Variable-weight curve stable car following control method considering road adhesion conditions

Abstract

The invention discloses a variable-weight curve stable vehicle following control method considering road adhesion conditions, which comprises the following steps of: estimating a tire-road adhesion coefficient according to a yaw velocity and a lateral acceleration measured by a vehicle sensor by utilizing an extended Kalman filter and combining a kinetic equation of a four-wheel independent drive vehicle, a Dugoff tire model and a random walk model; constructing an improved adaptive cruise control (ACC) safe distance model according to the estimated tire-road adhesion coefficient; s2, on the basis of the improved ACC safe distance model constructed in the step S2, constructing an ACC system containing direct yawing moment control intervention, and optimizing the ACC system by adopting a model prediction control method; the acceleration of the front vehicle is predicted, and a weighting matrix is established according to the change of the acceleration of the front vehicle, so that the intervention degree and the longitudinal vehicle following performance of direct yawing moment control under the low road surface attachment condition are adjusted, and stable vehicle following control over the curve under different road surface attachment conditions is achieved.

Description

Technical Field

[0001] The invention relates to a variable-weight curve stable following control method taking road adhesion conditions into consideration. Background Art

[0002] Currently, adaptive cruise control (ACC) has become an important means of improving vehicle safety and driving experience, with cornering conditions being a key area of ​​focus. Drivers' skills and habits vary significantly when following a car on a curve. When the tire-road adhesion coefficient fluctuates dramatically, the vehicle's lateral stability is difficult to maintain, and serious accidents can easily occur. Four-wheel independent drive electric vehicles (FWDs) have independently controlled wheel torques. The resulting additional yaw moment can effectively improve the vehicle's lateral stability, making them a mainstream trend in modern vehicle development. This approach is known as direct yaw moment control (DYC).

[0003] Existing research on following a vehicle on a curve often assumes ideal road adhesion conditions and, based on this assumption, introduces DYC to improve the vehicle's lateral stability. It's important to note that, due to climatic conditions, actual roads are often slippery or even partially icy. Continuing to use the aforementioned approach for following a vehicle on a curve would create numerous safety hazards. Specifically, when the leading vehicle accelerates, the host vehicle must also accelerate through the curve, putting its lateral stability to the test. Conversely, when the leading vehicle decelerates, the host vehicle must also decelerate through the curve, increasing the risk of a rear-end collision due to the low tire-road adhesion coefficient. Therefore, it's necessary to design a variable-weighted, stable following control method for curves, taking into account road adhesion conditions and adapting to changes in the leading vehicle's acceleration. Summary of the Invention

[0004] The present invention aims to solve the above problems in the prior art and provides a variable-weight curve stability following control method taking road adhesion conditions into consideration.

[0005] The technical solutions adopted in the present invention are:

[0006] A variable-weight curve stability following control method considering road adhesion conditions comprises the following steps:

[0007] S1: Using an extended Kalman filter, combined with the dynamic equations of a four-wheel independent drive vehicle, the Dugoff tire model, and the random walk model, the tire-road adhesion coefficient is estimated based on the yaw rate and lateral acceleration measured by the vehicle sensors.

[0008] S2: Construct an improved adaptive cruise control (ACC) safety distance model based on the estimated tire-road adhesion coefficient;

[0009] S3: Based on the improved ACC safety distance model constructed in step S2, an ACC system with direct yaw moment control intervention is constructed, and the ACC system is optimized using a model predictive control method;

[0010] S4: Predict the acceleration of the leading vehicle and establish a weighted matrix based on the change in the leading vehicle's acceleration to adjust the degree of intervention of direct yaw moment control and longitudinal following performance under low road adhesion conditions, thereby achieving stable following control in corners under different road adhesion conditions.

[0011] Furthermore, in S1, the tire-road adhesion coefficient estimation process is:

[0012] When the front wheel angle δ satisfies sinδ≈0 and cosδ≈1, the dynamic equation of the four-wheel independent drive vehicle is established considering the longitudinal, lateral and yaw motions:

[0013]

[0014] Among them, v x 、v y are the longitudinal and lateral velocities of the vehicle, r is the yaw angular velocity of the vehicle, m is the vehicle mass, and F x 、F y They represent longitudinal tire force and lateral tire force respectively. The subscripts fl, fr, rl, and rr represent the left front, right front, left rear, and right rear wheels respectively. z represents the moment of inertia of the vehicle around the z-axis, a and b are the distances from the center of mass to the front and rear axles, respectively, and t w is the vehicle wheelbase;

[0015] The Dugoff tire model is used to express the relationship between the tire longitudinal and lateral forces and the tire-road adhesion coefficient μ:

[0016]

[0017] in, Denote the normalized longitudinal and lateral tire forces, F z is the vertical load of the tire, C x is the tire longitudinal slip stiffness, C y is the tire cornering stiffness, λ is the slip rate, ε is the speed influence factor, and α is the tire slip angle;

[0018] The error characteristics of the tire-road adhesion coefficient μ are characterized by a random walk model:

[0019]

[0020] Among them, w μ is a two-dimensional white noise vector;

[0021] Combining formulas (1) and (5), we obtain the following state equation and measurement equation:

[0022]

[0023] Among them, the state vector Measurement vector z(t)=[a ym r m ] T , a ym and r m are the vehicle yaw rate and lateral acceleration measured by the sensor, f(x(t),u(t)) and h(x(t),u(t)) are the nonlinear functions in the state equation and measurement equation, respectively. The control input u(t) = δ, w(t) and v(t) are the process noise and measurement noise, respectively.

[0024] Discretization is performed on the basis of the state equation and the measurement equation, and the tire-road adhesion coefficient μ is estimated by an extended Kalman filter.

[0025] Furthermore, S2 is specifically:

[0026] Define the relative distance between the main vehicle and the preceding vehicle as Δd, the relative speed as Δv, and the acceleration of the main vehicle as a x , the actual distance between vehicles is d, and the speed of the preceding vehicle is v p , the delay time of the main vehicle transmission system is T s , the inertia gain is K1, and the expected longitudinal acceleration of the main vehicle is a xdes , considering the tire-road adhesion coefficient, the safe distance is d des ; Construct an improved ACC safety distance model according to the following formula:

[0027]

[0028] d des =f(μ)v x t h -kv rel +d min (8),

[0029] Among them, t h is the headway, v rel is the relative speed between the main vehicle and the preceding vehicle, k is the adjustment constant, d min is the minimum safety distance, f(μ) is the added friction adjustment coefficient, and its expression is:

[0030]

[0031] Among them, the subscripts norm and min represent the calibration value and the minimum value respectively, μ normand μ min are the calibrated adhesion coefficient and the minimum adhesion coefficient, respectively, f(μ min ) and f(μ norm ) corresponds to μ norm and μ min The friction adjustment coefficient setting value at .

[0032] Furthermore, in S3, an ACC system with direct yaw moment control intervention is constructed, specifically:

[0033] Based on a single-track vehicle model, the improved ACC safety distance model is combined with direct yaw moment control to form an ACC system with direct yaw moment control intervention. The system is expressed as:

[0034]

[0035] Wherein, the state vector x=[βrΔdΔv a x ] T , control vector u=[F xfl F xfr F xrl F xrr a xdes ] T , interference vector w=[δa p ] T , a p is the acceleration of the preceding vehicle, and the coefficient matrices A, B, and C are:

[0036]

[0037] Among them, C f and C r are the total cornering stiffness of the front and rear wheels, respectively.

[0038] Furthermore, in S3, the ACC system is optimized using a model predictive control method, specifically:

[0039] Set the cost function:

[0040]

[0041] Among them, x ref is the expected value of the state, N p and N c are the prediction time domain and the control time domain respectively, k+i|k represents the k+i moment value predicted at time k, Q(k) and R(k) are the weighted matrices corresponding to the state quantity and control quantity at time k, respectively, corresponding to:

[0042]

[0043] Among them, ωΔd 、ω Δv 、 ω r 、ω β as well as are the weight coefficients related to relative distance, relative speed, longitudinal acceleration of the main vehicle, expected acceleration of the main vehicle, yaw rate, sideslip angle of the center of mass, and longitudinal force of the wheel;

[0044] The difference between the actual and expected values ​​of the vehicle's sideslip angle and yaw rate is used as an evaluation index for the vehicle's lateral stability. The reference values ​​of the yaw rate and sideslip angle are expressed as follows:

[0045]

[0046] in, Wheelbase l = a + b, g is the gravitational acceleration constant, vehicle steady-state parameters

[0047] The constraints corresponding to the cost function are as follows:

[0048] Desired longitudinal acceleration and desired yaw rate constraints:

[0049]

[0050] Driving safety and comfort constraints:

[0051]

[0052] Where Δd max and Δv max are the upper limits of distance error and relative speed respectively. The driver’s error sensitivity SDE and SVE to tracking distance and speed are:

[0053]

[0054] Among them, k SDE d SDE 、k SVE and d SVE All are constant coefficients.

[0055] Furthermore, in S4, under low road adhesion conditions, the degree of direct yaw moment control intervention and the longitudinal following performance are adjusted as follows:

[0056] Select the time interval [t0,t0+T d ]Predict the acceleration of the preceding vehicle and fit the longitudinal acceleration of the preceding vehicle within the time interval. The fitting formula is:

[0057] a p (κ)=p0+p1+p2κ2 +…+p n κ n (17),

[0058] Among them, t0 is the current time, T d is the prediction range, n is the order of the polynomial, p i (i=0,1,2,…,n) is the coefficient to be fitted;

[0059] According to the different acceleration conditions of the preceding vehicle, the weighting matrices Q and R are adjusted according to the preset scheme to adapt to different driving conditions. The weighting matrix is ​​specifically expressed as follows:

[0060]

[0061] in, s i (i=1,2,…,8), σ1 and σ2 are all set coefficients.

[0062] The present invention first uses an extended Kalman filter to estimate the tire-road adhesion coefficient of the driving road in real time. Secondly, an ACC safety distance model that takes the tire-road adhesion coefficient into consideration is designed. Then, an ACC system with DYC intervention is constructed and optimized using the model predictive control (MPC) method. Finally, the acceleration of the leading vehicle is predicted, and a weighted matrix that is adjusted according to the change in the leading vehicle's acceleration is established to achieve regulation of the DYC intervention level and longitudinal following performance under low road adhesion conditions. When the main vehicle follows the leading vehicle and accelerates through a corner, this solution can provide protection for the lateral stability of the main vehicle; when the main vehicle follows the leading vehicle and accelerates through a corner at a constant speed or decelerates, this solution can reduce the risk of the main vehicle rear-ending. The resulting beneficial effects are:

[0063] (1) The present invention uses an extended Kalman filter to achieve real-time estimation of the tire-road adhesion coefficient, which can provide key parameters for curve following operations to optimize vehicle control effects;

[0064] (2) The present invention adopts an ACC safety distance model that takes into account the tire-road adhesion coefficient, thereby improving the vehicle's robustness in dealing with different road adhesion conditions;

[0065] (3) The present invention proposes a weighted matrix adjustment scheme for the acceleration change of the leading vehicle, which improves the lateral stability and longitudinal safety of the vehicle when following a vehicle on a curve under low adhesion conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a flow chart of the present invention.

[0067] Figure 2 Schematic diagram of the tire-road adhesion coefficient estimation.

[0068] Figure 3 Schematic diagram of the curve working condition. DETAILED DESCRIPTION

[0069] The present invention will be further described below with reference to the accompanying drawings.

[0070] like Figures 1 to 3 The present invention provides a variable weight curve stable following control method considering road adhesion conditions, and the specific steps are as follows:

[0071] Step 1: Estimation of the tire-road adhesion coefficient based on the extended Kalman filter.

[0072] When the front wheel angle δ satisfies sinδ≈0 and cosδ≈1, the dynamic equation of the four-wheel independent drive vehicle is established considering the longitudinal, lateral and yaw motions:

[0073]

[0074] Among them, v x 、v y are the longitudinal and lateral velocities of the vehicle, r is the yaw angular velocity of the vehicle, m is the vehicle mass, and F x 、F y They represent longitudinal tire force and lateral tire force respectively. The subscripts fl, fr, rl, and rr represent the left front, right front, left rear, and right rear wheels respectively. z represents the moment of inertia of the vehicle around the z-axis, a and b are the distances from the center of mass to the front and rear axles, respectively, and t w The wheelbase of the vehicle.

[0075] The Dugoff tire model is used to express the relationship between the tire longitudinal and lateral forces and the tire-road adhesion coefficient μ:

[0076]

[0077] in, Denote the normalized longitudinal and lateral tire forces, F z is the vertical load of the tire, C x is the tire longitudinal slip stiffness, C y is the tire cornering stiffness, λ is the slip rate, ε is the speed influence factor, and α is the tire slip angle.

[0078] The error characteristics of the tire-road adhesion coefficient μ are characterized by a random walk model:

[0079]

[0080] Among them, w μ is a two-dimensional white noise vector.

[0081] Combining equations (1) and (5), and considering the composition of actual sensor signals, we have the following state equations and measurement equations:

[0082]

[0083] Among them, the state vector Measurement vector z(t)=[a ym r m ] T , a ym and r m where yaw rate and lateral acceleration are the vehicle's yaw rate and lateral acceleration measured by the sensors, respectively. f(x(t), u(t)) and h(x(t), u(t)) are the nonlinear functions in the state equation and measurement equation, respectively. The control input u(t) = δ, and w(t) and v(t) are the process noise and measurement noise, respectively. Based on Equation (6), the tire-road adhesion coefficient μ can be estimated using an extended Kalman filter.

[0084] Step 2: Design an ACC safety distance model that takes into account the tire-road adhesion coefficient.

[0085] When a vehicle using ACC follows a preceding vehicle into a curve, if the preceding vehicle's speed changes, the vehicle will also accelerate or decelerate accordingly. However, rash acceleration and deceleration in curves can put the vehicle's lateral stability at risk. In particular, when the tire-road adhesion coefficient μ is low, the vehicle faces a high risk of skidding or even instability. Therefore, an improved ACC safety distance model is constructed based on the following formula:

[0086]

[0087] Among them, Δd and Δv are the relative distance and relative speed between the main vehicle and the preceding vehicle, respectively. x is the acceleration of the main vehicle, d is the actual vehicle distance, v p is the speed of the preceding vehicle, T s is the delay time of the main vehicle transmission system, K1 is the inertia gain, a xdes The expected longitudinal acceleration of the host vehicle, d des The safe distance considering the tire-road adhesion coefficient is:

[0088] d des =f(μ)v x t h -kv rel +d min (8),

[0089] Among them, t h is the headway, v rel is the relative speed between the main vehicle and the preceding vehicle, k is the adjustment constant, dmin is the minimum safety distance, and f(μ) is the added friction adjustment coefficient:

[0090]

[0091] Among them, the subscripts norm and min represent the calibration value and the minimum value respectively, μ norm and μ min are the calibrated adhesion coefficient and the minimum adhesion coefficient, respectively, f(μ min ) and f(μ norm ) corresponds to μ norm and μ min The friction adjustment coefficient setting value at .

[0092] Step 3: Build an ACC system with DYC intervention.

[0093] In cornering conditions, timely DYC intervention can improve the vehicle's lateral stability. Furthermore, by adjusting the weight coefficient of driving force distribution according to changes in vehicle speed, the corresponding driving objectives can be optimized.

[0094] Using a single-track vehicle model, the ACC system with DYC intervention can be expressed as:

[0095]

[0096] Wherein, the state vector x=[βrΔdΔv a x ] T , control vector u=[F xfl F xfr F xrl F xrr a xdes ] T , interference vector w=[δa p ] T , a p is the acceleration of the preceding vehicle, and the coefficient matrices A, B, and C are:

[0097]

[0098]

[0099] C f and C r are the total cornering stiffness of the front and rear wheels, respectively.

[0100] Step 4: Use the MPC method to optimize the ACC system with DYC intervention.

[0101] Set the following cost function:

[0102]

[0103] Among them, x ref is the expected value of the state, N p and N c are the prediction time domain and the control time domain, respectively. k+i|k represents the k+i moment value predicted at time k. Q(k) and R(k) are the weighted matrices corresponding to the state quantity and control quantity at time k, respectively:

[0104]

[0105] Among them, ω Δd 、ω Δv 、 ω r 、ω β as well as are the weight coefficients related to relative distance, relative speed, longitudinal acceleration of the main vehicle, expected acceleration of the main vehicle, yaw angular velocity, sideslip angle of the center of mass, and longitudinal force of the wheel.

[0106] The sideslip angle and yaw rate are important parameters for evaluating a vehicle's lateral stability. The difference between the actual and expected values ​​of the vehicle's sideslip angle and yaw rate can be used as an evaluation indicator of the vehicle's lateral stability. The reference values ​​for yaw rate and sideslip angle are expressed as follows:

[0107]

[0108] in, Wheelbase l = a + b, g is the gravitational acceleration constant, vehicle steady-state parameters

[0109] The constraints corresponding to the cost function are as follows:

[0110] Desired longitudinal acceleration and desired yaw rate constraints:

[0111]

[0112] Driving safety and comfort constraints:

[0113]

[0114] Where Δd max and Δv max are the upper limits of distance error and relative speed, respectively. The driver’s sensitivity to tracking distance and speed errors SDE (sensitivity to distance error) and SVE (sensitivity to velocity error) are:

[0115]

[0116] Among them, k SDE d SDE 、k SVE and d SVE All are constant coefficients.

[0117] Step 5: Under low road adhesion conditions, adjust the degree of DYC intervention and longitudinal following performance.

[0118] When the tire-road adhesion coefficient μ is low, changes in the leading vehicle's speed can affect the safety of the host vehicle following a curve: If the leading vehicle accelerates, the host vehicle, under the action of ACC, will also accelerate, posing the risk of roll or even instability. In this case, DYC should be deeply involved to prioritize lateral stability. If the leading vehicle is traveling at a constant speed or decelerating, to avoid rear-end collisions, the host vehicle should prioritize longitudinal safety. Specifically, the weighting matrices Q and R can be adjusted according to the following scheme to adapt to different driving conditions:

[0119] Select a certain time interval to make a reasonable prediction of the acceleration of the vehicle ahead. Assume that t0 is the current moment, T d is the prediction range, then the time interval [t0,t0+T d ] can be fitted as follows:

[0120] a p (κ)=p0+p1+p2κ 2 +…+p n κ n (17),

[0121] Where n is the order of the polynomial, p i (i = 0, 1, 2, ..., n) are the coefficients to be fitted. Considering that an increase in the order n will significantly increase the amount of computation and reduce the real-time performance of the corresponding system, n∈[3, 6] is usually used. On this basis, the weighted matrix for the change in the speed of the preceding vehicle can be specifically expressed as follows:

[0122]

[0123] in, s i (i=1,2,…,8), σ1 and σ2 are all set coefficients.

[0124] When the tire-road adhesion coefficient μ is high, the ground can exert greater longitudinal force on the vehicle, improving the leading vehicle's longitudinal following performance when following a vehicle in a curve. However, the leading vehicle's lateral stability is still under test. In this case, DYC will intervene deeply at all times, focusing on meeting lateral stability requirements. The degree of intervention is the same as when μ is low and the leading vehicle is accelerating.

[0125] The above description is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements can be made without departing from the principles of the present invention. These improvements should also be regarded as the scope of protection of the present invention.

Claims

1. A variable-weight curve stability following control method that takes into account road adhesion conditions, characterized by: The following steps are involved: S1: Using an extended Kalman filter, combined with the dynamic equations of a four-wheel independent drive vehicle, the Dugoff tire model, and the random walk model, the tire-road adhesion coefficient is estimated based on the yaw rate and lateral acceleration measured by the vehicle sensors. S2: Construct an improved adaptive cruise control (ACC) safety distance model based on the estimated tire-road adhesion coefficient; S3: Based on the improved ACC safety distance model constructed in step S2, an ACC system with direct yaw moment control intervention is constructed, and the ACC system is optimized using a model predictive control method; S4: Predict the acceleration of the leading vehicle and establish a weighted matrix based on the change in the leading vehicle's acceleration to adjust the degree of intervention of direct yaw moment control and longitudinal following performance under low road adhesion conditions, thereby achieving stable following control in corners under different road adhesion conditions.

2. The variable-weight curve stability following control method considering road adhesion conditions according to claim 1, characterized in that: In S1, the tire-road adhesion coefficient estimation process is: When the front wheel angle δ satisfies sinδ≈0 and cosδ≈1, the dynamic equation of the four-wheel independent drive vehicle is established considering the longitudinal, lateral and yaw motions: Among them, v x 、v y are the longitudinal and lateral velocities of the vehicle, r is the yaw angular velocity of the vehicle, m is the vehicle mass, and F x 、F y They represent longitudinal tire force and lateral tire force respectively. The subscripts fl, fr, rl, and rr represent the left front, right front, left rear, and right rear wheels respectively. z represents the moment of inertia of the vehicle around the z-axis, a and b are the distances from the center of mass to the front and rear axles, respectively, and t w is the vehicle wheelbase; The Dugoff tire model is used to express the relationship between the tire longitudinal and lateral forces and the tire-road adhesion coefficient μ: in, Denote the normalized longitudinal and lateral tire forces, F z is the vertical load of the tire, C x is the tire longitudinal slip stiffness, C y is the tire cornering stiffness, λ is the slip rate, ε is the speed influence factor, and α is the tire slip angle; The error characteristics of the tire-road adhesion coefficient μ are characterized by a random walk model: Among them, w μ is a two-dimensional white noise vector; Combining formulas (1) and (5), we obtain the following state equation and measurement equation: Among them, the state vector Measurement vector z(t)=[a ym r m ] T , a ym and r m are the vehicle yaw rate and lateral acceleration measured by the sensor, f(x(t),u(t)) and h(x(t),u(t)) are the nonlinear functions in the state equation and measurement equation, respectively. The control input u(t) = δ, w(t) and v(t) are the process noise and measurement noise, respectively. Discretization is performed on the basis of the state equation and the measurement equation, and the tire-road adhesion coefficient μ is estimated by an extended Kalman filter.

3. The variable-weight curve stability following control method considering road adhesion conditions according to claim 1, characterized in that: S2 is specifically: Define the relative distance between the main vehicle and the preceding vehicle as Δd, the relative speed as Δv, and the acceleration of the main vehicle as a x , the actual distance between vehicles is d, and the speed of the preceding vehicle is v p , the delay time of the main vehicle transmission system is T s , the inertia gain is K1, and the expected longitudinal acceleration of the main vehicle is a xdes , considering the tire-road adhesion coefficient, the safe distance is d des ; The improved ACC safety distance model is constructed according to the following formula: the des =f(μ)v x t h -kv rel +d min (8), Among them, t h is the headway, v rel is the relative speed between the main vehicle and the preceding vehicle, k is the adjustment constant, d min is the minimum safety distance, f(μ) is the added friction adjustment coefficient, and its expression is: Among them, the subscripts norm and min represent the calibration value and the minimum value respectively, μ norm and μ min are the calibrated adhesion coefficient and the minimum adhesion coefficient, respectively, f(μ min ) and f(μ norm ) corresponds to μ norm and μ min The friction adjustment coefficient setting value at .

4. The variable-weight curve stability following control method considering road adhesion conditions according to claim 1, characterized in that: In S3, an ACC system with direct yaw moment control intervention is constructed, specifically: Based on a single-track vehicle model, the improved ACC safety distance model is combined with direct yaw moment control to form an ACC system with direct yaw moment control intervention. The system is expressed as: Wherein, the state vector x=[βrΔdΔv a x ] T , control vector u=[F xfl F xfr F xrl F xrr a xdes ] T , interference vector w=[δa p ] T , a p is the acceleration of the preceding vehicle, and the coefficient matrices A, B, and C are: Among them, C f and C r are the total cornering stiffness of the front and rear wheels, respectively.

5. The variable-weight curve stability following control method considering road adhesion conditions as claimed in claim 4, characterized in that: In S3, the ACC system is optimized using a model predictive control method, specifically: Set the cost function: Among them, x ref is the expected value of the state, N p and N c are the prediction time domain and the control time domain respectively, k+i|k represents the k+i moment value predicted at time k, Q(k) and R(k) are the weighted matrices corresponding to the state quantity and control quantity at time k, respectively, corresponding to: Among them, ω Δd 、ω Δv 、 ω r 、ω β as well as are the weight coefficients related to relative distance, relative speed, longitudinal acceleration of the main vehicle, expected acceleration of the main vehicle, yaw rate, sideslip angle of the center of mass, and longitudinal force of the wheel; The difference between the actual and expected values ​​of the vehicle's sideslip angle and yaw rate is used as an evaluation index for the vehicle's lateral stability. The reference values ​​of the yaw rate and sideslip angle are expressed as follows: in, Wheelbase l = a + b, g is the gravitational acceleration constant, vehicle steady-state parameters The constraints corresponding to the cost function are as follows: Desired longitudinal acceleration and desired yaw rate constraints: Driving safety and comfort constraints: Where Δd max and Δv max are the upper limits of distance error and relative speed respectively. The driver’s error sensitivity SDE and SVE to tracking distance and speed are: Among them, k SDE d SDE 、k SVE and d SVE All are constant coefficients.

6. The variable-weight curve stability following control method considering road adhesion conditions according to claim 1, characterized in that: In S4, under low road adhesion conditions, the process of adjusting the degree of direct yaw moment control intervention and longitudinal following performance is as follows: Select the time interval [t0,t0+T d ]Predict the acceleration of the preceding vehicle and fit the longitudinal acceleration of the preceding vehicle within the time interval. The fitting formula is: a p (k)=p0+p1+p2k 2 +…+p n k n (17), Among them, t0 is the current time, T d is the prediction range, n is the order of the polynomial, p i (i=0,1,2,…,n) is the coefficient to be fitted; According to the different acceleration conditions of the preceding vehicle, the weighting matrices Q and R are adjusted according to the preset scheme to adapt to different driving conditions. The weighting matrix is ​​specifically expressed as follows: in, s i (i=1,2,…,8), σ1 and σ2 are all set coefficients.

Images