A stability control method and system for a distributed drive electric vehicle

Abstract

The application discloses a kind of distributed drive electric vehicle stability control method and system, desired centroid side slip angle and yaw angular velocity are calculated as control target according to vehicle dynamics model, actual centroid side slip angle and yaw angular velocity are estimated in real time by Kalman filtering.Further through stability judging law, vehicle state is identified using phase plane method.According to vehicle state, it is divided into: stable domain, coordination domain and non-stable domain.Further control target value and actual value are input into model reference adaptive controller, and the size of additional yaw moment and front wheel angle is respectively solved.The weight size of the two is calculated by the vehicle stability region determined in stability law.Additional front wheel angle can be directly input into vehicle actuator, if direct torque control is needed, additional yaw moment size needs to be input into torque distribution module for real-time solution, and finally left and right wheel torque value of front axle is output to vehicle actuator.

Description

TECHNICAL FIELD

[0001] The present application relates to the chassis coordinated control technology of the distributed drive electric vehicle, and particularly relates to the compensation of the front wheel rotation angle, the additional yaw moment and the weight coordination design between the AFS and the DYC. BACKGROUND

[0002] The handling stability of the automobile refers to the ability that the automobile can follow the direction given by the driver without feeling nervous and tired, and can resist the interference and keep stable driving when encountering external interference. With the coming of the "four modernizations" of the automobile industry and the change of the automobile power source layout, the research on the control method suitable for the handling stability of the electric vehicle has attracted widespread attention. The current driving form of the automobile mainly has three types (such as the attached Figure 2 ): the centralized drive electric vehicle, the wheel edge motor drive automobile and the wheel hub motor drive automobile. For the traditional automobile, the vehicle posture is usually corrected by increasing or decreasing the braking force inside each wheel to generate a yaw moment, but with the emergence of the distributed drive automobile, each wheel can exist as a separate power source, so the control is different from that of the traditional automobile.

[0003] Currently, the stability control of the distributed drive electric vehicle mainly reflects in the following several aspects: (1) active front wheel steering control (AFS): in the process of driver steering, the side slip angle of the wheel is changed by compensating the front wheel steering angle, and then an additional lateral force is generated to adjust the attitude of the vehicle body. However, this control method has poor robustness in the face of extreme weather or irregular road surface. (2) Direct yaw moment control (DYC): compared with the traditional vehicle, the distributed drive vehicle directly installs the power source on the wheel hub, and thus the response is fast and the energy loss in the rotation process is avoided. Since each wheel hub can provide power, when the vehicle needs to be intervened for stability control, an additional yaw moment can be generated by adjusting the braking / driving torque of the left and right wheels to improve the stability of the vehicle during steering. Direct additional yaw moment control responds quickly, but in the process of directly adjusting the yaw moment, the vehicle body attitude changes too violently, thereby reducing the driving experience. (3) Active rear wheel steering (ARS): by compensating the rear wheel with the same steering angle as the front wheel, the corresponding side slip force is generated, which can control the vehicle stability to a certain extent. At low speed, the rear wheel turns in the same direction as the front wheel, increasing the steering flexibility; at high speed, the rear wheel turns in the opposite direction of the front wheel, improving the stability. This method can significantly improve the vehicle handling and stability by coordinating the steering angles of the front and rear wheels, especially in high-speed lane changing and sharp turning situations. However, the ARS algorithm is complex, and the real-time response and precision of the system are high. In summary, if only DYC or AFS / ARS is used for vehicle stability control, it often cannot achieve good results. Therefore, starting from the vehicle handling stability and taking the driving comfort as the auxiliary, a model reference adaptive controller of DYC / AFS cooperative control is proposed, which judges whether the vehicle is unstable and the degree of instability through the phase plane method based on the centroid side slip angle, and then distributes the yaw moment and the front wheel compensation angle. SUMMARY

[0004] The present application provides a distributed drive electric vehicle stability control method and system. The method considers the changes of the vehicle mass and the road adhesion coefficient in the actual driving conditions as unknown parameters. The AFS / DYC controller is established by a linear two-degree-of-freedom model, and the additional front wheel steering angle and the additional yaw moment are taken as the control output. The distribution rule and the switching mode are determined by the phase plane method based on the centroid side slip angle. In the torque distribution, the quadratic optimal solution method is used to obtain the optimal torque with the minimum tire load rate as the objective function, and finally the obtained torque is distributed to the left and right wheels of the front axle to improve the vehicle stability.

[0005] A distributed drive electric vehicle stability control method, specifically comprising the following steps:

[0006] Step one, real-time acquisition of vehicle driving state information from vehicle sensors, including longitudinal vehicle speed, yaw rate, front wheel angle and other information, to provide reliable data support for subsequent steps.

[0007] Step two, the whole vehicle dynamics model is established and simplified as a "bicycle model", and the vehicle stability is described; the ideal center side slip angle and yaw rate module are obtained by analyzing the constraint conditions, and they are taken as the control target.

[0008] Step three, through the information obtained in step one and the two-degree-of-freedom vehicle model established in step two, considering that there are corresponding sensors for vehicle speed estimation and yaw rate, but there is no suitable sensor for the estimation of the center side slip angle, a two-degree-of-freedom Kalman filter is designed based on the actual situation, and the center side slip angle is estimated in real time through the input front wheel angle and current vehicle speed.

[0009] Step four, based on the ideal value obtained in step two and the data information obtained in step three, the additional front wheel angle and additional yaw rate are calculated through the model reference adaptive controller and input into the weight module. At the same time, the actual values of the center side slip angle and the yaw rate output from the Kalman filter observer enter the phase plane analysis module for stability judgment, and the specific control strategy is determined according to the stability judgment result.

[0010] Step five, based on the judgment in step four, if direct yaw moment control strategy is needed, a target function with minimum tire load rate is established, and the constraints of the motor and the road adhesion are considered, and then the left and right front wheel torques are further solved to realize the yaw moment control.

[0011] Further, the data output by step five is fed back to correct the current vehicle to form a closed-loop control, so as to maintain the stable state of the vehicle.

[0012] The beneficial effects of the present application are:

[0013] 1. The present application proposes an integrated multi-input multi-output (MIMO) adaptive control method for solving the parameter uncertainty between vehicle mass and road adhesion coefficient. Through accurate parameter estimation and dynamic adjustment, the adaptive control of the vehicle under different working conditions is realized, and the driving stability and safety of the vehicle are effectively improved.

[0014] 2、The application realizes the coordinated control between AFS and DYC based on the gain adjustment mechanism of stability index, and can be divided into three modes: (1) active front wheel steering control; (2) active front wheel steering and direct yaw moment collaborative control; (3) direct yaw moment control. The three control modes can be flexibly switched according to the real-time state of the vehicle and the working condition requirements, and are especially suitable for the driving requirements of the vehicle in the critical unstable state and extreme working conditions.

[0015] 3、The application is based on A phase plane method proposes an AFS / DYC collaborative distribution mode, which divides the system into stable domain, coordination domain and unstable domain, and proposes different control strategies for the three different modules. In the stable domain, the basic stability of the vehicle is mainly concerned; in the coordination domain, the coordination between AFS and DYC is focused on; in the unstable domain, precise control is carried out according to the steering demand. Through this domain control strategy, the vehicle control mode is flexibly adjusted to realize the optimization of vehicle dynamics under complex working conditions. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A two-degree-of-freedom model is provided for the vehicle;

[0017] Figure 2 A driving form is provided for the vehicle;

[0018] Figure 3 An overall flowchart is provided. DETAILED DESCRIPTION

[0019] The application will be further described below with reference to the accompanying drawings.

[0020] As shown in the accompanying Figure 3 According to the driver output throttle signal and steering wheel angle, the corresponding front wheel angle and longitudinal vehicle speed are obtained through the vehicle model, and then the actual front wheel angle and longitudinal vehicle speed are input into the two-degree-of-freedom model of the vehicle to output the desired center of mass side slip angle and yaw angular velocity as control target values, and the center of mass side slip angle and yaw angular velocity output from the Kalman filter are used as real values for control. The real value of the center of mass side slip angle is input into the stable domain module to determine whether the vehicle is in the stable domain, the coordination domain or the unstable domain, and the determination result is input into the weight module, and the result calculated in the model reference adaptive controller is input into the weight module. According to the current mode of the vehicle, the corresponding rule is used for distribution, if there is an additional front wheel angle, it is directly fed back to the vehicle model, if the additional yaw moment needs to be distributed, it needs to be further input into the moment distribution module, and finally the torque of the left and right wheels of the front axle is output through the calculation rule in the moment distribution module and fed back to the vehicle model, so as to realize closed-loop control of the system.

[0021] The application relates to a distributed drive electric vehicle stability control method, which mainly comprises the following steps.

[0022] Step one, obtaining vehicle state information from the outside world through a vehicle-mounted sensor device, wherein the vehicle state information comprises vehicle longitudinal speed v x , yaw angular velocity omega, and front wheel steering angle delta f .

[0023] Step two, according to the above-mentioned analysis, the following assumptions are made: Figure 1

[0024] (1) Assuming that the vehicle moves in a plane, the movement in the vertical direction of the plane is ignored;

[0025] (2) Assuming that the distance between each point in the vehicle body remains unchanged during the movement;

[0026] (3) Assuming that the vehicle steering system allows the front axle to rotate around the longitudinal axis of the vehicle, and the rear axle steering or the rear wheel angle with the vehicle body is constant is ignored;

[0027] (4) Assuming that the load in the vertical direction of the vehicle is constant, and the influence of the suspension system on the height of the vehicle gravity center is not considered;

[0028] (5) Assuming that the response of the tire in the lateral force aspect is approximately linear.

[0029] Therefore, a two-degree-of-freedom vehicle model is established according to the above assumptions to describe the vehicle stability:

[0030]

[0031] Wherein, beta is the mass center side slip angle, omega is the yaw angular velocity, v x is the vehicle longitudinal speed, v y is the vehicle lateral speed, a and b are the distances between the vehicle mass center and the front and rear axles, I z is the vehicle moment of inertia, delta f is the front wheel steering angle, and k1 and k2 are the front and rear axle side stiffness.

[0032] When the vehicle is in a stable state, according to the steady-state condition , the ideal yaw angular velocity and mass center side slip angle can be obtained:

[0033]

[0034] Wherein, m is the vehicle mass, L is the wheelbase, K is the stability factor,

[0035] ​But because the vehicle is affected by the adhesion force in the high-speed driving state, the tire is easy to reach the "friction ellipse" limit and often in the nonlinear region, so the boundary conditions should be fully considered in the design and optimization:

[0036]

[0037] Where, μ is the road adhesion coefficient, g is the gravity acceleration, ω dmax is the ideal maximum yaw rate, β dmax is the ideal maximum yaw rate, β dmax is the ideal maximum yaw rate, β dmax .

[0038] According to formula (2-2) and formula (2-3), the ideal yaw rate and yaw rate of the center of mass can be obtained:

[0039] ω d =min{ω d |,ω dmax |},β d =min{β d |,|β dmax}(2-4)

[0040] Step three, the design of Kalman filter observer based on the two-degree-of-freedom vehicle model in step two, because the current sensor can easily measure the longitudinal speed, yaw rate and front wheel angle during the vehicle driving process, but cannot directly obtain the size of the center of mass side slip angle, only relying on known parameters to estimate unknown parameters. Because the Kalman filter observer can optimize the estimation of the current system or predict the future state from the limited, noisy and biased observation state, therefore the Kalman filter is selected to measure the system parameters.

[0041] Assume that the state transition equation of the system is as follows:

[0042] x k =Ax k-1 +Bu k-1 +w k-1 (3-1)

[0043] Assume that the observation equation is as follows:

[0044] y k =Hx k-1 +v k (3-2)

[0045] In the formula, y k is the current state observation, x k is the current state quantity, x k-1 is the state quantity at the last time, u k-1 is the system input quantity at the last time, w k-1 , vk A is a state transition matrix, B is an input matrix, and H is a system observation matrix.

[0046] Kalman filtering is mainly divided into two stages: prediction stage and update stage. The two stages are combined through recursion, so that the system state can be effectively updated in a noisy environment. The specific equations are as follows:

[0047] (1) Prediction equation stage

[0048] X(k|k-1)=AX(k-1|k-1)+BU(k)(3-3)

[0049] P(k|k-1)=AP(k-1|k-1)A T +Q(3-4)

[0050] (2) Update equation stage

[0051] X(k|k)=X(k|k-1)+K(k)[y k -HX(k|k-1)](3-5)

[0052]

[0053] P(k|k)=[I-K(k)H]P(k|k-1)(3-7)

[0054] where X(k|k-1) is the state estimate at time k, A is the state transition matrix, which describes the state change of the system from time k-1 to time k, X(k-1|k-1) is the state estimate at time k-1, B is the input matrix, which describes the influence of the control input on the system state, U(k) is the control input at time k; P(k|k-1) is the state covariance matrix at time k, P(k-1|k-1) is the state covariance matrix at time k-1, Q is the process noise covariance matrix; X(k|k) is the state estimate at time k (after updating), y k -HX(k|k-1) is the observation residual, K(k) is the Kalman gain matrix at time k, R is the observation covariance matrix, P(k|k) is the state covariance transition matrix at time k after updating, and I is the identity matrix.

[0055] Further, the deformation and discretization of the vehicle two-degree-of-freedom equation (2-1) is made to conform to the form of Kalman filtering, and the discretized equation is as follows:

[0056]

[0057] where △t is the sampling interval, β(k) is the actual value of the mass center side slip angle at time k, ω(k) is the actual value of the yaw rate at time k, vx (k) is the longitudinal vehicle speed at time k, δ f (k) is the front wheel steering angle at time k.

[0058] Further, the above equation is converted into Kalman filter form equation:

[0059]

[0060] where, W(k), V(k) are the system noise at time k, is the observation of the system at time k.

[0061] Finally, the initial values are set according to the real conditions, and the estimated value of the centroid side slip angle is obtained by recursive calculation.

[0062] Step four, the model reference adaptive controller equation is derived through the information obtained in step three and step two. When there is a disturbance torque, the nonlinear equations of the vehicle in the lateral and longitudinal directions are represented as:

[0063]

[0064] In equation (4-1), F xf is the front wheel longitudinal force, F yf is the front wheel lateral force, F yr is the rear wheel lateral force, R is the wheel radius, δ is the front wheel steering angle, M dz is the disturbance torque, B is the wheel base, T dfl , T dfr are the driving torques of the left and right front wheels, respectively, and the subscripts fl, fr, rl, and rr represent the left front wheel, the right front wheel, the left rear wheel, and the right rear wheel, respectively.

[0065] At the same time, F yij = k ij α ij and T dij = RF xij , α ij is the wheel side slip angle, F xij represents the longitudinal force of the four wheels, and T dij is the driving force of each wheel ij = fl, fr, rl, rr, so the additional yaw moment can be represented as: where ΔF x is the difference between the longitudinal forces of the left and right wheels on the same axle.

[0066] By linearizing equation (4-1), a linear dynamic model of the vehicle in the longitudinal and lateral directions is obtained as a reference model:

[0067]

[0068] The upper controller is designed by using the concept of model reference adaptive controller, which aims to achieve vehicle handling and stability. The following matrix is defined according to equation (4-2):

[0069]

[0070] In equation (4-2), δ is the additional front wheel steering angle, δ d is the driver input front wheel steering angle, δ f = δ d + δ, M z is the additional yaw moment.

[0071] The input error is defined as:

[0072]

[0073] where, is the predicted value, is the parameter estimation error between the predicted value and the true value, m is the mass of the vehicle, μ is the road adhesion coefficient,

[0074] The weighted square sum of errors in the system is expressed using Lyapunov equation, so e and

[0075]

[0076] where, Γ = Γ T > 0 is a symmetric positive definite matrix, which represents the adaptive rate, so the derivative of equation (4-5) is:

[0077]

[0078] Considering that A2 and Γ are diagonal matrices, and then it can be deduced that:

[0079]

[0080] The control rate function is designed using equations (4-3), (4-4) as follows:

[0081]

[0082] where, are the predicted values of A2, B2, C2 respectively, and the system state error is: Q is a symmetric positive definite matrix.

[0083] By formula (4-6), (4-7), (4-8) can be obtained:

[0084]

[0085] In order to simplify the equation and make adaptive rate, let: Here P is a 2*2 order matrix containing state variables and reference model state (formula 4-2). Therefore, the derivative of Lyapunov function is as follows:

[0086]

[0087] Only when The system tends to be asymptotically stable, so the adaptive rate is defined as: In order to calculate P, it can be written as:

[0088]

[0089] Where, And because Therefore

[0090]

[0091] Therefore, is the error between the predicted value of vehicle mass and the actual value of road adhesion coefficient.

[0092] Define the weight matrix The values of q1, q2 are obtained by trial and debugging, and are brought into the control rate. Finally, the output can be obtained as:

[0093]

[0094] In the formula, is the predicted value of vehicle mass, is the error between the predicted value of vehicle mass and the actual value, is the predicted value of road adhesion coefficient, is the error between the predicted value of road adhesion coefficient and the actual value.

[0095] Step five, since AFS and DYC are all through indirect or direct ways to obtain additional yaw moment to control vehicle driving stability. The correction of AFS to front wheel angle only plays a role in small range instability, but if in extreme working conditions, the control of lateral stability is limited, if in critical instability, DYC should intervene in time to prevent the vehicle from turning too much. Accordingly, according to the method of Stability region division, the specific division is as follows: Stable domain: Coordination domain:

[0096]

[0097] B1=-9.1μ 2 B2=0.03371μ 2 c1B1+c2B2, c1 is a weighting factor and 0

[0098] The weighting factor of AFS is S, and the weighting factor of DYC is T. When in the stable region, only AFS works, S=1 and T=0. When in the non-stable region, only DYC works, S=0 and T=1.

[0099] The stability index x in the coordination region is defined as:

[0100]

[0101] The expression of the weighting factor S is as follows:

[0102]

[0103] i a =c1B2, i b =c2B2, m>0, n>0 are adjustment parameters and m+n=4.

[0104] The weighting factor of AFS is S=1-T.

[0105] Therefore, the additional front wheel angle and the additional yaw moment output by the AFS subsystem and the DYC subsystem respectively in the coordination region are:

[0106]

[0107] Step six, the additional yaw moment corresponding to the region divided in step five is used to design the lower layer actuator. The present application adopts optimal distribution and simplifies the objective function to reduce the calculation amount and improve the calculation accuracy. First, the stability margin η of each wheel is expressed by using the concept of friction ellipse i :

[0108]

[0109] F xi is the longitudinal force of each wheel, F yi is the lateral force of each wheel, F zi is the vertical load between each wheel, μ is the road adhesion coefficient, and i=1, 2, 3, 4 are the front axle left and right wheels and the rear axle left and right wheels.

[0110] But the center position changes due to acceleration and deceleration during driving, so the weight coefficient c is needed i , and the target function J of the whole vehicle load rate is obtained:

[0111]

[0112] Therefore, the stability margin of the whole vehicle can be represented by the target function J. The stability margin measures the ability of the vehicle to maintain a stable state under various driving conditions. The stability margin is negatively correlated with the load rate of the whole vehicle, that is, when the stability margin of the whole vehicle is greater, the load rate is smaller, that is, the possibility of vehicle instability is low. Conversely, the stability margin of the whole vehicle is lower, the load rate is greater, and the possibility of vehicle instability is greater. Therefore, the target function for optimizing the distribution is:

[0113]

[0114] Neglecting the control of the lateral force, the target function is simplified as:

[0115]

[0116] According to the relationship between the torque of each wheel and the longitudinal force, it can be known that:

[0117]

[0118] In the formula, r is the wheel radius.

[0119] The final target function equation can be obtained as:

[0120]

[0121] After the optimization function is finally determined, the influence of the maximum output torque of the motor and the road adhesion coefficient needs to be considered according to the actual situation, and the specific constraints are as follows:

[0122]

[0123] In the formula, T1, T2, T3, and T4 are the output torques of the front and rear wheels, B is the wheel track, T is the sum of the output torques of the four wheels, T max is the maximum output torque of the motor, and △M is the additional yaw moment (formula 5-2).

[0124] After the target function and the constraints are determined, the required torques of the left and right front wheels are further calculated as follows:

[0125]

[0126] The formula (5-10) is brought into the formula (5-8), and the following can be obtained:

[0127]

[0128] The partial derivative of formula (5-11) to T3, T4 can be obtained as follows:

[0129]

[0130] Let (5-12) be 0, the minimum value of the rear axle left and right wheel torque can be solved as follows:

[0131]

[0132] The driving torque of the front axle left and right wheels can be solved by bringing formula (5-13) back to formula (5-10):

[0133]

[0134] Based on the above control method, the application further provides a distributed driving electric vehicle stability control system, comprising an information acquisition module and an information processing module; the information acquisition module realizes the content of the above step one; the information processing module realizes the content of steps two to six.

[0135] The above series of detailed descriptions are only specific descriptions of the feasible embodiments of the application, and are not used to limit the protection scope of the application, and any equivalent means or changes without departing from the technology of the application should be included in the protection scope of the application.

Claims

1. A stability control method for a distributed drive electric vehicle, characterized by, The method comprises the following steps: S1, acquiring vehicle driving state information in real time, wherein the information comprises longitudinal vehicle speed, yaw rate and front wheel steering angle; S2, establishing a vehicle dynamics model, analyzing vehicle stability, and obtaining ideal vehicle mass center side slip angle and yaw rate through analysis of constraint conditions, and taking the ideal vehicle mass center side slip angle and yaw rate as control targets; S3, designing a Kalman filter based on the vehicle dynamics model, and estimating the vehicle mass center side slip angle in real time through input of the front wheel steering angle and current vehicle speed; S4, calculating additional front wheel steering angle and additional yaw rate through a model reference adaptive controller based on the ideal values obtained in S2 and the estimated values of the vehicle mass center side slip angle obtained in S3; S5, dividing a stability region by using a phase plane analysis method based on the results calculated in S4 and the actual values of the vehicle mass center side slip angle and yaw rate estimated in S3, designing weight coefficients of AFS and DYC, and selecting a control strategy according to a stability judgment result; S6, designing a lower-layer actuator according to the additional yaw moment output in the region divided in S5, and realizing torque distribution of the vehicle; In S2, the vehicle dynamics model adopts a two-degree-of-freedom model, and the two-degree-of-freedom model is as follows: where m is the vehicle mass, β is the side slip angle, ω is the yaw rate, v x is the vehicle longitudinal speed, a and b are the distances between the vehicle mass center and the front and rear axles, I z is the vehicle moment of inertia, δ f is the front wheel steering angle, k1 and k2 are the front and rear axle cornering stiffnesses, v y is the vehicle lateral speed; The design of the Kalman filter in S3 comprises the following steps: Discretize the two-degree-of-freedom equation of the vehicle to obtain a discretized equation as follows: Convert the above equation into a Kalman filter form equation: wherein, W(k), V(k) are system noises, is the observation value of the system at time k, Δt is the sampling interval, β(k) is the actual value of the centroid lateral deviation angle at time k, ω(k) is the actual value of the yaw rate at time k, v x (k) is the longitudinal vehicle speed at time k; Finally, set initial values according to actual conditions, and obtain the estimated value of the vehicle mass center side slip angle through recursive calculation; The method for dividing the region by using the phase plane analysis method in S5 is as follows: According to the method of stability region partitioning includes: Stable domain: Coordination domain: Non-stable region: where B1 = -9.1 μ 2 + 14.17 μ + 1.2, B2 = 0.03371 μ 2 + 0.02 μ + 0.01595, c1 is a weighting factor and 0 < c1 < 1, c2 is a weighting factor and c2 > 1, β is a centric side slip angle, and μ is a road adhesion coefficient.

2. The distributed drive electric vehicle stability control method of claim 1, wherein, The ideal vehicle mass center side slip angle and yaw rate in S2 are obtained by the following method: According to the vehicle steady state condition Deriving ideal yaw rate and center of mass side slip angle: where L is wheel base and K is a stability factor, Set boundary conditions: wherein μ is the road adhesion coefficient, ω dmax is the maximum value of the ideal yaw rate, β dmax is the maximum value of the ideal side slip angle Finally, the ideal vehicle mass center side slip angle and yaw rate are as follows: ω d = min{|ω d |, |ω dmax |}, β d = min{|β d |, |β dmax |} (2-4).

3. The distributed drive electric vehicle stability control method of claim 1, wherein, The implementation of S4 comprises the following steps: The nonlinear equations of the vehicle in the lateral and longitudinal directions are represented as follows: where F xf is the front wheel longitudinal force, F yf is the front wheel lateral force, F yr is the rear wheel lateral force, r is the wheel radius, δ f is the front wheel steering angle, M dz is the disturbance moment, B is the wheel base, T dfl , T dfr are the left and right front wheel driving torques, respectively. F yij = k ij α ij + r dij F xij , k ij represents the cornering stiffness of the wheel, α ij is the cornering angle of the wheel, F xij is the longitudinal force of the four wheels, and ij = fl, fr, rl, rr respectively represent the left and right wheels of the front axle and the left and right wheels of the rear axle. The additional yaw moment can be expressed as: where ΔF x is the difference between the longitudinal forces of the left and right wheels of the same axle. Linearization is performed on the equation (4-1) to obtain a linear dynamic model of the vehicle in the longitudinal and lateral directions, which is taken as a reference model: An upper-layer controller is designed by using the concept of the model reference adaptive controller, and the purpose is to realize vehicle maneuverability and stability. wherein δ is the additional front wheel steering angle, δ d is the driver input front wheel steering angle, δ f = δ d + δ, M z is the additional yaw moment; The input error is defined as: wherein, is the predicted value, is the parameter estimation error between the predicted value and the true value, and μ is the road adhesion coefficient, The weighted sum of squares of errors in the system is represented by using the Lyapunov equation: where Γ = Γ T >0 is a symmetric positive definite matrix representing the adaptation rate, so that the derivative with respect to both ends can be obtained: Also, considering that A2 and Γ are diagonal matrices, and It is derived that: The control rate functions are designed by using the formulas (4-3) and (4-4) as follows: wherein are the estimates of A2, B2, C2, respectively, and the system state error is: Q is a symmetric positive definite matrix, The formulas (4-6), (4-7) and (4-8) are obtained as follows: To simplify the equations and develop an adaptive rate, let Here P is a 2*2 matrix containing the state variables and the reference model states, so the derivative of the Lyapunov function is as follows: Only The system tends to be asymptotically stable, so the adaptive rate is defined as: To calculate P, one can write: wherein Further, because Therefore So, is the error between the predicted and actual values of the vehicle mass and road adhesion coefficient. Defining the weight matrix And into the control rate, the final output is: wherein is a vehicle mass prediction value, is an error between the vehicle mass prediction value and the true value, is a road surface adhesion coefficient prediction value, is an error between the road surface adhesion coefficient prediction value and the true value.

4. The distributed drive electric vehicle stability control method of claim 1, wherein, In S5, the weight coefficients of AFS and DYC are designed as follows: The weight coefficient of AFS is S, and the weight coefficient of DYC is T. When in a stable region, only AFS works, S=1 and T=0. When in a non-stable region, only DYC works, S=0 and T=1. wherein i a = c1B2, i b = c2B2, m > 0, n > 0 are adjustment parameters and m + n = 4; The stability index in the coordination region is defined as: The expression of the weight coefficient Q is designed as follows:

5. The distributed drive electric vehicle stability control method of claim 4, wherein, The weight coefficient of AFS S=1-T. The additional front wheel steering angle and additional yaw moment output by the AFS subsystem and the DYC subsystem in the coordination region are as follows: where F xi is the longitudinal force of each wheel, F yi is the lateral force of each wheel, F zi is the vertical load between each wheel, μ is the road adhesion coefficient, and i = 1, 2, 3, 4 are the front left and right wheels and the rear left and right wheels, respectively. The implementation of S6 comprises design of a target function, and is specifically as follows: c i are weight coefficients; Firstly, the stability margin of each wheel is represented by using the friction ellipse concept: A vehicle load rate target function is designed: The target function for optimized distribution is designed as follows: Neglecting the control of lateral force, the objective function is simplified as: According to the relationship between the wheel torque and the longitudinal force, we have: Where r is the wheel radius; The final objective function is:

6. The distributed drive electric vehicle stability control method of claim 5, wherein, The implementation of S6 also includes designing specific constraints, as follows: In the formula, T1, T2, T3, T4 are output torques of the front and rear axles and left and right wheels respectively, T is the sum of the output torques of the four wheels, T max is the maximum output torque of the motor, and ΔM is an additional yaw moment output by the weight distribution module. After determining the objective function and the constraints, the required torques of the left and right front wheels are calculated as follows: Substituting equation (5-10) into equation (5-8), we have: Taking the partial derivative of equation (5-11) with respect to T3 and T4, we have: Setting equation (5-12) to 0, we have the minimum values of the torques of the left and right rear wheels: Substituting equation (5-13) into equation (5-10), we have the torques of the left and right front wheels:

7. A distributed drive electric vehicle stability control system, characterized in that, The information processing module realizes the contents of S2-S6 in the stability control method of the distributed drive electric vehicle according to any one of claims 1-6.

Images