Intelligent electric vehicle stability control method based on stability probability spectrum

Abstract

The invention discloses an intelligent electric vehicle stability control method based on a stability probability spectrum. A data acquisition module, a density-based clustering algorithm module, a probability spectrum generation module based on a convolutional neural network, a self-adaptive model prediction control module and an actuator cooperative distribution control module are included. According to the invention, the data acquisition module is used for acquiring data to calculate the lateral load transfer rate, and the density-based clustering algorithm is started to classify the stability state of the vehicle; and then, generating a stability probability spectrum through a convolutional neural network, dynamically adjusting a weight factor in a minimum cost function in a control target and model prediction control based on the stability probability spectrum, and coordinating response priorities of a plurality of actuators by judging the stability state of the vehicle so as to enhance the transverse and longitudinal coupling control capability. The method can achieve the dynamic recognition and hierarchical control of the stability state of the vehicle, and effectively improves the operation stability and control robustness of the whole vehicle under a complex working condition.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent electric vehicle control, and particularly relates to a multi-actuator coordinated control strategy combining stability probability spectrum analysis, and is particularly applicable to a stability enhancement control system for a four-wheel independent drive intelligent electric vehicle. Background Art

[0002] With the rapid development of intelligent electric vehicles, their driving stability has become a key indicator of safety performance. Traditional vehicle stability control methods mostly rely on a single controller, such as ABS or ESC, and it is difficult to effectively handle the non-linear dynamic response of vehicles under complex working conditions, especially in the extreme handling area. In recent years, intelligent control methods that integrate artificial intelligence and model predictive control have received attention, but there is still a lack of an early prediction and response mechanism for the instability trend of vehicle states, resulting in a lag in the response of control strategies and low execution efficiency. Therefore, there is an urgent need for a stability control system that can dynamically sense the change trend of vehicle stability states and coordinate the common action of each control actuator. Summary of the Invention

[0003] The purpose of the present invention is to provide a stability control method for intelligent electric vehicles based on a stability probability spectrum, to achieve dynamic identification and hierarchical control of vehicle stability states, and effectively improve the handling stability performance and control robustness of the whole vehicle under complex working conditions.

[0004] To achieve the above purpose, the present invention provides the following solutions:

[0005] A stability control method for intelligent electric vehicles based on a stability probability spectrum, including a data acquisition module, a density-based clustering algorithm module, a probability spectrum generation module based on a convolutional neural network, an adaptive model predictive control module, and an actuator collaborative allocation control module;

[0006] The data acquisition module is used to collect vehicle state data; the vehicle state data includes longitudinal speed u c , yaw rate r, roll angle of the center of mass lateral acceleration a y and the vertical loads of the four wheels; calculate the lateral load transfer rate according to the vertical loads of the four wheels, and the calculation formula is:

[0007]

[0008] In the formula, L TR is the lateral load transfer rate, F ZL1 is the vertical load of the left front wheel, F ZR1 is the vertical load of the right front wheel, F ZL2 is the vertical load of the left rear wheel, F ZR2 is the vertical load of the right rear wheel, h cis the vehicle's center of mass height, a y is the lateral acceleration, k s is the suspension stiffness, is the angular velocity of the center of mass roll, c s is the suspension damping coefficient, is the angular acceleration of the center of mass roll, F d is the aerodynamic downforce, t w is the vehicle's track width, λ1 is the influence factor of lateral acceleration and center of mass height, λ2 is the influence factor of suspension stiffness, λ3 is the influence factor of suspension damping, λ4 is the influence factor of aerodynamics, λ5 is the track width correction factor;

[0009] The calculation formula for the influence factor λ1 of lateral acceleration and center of mass height is:

[0010]

[0011] In the formula, m is the vehicle's total mass, h c is the vehicle's center of mass height, t w is the vehicle's track width;

[0012] The calculation formula for the influence factor λ2 of suspension stiffness is:

[0013]

[0014] In the formula, K φ is the equivalent roll stiffness of the suspension, t w is the vehicle's track width;

[0015] The calculation formula for the influence factor λ3 of suspension damping is:

[0016]

[0017] In the formula, c φ is the equivalent roll damping of the suspension, t w is the vehicle's track width;

[0018] The calculation formula for the influence factor λ4 of aerodynamics is:

[0019]

[0020] In the formula, ρ is the air density, A is the vehicle's frontal area, C L is the lift coefficient, u c is the longitudinal velocity;

[0021] The calculation formula for the track width correction factor λ5 is:

[0022]

[0023] In the formula, l ris the distance from the rear axle to the center of mass, and L is the wheelbase.

[0024] The density-based clustering algorithm module is used to cluster the vehicle state data to generate four types of stability state labels, namely stable state, trend stable state, trend unstable state, and unstable state, including:

[0025] Calculate the total weight W of the vehicle state feature point path according to the following formula:

[0026]

[0027] In the formula, p and q are respectively vehicle state feature points, D(p, q) is the original distance between vehicle state feature points, S is the non-linear strength control factor, D hybrid (p, q) is the mixed distance between vehicle state feature points, n is the number of vehicle state feature points, ω is the attenuation coefficient, D hybrid (p i , p i+1 ) is the mixed distance between adjacent vehicle state feature points;

[0028] Judge the stability state according to the following rules:

[0029]

[0030] In the formula, θ k is the k-th type of stability state, θ1 is the probability spectral density of the stable state, θ2 is the probability spectral density of the trend stable state, θ3 is the probability spectral density of the trend unstable state, θ4 is the probability spectral density of the unstable state, W is the total weight of the vehicle state feature point path, and ψ is the reachability threshold.

[0031] The fully connected layer of the probability spectrum generation module based on the convolutional neural network is a 4-layer fully connected network, and the output layer generates four types of stability probability spectral densities θ k according to the following formula to calculate the output of the fully connected layer:

[0032]

[0033] In the formula, Z k is the original output value of the k-th type of node in the fully connected layer, is the weight matrix of the k-th type in the fourth layer of the fully connected layer, h (f3) is the output activation value of the third layer of the fully connected layer, b k is the bias term of the k-th type;

[0034] Calculate the stability probability spectral density θ of the output layer according to the following formula k :

[0035]

[0036] In the formula, Z k is the original output value of the k-th type of node in the fully connected layer, η is the exponential gain factor, τ is the temperature coefficient, θ1 is the probability spectral density of the stable state, θ2 is the probability spectral density of the trend stable state, θ3 is the probability spectral density of the trend unstable state, and θ4 is the probability spectral density of the unstable state.

[0037] The adjustment of the weight factor in the adaptive model predictive control module includes:

[0038] Calculate the centroid roll angle weight factor according to the following formula

[0039]

[0040] In the formula, k c is the amplification factor, α is the nonlinear gain factor, s1 is the weight coefficient, and θ3 is the probability spectral density of the trend unstable state;

[0041] Calculate the lateral velocity weight factor according to the following formula

[0042]

[0043] In the formula, k c is the amplification factor, σ is the nonlinear gain factor, s2 is the weight coefficient, ε is the nonlinear gain factor, and θ3 is the probability spectral density of the trend unstable state;

[0044] Calculate the yaw rate weight factor Γ according to the following formula r :

[0045]

[0046] In the formula, k c is the amplification factor, γ is the nonlinear gain factor, s3 is the weight coefficient, and θ4 is the probability spectral density of the unstable state.

[0047] The objective function in the adaptive model predictive control module is:

[0048]

[0049] In the formula, v y,ref is the ideal lateral velocity reference value, r ref is the ideal yaw rate reference value, v y (k) is the lateral velocity at the k-th moment, r(k) is the yaw rate at the k-th moment, is the centroid roll angle at the k-th moment, m is the vehicle mass, and l f is the distance from the front axle to the centroid, L is the wheelbase, and K yfis the cornering stiffness of the front wheels, u c is the longitudinal speed, μ is the road surface adhesion coefficient, σ f is the front wheel steering angle, g is the acceleration due to gravity, F is the stability factor, N p is the prediction time domain step size, Γ is the weight matrix, ΔU(k) is the control input increment at the k-th moment, M is the control input weight matrix;

[0050] The stability factor F is calculated according to the following formula:

[0051]

[0052] In the formula, m is the vehicle mass, l f is the distance from the front axle to the center of mass, l r is the distance from the rear axle to the center of mass, L is the wheelbase, K yf is the cornering stiffness of the front wheels, K yr is the cornering stiffness of the rear wheels;

[0053] The weight matrix Γ is calculated according to the following formula:

[0054]

[0055] In the formula, is the roll angle weight factor of the center of mass, is the lateral speed weight factor, Γ r is the yaw rate weight factor.

[0056] The actuator collaborative distribution control module is used to dynamically distribute the execution priorities of the active front-wheel steering control system, the direct yaw moment control system, and the active suspension control system to achieve longitudinal, lateral, and vertical collaborative control, including:

[0057] The front wheel steering angle δ of the active front-wheel steering control system is calculated according to the following formula f :

[0058]

[0059] In the formula, δ f0 is the basic front wheel steering angle, ξ is the instability state suppression coefficient, θ4 is the probability spectral density of the instability state, k β is the cornering angle compensation gain, β is the actual cornering angle, β max is the maximum allowable cornering angle, b is the numerical stability constant, R is the trend instability state response coefficient, θ3 is the probability spectral density of the trend instability state;

[0060] The yaw moment M of the direct yaw moment control system is calculated according to the following formula z :

[0061]

[0062] Wherein, M z0 is the basic yaw moment, is the instability probability gain coefficient, k r is the yaw rate error gain, r ref is the ideal yaw rate, r is the yaw rate, T is the torque amplification coefficient, and θ4 is the probability spectral density of the instability state;

[0063] The suspension force ΔF of the active suspension control system is calculated according to the following formula z :

[0064]

[0065] Wherein, ΔF z0 is the basic suspension force, is the roll angle of the center of mass, the maximum allowable roll angle of the center of mass, N is the nonlinear index, Q is the trend stability state decay coefficient, and θ2 is the probability spectral density of the trend stability state, is the roll damping gain, is the roll angular velocity, α is the sensitivity coefficient of the trend instability state, and θ3 is the probability spectral density of the trend instability state.

[0066] The beneficial effects of the present invention are as follows:

[0067] The present invention provides an intelligent electric vehicle stability control method based on a stability probability spectrum, which realizes the early prediction and identification of the instability trend through a data acquisition module, a density-based clustering algorithm module, a probability spectrum generation module based on a convolutional neural network, an adaptive model predictive control module, and an actuator cooperative allocation control module; dynamically adjusts the control target and weight based on the probability spectrum to achieve optimal control; coordinates the response priorities of multiple actuators, enhances the longitudinal and lateral coupling control ability, realizes the dynamic identification and hierarchical control of the vehicle stability state; and effectively improves the handling stability performance and control robustness of the whole vehicle under complex working conditions. Description of the Drawings

[0068] Figure 1 is a framework diagram of an intelligent electric vehicle stability control method based on a stability probability spectrum proposed by the present invention. Detailed Embodiment

[0069] Referring to Figure 1 , the intelligent electric vehicle stability control method based on a stability probability spectrum described in the present invention includes a data acquisition module, a density-based clustering algorithm module, a probability spectrum generation module based on a convolutional neural network, an adaptive model predictive control module, and an actuator cooperative allocation control module;

[0070] The data acquisition module is used to acquire vehicle state data; the vehicle state data includes longitudinal speed u c , yaw rate r, roll angle of the center of mass lateral acceleration a y and the vertical loads of the four wheels; calculate the lateral load transfer ratio according to the vertical loads of the four wheels, and the calculation formula is:

[0071]

[0072] In the formula, L TR is the lateral load transfer ratio, F ZL1 is the vertical load of the left front wheel, F ZR1 is the vertical load of the right front wheel, F ZL2 is the vertical load of the left rear wheel, F ZR2 is the vertical load of the right rear wheel, h c is the height of the vehicle's center of mass, a y is the lateral acceleration, k s is the suspension stiffness, is the roll angular velocity of the center of mass, c s is the suspension damping coefficient, is the roll angular acceleration of the center of mass, F d is the aerodynamic downforce, t w is the vehicle track, λ1 is the influence factor of lateral acceleration and center of mass height, λ2 is the influence factor of suspension stiffness, λ3 is the influence factor of suspension damping, λ4 is the influence factor of aerodynamics, and λ5 is the track correction factor;

[0073] The calculation formula of the influence factor λ1 of lateral acceleration and center of mass height is:

[0074]

[0075] In the formula, m is the total vehicle mass, h c is the height of the vehicle's center of mass, t w is the vehicle track;

[0076] The calculation formula of the influence factor λ2 of suspension stiffness is:

[0077]

[0078] In the formula, K φ is the equivalent roll stiffness of the suspension, t w is the vehicle track;

[0079] The calculation formula of the influence factor λ3 of suspension damping is:

[0080]

[0081] In the formula, c φis the equivalent roll damping of the suspension, t w is the vehicle track width;

[0082] The calculation formula for the aerodynamic influence factor λ4 is:

[0083]

[0084] In the formula, ρ is the air density, A is the vehicle frontal area, C L is the lift coefficient, u c is the longitudinal speed;

[0085] The calculation formula for the track width correction factor λ5 is:

[0086]

[0087] In the formula, l r is the distance from the rear axle to the center of mass, and L is the wheelbase.

[0088] The density-based clustering algorithm module is used to cluster the vehicle state data to generate four types of stability state labels, namely stable state, trend stable state, trend unstable state, and unstable state, including:

[0089] Calculate the total weight W of the vehicle state feature point path according to the following formula:

[0090]

[0091] In the formula, p and q are respectively vehicle state feature points, D(p,q) is the original distance between vehicle state feature points, S is the nonlinear strength control factor, D hybrid (p,q) is the mixed distance between vehicle state feature points, n is the number of vehicle state feature points, ω is the attenuation coefficient, D hybrid (p i ,p i+1 ) is the mixed distance between adjacent vehicle state feature points;

[0092] Judge the stability state according to the following rules:

[0093]

[0094] In the formula, θ k is the kth type of stability state, θ1 is the probability spectral density of the stable state, θ2 is the probability spectral density of the trend stable state, θ3 is the probability spectral density of the trend unstable state, θ4 is the probability spectral density of the unstable state, W is the total weight of the vehicle state feature point path, and ψ is the reachability threshold.

[0095] The fully connected layer of the probability spectrum generation module based on the convolutional neural network is a four-layer fully connected network, and the output layer generates four types of stability probability spectrum densities θ by the Softmax function. k , calculate the output of the fully connected layer according to the following formula:

[0096]

[0097] In the formula, Z k is the original output value of the k-th type of node in the fully connected layer, is the weight matrix of the k-th type in the fourth fully connected layer, h (f3) is the output activation value of the third fully connected layer, b k is the bias term of the k-th type;

[0098] Calculate the stability probability spectrum density θ of the output layer according to the following formula k :

[0099]

[0100] In the formula, Z k is the original output value of the k-th type of node in the fully connected layer, η is the exponential gain factor, τ is the temperature coefficient, θ1 is the stable state probability spectrum density, θ2 is the trend stable state probability spectrum density, θ3 is the trend instability state probability spectrum density, and θ4 is the instability state probability spectrum density.

[0101] The adjustment of the weight factor in the adaptive model predictive control module includes:

[0102] Calculate the centroid roll angle weight factor according to the following formula

[0103]

[0104] In the formula, k c is the amplification factor, α is the nonlinear gain factor, s1 is the weight coefficient, and θ3 is the trend instability probability spectrum density;

[0105] Calculate the lateral velocity weight factor according to the following formula

[0106]

[0107] In the formula, k c is the amplification factor, σ is the nonlinear gain factor, s2 is the weight coefficient, ε is the nonlinear gain factor, and θ3 is the trend instability probability spectrum density;

[0108] Calculate the yaw rate weight factor Γ according to the following formula r :

[0109]

[0110] In the formula, k c is the amplification factor, γ is the nonlinear gain factor, s3 is the weight coefficient, and θ4 is the instability probability spectral density.

[0111] The objective function in the adaptive model predictive control module is as follows:

[0112]

[0113] In the formula, v y,ref is the ideal lateral velocity reference value, r ref is the ideal yaw rate reference value, v y (k) is the lateral velocity at the k-th moment, r(k) is the yaw rate at the k-th moment, is the sideslip angle of the center of mass at the k-th moment, m is the vehicle mass, l f is the distance from the front axle to the center of mass, L is the wheelbase, K yf is the front wheel cornering stiffness, u c is the longitudinal velocity, μ is the road surface adhesion coefficient, σ f is the front wheel steering angle, g is the gravitational acceleration, F is the stability factor, N p is the prediction time domain step, Γ is the weight matrix, ΔU(k) is the control input increment at the k-th moment, and M is the control input weight matrix;

[0114] Calculate the stability factor F according to the following formula:

[0115]

[0116] In the formula, m is the vehicle mass, l f is the distance from the front axle to the center of mass, l r is the distance from the rear axle to the center of mass, L is the wheelbase, K yf is the front wheel cornering stiffness, K yr is the rear wheel cornering stiffness;

[0117] Calculate the weight matrix Γ according to the following formula:

[0118]

[0119] In the formula, is the sideslip angle of the center of mass weight factor, is the lateral velocity weight factor, Γ r is the yaw rate weight factor.

[0120] The actuator cooperative allocation control module is used to dynamically allocate the execution priorities of the active front wheel steering control system, the direct yaw moment control system, and the active suspension control system to achieve longitudinal, lateral, and vertical cooperative control, including:

[0121] When the vehicle stability state is in a tendency of instability, the strategy of the actuator collaborative allocation control module is to execute the active front-wheel steering control system and the active suspension control system, and calculate the front-wheel steering angle δ of the active front-wheel steering control system according to the following formula f :

[0122]

[0123] In the formula, δ f0 is the basic front-wheel steering angle, ξ is the instability state suppression coefficient, θ4 is the probability spectral density of the instability state, k β is the side slip angle compensation gain, β is the actual side slip angle, β max is the maximum allowable side slip angle, b is the numerical stability constant, R is the response coefficient of the tendency of instability state, θ3 is the probability spectral density of the tendency of instability state;

[0124] Calculate the suspension force ΔF of the active suspension control system according to the following formula z :

[0125]

[0126] In the formula, ΔF z0 is the basic suspension force, is the roll angle of the center of mass, the maximum allowable roll angle of the center of mass, N is the non-linear index, Q is the attenuation coefficient of the tendency stable state, θ2 is the probability spectral density of the tendency stable state, is the roll damping gain, is the roll angular velocity, α is the sensitivity coefficient of the tendency of instability state, θ3 is the probability spectral density of the tendency of instability state.

[0127] When the vehicle stability state is in an unstable state, the strategy of the actuator collaborative allocation control module is to execute the active front-wheel steering control system, the direct yaw moment control system, and the active suspension control system, and calculate the yaw moment M of the direct yaw moment control system according to the following formula z :

[0128]

[0129] In the formula, M z0 is the basic yaw moment, is the instability probability gain coefficient, k r is the yaw angular velocity error gain, r ref is the ideal yaw angular velocity, r is the yaw angular velocity, T is the torque amplification coefficient, θ4 is the probability spectral density of the instability state.

Claims

1. An intelligent electric vehicle stability control method based on the stability probability spectrum, characterized in that, This method is applicable to intelligent electric vehicles with four-wheel independent drive, and includes a data acquisition module, a density-based clustering algorithm module, a probability spectrum generation module based on a convolutional neural network, an adaptive model predictive control module, and an actuator collaborative allocation control module; The data acquisition module is used to acquire vehicle state data; the vehicle state data includes longitudinal speed u c , yaw rate r, roll angle of the center of mass lateral acceleration a y and the vertical loads of the four wheels; calculate the lateral load transfer ratio according to the vertical loads of the four wheels, and the calculation formula is: Where, L TR is the lateral load transfer ratio, F ZL1 is the vertical load of the left front wheel, F ZR1 is the vertical load of the right front wheel, F ZL2 is the vertical load of the left rear wheel, F ZR2 is the vertical load of the right rear wheel, h c is the height of the vehicle's center of mass, a y is the lateral acceleration, k s is the suspension stiffness, is the angular velocity of the center of mass roll, c s is the suspension damping coefficient, is the angular acceleration of the center of mass roll, F d is the aerodynamic downforce, t w is the vehicle track width, λ1 is the influence factor of lateral acceleration and the height of the center of mass, λ2 is the influence factor of suspension stiffness, λ3 is the influence factor of suspension damping, λ4 is the influence factor of aerodynamics, λ5 is the track width correction factor; Calculate the lateral acceleration and the centroid height influence factor λ1 according to the following formula: where m is the vehicle mass and h c is the height of the vehicle's center of mass, and t w is the vehicle track width; Calculate the suspension stiffness influence factor λ2 according to the following formula: where K φ is the equivalent roll stiffness of the suspension, and t w is the wheelbase of the vehicle; Calculate the suspension damping influence factor λ3 according to the following formula: where c φ is the equivalent roll damping of the suspension, and t w is the track width of the vehicle; Calculate the aerodynamic influence factor λ4 according to the following formula: where ρ is the air density, A is the frontal area of the vehicle, C L is the lift coefficient, and u c is the longitudinal velocity; Calculate the track correction factor λ5 according to the following formula: where \(l\) r is the distance from the rear axle to the center of mass, and \(L\) is the wheelbase.

2. The intelligent electric vehicle stability control method based on the stability probability spectrum according to claim 1, characterized in that, The density-based clustering algorithm module is used to cluster the vehicle state data to generate four types of stability state labels, namely stable state, trend stable state, trend unstable state, and unstable state, including: Calculate the total weight W of the vehicle state characteristic point path according to the following formula: Wherein, p and q are respectively vehicle state feature points, D(p, q) is the original distance between vehicle state feature points, S is a non-linear intensity control factor, and D hybrid (p, q) is the mixed distance between vehicle state feature points, n is the number of vehicle state feature points, ω is an attenuation coefficient, and D hybrid (p i , p i+1 ) is the mixed distance between adjacent vehicle state feature points; Judge the stability state according to the following rules: where θ k is the probability spectral density of the k-th type of stability state, θ1 is the probability spectral density of the stable state, θ2 is the probability spectral density of the trend-stable state, θ3 is the probability spectral density of the trend-unstable state, θ4 is the probability spectral density of the unstable state, W is the total weight of the vehicle state characteristic point path, and ψ is the reachability threshold.

3. The intelligent electric vehicle stability control method based on the stability probability spectrum according to claim 1, characterized in that, The fully connected layer of the probability spectrum generation module based on the convolutional neural network is a four-layer fully connected network, and the output layer generates four types of stability probability spectral densities θ by the Softmax function k , and the output of the fully connected layer is calculated according to the following formula: Wherein, Z k is the original output value of the k-th type of node in the fully connected layer, is the weight matrix of the k-th type in the fourth fully connected layer, h (f3) is the output activation value of the third fully connected layer, b k is the bias term of the k-th type; Calculate the stability probability spectral density θ of the output layer according to the following formula k :[[]]END]] where Z k is the original output value of the k-th node in the fully connected layer, η is the exponential gain factor, τ is the temperature coefficient, θ1 is the probability spectral density of the stable state, θ2 is the probability spectral density of the trend stable state, θ3 is the probability spectral density of the trend unstable state, and θ4 is the probability spectral density of the unstable state.

4. An intelligent electric vehicle stability control method based on a stability probability spectrum according to claim 1, characterized in that, The adjustment of the weight factor in the adaptive model predictive control module includes: Calculate the centroid roll angle weighting factor according to the following formula where k c is the amplification factor, α is the nonlinear gain factor, s1 is the weight coefficient, and θ3 is the spectral density of the trend instability probability; Calculate the lateral velocity weight factor according to the following formula where k c is the amplification factor, σ is the nonlinear gain factor, s2 is the weight coefficient, ε is the nonlinear gain factor, and θ3 is the probability spectral density of trend instability; Calculate the yaw rate weighting factor Γ according to the following formula r :[[]]END]] where k c is the amplification factor, γ is the nonlinear gain factor, s3 is the weight coefficient, and θ4 is the probability spectral density of instability.

5. An intelligent electric vehicle stability control method based on a stability probability spectrum according to claim 1, characterized in that, The objective function in the adaptive model predictive control module is: where, v y,ref is the ideal lateral velocity reference value, r ref is the ideal yaw rate reference value, v y (k) is the lateral velocity at the k-th moment, r(k) is the yaw rate at the k-th moment, is the sideslip angle of the center of mass at the k-th moment, m is the vehicle mass, l f is the distance from the front axle to the center of mass, L is the wheelbase, K yf is the cornering stiffness of the front wheels, u c is the longitudinal velocity, μ is the road adhesion coefficient, σ f is the front wheel steering angle, g is the gravitational acceleration, F is the stability factor, N p is the prediction time domain step size, Γ is the weight matrix, ΔU(k) is the control input increment at the k-th moment, M is the control input weight matrix; Calculate the stability factor F according to the following formula: where m is the vehicle mass, l f is the distance from the front axle to the center of mass, l r is the distance from the rear axle to the center of mass, L is the wheelbase, K yf is the cornering stiffness of the front wheels, K yr is the cornering stiffness of the rear wheels; Calculate the weight matrix Γ according to the following formula: In the formula, is the weight factor of the roll angle of the center of mass, is the weight factor of the lateral velocity, and Γ r is the weight factor of the yaw rate.

6. The intelligent electric vehicle stability control method based on the stability probability spectrum according to claim 1, wherein The actuator collaborative allocation control module is used to dynamically allocate the execution priorities of the active front-wheel steering control system, the direct yaw moment control system, and the active suspension control system to achieve longitudinal, lateral, and vertical collaborative control, including: When the vehicle stability state is in a trend of instability, the strategy of the actuator collaborative distribution control module is to execute the active front-wheel steering control system and the active suspension control system, and calculate the front-wheel steering angle δ of the active front-wheel steering control system according to the following formula f : where δ f0 is the basic front wheel steering angle, ξ is the instability state suppression coefficient, θ4 is the probability spectral density of the instability state, k β is the cornering angle compensation gain, β is the actual cornering angle, β max is the maximum allowable cornering angle, b is the numerical stability constant, R is the response coefficient of the trend instability state, and θ3 is the probability spectral density of the trend instability state; Calculate the suspension force ΔF of the active suspension control system according to the following formula z : where, ΔF z0 is the basic suspension force, is the roll angle of the center of mass, the maximum allowable roll angle of the center of mass, N is the nonlinear index, Q is the attenuation coefficient of the trend stable state, θ2 is the probability spectral density of the trend stable state, is the roll damping gain, is the roll angular velocity, α is the sensitivity coefficient of the trend unstable state, θ3 is the probability spectral density of the trend unstable state; When the vehicle stability state is in an unstable state, the strategy of the actuator collaborative distribution control module is to execute the active front wheel steering control system, the direct yaw moment control system, and the active suspension control system, and calculate the yaw moment M of the direct yaw moment control system according to the following formula z : Where, M z0 is the basic yaw moment, is the instability probability gain coefficient, k r is the yaw rate error gain, r ref is the ideal yaw rate, r is the yaw rate, T is the torque amplification coefficient, and θ4 is the probability spectral density of the instability state.

Images