A preset performance control method for an intelligent vehicle steering system

Abstract

A preset performance control method for an intelligent vehicle steering system is proposed, which studies the preset performance tracking control problem of an intelligent vehicle steering system with model nonlinearity and parameter uncertainty. First, a radial basis function neural network is used to online approximate the nonlinearities in the steering system that are difficult to model, including friction torque and return torque. Then, the barrier Lyapunov function technology is combined to design a preset performance controller for the intelligent vehicle's wire-controlled steering system. In the controller design, dynamic gain technology is used to compensate for the influence of unknown control gain on the system control performance. Finally, combined with the Lyapunov stability theorem analysis, it is verified that the designed controller can achieve the tracking error of the front wheel angle of the intelligent vehicle steering system to converge to the preset neighborhood of the origin within a preset time. Compared with the existing control methods, the present invention greatly shortens the time of the controllable process and improves work efficiency.

Description

Technical Field

[0001] The invention belongs to the field of angle tracking control of an intelligent vehicle steering system, and in particular relates to a preset performance control method of an intelligent vehicle steering system with model nonlinearity and parameter uncertainty. Background Art

[0002] With the development of current automobile technology, smart cars have attracted widespread attention because of their ability to improve traffic efficiency, safety and ride comfort. As one of the parts of the smart vehicle's controlled-by-wire chassis, the controlled-by-wire steering system can control the front wheel steering to make the smart vehicle drive along the planned path, thereby ensuring traffic efficiency and driving safety. At present, there are problems in the controlled-by-wire steering system, such as the difficulty in determining the type of friction model, the rationality of the assumptions based on which the tire self-aligning torque model is established, and the difficulty in accurately identifying the model parameters of the steering system in practice. This makes it difficult for the conventional controller design method based on Lyapunov stability theory to be applied in the controlled-by-wire steering system.

[0003] Obtaining the preset tracking accuracy of the steer-by-wire system (i.e., the tracking error of the steering angle converges to the preset range of the origin within a specified time) while ensuring the stability of the closed-loop system is an important factor in ensuring driving safety. Currently, there is an urgent need for a preset performance controller to meet the preset tracking performance requirements of the steering angle for driving safety. In the design of the preset tracking controller for the front wheel angle, this patent fully considers the impact of unknown nonlinearities on control performance, uses a radial basis function neural network to approximate the unknown nonlinearities of the system online, and considers the impact of unknown control gain on control performance, and uses dynamic gain technology to compensate for the impact of unknown control gain on system control performance. Summary of the invention

[0004] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art and propose a preset performance control method for an intelligent vehicle steering system, which can effectively avoid the influence of model nonlinearity and parameter uncertainty on the system, provide transient and steady-state performance of preset tracking errors, and combine radial basis function neural network technology and dynamic gain technology to estimate and compensate for the uncertain nonlinearity and unknown control gain of the system, and finally complete the design of the tracking controller.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] The present invention discloses a preset performance control method for an intelligent vehicle steering system, comprising the following steps:

[0007] S1. Establish a mathematical model of the steer-by-wire system;

[0008] S2. Use radial basis function neural network to approximate unknown nonlinearity of steer-by-wire system online;

[0009] S3. Define the tracking error between the front wheel steering angle of the steering system and the desired signal, and define the preset performance function of the front wheel steering angle tracking error;

[0010] S4. Construct barrier Lyapunov function;

[0011] S5. Combining the obstacle Lyapunov function technology, a preset performance controller that meets the requirements is designed for the wire-controlled steering system of intelligent vehicles.

[0012] As a preferred technical solution, in step S1, the mathematical model of the steer-by-wire system is:

[0013] The mathematical model of the steering motor of the steer-by-wire system is expressed as:

[0014]

[0015] Among them, J m is the moment of inertia of the motor, δ m is the motor output angle, B m is the viscous friction coefficient of the motor, τ l is the load of the motor, τ m is the output torque of the motor, τ d is the disturbance torque of the motor.

[0016] The rotation of the steering front wheel about its vertical axis intersecting the wheel center can be described as:

[0017]

[0018] Among them, J f is the moment of inertia of the front wheel, δ f is the steering angle of the front wheels, τ f and τ e are the friction torque and self-aligning torque of the front wheel, τ s It is the torque transmitted from the motor output torque to the front wheels through mechanical connections and reducers. In addition, the transmission ratio between the steering motor and the front wheels can be expressed as:

[0019]

[0020] Where μ is the total transmission ratio between the steering motor and the front wheels.

[0021] Combining equations (1)-(3), the mathematical model of the steer-by-wire system can be expressed as the following state equation:

[0022]

[0023] Where x = [x 1 ,x2 ] T =[δ f ,δ f ] T is the state variable of the steer-by-wire system, f(x)=-μ 2 (B m x 2 +τ f +τ e ) / (J f +μ 2 J m ) is the unknown nonlinearity of the system, g = μ / (J f +μ 2 J m ) is the unknown control gain of the system, d(t) = μτ d / (J f +μ 2 J m ) is the time-varying disturbance of the system, u and y are the control input and output of the steer-by-wire system.

[0024] As a preferred technical solution, in step S2,

[0025] For any continuous function f(x) in the compact set Then there exists a set of optimal weight vectors This makes the radial basis function neural network have the following approximation accuracy:

[0026]

[0027] in, is an unknown positive constant, ξ(x)=[ξ 1 (x),...,ξ m (x)] T is the basis vector of the radial basis function neural network, where ξ j (x), j = 1, ..., m usually chooses a Gaussian function, that is:

[0028]

[0029] Among them, μ i =[μ 1i ,μ 2i ,...,μ mi ] T and φ j denote the node center and variance respectively.

[0030] As a preferred technical solution, in step S3, the following variables are defined:

[0031] z 1 =yy d(7)

[0032]

[0033] Among them, y d is the expected signal of the front wheel steering angle, and α is the virtual control law to be designed below. Define the tracking error z of the front wheel steering angle 1 The default performance function is:

[0034]

[0035] Among them, k 0 , k ∞ and T p are all designed normal numbers and must satisfy |z 1 (0)|<k 0 .

[0036] As a preferred technical solution, in step S4, the following obstacle Lyapunov function is constructed:

[0037]

[0038] As a preferred technical solution, in step S5, the virtual control law α is designed as:

[0039]

[0040] Among them, η 0 and η 1 are designed normal numbers.

[0041] The designed control law is:

[0042]

[0043]

[0044] Among them, η 2 and ε is a designed positive constant, H=[1,|u 0 |] T , is the unknown parameter vector The approximation result of , the unknown normal number satisfy d(t) is the external time-varying disturbance.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. The present invention proposes to design a preset performance controller for the wire-controlled steering system of an intelligent vehicle by combining the obstacle Lyapunov function technology, so that the wire-controlled steering system with model uncertainty and external time-varying disturbance converges to a preset neighborhood of the origin within a preset time, and can meet the transient and steady-state performance of the pre-specified tracking error, thereby improving the tracking accuracy of the steering system.

[0047] 2. The present invention uses radial basis function neural network technology to online approximate the unknown nonlinearity of the wire-controlled steering system, and combines dynamic gain technology to compensate for the influence of unknown control gain on system control performance, thereby enhancing the practicality of the control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 The invention is a schematic diagram of the structure of the wire control steering system.

[0049] Figure 2 A preset tracking performance map for the front wheel turning angle of the invention.

[0050] Figure 3 A structural block diagram of the tracking control system preset for the steering system of the invention.

[0051] Figure 4 Tracking performance of the invented steer-by-wire system.

[0052] Figure 5 Tracking error of the invented steer-by-wire system d Schematic diagram of .

[0053] Figure 6 Schematic diagram of the control input u of the invented steer-by-wire system.

[0054] Figure 7 A schematic diagram of the adaptive results of the θ vector in the invented radial basis function neural network.

[0055] Figure 8 Schematic diagram of the estimation result of the W vector in the controller of the invented steer-by-wire system. DETAILED DESCRIPTION

[0056] The present invention will be further described in detail below in conjunction with the accompanying drawings and examples, but the present invention is not limited to the following embodiments.

[0057] Example

[0058] In the tracking control of the intelligent vehicle steering system, the nonlinearity of the system model and the uncertainty of parameters are considered. The schematic diagram of the steering system is shown in Figure 1 A preset performance control method for an intelligent vehicle steering system, the detailed implementation process of which includes:

[0059] S1. Establish a mathematical model of the steer-by-wire system:

[0060] The mathematical model of the steering motor of the steer-by-wire system is expressed as:

[0061]

[0062] Among them, J m is the moment of inertia of the motor, δ m is the motor output angle, B m is the viscous friction coefficient of the motor, τ l is the load of the motor, τ m is the output torque of the motor, τ d is the disturbance torque of the motor.

[0063] The rotation of the steering front wheel about its vertical axis intersecting the wheel center can be described as:

[0064]

[0065] Among them, J f is the moment of inertia of the front wheel, δ f is the steering angle of the front wheels, τ f and τ e are the friction torque and self-aligning torque of the front wheel, τ s It is the torque transmitted from the motor output torque to the front wheels through mechanical connections and reducers. In addition, the transmission ratio between the steering motor and the front wheels can be expressed as:

[0066]

[0067] Where μ is the total transmission ratio between the steering motor and the front wheels.

[0068] Combining equations (1)-(3), the mathematical model of the steer-by-wire system can be expressed as the following state equation:

[0069]

[0070] Where x = [x 1 ,x 2 ] T =[δ f ,δ f ] T is the state variable of the steer-by-wire system, f(x)=-μ 2 (B m x 2 +τ f +τ e ) / (J f +μ 2 J m ) is the unknown nonlinearity of the system, g = μ / (Jf +μ 2 J m ) is the unknown control gain of the system, d(t) = μτ d / (J f +μ 2 J m ) is the time-varying disturbance of the system, u and y are the control input and output of the steer-by-wire system.

[0071] In this example, the model parameters of the steer-by-wire system are selected as follows:

[0072] Friction torque τ of steer-by-wire system f and the return torque τ e The simulation models are selected as follows:

[0073] τ f =0.25(tanh(100x 2 )-tanh(x 2 ))+30tanh(100x 2 )+10x 2 (5)

[0074]

[0075] Among them, β and γ can be obtained through the following two-degree-of-freedom model, namely:

[0076]

[0077] The parameters of the steer-by-wire system are selected as follows:

[0078] t p =0.023m,t m =0.016m, m=2000kg, I z =1300N·m,l f =1.2m,l f =1.05m, J w =3.8kg·m 2 , J m =0.0035kg·m 2 , B m =0.018·s / rad, μ=18, v=10, C f =C r =6×10 4 The initial value of the state variable is selected as x(0) = [-0.1, 0] T ,y d =0.35sin(0.3t)rad, the simulation step size is selected as 0.001s. In the simulation, the time-varying disturbance d(t) is set as

[0079]

[0080] S2. Use radial basis function neural network to approximate unknown nonlinearity of steer-by-wire system online;

[0081] For any continuous function f(x) in the compact set Then there exists a set of optimal weight vectors This makes the radial basis function neural network have the following approximation accuracy:

[0082]

[0083] in, is an unknown positive constant, ξ(x)=[ξ 1 (x),...,ξ m (x)] T is the basis vector of the radial basis function neural network, where ξ j (x), j = 1, ..., m usually chooses a Gaussian function, that is:

[0084]

[0085] Among them, μ i =[μ 1i ,μ 2i ,…,μ mi ] T and φ j denote the node center and variance respectively.

[0086] The membership function of the radial basis function neural network system is selected as:

[0087]

[0088] The initial value of the optimal weight vector is selected as: θ(0) = zeros(4,1).

[0089] S3. Define the tracking error between the front wheel steering angle of the steering system and the desired signal, and define the preset performance function of the front wheel steering angle tracking error;

[0090] Define the following variables:

[0091] z 1 =yy d (12)

[0092]

[0093] Among them, y d is the expected signal of the front wheel steering angle, and α is the virtual control law to be designed below.

[0094] Figure 2 is the preset tracking performance map of the front wheel steering angle, and defines the tracking error z of the front wheel steering angle 1 The default performance function is:

[0095]

[0096] Among them, k 0 , k ∞ and T p are all designed normal numbers and must satisfy |z 1 (0)|<k 0 The parameters of the preset performance function (14) are selected as k 0 =0.4, k ∞ =1×10 -3 and T p =0.1s.

[0097] S4. Construct barrier Lyapunov function;

[0098] In step S4, the following barrier Lyapunov function is constructed:

[0099]

[0100] S5. Combined with the obstacle Lyapunov function technology, a preset performance controller that meets the requirements is designed for the steer-by-wire system of intelligent vehicles;

[0101] In step S5, the virtual control law α is designed as:

[0102]

[0103] Among them, η 0 and η 1 are designed normal numbers, η 0 =20 and η 1 =15.

[0104] The designed control law is:

[0105]

[0106]

[0107] Among them, η 2 and ε is a designed normal constant, η 2 = 10 and ε = 1. H = [1, |u 0 |] T , is the unknown parameter vector The approximation result of , the unknown normal number satisfy d(t) is the external time-varying disturbance.

[0108] The adaptive law design of the optimal weight vector θ and the unknown parameter vector W is:

[0109]

[0110]

[0111] The parameters of the adaptive law (19)-(20) are chosen as γ 1 =500,σ 1 =1, θ(0)=zeros(4,1), γ 2 =1×10 -2 , σ 2 =1, and

[0112] Figure 3 The control system structure diagram of the steering system is shown in Figure 2. The control parameters are set as before, and the system simulation is performed. Figure 4 Tracking performance of the invented steer-by-wire system. Figure 5 Tracking error of the invented steer-by-wire system d Schematic diagram, from Figure 4 and Figure 5 It can be seen that the steer-by-wire system can obtain the preset control performance under the action of the designed controller, that is, the tracking error can converge to |yy within 0.1s. d |≤10 -3 . Figure 6 Schematic diagram of the control input u of the invented steer-by-wire system. Figure 7 A schematic diagram of the adaptive results of the θ vector in the invented radial basis function neural network. Figure 8 Schematic diagram of the estimation result of the W vector in the controller of the invented steer-by-wire system.

[0113] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A preset performance control method for an intelligent vehicle steering system, characterized in that: The following steps are included: S1. Establish a mathematical model of the steer-by-wire system; S2. Use radial basis function neural network to approximate unknown nonlinearity of steer-by-wire system online; S3. Define the tracking error between the front wheel steering angle of the steering system and the desired signal, and define the preset performance function of the front wheel steering angle tracking error; S4. Construct barrier Lyapunov function; S5. Combined with the obstacle Lyapunov function technology, a preset performance controller that meets the requirements is designed for the steer-by-wire system of intelligent vehicles; Among them, in step S1, The mathematical model of the steer-by-wire system is: The mathematical model of the steering motor of the steer-by-wire system is expressed as: Among them, J m is the moment of inertia of the motor, δ m is the motor output angle, B m is the viscous friction coefficient of the motor, τ l is the load of the motor, τ m is the output torque of the motor, τ d is the disturbance torque of the motor; The rotation of the steering front wheel about its vertical axis intersecting the wheel center can be described as: Among them, J f is the moment of inertia of the front wheel, δ f is the steering angle of the front wheels, τ f and τ e are the friction torque and self-aligning torque of the front wheel, τ s It is the torque transmitted from the motor output torque to the front wheels through mechanical connections and reducers. In addition, the transmission ratio between the steering motor and the front wheels can be expressed as: Where μ is the total transmission ratio between the steering motor and the front wheels; Combining equations (1)-(3), the mathematical model of the steer-by-wire system can be expressed as the following state equation: Where x = [x1, x2] T =[δ f ,δ f ] T is the state variable of the steer-by-wire system, f(x)=-μ 2 (B m x2+τ f +τ e ) / (J f +μ 2 J m ) is the unknown nonlinearity of the system, g = μ / (J f +μ 2 J m ) is the unknown control gain of the system, d(t) = μτ d / (J f +μ 2 J m ) is the time-varying disturbance of the system, u and y are the control input and output of the steer-by-wire system; Among them, in step S2, For any continuous function f(x) in the compact set Then there exists a set of optimal weight vectors This makes the radial basis function neural network have the following approximation accuracy: in, is an unknown positive constant, ξ(x)=[ξ 1 (x),…,ξ m (x)] T is the basis vector of the radial basis function neural network, where ξ j (x), j = 1, ..., m usually selects a Gaussian function, that is: Among them, μ i =[μ 1i ,μ 2i ,…,μ mi ] T and φ j denote the node center and variance respectively; Among them, in step S3, For the steer-by-wire system (4), define the following variables: z1=y-y d (7) Among them, y d is the expected signal of the front wheel steering angle, α is the virtual control law to be designed below; the preset performance function of the tracking error z1 of the front wheel steering angle is defined as: Among them, k0, k ∞ and T p All are designed normal numbers and must satisfy |z1(0)|<k0; Among them, in step S5, Design the virtual control law α as: Among them, η0 and η1 are both designed normal numbers; The designed control law is: Among them, η2 and ε are designed normal constants, H=[1,|u0|] T , is the unknown parameter vector The approximation result of , the unknown normal number satisfy d(t) is the external time-varying disturbance.

2. The preset performance control method of an intelligent vehicle steering system according to claim 1, characterized in that: In step S4, Construct the following barrier Lyapunov function:

Images