A vehicle platoon adaptive fault-tolerant control method based on predetermined performance

Abstract

The application discloses a kind of vehicle formation adaptive fault-tolerant control methods based on predetermined performance, it is related to the adaptive sliding mode control technical field of vehicle formation, it includes, obtaining vehicle dynamic equation, set brake safety distance between vehicles, obtain spacing error equation, and set spacing error constraint, to obtain conversion spacing error equation;Nonlinear dynamics in third-order vehicle model is fitted using radial basis neural network, according to the error equation after conversion, design parameter adaptive law, obtain vehicle fault estimation value, neural network weight estimation value, the estimation value of interference and error, obtain adaptive fault-tolerant controller, realize fault-tolerant control to vehicle formation.The application can effectively ensure the safety of vehicle formation driving, improve fault tolerance rate when fault occurs.The application can be applied in the technical field of vehicle formation fault-tolerant control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle platoon fault-tolerant control, in particular to a vehicle platoon adaptive fault-tolerant control method based on predetermined performance. BACKGROUND

[0002] In recent years, with the rapid development of the automobile industry and the improvement of people's living standards, cars have entered thousands of households, bringing convenience to people's work and life. However, the rapid increase in the number of cars has also brought a series of problems. A large number of vehicles on the road will cause energy waste and air pollution, and also greatly increase the probability of traffic accidents, threatening people's life and property safety. Therefore, it is urgent to develop a new intelligent transportation system to solve the problems in the existing urban traffic.

[0003] The existing control research on vehicle platoon generally only considers the steady-state performance of vehicle platoon, that is, the spacing error will converge to the vicinity of zero, ignoring the research on transient performance (such as convergence speed). The predetermined performance control method is an effective tool to solve the spacing error constraint, which can make the spacing error strictly within a predefined region bounded by a decay time function. Verginis and Bechlioulis applied the predetermined performance control method to a second-order uncertain vehicle platoon system and achieved robust platoon control with predetermined performance under collision constraints and connectivity constraints. However, Verginis and Bechlioulis only guarantee the stability of a single vehicle under predetermined performance, and how to guarantee the stability of vehicle platoon under predetermined performance is still a challenge.

[0004] When the vehicle actuator fails, how to ensure the safe operation of the vehicle platoon, avoid collision, and keep the vehicle state consistent in the platoon is a problem that needs to be solved urgently. SUMMARY

[0005] In view of the shortcomings of the prior art, the present application provides a vehicle platoon adaptive fault-tolerant control method based on predetermined performance, which realizes the safe operation of vehicle platoon on the road, improves the fault tolerance rate when the actuator fails, and at the same time guarantees the stability of a single vehicle and vehicle platoon.

[0006] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows: a vehicle platoon adaptive fault-tolerant control method based on predetermined performance, comprising:

[0007] According to the dynamic characteristics and engine characteristics of the vehicle during operation, a third-order vehicle model is established;

[0008] Set the braking safety distance between vehicles, obtain the spacing error equation, and set the spacing error constraint to obtain the spacing error equation converted into an unconstrained equation;

[0009] The nonlinear dynamics in the third-order vehicle model are fitted by using a radial basis neural network, a parameter adaptive law is designed according to the converted unconstrained range error equation, vehicle fault estimation, radial basis neural network weight estimation and disturbance and error estimation are obtained on line, an adaptive fault-tolerant controller is obtained, and fault-tolerant control on the vehicle formation is realized.

[0010] Further improvement of the present application is that the method for establishing the third-order vehicle model is:

[0011] The vehicle formation is composed of one leading vehicle and N following vehicles, i is the vehicle label, i=0,1,…,N, the label of the leading vehicle is 0, and the dynamic model of the vehicle is:

[0012]

[0013] In the formula, respectively represent the derivative of the position of the vehicle, the speed, the derivative of the speed and the acceleration, t represents time, m i is the mass of the vehicle, F e,i (t) is the driving force generated by the engine, F r,i (t) = r i m i gcosθ is the sliding friction of the vehicle, F g,i (t) = m i gsinθ is the horizontal component of the gravity of the vehicle, D i (t) is an unknown disturbance in the outside world, r i is the sliding friction coefficient, g is the gravity acceleration, θ is the road slope, A i is the cross-sectional area of the vehicle, C di is the drag coefficient;

[0014] The engine model of the following vehicle is:

[0015]

[0016] In the formula, is the derivative of the driving force, t i is the engine constant, u i (t) is the pedal input torque;

[0017] The following third-order vehicle model is obtained from the dynamic model of the vehicle and the engine model of the following vehicle:

[0018]

[0019]

[0020]

[0021] wherein, is a nonlinear dynamic, is the derivative of the vehicle acceleration, w i (t) is an external disturbance.

[0022] A further improvement of the invention is that the safety distance between the vehicles is given by:

[0023]

[0024] wherein, D des,i is the ideal distance between two adjacent vehicles, d i is the safety distance between two adjacent vehicles at rest, σ is a safety factor based on road or weather conditions, A m is the absolute value of the maximum possible acceleration, M and r are two positive constants.

[0025] A further improvement of the invention is that the distance error equation is given by:

[0026]

[0027] wherein, is the distance error, its first derivative and its second derivative, respectively, D act,i is the actual distance between two adjacent vehicles, x i-1 , x i is the position of the vehicle, L i is the length of the vehicle, v i-1 , v i is the velocity of the vehicle, a i-1 , a i is the acceleration of the vehicle, is the derivative of the acceleration.

[0028] A further improvement of the invention is that the distance error constraint is given by:

[0029]

[0030] wherein, are two positive parameters, q i (t) is a strictly non-increasing performance function, e i (t) is the distance error and satisfies q i (t) > 0.

[0031] The performance function q i (t) is described by:

[0032]

[0033] where q 0i = q i (t) t＝0 > 0, q ∞i = q i (t) t→∞ > 0, φ i > 0, and q 0i > q ∞i .

[0034] Further improvement of the present application is that the unconstrained range error equation acquisition method is:

[0035] A conversion function is used to convert the constrained range error equation e i (t) into the unconstrained range error equation ε i (t). A continuous smooth and monotonically increasing function ψ i (x) is selected, and its inverse function is obtained, so that e i (t) and ε i (t) satisfy the following equation:

[0036]

[0037] wherein, The adopted ψ i (x) function is:

[0038]

[0039] Then the following unconstrained range error equation can be obtained:

[0040]

[0041] wherein, respectively, are the first and second derivatives of the unconstrained range error, M1, M2, M3 are three variables, respectively, are the first and second derivatives of the performance function.

[0042] Further improvement of the present application is that the parameter adaptive law is:

[0043]

[0044] wherein, b i , m i , a i , β i , γ i , σ η,i , σ θ,iand sigma φ,i is a positive parameter, S i is a sliding surface, Z i =[v i ,a i ] is a network input, h i (Z i ) is a Gaussian kernel function, and are the 2-norm of the RBF neural network weight estimation value and its derivative, and are the estimation value of the external disturbance and the RBF reconstruction error and its derivative, and are the estimation value of the actuator fault degree and its derivative, lambda and K d are parameters.

[0045] Further improvements of the present application are that the adaptive fault-tolerant controller is:

[0046]

[0047] In the formula, u i is a controller, is an intermediate variable, k i , K p , K i , Delta i are positive parameters, is the derivative of the sliding surface.

[0048] Further improvements of the present application are that each following vehicle in the vehicle platoon is configured with an adaptive fault-tolerant controller.

[0049] Thanks to the above technical solutions, the present application has achieved the following technical progress: a vehicle platoon adaptive fault-tolerant control method based on predetermined performance is disclosed, which comprises the following steps: obtaining a vehicle dynamic equation, setting a braking safety distance between vehicles, obtaining a distance error equation, and setting a distance error constraint, thereby obtaining a converted distance error equation; using a radial basis neural network to fit the nonlinear dynamics in a third-order vehicle model, designing a parameter adaptive law according to the converted error equation, obtaining online vehicle fault estimation value, neural network weight estimation value, disturbance and error estimation value, obtaining an adaptive fault-tolerant controller, realizing fault-tolerant control of the vehicle platoon, and ensuring the safety of the vehicle platoon driving. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to illustrate the technical solutions adopted in the embodiments of the present application, the drawings involved in the technical solutions are briefly introduced as follows. In order to more completely understand the present application and its advantages, the following description is made in conjunction with the drawings, in which:

[0051] Figure 1 A schematic diagram of a vehicle fleet consisting of seven vehicles is provided for the embodiments of the present application.

[0052] Figure 2 A flowchart of a vehicle platoon adaptive fault-tolerant control method based on predetermined performance control is provided for the embodiments of the present application.

[0053] Figure 3 Simulation diagrams of spacing errors of vehicles in a vehicle fleet at different times and performance functions are provided for the embodiments of the present application.

[0054] Figure 4 Simulation diagrams of spacing errors of vehicles in a vehicle fleet at different times are provided for the embodiments of the present application.

[0055] Figure 5 Simulation diagrams of positions of vehicles in a vehicle fleet at different times are provided for the embodiments of the present application.

[0056] Figure 6 Simulation diagrams of speeds of vehicles in a vehicle fleet at different times are provided for the embodiments of the present application.

[0057] Figure 7 Simulation diagrams of accelerations of vehicles in a vehicle fleet at different times are provided for the embodiments of the present application. DETAILED DESCRIPTION

[0058] The present application is described in detail below in conjunction with the drawings and embodiments, and the advantages and outstanding features of the present application are apparent to those skilled in the art.

[0059] In the present specification, the embodiments used to describe the principles of the present application are only illustrative and are not in any way construed as limiting the scope of the present application, and therefore the following detailed description is provided to help understand the exemplary embodiments of the present application defined by the claims and their equivalents. In addition, the technical terms and terms used herein are the same as those commonly understood by those skilled in the art. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0060] The application provides a self-adaptive fault-tolerant control method for a vehicle formation when an actuator fault occurs, adopts an online fault estimation control method, realizes fault-tolerant control, and studies the transient performance and the steady-state performance of the vehicle based on a predetermined performance control method. The method comprises a parameter self-adaptive law and an adaptive sliding mode controller based on the predetermined performance control. The parameter self-adaptive law estimates the actuator fault degree, the RBF neural network weight related parameters, the disturbance and the reconstruction error online through the longitudinal dynamics of the vehicle. The adaptive sliding mode controller converts the constrained tracking error into an unconstrained tracking error by setting a preset function, introduces a PID type sliding mode surface, and realizes the state synchronization of the vehicle formation and the completion of the cooperative control task in a limited time based on the conversion distance error equation when the fault occurs. The control mechanism provided by the application improves the safety of the vehicle formation system and has high practical value.

[0061] As shown in the embodiment of the application, Figure 1 a vehicle formation composed of one leading vehicle and six following vehicles is taken as an example, and the adaptive fault-tolerant control method for the vehicle formation based on the predetermined performance control is used to control the vehicle formation.

[0062] An adaptive fault-tolerant control method for a vehicle formation based on predetermined performance control, as shown in the formula, Figure 2 the method comprises the following steps:

[0063] Step 1: a three-order vehicle model is established according to the dynamic characteristics and the engine characteristics when the vehicle is running.

[0064] The vehicle formation is composed of one leading vehicle and six following vehicles, and i is the vehicle label i=0, 1, …, 6, wherein the label of the leading vehicle is 0, and the dynamic model of the vehicle is:

[0065]

[0066] In the formula, respectively represent the derivative of the position of the vehicle, the speed, the derivative of the speed, the acceleration, t represents time, m i is the mass of the vehicle, F e,i (t) is the driving force generated by the engine, F r,i (t) = r i m i gcosθ is the sliding friction of the vehicle, F g,i (t) = m i gsinθ is the horizontal component of the gravity of the vehicle, D i (t) is the air resistance, r i is the sliding friction coefficient, g is the gravity acceleration, and θ is the road slope. A for air damping i C for vehicle cross-sectional area di μ for drag coefficient

[0067] The following vehicle engine model is:

[0068]

[0069] where, τ is the derivative of driving force i u is engine constant i (t) is pedal input torque

[0070] Combining the engine model, the vehicle dynamics model can be rewritten as:

[0071]

[0072]

[0073]

[0074] where, w is the nonlinear dynamics i (t) is external disturbance

[0075] Step 2, set the braking safety distance between vehicles, obtain the distance error equation, and set the distance error constraint, so as to obtain the conversion distance error equation;

[0076] The braking safety distance between vehicles is a variable distance, which is:

[0077]

[0078] where, des,i D is the ideal distance between two adjacent vehicles i d is the safety distance between two adjacent vehicles at rest m σ is the safety factor based on road or weather conditions i M and r are two constants greater than zero act,i v is the speed of the vehicle

[0079] The distance error equation is:

[0080]

[0081] where, D is the distance error, the first derivative of the distance error, and the second derivative of the distance error, respectively act,i x is the actual distance between two adjacent vehicles i-1 x iL is the position of the vehicle, i L is the length of the vehicle, i-1 L is the length of the vehicle, i v is the velocity of the vehicle, i-1 v is the velocity of the vehicle, i a is the acceleration of the vehicle, a is the acceleration of the vehicle.

[0082] The spacing error constraint is given by:

[0083]

[0084] where, e(t) are two positive parameters, i q(t) is the spacing error, i q(t) is a strictly non-increasing performance function and satisfies q i (t) > 0.

[0085] The performance function q i (t) is described as:

[0086]

[0087] where q 0i = q i (t) | t＝0 > 0, q ∞i = q i (t) | t→∞ > 0, φ i > 0 are parameter variables and satisfy q 0i > q ∞i .

[0088] To solve the problem of controller design difficulty, it is necessary to obtain an unconstrained conversion spacing error equation. The present application uses a conversion function to convert the constrained spacing error equation e i (t) into an unconstrained spacing error equation ε i (t). Here a continuous smooth and monotonically increasing function ψ i (·) is selected, and its inverse function is taken to make e i (t) and ε i (t) satisfy the following equation:

[0089]

[0090] where,

[0091] The ψ i (·) function is:

[0092]

[0093] The conversion distance error equation is:

[0094]

[0095] In the formula, Unconstrained distance error, first derivative and second derivative of unconstrained distance error, M1, M2 and M3 are three variables, First derivative and second derivative of performance function.

[0096] Step 3, the RBF neural network is used to fit the nonlinear dynamics in the third-order vehicle model, the parameter adaptive law is designed according to the converted error equation, the vehicle fault estimation value, the neural network weight estimation value, the disturbance and the error estimation value are obtained online, the adaptive sliding mode controller is obtained, and the fault-tolerant control of the vehicle formation is realized.

[0097] The RBF neural network fits the nonlinear dynamics in the third-order vehicle model, and the specific process is as follows:

[0098] f i (v i ,a i ,t)=W i *T h i (Z i )+ε i

[0099] In the formula, Is the ideal weight, h i (Z i ) is the output of the Gaussian function, and epsilon i Is the reconstruction error.

[0100] The sliding mode controller is used to control the vehicle, and the selected PID type sliding surface contains the converted distance error, and the description is as follows:

[0101]

[0102] In the formula, K p >0, K i >0 and K d >0 respectively represent the proportional, integral and differential coefficients, and s i Is the sliding surface.

[0103] The coupling sliding surface is used to establish the relationship between the i-th vehicle and the i+1-th vehicle in the vehicle formation, and the coupling sliding surface is as follows:

[0104]

[0105] In the formula, Si (t) is the coupling slip surface, l > 0 is a weight factor, because there are only N following vehicles, so s N+1 is absent, in the above equation, s N+1 = 0.

[0106] The designed parameter estimation law is:

[0107]

[0108] In the equation, b i , μ i , a i , β i , γ i , σ η,i , σ θ,i and σ φ,i are positive parameters, S i is the slip surface, Z i = [v i , a i ] is the network input, h i (Z i ) is the Gaussian kernel function, and are the 2-norm and its derivative of the RBF neural network weight estimation value respectively, and are the estimation value and its derivative of the external disturbance and the RBF reconstruction error respectively, and are the estimation value and its derivative of the actuator fault degree respectively, λ and K d are parameters.

[0109] The designed adaptive fault-tolerant controller is:

[0110]

[0111] In the equation, u i is the controller, R i is an intermediate variable, k i , K p , K i , Δ i are positive parameters, is the derivative of the slip surface.

[0112] This embodiment assumes that the parameters of all vehicles are: vehicle mass m i = 1500 kg, inter-vehicle distance d i = 6.5 m when stationary, maximum acceleration A m = 7 m / s 2, M = 0.5, r = 3, engine constant τ i = 0.2 s. Vehicle length L i (m) = [4.5 45 4.5 5.5 4], external disturbance is set as w i = 0.01 sin(t). Safety factor σ = 0.2, air density Vehicle cross-sectional area A i = 2.5 m 2 , resistance coefficient C di = 0.36, initial position of vehicle x i (0) = [150 135 125.5 112.5 99.5 87 75.5], initial speed of vehicle v i (0) = [1 4 2 0 5 3 1], initial acceleration of vehicle a i (0) = [0 1 5 2 1 3 1].

[0113] In this embodiment, the ideal acceleration of the lead vehicle is set as:

[0114]

[0115] In this embodiment, the neural network contains 20 neurons, Z i (t) is the input of vehicle speed v i (t), the center points of the Gaussian basis functions are averagely distributed in the interval [-2, 5], the standard deviation of the basis functions is set as 2, and the driving efficiency function is set as h eff,i (t) = 0.7 + 0.3 sin(0.1it), the bias fault function is set as h bia,i (t) = 0.01 sin(t). The performance function q i (t) = (10-0.2) exp(-0.1t) + 0.2, the PID parameters are set as K p = 5, K i = 3, K d = 0.4, and the adaptive initial value is set as The adaptive parameters are set as a i = 15, β i = 20, γ i = 10, and σ ηi = σ θi = σ φi = 0.01 e -10t , the adjustment parameters are set as b i = 2, μ i = 0.5, Δ i = 0.5, k i = 45, λ = 0.6. The simulation results are as followsFigures 3-7 is shown. Figure 3 is the variation curve of the distance error and the performance function of each vehicle, which shows that the vehicle platoon can achieve the predetermined performance control; Figure 4 is the distance error curve, which shows that the distance error will tend to zero in a limited time; Figure 5 is the position curve, Figure 6 is the speed curve, Figure 7 is the acceleration curve.

[0116] Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present application, and are not limited thereto; although the present application has been described in detail with reference to the foregoing examples, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A vehicle platoon adaptive fault-tolerant control method based on predetermined performance, characterized in that, The application relates to a method for establishing a third-order vehicle model, a method for obtaining an unconstrained spacing error equation, a method for obtaining a parameter adaptive law, and an adaptive fault-tolerant controller. The third-order vehicle model is established according to the dynamic characteristics and engine characteristics of the vehicle in operation; A braking safety spacing between vehicles is set, a spacing error equation is obtained, and a spacing error constraint is set, so that the spacing error equation is converted into an unconstrained spacing error equation; The spacing error constraint is: wherein are two positive parameters, q i (t) is a strictly non-increasing performance function, e i (t) is a spacing error, and satisfies q i (t) > 0; The performance function q i (t) is described as: where q 0i = q i (t) t＝0 > 0, q ∞i = q i (t) t→∞ > 0, φ i > 0, and q 0i > q ∞i ; The unconstrained spacing error equation is obtained by: A conversion function is used to convert the constrained range error equation e i (t) into an unconstrained range error equation ε i (t); Select a continuous smooth and monotonically increasing function ψ i (·), take its inverse function Make e i (t) and ε i (t) satisfy the following equation: wherein The employed ψ i The (·) function is: Then the following unconstrained spacing error equation can be obtained: where ε i , are the unconstrained range error, the first and second derivatives of the unconstrained range error, respectively, and M1, M2, M3 are three variables, are the first and second derivatives of the performance function, respectively; A radial basis neural network is used to fit the nonlinear dynamics in the third-order vehicle model, a parameter adaptive law is designed according to the converted unconstrained spacing error equation, online vehicle fault estimation values, radial basis neural network weight estimation values and disturbance and error estimation values are obtained, an adaptive fault-tolerant controller is obtained, and fault-tolerant control of the vehicle formation is realized.

2. The adaptive fault-tolerant control method for vehicle platoon based on predetermined performance according to claim 1, wherein, The method for establishing the third-order vehicle model is: The vehicle formation is composed of one lead vehicle and N follower vehicles, i is taken as a vehicle label, i=0, 1,..., N, the label of the lead vehicle is 0, and the dynamic model of the vehicle is: In the formula, v i (t), a i (t) represents the derivative of the vehicle's position, velocity, derivative of velocity, and acceleration, respectively, where t represents time, m i For the mass of the vehicle, F e,i (t) represents the driving force generated by the engine, F r,i (t)=r i m i gcosθ is the sliding friction force of the vehicle, F g,i (t)=m i gsinθ represents the horizontal component of the vehicle's weight. For air resistance, D i (t) represents unknown external interference, r i Let g be the coefficient of sliding friction, g be the acceleration due to gravity, and θ be the road slope. For air damping, A i Let C be the cross-sectional area of ​​the vehicle. di This is the drag coefficient; The engine model of the follower vehicle is: wherein is the derivative of the driving force, τ i is an engine constant, u i (t) is the pedal input torque; The third-order vehicle model is obtained from the dynamic model of the vehicle and the engine model of the follower vehicle: wherein is a non-linear dynamic, is the derivative of the vehicle acceleration, w i (t) is an external disturbance.

3. The adaptive fault-tolerant control method for vehicle platoon based on predetermined performance according to claim 1, wherein, The braking safety spacing is as follows: where D des,i is the ideal distance between two adjacent vehicles, d i is the safety distance between two adjacent vehicles at rest, σ is a safety factor based on road or weather conditions, A m is the absolute value of the maximum possible acceleration, M and r are two positive constants.

4. The adaptive fault-tolerant control method for vehicle platoon based on predetermined performance according to claim 1, wherein, The spacing error equation is: where e i , are the spacing error, the first derivative of the spacing error and the second derivative of the spacing error, respectively, D act,i is the actual spacing between two adjacent vehicles, x i-1 ,x i is the position of the vehicle, L i is the vehicle length, v i-1 ,v i is the speed of the vehicle, a i-1 ,a i is the acceleration of the vehicle, is the derivative of the acceleration.

5. The adaptive fault-tolerant control method for vehicle platoon based on predetermined performance according to claim 1, wherein, The parameter adaptive law is: where b i , μ i , α i , β i , γ i , σ η,i , σ θ,i and σ φ,i are positive parameters, S i is the sliding surface, Z i = [v i , a i ] is the network input, h i (Z i ) is the Gaussian kernel function, and are the 2-norm and its derivative of the RBF neural network weight estimation , respectively, and are the estimation of the external disturbance and the RBF reconstruction error and its derivative, respectively, and are the estimation of the actuator fault degree and its derivative, respectively, λ and K d are parameters.

6. The adaptive fault-tolerant control method for vehicle platoon based on predetermined performance according to claim 1, wherein, The adaptive fault-tolerant controller is: where u i is a controller, R i is an intermediate variable, k i , K p , K i , Δ i is a positive parameter, is the derivative of the sliding surface.

7. The adaptive fault-tolerant control method for vehicle platoon based on predetermined performance according to claim 6, wherein, Each follower vehicle in the vehicle formation is configured with an adaptive fault-tolerant controller.

Images