Automobile intelligent chassis AFS / DYC cooperative controller and control method based on extension theory

Abstract

The invention discloses an automobile intelligent chassis AFS / DYC cooperative controller and a control method based on an extension theory, an extension theory method is innovatively introduced in a vehicle chassis control technology, an extension domain and a correlation degree function are constructed through coordination control of an AFS and a DYC, and the AFS and the DYC are coordinated to realize cooperative control of the AFS and the DYC. The adaptive capacity of the dynamic characteristics of the vehicle and the stability of system mode switching under the complex working condition are improved, and the problems of dynamic characteristic changes and stability requirements of the vehicle under the complex working condition can be accurately solved on the premise that it is guaranteed that the system is easy to control. A K-means clustering method is introduced, the optimal division problem of a classical domain, an extension domain and a non-domain is solved, and the algorithm avoids the defect that a traditional method is prone to falling into local optimum through rapid iteration and high global search capacity. Meanwhile, according to the method, the determination process of the boundary value is simplified, the stability and the control performance of the vehicle are remarkably improved through clear division of the clustering result, and effective support is provided for system optimization under the complex dynamic condition.

Description

Technical Field

[0001] The invention belongs to the technical field of intelligent automobile stability control, and particularly relates to an auxiliary driving vehicle chassis AFS / DYC body stability controller and a control method based on extension theory. Background Art

[0002] The stability control of the vehicle chassis control execution link and the coordinated control between chassis systems are the main directions and research difficulties that reflect intelligence. The active front steering system (AFS) mainly applies an additional steering angle to the steering wheel to fine-tune the vehicle's posture. Therefore, AFS can control the stability of the vehicle while ensuring the vehicle's driving state. However, when the lateral force of the vehicle tire is in a nonlinear range, it is difficult for a single AFS system to achieve stable control. Direct yaw moment control (DYC) mainly controls the stability of the vehicle by distributing the additional yaw moment calculated by the controller to each wheel through a torque distribution module. However, DYC will affect the vehicle's speed to a certain extent. Combining the above-mentioned characteristics of AFS and DYC, the advantages and disadvantages of the two systems can be complemented through a coordinated control system, which can improve vehicle stability while optimizing the vehicle's handling performance and ensuring driving safety.

[0003] Many scholars have proposed a variety of methods to study and explore the AFS / DYC coordinated control problem: logic switching; fuzzy control; multi-agent method; proportional integral control (PI), etc. However, the above-mentioned coordinated control methods all have some problems: the large mutations caused by logic switching and fuzzy control during mode switching may affect the control effect of the vehicle; the multi-agent method requires information interaction between the two systems, so the calculation is large and the control is complex; the PI control method only coordinates from the perspective of the system, so the coordination allocation is not targeted enough. The extension theory proposed by Chinese scholar Professor Cai Wen is an effective method for dealing with contradictory problems. It has been widely used in many fields. It has the advantages of clear logic, systematic methods, and strong adaptability. It can effectively solve the coordination problems between subsystems in complex systems and has important potential in solving the AFS / DYC chassis coordinated control problem.

[0004] The basic idea of ​​extension control theory: Extension theory refers to the use of extension theory to study the laws and methods of dealing with contradictory problems in the control process. It describes the process of mutual conversion and quantitative and qualitative changes between "yes" and "no" of things, and mainly includes the extraction of feature quantities, two-dimensional extension sets, correlation calculation, pattern recognition, and control output calculation. The extraction of feature quantities mainly involves extracting the state quantities that the system needs to focus on. Then, according to the judgment conditions of the system performance defined in advance, these feature quantities are divided into classical domains, extension domains, and non-domains, which respectively indicate that the system is in a stable state, stable boundary, and unstable state. The following is the calculation of the correlation of feature quantities. By monitoring the state through the correlation function, the control process can smoothly transition between the classical domain, extension domain, and non-domain, and the high-precision control area can be expanded from the classical domain to the extension domain. Through this process, real-time and accurate monitoring and dynamic adjustment of the system state can be achieved, thereby significantly improving the stability and control accuracy of complex systems. Summary of the invention

[0005] In view of the problems existing in the existing intelligent vehicle chassis control technology, the present invention proposes a design of an assisted driving vehicle chassis AFS / DYC body stability controller under the framework of extension theory, and applies the extension theory to the body stability control to meet the requirements of the vehicle to quickly and accurately adjust the body posture and maintain stability under various working conditions.

[0006] The technical solution of the present invention is: an assisted driving vehicle chassis AFS / DYC body stability control method based on extension theory, comprising the following steps:

[0007] s1: Establish a two-degree-of-freedom dynamic model of the vehicle, and establish a state space expression based on the vehicle dynamics model, where the state quantity is the lateral velocity v of the vehicle y , the vehicle's yaw rate ω. Based on the idea of ​​rolling optimization of the MPC control strategy, the additional front wheel steering angle Δδ required to maintain vehicle stability is calculated according to the constructed cost function f* and the additional yaw moment ΔM z* .

[0008] s2: Selection of characteristic quantities based on extension theory. The characteristic quantities are quantitative descriptions of the vehicle driving state according to the energy method, specifically RT energy ratio and SY energy ratio.

[0009] s3: Determination of the boundary of the extension theory area. The boundary is selected by using the K-means clustering method. The area is divided through cluster analysis, and finally the boundary of the extension partition is obtained, that is,

[0010] s4: Select the vehicle's real-time RT energy ratio and the rate of change of the RT energy ratio and as feature quantities. A two-dimensional region is constructed based on the two selected feature quantities, and the boundary is determined by the K-means clustering method. Divided into classical domain, extension domain and non-domain, the correlation function K(S) is calculated AFS When K(S) AFS ≥1, the vehicle is in the classic domain, the vehicle body is relatively stable, and no AFS intervention is required, so KA=0; when 0≤K(S) AFS When KA is less than 1, the vehicle is in the extension range and the vehicle body state needs to be adjusted in time, KA = K(S) AFS ; When K(S) AFS When KA is less than 0, the vehicle is in the non-domain and the vehicle body state is extremely unstable, requiring AFS to fully intervene, so KA=1.

[0011] s5: Select the vehicle's real-time SY energy ratio and SY energy ratio change rate at the same time and as feature quantities. A two-dimensional region is constructed based on the two selected feature quantities, and the boundary is determined by the K-means clustering method. Divided into classical domain, extension domain and non-domain, the correlation function K(S) is calculated DYC When K(S) DYC ≥1, the vehicle is in the classic domain, the vehicle body is relatively stable, and DYC intervention is not required, so KD=0; when 0≤K(S) DYC When <1, the vehicle is in the extension domain and the vehicle body state needs to be adjusted in time, KD=K(S) DYC ; When K(S) DYC When KD is less than 0, the vehicle is in the non-domain and the vehicle body state is extremely unstable, requiring DYC to fully intervene, so KD = 1.

[0012] S6: Finally, the output of the extension collaborative controller Δδ is obtained f* *KA and ΔM z* *KD.

[0013] Furthermore, the specific method of establishing the vehicle state space equation in S1 is as follows:

[0014] The two-degree-of-freedom dynamic model is established as follows:

[0015]

[0016] Where m represents the mass of the vehicle, v x and v y Represent the longitudinal speed and lateral speed respectively, F yf , Fyr are the lateral forces on the front and rear axles, respectively, generated by the side slip angles of the front and rear wheels, I z is the moment of inertia of the vehicle about its center of mass, lf , l r are the distances from the center of mass to the front and rear axles, ΔM z is the additional yaw moment;

[0017] Model the lateral forces of the front and rear wheels:

[0018]

[0019] Among them C f , C r are the cornering stiffness of the front and rear wheels, α f and α r are the sideslip angles of the front and rear wheels respectively;

[0020] The sideslip angles of the front and rear wheels can be expressed by the state of the vehicle as:

[0021]

[0022] where δ f Represents the total front wheel steering angle, which includes the front wheel steering angle δ input by the driver fs and the additional front wheel steering angle Δδ f ; That is, δf = δ fs +Δδ f , Δδ f Provided by AFS system;

[0023] Define the state variable as the lateral velocity v of the vehicle y and yaw rate ω, used to measure the stability of the vehicle, namely

[0024]

[0025] Define the control input as the total front wheel steering angle δ f and the additional yaw moment ΔM z ;

[0026] Right now

[0027] According to the vehicle lateral motion equation, we can get:

[0028]

[0029] According to the vehicle's yaw motion equation, we can get:

[0030]

[0031] Write the above two equations in matrix form:

[0032]

[0033] in is the state vector; A is the system matrix, which describes the dynamic behavior of the state variables; B1 and B2 are the influence matrices of the additional front wheel angle and additional yaw moment on the state change, respectively, as follows:

[0034]

[0035] Furthermore, the design of the MPC controller of S1 is as follows:

[0036] Set the ideal reference state as:

[0037]

[0038] The ideal yaw rate ω ref It is expressed as:

[0039]

[0040] δ fs The ideal lateral speed v is the front wheel turning angle input by the driver yref Set to 0;

[0041] The cost function of MPC is used to minimize the difference between the vehicle state and the ideal state, while controlling the total front wheel angle and the additional yaw moment. The cost function is defined as

[0042]

[0043] Where Q represents the weight matrix of the state error, which is used to control the deviation of the lateral velocity and yaw rate from the reference state; R1 is the control weight matrix of the total front wheel turning angle; R2 is the control weight matrix of the additional yaw moment;

[0044] After the model is discretized, the MPC controller predicts the vehicle state in the next N steps in each time step by optimizing the cost function J and calculates the optimal total front wheel steering angle δ at each moment. f* and the additional yaw moment ΔM z* ,This optimization problem can be solved in real time by a quadratic programming solver (QP);

[0045] According to the total front wheel turning angle δ f* Find the optimal additional front wheel steering angle control input at each moment: Δδ f* =δ f* -δ fs ;

[0046] Δδ f* and ΔM z* As the input of extension coordinated control.

[0047] Furthermore, the RT energy ratio in S2 represents the ratio of turning kinetic energy to forward kinetic energy, which can be specifically expressed by the formula:

[0048]

[0049] in is the radius of inertia, β is the side slip angle of the center of mass of the vehicle, and it can usually be considered that tanβ≈β, m represents the mass of the vehicle, v represents the vehicle speed, and ω represents the yaw angular velocity of the vehicle, so the formula can be written as:

[0050]

[0051] The SY energy ratio represents the ratio of the lateral kinetic energy to the lateral sway kinetic energy, which can be specifically expressed by the formula:

[0052]

[0053] I z Represents the vehicle's moment of inertia about the z-axis.

[0054] Further, the implementation of S3 includes:

[0055] Collect the SY energy ratio and the change rate of the SY energy ratio when the vehicle is running RT energy ratio, rate of change of RT energy ratio Data, ensuring that the data covers normal and abnormal states;

[0056] To SY, RT, Normalize the data to be between 0 and 1:

[0057]

[0058] Among them, SY min , SY max Respectively represent the minimum and maximum values ​​of the SY energy ratio in all data, Respectively represent the minimum and maximum values ​​of the SY energy ratio change rate in all data, RT min , RT max Respectively represent the minimum and maximum RT energy ratios in all data, They represent the minimum and maximum values ​​of the RT energy ratio change rate in all data respectively;

[0059] Next, we determine SY and The extension theory boundary SY1, SY2,

[0060] First randomly select 3 containing SY and The information data points are used as the initial centroids, namely c1(SY a1 , ), c2(SY a2 , ), c3(SY a3 , ), for each remaining data point x i (SY i , ), calculate its distance from all centroids c k The Euclidean distance of:

[0061]

[0062] Among them, SY ak and Respectively represent the two eigenvalues ​​of the kth centroid;

[0063] According to the calculation results, each data point x i Assign them to the cluster corresponding to the centroid with the smallest distance. After the data points in each cluster are assigned, calculate the average position of all data points in the cluster and update the centroid of the cluster:

[0064]

[0065] Where N k represents the number of data points in the Kth cluster, S k represents all data points in the Kth cluster, x j Represents the coordinates of the jth data point in the Kth cluster;

[0066] Then repeat the above steps until the center of mass is stable and stop iterating; finally, a stable center of mass c is obtained. 1s (SY a1s , ), c 2s (SY a2s , ), c 3s (SY a3s , );

[0067] Finalized boundaries:

[0068]

[0069] Similarly, determine RT and The extension theory boundary RT1, RT2,

[0070] Further, in S4, the boundary The division into classic domain, extension domain and non-domain is as follows:

[0071] Classical domain: RT1 and Determined boundary composition;

[0072] Extension domain: RT2 and Determined boundaries;

[0073] Non-domain: The area beyond the first two boundaries.

[0074] 7. The AFS / DYC collaborative control method for an intelligent chassis of an automobile based on extension theory according to claim 6 is characterized in that the correlation function K(S) in S4 is AFS It is used to reflect the relationship between the current vehicle status and the AFS intervention requirements, as follows:

[0075] Assume that the characteristic quantity RT and At point P3 in the two-dimensional region, the extension distances from point P3 to the classical domain and the extension domain are ρ(P3,(P4,P1)) and ρ(P3,(P5,P2)) respectively. Considering it as a piecewise function, the calculation results are as follows:

[0076]

[0077] Correlation function K(S) AFS It reflects the stability state of the entire control process. The final value directly shows the gap between the current state and the optimal state. The calculation formula is as follows:

[0078]

[0079] in

[0080] D(P3,(P5,P2),(P4,P1))=ρ(P3,(P5,P2))-ρ(P3,(P4,P1)) (23)

[0081] When K(S) AFS When ≥1, the vehicle is in the classic domain, that is, in a relatively stable state, and does not require the intervention of AFS. At this time, KA=0 is set, that is, AFS does not participate;

[0082] When 0≤K(S) AFS <1, the vehicle is in the extension domain, that is, the state of the vehicle begins to deviate from the stable state and requires a certain degree of adjustment. At this time, KA=K(S) AFS , indicating that AFS will partially intervene to help adjust the vehicle status;

[0083] If K(S) AFS<0, the vehicle enters the non-domain, which means that the vehicle is in an extremely unstable state. At this time, AFS needs to fully intervene and make full-scale state correction. At this time, KA=1 is set, indicating that AFS is fully involved to ensure the safety and stability of the vehicle.

[0085] Further, the S5 determines the boundary It is divided into classic domain, extension domain and non-domain. The specific divisions are as follows:

[0086] Classical domain: consists of boundary values ​​SY1 and The defined area consists of a relatively stable state of the vehicle;

[0087] Extension domain: consists of boundary value SY2 and The defined area consists of a certain transition state where the vehicle has deviated from the stable state and needs to be adjusted;

[0088] Non-domain: The vehicle is in an area outside the boundary, indicating that the vehicle's state is unstable and requires emergency intervention.

[0089] Furthermore, the correlation function K(S) of S5 is DYC The calculation method of the correlation function K(S) in step S4 AFS The calculation method is the same as that of , which is used to reflect the relationship between the current state of the vehicle and the intervention requirements of the DYC system, as follows:

[0090] When K(S) DYC When ≥1, the vehicle is in the classic domain, the vehicle body is relatively stable, and DYC intervention is not required. Therefore, KD=0 is set, which means that DYC does not participate.

[0091] When 0≤K(S) DYC <1, the vehicle is in the extension domain, indicating that the vehicle's state begins to deviate from the stable state and needs appropriate adjustment. At this time, the DYC system partially intervenes to help the vehicle adjust its state, so KD=K(S) is set. DYC , that is, DYC intervenes and adjusts according to the degree of state deviation;

[0092] When K(S) DYC When KD is <0, it means that the vehicle has entered the non-domain and is in an extremely unstable state. The dynamic performance of the vehicle may be abnormally serious and requires the full intervention of the DYC system to restore the stability of the vehicle. Therefore, setting KD=1 means that the DYC system is fully involved to ensure vehicle safety.

[0093] The present invention also proposes an automobile intelligent chassis AFS / DYC collaborative controller based on extension theory, which can execute the contents of the above-mentioned automobile intelligent chassis AFS / DYC collaborative control method based on extension theory.

[0094] Beneficial effects of the present invention:

[0095] (1) The present invention innovatively introduces the extension theory method into the vehicle chassis control technology. By coordinating the control of AFS and DYC, the extension domain and correlation function are constructed, which improves the adaptability of the vehicle dynamic characteristics under complex working conditions and the smoothness of the system mode switching. It can accurately respond to the changes in vehicle dynamic characteristics and stability requirements under complex working conditions while ensuring that the system is easy to control.

[0096] (2) The K-means clustering method is introduced to solve the optimal division problem of the classical domain, the extension domain and the non-domain. The algorithm avoids the disadvantage of traditional methods that they are prone to fall into local optimality with its fast iteration and strong global search ability. At the same time, this method simplifies the process of determining the boundary value. Through the clear division of clustering results, the vehicle stability and control performance are significantly improved, providing effective support for system optimization under complex dynamic conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0097] Figure 1 This is a schematic diagram of the control framework

[0098] Figure 2 is the correlation function K(S) AFS Calculation diagram

[0099] Figure 3 is the correlation function K(S) DYC Calculation diagram DETAILED DESCRIPTION

[0100] In order to facilitate the understanding of those skilled in the art, the present invention is further described below in conjunction with embodiments and drawings. The contents mentioned in the implementation modes are not intended to limit the present invention.

[0101] The control method framework of the present invention is as follows Figure 1 As shown, the control method includes the following steps:

[0102] Step 1: Establish the two-degree-of-freedom dynamic differential equation of the whole vehicle, considering the two degrees of freedom of lateral and yaw

[0103]

[0104] Where m represents the mass of the vehicle, v x and v y Represent the longitudinal speed and lateral speed respectively, F yf , Fyr are the lateral forces on the front and rear axles, respectively, generated by the side slip angles of the front and rear wheels. z is the moment of inertia of the vehicle about its center of mass, l f , l r are the distances from the center of mass to the front and rear axles, respectively. ΔMz is the additional yaw moment, and ω represents the yaw angular velocity.

[0105] By default, the lateral force of the tire is proportional to its sideslip angle (in the approximate linear region), so the lateral force of the front and rear wheels is:

[0106]

[0107] Among them C f , C r are the cornering stiffness of the front and rear wheels, α f and α r are the sideslip angles of the front and rear wheels respectively.

[0108] The sideslip angles of the front and rear wheels can be expressed by the state of the vehicle as:

[0109]

[0110] where δ f Represents the total front wheel steering angle, which includes the front wheel steering angle δ input by the driver fs and the additional front wheel steering angle Δδ f That is, δf=δ fs +Δδ f , and Δδ f Provided by AFS system.

[0111] Define the state variable as the lateral velocity v of the vehicle y and yaw rate ω, which are used to measure the stability of the vehicle.

[0112] Right now

[0113] Define the control input as the total front wheel steering angle δ f and the additional yaw moment ΔM z .

[0114] Right now

[0115] According to the vehicle lateral motion equation, we can get:

[0116]

[0117] According to the vehicle's yaw motion equation, we can get:

[0118]

[0119] Write the above two equations in matrix form:

[0120]

[0121] in is the state vector, A is the system matrix, describing the dynamic behavior of the state variables, B1 and B2 are the influence matrices of the additional front wheel angle and additional yaw moment on the state change, respectively, as follows:

[0122]

[0123] The MPC controller is established below. The purpose of the MPC controller is to control the total front wheel steering angle δ f and the additional yaw moment ΔM z , so that the vehicle's yaw rate ω and lateral velocity v y As close to the ideal steady state as possible.

[0124] Set the ideal reference state as:

[0125]

[0126] The ideal yaw rate ω ref It can be expressed as:

[0127]

[0128] δ fs is the front wheel turning angle input by the driver, the ideal lateral speed v y,ref Set to 0, l represents the vehicle wheelbase, K u is the stability factor.

[0129] The cost function of MPC is used to minimize the difference between the vehicle state and the ideal state, while controlling the total front wheel angle and the size of the additional yaw moment. The cost function J is defined as

[0130]

[0131] Where Q represents the weight matrix of state error, which is used to control the deviation of lateral velocity and yaw rate from the reference state; R1 is the control weight matrix of the total front wheel steering angle; R2 is the control weight matrix of the additional yaw moment; x(k) represents the real-time state matrix of the vehicle, u1(k) represents the real-time front wheel steering angle input of the vehicle, and u2(k) represents the real-time additional yaw moment input of the vehicle.

[0132] After discretizing the vehicle dynamics model, the MPC controller predicts the vehicle state in the next N steps in each time step by optimizing the cost function J and calculates the optimal total front wheel steering angle δ at each moment. f* and the additional yaw moment ΔM z* This optimization problem can be solved in real time by an existing quadratic programming solver (QP).

[0133] Then the optimal auxiliary front wheel steering angle control input at each moment can be obtained: Δδf* =δ f* -δ fs .

[0134] Δδ f* and ΔM z* As the input of the extension coordinated control part.

[0135] Step 2: Selection of extension theory feature quantities

[0136] The energy method is used to quantitatively describe the vehicle's driving state, specifically the RT energy ratio and the SY energy ratio.

[0137] RT energy ratio, which represents the ratio of turning kinetic energy to forward kinetic energy, can be expressed by the formula:

[0138]

[0139] Among them, RKE is turning kinetic energy, TKE is forward kinetic energy, is the radius of inertia, β is the side slip angle of the center of mass of the vehicle, and it can usually be considered that tanβ≈β, m represents the mass of the vehicle, v represents the vehicle speed, and ω represents the yaw angular velocity of the vehicle, so the formula can be written as:

[0140]

[0141] SY energy ratio, which indicates the ratio of the lateral kinetic energy to the lateral sway kinetic energy, can be expressed by the formula:

[0142]

[0143] I z It represents the moment of inertia of the vehicle around the z-axis, SKE represents the lateral kinetic energy, and YKE represents the lateral kinetic energy.

[0144] Step 3: Determination of the boundary of the extension theory region

[0145] Collect the SY energy ratio and SY energy ratio change rate of a large number of vehicles during operation RT energy ratio, rate of change of RT energy ratio Data, to ensure that the data covers normal, abnormal and other states.

[0146] In order to eliminate the influence of the eigenvalue size, SY, RT, Normalization is performed to scale the data between 0 and 1.

[0147]

[0148] Among them, SY min , SY maxThey represent the minimum and maximum SY energy ratios in all data, respectively. Respectively represent the minimum and maximum values ​​of the SY energy ratio change rate in all data. min , RT max They represent the minimum and maximum RT energy ratios in all data, respectively. They represent the minimum and maximum values ​​of the RT energy ratio change rate in all data respectively.

[0149] Next, we determine SY and The extension theory boundary SY1, SY2,

[0150] In the present invention, k is 3, that is, 3 random numbers including SY and The information data points are used as the initial centroids, namely c1(SY a1 , ), c2(SY a2 , ), c3(SY a3 , ), for each remaining data point x i (SY i ,

[0151] ), calculate its distance from all centroids c k The Euclidean distance of:

[0152]

[0153] Among them, SY ak and They represent the two eigenvalues ​​of the kth centroid, k=3.

[0154] According to the calculation results, each data point x i Assign it to the cluster corresponding to the centroid with the smallest distance. After the data points in each cluster are assigned, calculate the average position of all data points in the cluster and update the centroid of the cluster:

[0155]

[0156] Where N k represents the number of data points in the kth cluster, S k represents all data points in the kth cluster, x j Represents the coordinates of the jth data point in the kth cluster.

[0157] Then repeat the above steps until the center of mass is stable and stop iterating. Finally, we get a stable center of mass c 1s (SY a1s , ), c 2s (SY a2s , ), c 3s (SY a3s , ).

[0158] Finalized boundaries:

[0159]

[0160] Similarly, RT and The extension theory boundary RT1, RT2,

[0161] Step 4: First, select the vehicle's RT energy ratio and the rate of change of the RT energy ratio As feature quantities. These two feature quantities constitute a two-dimensional feature space, which is used to describe the dynamic behavior and state of the vehicle. Then, the K-means clustering method is used to delineate RT and These boundary values ​​define three typical state areas of the vehicle: 1. Classical domain: composed of RT1 and 2. Extendable domain: RT2 and 3. Non-domain: The area beyond the first two boundaries. By dividing these areas, a function K(S) associated with the AFS system can be calculated. AFS ,

[0162] like Figure 2 As shown, let the characteristic quantities of the current vehicle (RT and ) is located at point P3 in the two-dimensional region. The extension distances from point P3 to the classical domain and the extension domain are ρ(P3,(P4,P1)) and ρ(P3,(P5,P2)) respectively. They can be regarded as piecewise functions. The calculation results are as follows:

[0163]

[0164] Correlation function K(S) AFS It reflects the stability state of the entire control process. The final value directly shows the gap between the current state and the optimal state. The calculation formula is as follows:

[0165]

[0166] in

[0167] D(P3,(P5,P2),(P4,P1))=ρ(P3,(P5,P2))-ρ(P3,(P4,P1)) (46)

[0168] This function reflects the relationship between the current vehicle status and the AFS intervention requirement.

[0169] When K(S) AFS When ≥1, the vehicle is in the classic domain, that is, in a relatively stable state, and does not require the intervention and adjustment of AFS. Therefore, KA=0 is set, that is, the weight coefficient of AFS is 0, and AFS does not participate.

[0170] When 0≤K(S) AFS <1, the vehicle is in the extension domain, that is, the state of the vehicle begins to deviate from the stable state and requires a certain degree of adjustment. At this time, KA=K(S) AFS , indicating that AFS will partially intervene to help adjust the vehicle status.

[0171] If K(S) AFS <0, the vehicle enters the non-domain, which means that the vehicle is in an extremely unstable state. At this time, AFS needs to fully intervene and make full-scale state correction. Therefore, KA=1 is set, indicating that AFS is fully involved to ensure the safety and stability of the vehicle.

[0172] Step 5: At the same time, select the actual SY energy ratio of the vehicle and the rate of change of the SY energy ratio As feature quantities. These two feature quantities constitute a two-dimensional feature space, which can effectively describe the dynamic performance of the vehicle in terms of yaw moment and lateral moment. Then, the boundary is determined based on the K-means clustering method, and the feature quantities SY and The boundary value SY1, SY2, The vehicle state area is divided into the following three areas: 1. Classical domain: composed of boundary values ​​SY1 and The area defined by SY2 represents that the vehicle is in a relatively stable state. 2. Extension domain: It is composed of boundary value SY2 and The defined area represents that the vehicle is in a certain transition state, deviating from the stable state and requiring adjustment. 3. Non-domain: The area where the vehicle is outside the boundary indicates that the vehicle is unstable and requires emergency intervention. Through the division of the above areas, a correlation function K(S) related to the DYC system can be calculated. DYC The calculation process is the same as the correlation function K(S) in step 4. AFS The calculation method is consistent with that of Figure 3 . K(S) DYC Used to reflect the relationship between the vehicle's current status and the DYC system's intervention requirements.

[0173] When K(S) DYCWhen ≥1, the vehicle state is in the classical domain, the vehicle body is relatively stable, and DYC intervention is not required. Therefore, KD=0 is set, that is, the weight coefficient of DYC is 0, and DYC does not participate.

[0174] When 0≤K(S) DYC When KD is less than 1, the vehicle is in the extension domain, indicating that the vehicle's state begins to deviate from the stable state and needs appropriate adjustment. At this time, the DYC system partially intervenes to help the vehicle adjust its state, so KD is set to K(S). DYC , that is, DYC intervenes and adjusts according to the degree of state deviation.

[0175] When K(S) DYC When KD is <0, it means that the vehicle has entered the non-domain and is in an extremely unstable state. The dynamic performance of the vehicle may be extremely serious, and the DYC system needs to intervene completely to restore the stability of the vehicle. Therefore, setting KD = 1 means that the DYC system is fully involved to ensure vehicle safety.

[0176] Step 6: Get Δδ f* *KA and ΔM z* *KD, input to AFS and DYC distribution execution module to complete the vehicle body stability control.

[0177] An embodiment of the present invention further provides a controller, such as an on-board control device of a smart car, which can execute the above control method.

[0178] The series of detailed descriptions listed above are only specific descriptions of feasible implementation methods of the present invention. They are not intended to limit the scope of protection of the present invention. All equivalent methods or changes that do not deviate from the technical creation of the present invention should be included in the scope of protection of the present invention.

Claims

1. A collaborative control method of AFS / DYC for intelligent chassis of automobile based on extension theory, characterized in that: These include: S1: Establish a two-degree-of-freedom dynamic model of the vehicle, and establish the state space equation based on the vehicle dynamics model, where the state quantity is defined as the lateral velocity v of the vehicle y , yaw rate ω, and the control input is the total front wheel steering angle δ f and the additional yaw moment ΔM z Based on the MPC controller, the additional front wheel steering angle Δδ required to maintain vehicle stability is calculated according to the constructed cost function f* and the additional yaw moment ΔM z* ; S2: Select the extension theory feature quantity and quantitatively describe the vehicle driving state according to the energy method. The feature quantities are RT energy ratio and SY energy ratio; S3: Determine the boundary of the extension theory area. The boundary is selected using the K-means clustering method. The area is divided through cluster analysis, and finally the boundary of the extension partition is obtained, namely RT1. RT2, SY1, SY2, S4: Select the real-time RT energy ratio and the change rate of the RT energy ratio of the vehicle and as feature quantities, a two-dimensional region is constructed based on the two selected feature quantities, and the determined boundary RT1, RT2, Divided into classical domain, extension domain and non-domain, the correlation function K(S) of the AFS system is calculated AFS , according to the correlation function K(S) AFS The scope determines the degree of AFS system intervention control; S5: Select the vehicle's real-time SY energy ratio and SY energy ratio change rate and as feature quantities, a two-dimensional region is constructed based on the two selected feature quantities, and the determined boundary SY1, SY2, Divided into classical domain, extension domain and non-domain, the correlation function K(S) is calculated DYC , according to the correlation function K(S) DYC The scope determines the degree of DYC system intervention control; S6: The degree of AFS system intervention in step S4 and the degree of DYC system intervention in step S5 are combined to obtain the final extension collaborative control output Δδ f* *KA and ΔM z* *KD.

2. According to claim 1, the AFS / DYC collaborative control method for an intelligent chassis of an automobile based on extension theory is characterized in that: The specific method of establishing the vehicle state space equation in S1 is as follows: The two-degree-of-freedom dynamic model is established as follows: Where m represents the mass of the vehicle, v x and v y Represent the longitudinal speed and lateral speed respectively, F yf , Fyr are the lateral forces on the front and rear axles, respectively, generated by the side slip angles of the front and rear wheels, I z is the moment of inertia of the vehicle about its center of mass, l f , l r are the distances from the center of mass to the front and rear axles, ΔM z is the additional yaw moment; Model the lateral forces of the front and rear wheels: Among them C f , C r are the cornering stiffness of the front and rear wheels, α f and α r are the sideslip angles of the front and rear wheels respectively; The sideslip angles of the front and rear wheels can be expressed by the state of the vehicle as: where δ f Represents the total front wheel steering angle, which includes the front wheel steering angle δ input by the driver fs and the additional front wheel steering angle Δδ f ; That is, δf = δ fs +Δδ f , Δδ f Provided by AFS system; Define the state variable as the lateral velocity v of the vehicle y and yaw rate ω, used to measure the stability of the vehicle, namely Define the control input as the total front wheel steering angle δ f and the additional yaw moment ΔM z ; Right now According to the vehicle lateral motion equation, we can get: According to the vehicle's yaw motion equation, we can get: Write the above two equations in matrix form: in is the state vector; A is the system matrix, which describes the dynamic behavior of the state variables; B1 and B2 are the influence matrices of the additional front wheel angle and additional yaw moment on the state change, respectively, as follows:

3. The AFS / DYC collaborative control method for an intelligent chassis of an automobile based on extension theory according to claim 2 is characterized in that: The design of the MPC controller of S1 is as follows: Set the ideal reference state as: The ideal yaw rate ω ref It is expressed as: δ fs The ideal lateral speed v is the front wheel turning angle input by the driver y,ref Set to 0; l is the vehicle wheelbase, K u is the stability factor. The cost function of MPC is used to minimize the difference between the vehicle state and the ideal state, while controlling the total front wheel angle and the additional yaw moment. The cost function J is defined as Where Q represents the weight matrix of the state error, which is used to control the deviation of the lateral velocity and yaw rate from the reference state; R1 is the control weight matrix of the total front wheel steering angle; R2 is the control weight matrix of the additional yaw moment; After the vehicle dynamics model is discretized, the MPC controller predicts the vehicle state in the next N steps in each time step by optimizing the cost function J and calculates the optimal total front wheel steering angle δ at each moment. f* and the additional yaw moment ΔM z* ,This optimization problem can be solved in real time by a quadratic programming solver (QP); According to the total front wheel turning angle δ f* Find the optimal additional front wheel steering angle control input at each moment: Δδ f* =δ f* -δ fs ; Δδ f* and ΔM z* As the input of the extension coordinated control part.

4. The AFS / DYC collaborative control method for an intelligent chassis of an automobile based on extension theory according to claim 1 is characterized in that: The RT energy ratio in S2 represents the ratio of turning kinetic energy to forward kinetic energy, which can be specifically expressed by the formula: in is the radius of inertia, β is the sideslip angle of the vehicle’s center of mass, and it can usually be considered that tanβ≈β, m represents the mass of the vehicle, v represents the vehicle speed, and ω represents the vehicle’s yaw rate, so it can be written as: The SY energy ratio represents the ratio of the lateral kinetic energy to the lateral sway kinetic energy, which can be specifically expressed by the formula: I z Represents the vehicle's moment of inertia about the z-axis.

5. The AFS / DYC collaborative control method for an intelligent chassis of an automobile based on extension theory according to claim 1 is characterized in that: The implementation of S3 includes: Collect the SY energy ratio and the change rate of the SY energy ratio when the vehicle is running RT energy ratio, rate of change of RT energy ratio Data, ensuring that the data covers normal and abnormal states; To SY, RT, Normalize the data to be between 0 and 1: Among them, SY min , SY max Respectively represent the minimum and maximum values ​​of the SY energy ratio in all data, Respectively represent the minimum and maximum values ​​of the SY energy ratio change rate in all data, RT min , RT max Respectively represent the minimum and maximum RT energy ratios in all data, They represent the minimum and maximum values ​​of the RT energy ratio change rate in all data respectively; Next, we determine SY and The extension theory boundary SY1, SY2, First randomly select 3 containing SY and The information data points are used as the initial centroids, respectively For each of the remaining data points Calculate its relative position to all centroids c k The Euclidean distance of: Among them, SY ak and Respectively represent the two eigenvalues ​​of the kth centroid; According to the calculation results, each data point x i Assign them to the cluster corresponding to the centroid with the smallest distance. After the data points in each cluster are assigned, calculate the average position of all data points in the cluster and update the centroid of the cluster: Where N k represents the number of data points in the kth cluster, S k represents all data points in the kth cluster, x j Represents the coordinates of the jth data point in the kth cluster; Then repeat the above steps until the center of mass is stable and stop iterating; finally, a stable center of mass is obtained. Finalized boundaries: Similarly, determine RT and The extension theory boundary RT1, RT2, 6. The AFS / DYC collaborative control method for an intelligent chassis of an automobile based on extension theory according to claim 1 is characterized in that: In S4, the boundary RT1, RT2, The division into classic domain, extension domain and non-domain is as follows: Classical domain: RT1 and Determined boundary composition; Extension domain: RT2 and Determined boundaries; Non-domain: The area beyond the first two boundaries.

7. The AFS / DYC collaborative control method for an intelligent chassis of an automobile based on extension theory according to claim 6 is characterized in that: The correlation function K(S) in S4 is AFS It is used to reflect the relationship between the current vehicle status and the AFS intervention requirements, as follows: Assume that the characteristic quantity RT and At point P3 in the two-dimensional region, the extension distances from point P3 to the classical domain and the extension domain are ρ(P3,(P4,P1)) and ρ(P3,(P5,P2)) respectively. Considering it as a piecewise function, the calculation results are as follows: Correlation function K(S) AFS It reflects the stability state of the entire control process. The final value directly shows the gap between the current state and the optimal state. The calculation formula is as follows: in D(P3,(P5,P2),(P4,P1))=ρ(P3,(P5,P2))-ρ(P3,(P4,P1)) (23) When K(S) AFS When ≥1, the vehicle is in the classic domain, that is, in a relatively stable state, and does not require the intervention of AFS. At this time, KA=0 is set, that is, AFS does not participate; When 0≤K(S) AFS <1, the vehicle is in the extension domain, that is, the state of the vehicle begins to deviate from the stable state and requires a certain degree of adjustment. At this time, KA=K(S) AFS , indicating that AFS will partially intervene to help adjust the vehicle status; If K(S) AFS <0, the vehicle enters the non-domain, which means that the vehicle is in an extremely unstable state. At this time, AFS needs to fully intervene and make full-scale state correction. At this time, KA=1 is set, indicating that AFS is fully involved to ensure the safety and stability of the vehicle.

8. The AFS / DYC collaborative control method for an intelligent chassis of an automobile based on extension theory according to claim 1 is characterized in that: The S5 will determine the boundary of SY1, SY2, It is divided into classic domain, extension domain and non-domain. The specific divisions are as follows: Classical domain: consists of boundary values ​​SY1 and The defined area consists of a relatively stable state of the vehicle; Extension domain: consists of boundary value SY2 and The defined area consists of a certain transition state where the vehicle has deviated from the stable state and needs to be adjusted; Non-domain: The vehicle is in an area outside the boundary, indicating that the vehicle's state is unstable and requires emergency intervention.

9. The AFS / DYC collaborative control method for an intelligent chassis of an automobile based on extension theory according to claim 8 is characterized in that: The correlation function K(S) of S5 DYC The calculation method of the correlation function K(S) in step S4 AFS The calculation method is the same as that of , which is used to reflect the relationship between the current state of the vehicle and the intervention requirements of the DYC system, as follows: When K(S) DYC When ≥1, the vehicle is in the classic domain, the vehicle body is relatively stable, and DYC intervention is not required. Therefore, KD=0 is set, which means that DYC does not participate. When 0≤K(S) DYC <1, the vehicle is in the extension domain, indicating that the vehicle's state begins to deviate from the stable state and needs appropriate adjustment. At this time, the DYC system partially intervenes to help the vehicle adjust its state, so KD=K(S) is set. DYC , that is, DYC intervenes and adjusts according to the degree of state deviation; When K(S) DYC When KD is <0, it means that the vehicle has entered the non-domain and is in an extremely unstable state. The dynamic performance of the vehicle may be abnormally serious and requires the full intervention of the DYC system to restore the stability of the vehicle. Therefore, setting KD=1 means that the DYC system is fully involved to ensure vehicle safety.

10. An AFS / DYC collaborative controller for intelligent chassis of automobile based on extension theory, characterized in that: The controller is capable of executing the contents of the AFS / DYC collaborative control method for an intelligent chassis of an automobile based on extension theory as described in claim 1.

Images