (25) in It is a predetermined constant; The fault detection time is given by the following formula: (26) in Indicates the infimum; Due to the replacement delay, the backup actuator cannot be... It doesn't start running immediately, but rather replaces the delay time. Then it starts running; In order to achieve the required recovery time Internal error Restore to range In tracking error Introducing shift functions based on this Its expression is as follows: (27) in It is a constant. , express Time of the first The tracking error of each following vehicle; According to the shift function The property of the replacement moment Next, new error variables were designed for the faulty vehicles. as follows: (28) in ; At this time, the new error variable is It always stays within the preset performance boundaries, that is, it satisfies: ； in Indicates the faulty vehicle is in The new error variable at time 10:00 Indicates the faulty vehicle is in The shift function at time; Using transformation function For the new error variable Transformation is performed to achieve the preset performance. Therefore, the system output can be further transformed into: (29) in This indicates that after introducing a shift function transformation, the vehicle is following... The output at time t, after introducing a shift function transformation, leads the vehicle in Output at time ; Then there must exist a pseudo-partial derivative. , making It can be transformed into a compact-format dynamic linearized model, as shown in equation (30): (30) in , , For positive integers, To introduce a shift function transformation, the following vehicle is in Output at any moment; Combining step 4, design pseudo-partial derivatives. estimator for: ； in , For design parameters; Based on step 5, the reconfigurable fault-tolerant model-free adaptive controller based on the shift function is designed as follows: (31) (32) (33) in This represents a reconfigurable estimator. Indicates a reconfigurable observer. This represents a reconfigurable, fault-tolerant, model-free adaptive controller based on a shift function. , , For design parameters, It is a constant.

8. A data-driven vehicle platoon fault-tolerant tracking performance recovery control system, characterized in that, Includes the following modules: The model building module is used to build discrete system models of the leader and follower vehicles in a vehicle platoon system. The problem transformation module is used to introduce a preset performance function during the output transformation process of the vehicle queuing system, transforming complex time-varying inequality constraint problems into simple variable boundedness problems. The model linearization module is used to establish a compact form dynamic linearization model based on the output of the vehicle queuing system using the compact form dynamic linearization method. The controller design module is used to design an estimator to estimate the pseudo-partial derivatives, and at the same time introduces an observer to complete the design of a model-free adaptive controller for the vehicle queuing system under normal conditions. Based on the model-free adaptive controller under normal conditions, an attack indication function is introduced, and a model-free adaptive controller based on an attack compensation mechanism is designed to deal with non-periodic DoS attacks. A shift function is introduced into the tracking error, and a reconfigurable fault-tolerant model-free adaptive controller based on the shift function is designed to solve the problem of delayed replacement after actuator failure. And a tracking control and performance recovery module, used to implement tracking control of the vehicle queuing system using a model-free adaptive controller based on an attack compensation mechanism; The designed reconfigurable fault-tolerant model-free adaptive controller based on shift function is used to restore the tracking performance of the vehicle queuing system after a failure.

Images