Traction control method and system for high-low attachment butt joint road surface

Abstract

The invention relates to the technical field of automobile safety control, and discloses a traction control method and system for a high-low attachment butt joint road surface, and the method comprises the following steps: collecting vehicle dynamic information, and then carrying out the filtering processing of the vehicle dynamic information; calculating an initial docking road adhesion coefficient, compensating the initial docking road adhesion coefficient by using the filtered vehicle dynamic information to obtain a compensated road adhesion coefficient, and estimating a reference vehicle speed based on the filtered vehicle dynamic information and the compensated road adhesion coefficient; calculating a driving shaft speed control error according to the reference vehicle speed, and calculating a target driving torque based on the driving shaft speed control error; and calculating a driving wheel speed control deviation according to the reference vehicle speed, and calculating a target braking torque based on the driving wheel speed control deviation. The problem that in the prior art, traction control is difficult to carry out accurately is solved.

Description

Technical Field

[0001] This invention relates to the field of automotive safety control technology, specifically a traction control method and system for high and low adhesion road surfaces. Background Technology

[0002] Traction control system (TCS) is used to prevent wheel slippage when a vehicle accelerates. Slippage typically occurs on slippery surfaces with low coefficient of friction, such as snow, water, and off-road conditions, or when wheels suddenly slip while transitioning from a high-friction surface to a low-friction surface, and the wheels cannot generate enough traction to propel the vehicle.

[0003] When a vehicle travels on a transition surface (from a high-friction surface to a low-friction surface, or vice versa), the vehicle's friction coefficient is estimated. Changes in the friction coefficient identify whether the vehicle is transitioning from high to low friction or vice versa. If transitioning from high to low friction, torque is reduced according to the designed torque gradient to prevent excessive torque and slippage. Conversely, if transitioning from low to high friction, torque is increased according to the designed torque gradient to ensure adequate acceleration.

[0004] However, in the existing technology, due to the problem of traction control accuracy, the following may occur: when moving from high traction to low traction, the torque is not precisely controlled, causing the vehicle to continuously slip after entering the low traction state; or when moving from low traction to high traction, the torque increase is not timely, resulting in insufficient vehicle acceleration, or the vehicle will experience an impact when the torque increase is too large. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, the present invention provides a traction control method and system for high and low adhesion road surfaces, solving the problems of difficulty in accurately performing traction control in the prior art.

[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: A traction control method for high- and low-adhesion road surfaces includes the following steps: The system collects vehicle dynamic information and then filters it. The vehicle dynamic information includes one or more of the following: vehicle speed, wheel speed, lateral and longitudinal acceleration, yaw rate, steering angle, tire force, drive status information, and braking status information. Calculate the initial road surface adhesion coefficient, compensate the initial road surface adhesion coefficient using filtered vehicle dynamic information, and obtain the compensated road surface adhesion coefficient. Estimate the reference vehicle speed based on the filtered vehicle dynamic information and the compensated road surface adhesion coefficient. The drive axle speed control error is calculated based on the reference vehicle speed, and the target drive torque is calculated based on the drive axle speed control error; and the drive wheel speed control deviation is calculated based on the reference vehicle speed, and the target braking torque is calculated based on the drive wheel speed control deviation.

[0007] The beneficial effects of this invention are: This invention uses algorithms to predict relevant information about the vehicle's contact with the road surface, providing early warnings to the traction control system and offering better estimates of drive torque and braking torque, thereby improving the accuracy and performance of traction control on the contact surface. This invention can accurately identify information at the contact point between high and low adhesion surfaces, enabling precise control of the vehicle's drive torque and braking torque on the contact surface, eliminating the problems of continuous slippage when a vehicle with high adhesion enters a low adhesion surface, and poor acceleration and excessive impact when a vehicle with low adhesion enters a high adhesion surface.

[0008] Based on the above technical solution, the present invention can be further improved as follows.

[0009] As a preferred technical solution, the calculation of the initial adhesion coefficient of the road surface includes the following steps: Acquire road surface images; The road surface image is input into the Swin Transformer V2 model and the ConvNeXt V2 model for feature extraction; The extracted features are fused; The initial adhesion coefficient of the docking surface is calculated based on the characteristics after fusion.

[0010] The beneficial effects of adopting the above-mentioned preferred technical solution are: It facilitates the calculation of the initial adhesion coefficient of the road surface.

[0011] As a preferred technical solution, the formula for calculating the initial adhesion coefficient of the road surface is: ; Where μ represents the initial adhesion coefficient of the docking surface, and μ_0 represents the base friction coefficient. T_roughness represents the first weight, and T_roughness represents the road surface roughness feature. This indicates the second weight, and T_contrast represents the texture contrast feature. This indicates the third weight, and M_type indicates the road surface material type. This indicates the fourth weight, and W_level represents the estimated level of slipperiness. This indicates the fifth weight, and Slip_ratio represents the estimated slip ratio.

[0012] The beneficial effects of adopting the above-mentioned preferred technical solution are: It facilitates the calculation of the initial adhesion coefficient of the road surface.

[0013] As a preferred technical solution, the initial road surface adhesion coefficient is compensated using filtered vehicle dynamic information to obtain the compensated road surface adhesion coefficient, including the following steps: A vehicle dynamics model is constructed based on the filtered vehicle dynamics information, and the longitudinal adhesion coefficient and lateral adhesion coefficient are calculated respectively. By combining the longitudinal adhesion coefficient and the transverse adhesion coefficient, the compensated road surface adhesion coefficient is obtained, as shown in the formula: μ_fused=w_long·μ_long+w_lat·μ_lat; Where μ_fused represents the compensated road surface adhesion coefficient, w_long represents the longitudinal weight, μ_long represents the longitudinal adhesion coefficient, w_lat represents the lateral weight, and μ_lat represents the lateral adhesion coefficient.

[0014] The beneficial effects of adopting the above-mentioned preferred technical solution are: This makes it easier to obtain the compensated road surface adhesion coefficient.

[0015] As a preferred technical solution, the formula for calculating the longitudinal adhesion coefficient is: μ_long=(m·a_x+F_aero+F_roll) / (F_z_drive); Where μ_long represents the longitudinal adhesion coefficient, m represents the vehicle mass, a_x represents the longitudinal acceleration, F_aero represents air resistance, F_roll represents rolling resistance, and F_z_drive represents the vertical load on the drive shaft.

[0016] The beneficial effects of adopting the above-mentioned preferred technical solution are: It facilitates the calculation of the longitudinal adhesion coefficient.

[0017] As a preferred technical solution, the formula for calculating the lateral adhesion coefficient is: μ_lat=(v² / R) / [g·cos(θ)-(h_cg / t_w)·(v² / R)·sin(φ)]; Where μ_lat represents the lateral adhesion coefficient, v represents the actual vehicle speed, R represents the turning radius, g represents the gravitational acceleration, θ represents the road slope, h_cg represents the center of gravity height, t_w represents the wheel track, and φ represents the roll angle.

[0018] The beneficial effects of adopting the above-mentioned preferred technical solution are: It facilitates the calculation of the lateral adhesion coefficient.

[0019] As a preferred technical solution, estimating the reference vehicle speed based on filtered vehicle dynamic information and compensated road adhesion coefficient includes the following steps: Construct a system state model; In the system state model, the state variables are: X=[v_x,v_y,γ,ω_fl,ω_fr,ω_rl,ω_rr,κ_fl,κ_fr,κ_rl,κ_rr]^T; Where v_x represents the longitudinal reference vehicle speed, v_y represents the lateral vehicle speed, γ represents the yaw rate, ω_fl represents the angular velocity of the left front wheel, ω_fr represents the angular velocity of the right front wheel, ω_rl represents the angular velocity of the left rear wheel, ω_rr represents the angular velocity of the right rear wheel, κ_fl represents the longitudinal slip ratio of the left front wheel, κ_fr represents the longitudinal slip ratio of the right front wheel, κ_rl represents the longitudinal slip ratio of the left rear wheel, κ_rr represents the longitudinal slip ratio of the right rear wheel, and T represents transpose; The input variable is: u=[T_fl,T_fr,T_rl,T_rr,δ_f]^T; Where T_fl represents the driving torque of the left front wheel, T_fr represents the driving torque of the right front wheel, T_rl represents the driving torque of the left rear wheel, T_rr represents the driving torque of the right rear wheel, and δ_f represents the front wheel steering angle; The output variable is: y=[ω_fl_meas,ω_fr_meas,ω_rl_meas,ω_rr_meas,a_x_meas,a_y_meas,γ_meas]^T; Wherein, ω_fl_meas represents the wheel speed measurement value of the left front wheel, ω_fr_meas represents the wheel speed measurement value of the right front wheel, ω_rl_meas represents the wheel speed measurement value of the left rear wheel, ω_rr_meas represents the wheel speed measurement value of the right rear wheel, a_x_meas represents the longitudinal acceleration measurement value, a_y_meas represents the lateral acceleration measurement value, and γ_meas represents the yaw rate measurement value. The constraints include: F_x,fl=μ_fused·F_z,fl·C_x(κ_fl,α_fl), F_x,fr=μ_fused·F_z,fr·C_x(κ_fr,α_fr), F_x,rl=μ_fused·F_z,rl·C_x(κ_rl,α_rl), F_x,rr=μ_fused·F_z,rr·C_x(κ_rr,α_rr), F_y,fl=μ_fused·F_z,fl·C_y(κ_fl,α_fl), F_y,fr=μ_fused·F_z,fr·C_y(κ_fr,α_fr), F_y,rl=μ_fused·F_z,rl·C_y(κ_rl,α_rl), F_y,rr=μ_fused·F_z,rr·C_y(κ_rr,α_rr); Where F_x,fl represents the longitudinal force of the left front wheel, F_x,fr represents the longitudinal force of the right front wheel, F_x,rl represents the longitudinal force of the left rear wheel, F_x,rr represents the longitudinal force of the right rear wheel, F_y,fl represents the lateral force of the left front wheel, F_y,fr represents the lateral force of the right front wheel, F_y,rl represents the lateral force of the left rear wheel, F_y,rr represents the lateral force of the right rear wheel, F_z,fl represents the vertical load of the left front wheel, F_z,fr represents the vertical load of the right front wheel, F_z,rl represents the vertical load of the left rear wheel, F_z,rr represents the vertical load of the right rear wheel, C_x represents the longitudinal force coefficient function, C_y represents the lateral force coefficient function, α_fl represents the sideslip angle of the left front wheel, α_fr represents the sideslip angle of the right front wheel, α_rl represents the sideslip angle of the left rear wheel, and α_rr represents the sideslip angle of the right rear wheel. The optimal state sequence is calculated using the formula: X*=[x*(k|k),x*(k+1|k),...,x*(k+N-1|k)]; Where X* represents the optimal state sequence, k represents the sampling time number, N represents the total number of sampling times, x*(k|k) represents the optimal estimate of the state at sampling time k, and x*(k+N-1|k) represents the optimal estimate of the state at sampling time k+N-1. The estimated reference vehicle speed is calculated using the formula: v_x,est(k)=[1,0,0,...,0]·x*(k|k) Where v_x,est(k) represents the estimated reference vehicle speed at sampling time k.

[0020] The beneficial effects of adopting the above-mentioned preferred technical solution are: It facilitates the estimation of reference vehicle speed.

[0021] As a preferred technical solution, the formula for calculating the target driving torque is: T_d=T_l+B_d·ω_d+T_fd+J_d·[λ_d·e_d+k_d·sat(s_d / Φ_d)+ε_d·s_d]; Where T_d represents the target driving torque, T_l represents the load torque, B_d represents the driving shaft damping coefficient, ω_d represents the actual speed of the driving shaft, T_fd represents the driving shaft friction torque, J_d represents the equivalent moment of inertia of the driving shaft, λ_d represents the driving control sliding surface parameter, e_d represents the driving shaft speed control error, k_d represents the driving control switching gain, sat represents the saturation function, ε_d represents the driving control exponential approach coefficient, s_d represents the driving control sliding surface function, and Φ_d represents the driving control boundary layer thickness.

[0022] The beneficial effects of adopting the above-mentioned preferred technical solution are: This facilitates the calculation of the target driving torque.

[0023] As a preferred technical solution, the formula for calculating the target braking torque is: T_b=T_l-F_x·R-B_w·ω_w-T_fw+J_w·[λ_w·e_w+k_w·sat(s_w / Φ_w)+ε_w·s_w]; Where T_b represents the target braking torque, F_x represents the tire longitudinal force, B_w represents the wheel rotational damping coefficient, ω_w represents the actual speed of the drive wheel, T_fw represents the wheel bearing friction torque, J_w represents the wheel rotational inertia, λ_w represents the braking control sliding surface parameters, e_w represents the drive wheel speed control error, k_w represents the braking control switching gain, Φ_w represents the braking control boundary layer thickness, ε_w represents the braking control exponential approach coefficient, and s_w represents the braking control sliding surface function.

[0024] The beneficial effects of adopting the above-mentioned preferred technical solution are: This facilitates the calculation of the target braking torque.

[0025] Based on the above technical solutions, the present invention also provides a traction control system for high and low adhesion road surfaces.

[0026] A traction control system for high and low adhesion road surfaces includes a processor, a memory, and a computer program stored in the memory. When the computer program is executed by the processor, it implements the steps of the traction control method for high and low adhesion road surfaces.

[0027] Compared with the prior art, the present invention has the following advantages: This invention uses algorithms to predict relevant information about the vehicle's contact with the road surface, providing early warnings to the traction control system and offering better estimates of drive torque and braking torque, thereby improving the accuracy and performance of traction control on the contact surface. This invention can accurately identify information at the contact point between high and low adhesion surfaces, enabling precise control of the vehicle's drive torque and braking torque on the contact surface, eliminating the problems of continuous slippage when a vehicle with high adhesion enters a low adhesion surface, and poor acceleration and excessive impact when a vehicle with low adhesion enters a high adhesion surface. Attached Figure Description

[0028] Figure 1 This is one of the schematic diagrams for connecting road surfaces; Figure 2 This is the second schematic diagram of the road surface connection; Figure 3 This is a schematic diagram illustrating the steps of a traction control method for high and low adhesion road surfaces. Figure 4 A schematic diagram illustrating the steps of input signal preprocessing; Figure 5 A schematic diagram illustrating the steps for determining the connection point information of high and low adhesion road surfaces; Figure 6 This is a schematic diagram of the traction control process. Detailed Implementation

[0029] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0030] The principles and features of the present invention are described below. The embodiments given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0031] Example 1 like Figures 1 to 6 As shown, the details are as follows: like Figure 1 As shown, this invention mainly includes three modules: 1. An input signal preprocessing module, which performs validity judgment and filtering on the system input signal and related signals (signal messages received by the system from other nodes or the system) (e.g., vehicle speed, wheel speed, lateral and longitudinal acceleration, yaw rate, steering angle, etc.); 2. Information judgment at the junction of high and low adhesion surfaces, which judges relevant information at the junction of high and low adhesion surfaces based on the system input signal and in combination with relevant additional factors (e.g., distance between the wheel and the junction, ground adhesion coefficient at the junction, etc.); 3. A traction control module, which calculates the driving torque when the vehicle enters the junction surface based on the output information of the above modules, thereby achieving stable torque control of the vehicle.

[0032] 1. Input signal preprocessing module. For example... Figure 4As shown, this involves input signal filtering and input signal validity determination.

[0033] First, the input signal (such as linear signals like vehicle speed, wheel speed, lateral and longitudinal acceleration, yaw rate, and steering angle) is filtered. The filtering method can be specified according to requirements. For example, a first-order filter is: Y(n) = aX(n) + (1-a)Y(n-1), where Y(n) is the current filter output value; Y(n-1) is the previous filter output value; a is the filter coefficient; and X(n) is the current filter input signal.

[0034] 2) When the system receives signal messages from other nodes or systems, it assigns values ​​to determine the validity of relevant signals and messages (the values ​​are valid or invalid). Simultaneously, it checks for any hardware or software faults within the system, which also serves as a standard for system availability. In other words, the module's function is only enabled if the input signal is valid and there are no hardware or software faults within the system; otherwise, the function cannot be enabled.

[0035] 2. Information assessment at the junction of high and low adhesion road surfaces. For example... Figure 5 As shown, it includes three parts: 1) a docking point information recognition and early warning module, which uses an onboard front camera to perform image recognition processing, identifies and judges docking points with varying adhesion levels, and calculates docking point information (distance between the wheel and the docking point, initial docking point road surface adhesion coefficient, and docking point warning sign position); 2) docking point information estimation based on vehicle dynamic information, which, based on vehicle dynamics and vehicle dynamic signals (vehicle speed, wheel speed, lateral and longitudinal acceleration, yaw rate, tire force, drive braking status information, etc.), further predicts and estimates the road surface information at the docking point, and performs constraint compensation on the initial docking road surface adhesion coefficient in 1); 3) vehicle reference speed estimation, which, based on the MPC model predictive control algorithm, estimates the vehicle reference speed based on vehicle dynamic signals (vehicle speed, wheel speed, lateral and longitudinal acceleration, tire force, etc.) and the estimated adhesion coefficient.

[0036] 2.1 Dock Joint Information Recognition and Early Warning Module. This module processes image information captured by the vehicle's front-facing camera and identifies the uneven surface adhesion ahead of the vehicle. Image processing algorithms are used to obtain the distance between the wheels and the docking point, the initial adhesion coefficient of the docking surface, and output a docking surface warning marker. This module uses both SwinTransformer V2 and ConvNeXt V2 methods to extract features and designs a multi-task network to simultaneously perform docking surface recognition, distance estimation, and adhesion coefficient estimation. Its implementation principle and process are as follows: The processing flow is roughly as follows: input image → Swin Transformer V2 + ConvNeXt V2 feature extraction → feature fusion → result post-processing → output [road surface type, boundary, distance, adhesion coefficient].

[0037] The relevant processing principles are as follows: 1) Feature extraction network: The input is a single-frame image, and the output is Swing Transformer V2 features and ConvNeXt V2 features.

[0038] Feature extraction principle and mechanism: 2.1) Swin Transformer V2 principle Self-attention formula for shifted windows: Variable meaning: Q,K,V∈R^(M²×d): Query, key, and value matrix (M² is the number of patches in the window); d: Feature dimension; B∈R^(M²×M²): Relative position offset; M∈R^(M²×M²): Shift window mask, enabling cross-window connections.

[0039] Residual connectivity and layer normalization: SwinV2 block formula: Variable meaning: : Output of layer l; LN: Layer normalization; MLP: Multilayer perceptron; W-MSA / SW-MSA: Window / shift window multi-head self-attention.

[0040] 2.2) ConvNeXt V2 Local Texture Feature Extraction Depthwise separable convolution formula: F_convnext=DWConv_{k×k}(GN(GELU(Conv_{1×1}(F_in)))) Variable meaning: DWConv_{k×k}: k×k depthwise separable convolution (k=7); GN: group normalization; GELU: Gaussian error linear unit activation; Conv_{1×1}: 1×1 point convolution with channel blending.

[0041] Feature enhancement formula: g_x=∑_{j=1}^N(x_i^j)² x_i'=x_i / (g_x+α)×γ_i+β_i Variable meaning: x_i: Feature map of the i-th channel; g_x: Global feature response norm; α: Small constant to prevent division by zero; γ_i, β_i: Learnable scaling and offset parameters.

[0042] 3) Feature fusion and interaction module Interaction between Swin Transformer V2 and ConvNeXt V2: Q_conv=F_convnextW_Q^c,K_swin=F_swinW_K^s,V_swin=F_swinW_V^s Q_swin=F_swinW_Q^s,K_conv=F_convnextW_K^c,V_conv=F_convnextW_V^c Variable meaning: W_Q^c, W_K^c, W_V^c: Projection weights of the ConvNeXt branch; W_Q^s, W_K^s, W_V^s: Projection weights of the Swin branch; F_fused^c, F_fused^s: Features after bidirectional fusion.

[0043] Multi-scale features: FPN design: P5=Conv1×1(F_swin^5)+Upsample(Conv1×1(F_convnext^5)) P4=Conv1×1(F_swin^4)+Conv1×1(F_convnext^4)+Upsample(P5) P3=Conv1×1(F_swin^3)+Conv1×1(F_convnext^3)+Upsample(P4) P2=Conv1×1(F_swin^2)+Conv1×1(F_convnext^2)+Upsample(P3) Road surface property analysis: Road surface type segmentation: Y_seg=Softmax(Conv3×3(CA(P2,P3,P4,P5))); Where CA stands for cross-scale attention: CA(Q,K,V)=Softmax(QKᵀ / √d)V.

[0044] Enhanced boundary awareness features: F_boundary=σ(Conv1×1([P2,P3]))⊙P2 Y_boundary=Sigmoid(Conv3×3(F_boundary)) Variable meaning: σ: Sigmoid activation function; ⊙: Element-wise multiplication; [P2,P3]: Feature concatenation; 4) Estimation of the distance between the wheel and the docking point: Projective geometry based on camera model: [u_b,v_b]=argmax(Y_boundary) Inverse perspective transformation formula: X_c=(u_b-c_x)·Z / f_x Y_c=(v_b-c_y)·Z / f_y Z=h / tan(θ+arctan((v_b-v_h) / f_y)) Distance calculation: D = √((X_c - X_w)² + (Y_c - Y_w)²) Variable meaning: [u_b,v_b]: Coordinates of the boundary points of the image coordinate system; [X_c,Y_c,Z]: 3D coordinates in the camera coordinate system; [X_w,Y_w]: The position of the wheel in the camera coordinate system; h: Camera mounting height; θ: Camera pitch angle; f_x, f_y: Camera focal length; c_x, c_y: Coordinates of the principal point of the camera; v_h: The coordinate of the horizon v.

[0045] 5) Estimation of adhesion coefficient: Texture-material-geometry co-modeling: F_texture=GlobalMaxPool(F_convnext) (Texture feature) F_semantic=GlobalAvgPool(F_swin) (Semantic feature) F_geometry=[D,v,a] (Geometric motion characteristics) F_friction=MLP([F_texture;F_semantic;F_geometry]) Adhesion coefficient model based on physical constraints: Variable meaning: μ_0: Basic friction coefficient; T_roughness: Road surface roughness feature (extracted from ConvNeXt); T_contrast: Texture contrast feature; M_type: Road surface material type (classified from SwinV2) W_level: Estimated level of slipperiness; Slip_ratio: Estimated slip ratio; , , , , Learnable physical weights.

[0046] The above process uses image processing to identify whether there is a road surface connection ahead of the vehicle and estimates the distance between the wheels and the connection point, as well as the adhesion coefficient of the road surface. When the distance between the driving wheel and the connection point is less than a specified threshold, a road surface connection recognition warning sign will be activated. Depending on the specific needs, this sign can be used to prompt the driver to brake manually or for the system to intervene automatically. If the driver does not take action, the system will intervene automatically.

[0047] 2.2 Estimating docking information based on vehicle dynamic information. In this module, the estimated μ value in 2.1 is mainly compensated based on the real-time vehicle signal. According to the vehicle dynamics model, the longitudinal adhesion coefficient and the lateral adhesion coefficient are calculated respectively, and then fusion processing is performed.

[0048] 2.2.1 Estimation of adhesion coefficient based on longitudinal dynamics: Driving conditions: The slip ratio is calculated using the relationship between the slip ratio of the drive wheel and the coefficient of adhesion. λ_drive=(ω·rv) / max(ω·r,v)×100% Variable meaning: λ_drive: Drive slip ratio (%); ω: Drive wheel angular velocity (rad / s); r: Effective tire rolling radius (m); v: Vehicle actual speed (m / s), obtained through non-drive wheel speed or GPS.

[0049] Adhesion coefficient estimation: μ_long=(m·a_x+F_aero+F_roll) / (F_z_drive) in: a_x=dv / dt longitudinal acceleration F_aero=0.5·ρ·C_d·A·v² air resistance F_roll = f_roll·m·g·cos(θ) Rolling resistance F_z_drive=(m·g·l_r-m·a_x·h_cg) / L Drive shaft vertical load (rear drive) m: Vehicle mass (kg); ρ: Air density (kg / m³); C_d: Drag coefficient; A: Frontal area (m²); f_roll: Rolling resistance coefficient; θ: Road slope (rad); l_r: Distance from center of gravity to rear axle (m); h_cg: Center of gravity height (m); L: Wheelbase (m).

[0050] Braking conditions: Slip ratio: λ_brake=(v-ω·r) / v×100% Adhesion coefficient based on braking pressure: μ_brake=(F_brake-I_w·dω / dt) / (r·F_z_brake) Where F_brake = k_p·p_brake braking force; I_w: wheel rotational inertia (kg·m²); k_p: braking pressure-force conversion coefficient; p_brake: braking pressure (Pa); F_z_brake: brake axle vertical load.

[0051] 2.2.2 Estimation of adhesion coefficient based on lateral dynamics: Estimation of lateral adhesion coefficient considering steady-state steering conditions: μ_lat=a_y / g=(v² / R) / g Where a_y: lateral acceleration (m / s²); R: turning radius (m); g: gravitational acceleration (9.81m / s²).

[0052] The exact formula considering load transfer is as follows: μ_lat=(v² / R) / [g·cos(θ)-(h_cg / t_w)·(v² / R)·sin(φ)] Where t_w: wheelbase (m); φ: camber angle (rad).

[0053] Adhesion coefficient weighted fusion: μ_fused=w_long·μ_long+w_lat·μ_lat Where w_long is the vertical weight and w_lat is the horizontal weight.

[0054] 2.3 Estimating the vehicle's reference speed: In this module, based on the compensated road adhesion coefficient provided in section 2.2, as well as vehicle status and signals such as vehicle speed, wheel speed, and lateral and longitudinal acceleration, the vehicle reference speed is estimated using the MPC model predictive control algorithm. The calculation process is as follows: 1) System state model: State variable definition: X=[v_x,v_y,γ,ω_fl,ω_fr,ω_rl,ω_rr,κ_fl,κ_fr,κ_rl,κ_rr]^T Where, v_x: longitudinal reference vehicle speed (to be estimated); v_y: lateral vehicle speed; γ: yaw rate; ω_ij: angular velocity of the four wheels; κ_ij: longitudinal slip ratio of the four wheels; T: transpose.

[0055] Input variable: u=[T_fl,T_fr,T_rl,T_rr,δ_f]^T Where T_ij: driving torque of each wheel; δ_f: front wheel steering angle.

[0056] Output variable: y=[ω_fl_meas,ω_fr_meas,ω_rl_meas,ω_rr_meas,a_x_meas,a_y_meas,γ_meas]^T Wherein, ω_ij_meas: wheel speed measurement values ​​of w_fl_meas, w_fr_meas, w_rl_meas, w_rr_meas, a_x_meas: longitudinal acceleration measurement value, a_y_meas: lateral acceleration measurement value, γ_meas: yaw rate measurement value.

[0057] 2) Vehicle dynamics equations: Longitudinal dynamics: m·(a_x-v_y·γ)=Σ(F_x,ij·cosδ_ij-F_y,ij·sinδ_ij) Where m: vehicle mass; a_x: longitudinal acceleration; F_x,ij,F_y,ij: longitudinal and lateral forces of the tire; δ_ij: tire steering angle.

[0058] Lateral dynamics: m·(a_y+v_x·γ)=Σ(F_x,ij·sinδ_ij+F_y,ij·cosδ_ij) Horizontal yaw dynamics: I_z·dγ / dt=ΣM_z,ij Where I_z: vehicle yaw moment of inertia; M_z,ij: yaw moment of each tire about the vehicle's center of gravity.

[0059] Wheel dynamics: I_w·dω_ij / dt=T_ij-F_x,ij·R-T_b,ij Where I_w: wheel moment of inertia; R: effective rolling radius of the wheel; T_b,ij: braking torque.

[0060] Slip rate definition: κ_ij=|ω_ij·R-v_x,ij| / max(|ω_ij·R|,|v_x,ij|,ε) Where, v_x,ij: longitudinal velocity at the center of each tire; ε: a small constant to prevent division by zero.

[0061] Tire center speed: v_x,fl=(v_x-0.5·L_f·γ)·cosδ_f+(v_y+l_f·γ)·sinδ_f v_x,fr=(v_x+0.5·L_f·γ)·cosδ_f+(v_y+l_f·γ)·sinδ_f v_x,rl=v_x-0.5·L_r·γ v_x,rr=v_x+0.5·L_r·γ Where, l_f, l_r: distance from the center of mass to the front and rear axles; L_f, L_r: track width between the front and rear wheels.

[0062] 3) Tire model: Longitudinal force: F_x,ij=μ_fused·F_z,ij·f_x(κ_ij,α_ij) Lateral force: F_y,ij=μ_fused·F_z,ij·f_y(κ_ij,α_ij) Where F_z,ij: vertical load of each tire; α_ij: slip angle of each tire; C_x, C_y: dimensionless tire force coefficient functions, specifically, C_x represents the longitudinal force coefficient function and C_y represents the lateral force coefficient function.

[0063] Considering load transfer, vertical load calculation: F_z,fl=(m·g·l_r) / (2·l)-(m·a_x·h) / (2·l)-(m·a_y·h·l_r) / (L_f·l) F_z,fr=(m·g·l_r) / (2·l)-(m·a_x·h) / (2·l)+(m·a_y·h·l_r) / (L_f·l) F_z,rl=(m·g·l_f) / (2·l)+(m·a_x·h) / (2·l)-(m·a_y·h·l_f) / (L_r·l) F_z,rr=(m·g·l_f) / (2·l)+(m·a_x·h) / (2·l)+(m·a_y·h·l_f) / (L_r·l) Where h is the height of the centroid; l = l_f + l_r is the wheelbase.

[0064] 4) MPC state estimation algorithm: 4.1) Discretized state-space model: Discretize the continuous-time model: x(k+1)=f(x(k),u(k))+w(k) y(k)=h(x(k),u(k))+v(k) Where, k: sampling time number, x(k): system state vector at sampling time k, f(x(k),u(k)): state transition function, y(k): observation vector at sampling time k, h(x(k),u(k)): observation function, describing how the system state and input are mapped to the measurement output, w(k): process noise; v(k): measurement noise.

[0065] Specific format: v_x(k+1)=v_x(k)+T_s·[ΣF_x,ij(k) / m+v_y(k)·γ(k)]+w_vx(k) v_y(k+1)=v_y(k)+T_s·[ΣF_y,ij(k) / m-v_x(k)·γ(k)]+w_vy(k) γ(k+1)=γ(k)+T_s·[ΣM_z,ij(k) / I_z]+w_γ(k) ω_ij(k+1)=ω_ij(k)+T_s·[(T_ij(k)-F_x,ij(k)·R) / I_w]+w_ω,ij(k) Where T_s is the total sampling time, w_vx(k) is the process noise of the longitudinal velocity, and w_vy(k) is the process noise of the transverse velocity.

[0066] 4.2) MPC Optimization Solve the following optimization problem at sampling time k: Min J=Σ_{i=0}^{N-1}{ q_ω·Σ(ω_ij(k+i|k)-ω_ij,meas(k+i))² +q_ax·(a_x(k+i|k)-a_x,meas(k+i))² +q_ay·(a_y(k+i|k)-a_y,meas(k+i))² +q_γ·(γ(k+i|k)-γ_meas(k+i))² +r_T·ΣΔT_ij(k+i|k)²+r_δ·Δδ_f(k+i|k)² } +p_vx·(v_x(k|k)-v_x,prior(k))² Where, min: the function for finding the minimum value, J: the objective function, i: the index of the prediction step (i=0,1...N-1), N: the prediction time domain length (prediction steps), q_ω: wheel speed error weighting coefficient, ω_ij(k+i|k): the predicted rotational speed of the j-th wheel at sampling time k+i, ω_ij,meas(k+i): the actual measured rotational speed of the j-th wheel at sampling time k+i, q_ax: longitudinal acceleration error weighting coefficient, a_x(k+i|k): the predicted longitudinal acceleration, a_x,meas(k+i): the longitudinal acceleration measured by the IMU, q_ay: lateral acceleration error weighting coefficient, a_y(k+i|k): The predicted lateral acceleration, a_y,meas(k+i): lateral acceleration measured by the IMU, q_γ: yaw rate error weighting coefficient, γ(k+i|k): predicted yaw rate, γ_meas(k+i): yaw rate measured by the gyroscope, r_T: driving torque change weighting coefficient, ΔT_ij(k+i|k): change in driving torque, r_δ: front wheel angle change weighting coefficient, Δδ_f(k+i|k): change in front wheel angle, p_vx: longitudinal speed prior error weighting coefficient, v_x(k|k): longitudinal speed estimate at sampling time k, v_x,prior(k): prior estimate of longitudinal speed.

[0067] The constraints include: dynamic constraints, output constraints, state constraints, and tire force constraints.

[0068] 5) Estimation of reference vehicle speed: Based on the above, the optimal state sequence is obtained as: X*=[x*(k|k),x*(k+1|k),...,x*(k+N-1|k)] Reference vehicle speed estimate: v_x,est(k)=[1,0,0,...,0]·x*(k|k) The updated reference speed can be: v_x,est(k)=v_x,pred(k)+K_MPC(k)·[y_meas(k)-y_pred(k)] Where, x*(k|k): the optimal estimate of the state at sampling time k based on information up to and including time k; x*(k|k+i): the predicted value at time k+i based on information up to and including time k; v_x,pred(k): the state prediction based on the previous time; K_MPC(k): the MPC equivalent observer gain matrix (obtained implicitly through optimization); y_meas(k): the actual measured value; y_pred(k): the output predicted value based on the predicted state.

[0069] 3. Traction control module: In this module, based on the sliding mode control algorithm, the driving torque of the vehicle drive axle is estimated by the control deviation of the vehicle drive axle speed, and the braking torque of the vehicle drive wheels is estimated by the control deviation of the wheel speed.

[0070] Axle speed and wheel speed control deviation: based on Figure 6 As shown, the reference vehicle speed calculated in step 2, along with relevant additional factors (acceleration, road adhesion coefficient, etc.), is used to calculate the reference axle speed and reference wheel speed. These are then combined with axle speed deviation and wheel speed deviation to obtain the target axle speed and target wheel speed. Based on the vehicle's current actual axle speed and actual wheel speed, the control deviation is obtained. The target driving torque can then be calculated using the axle speed target deviation, and the target braking torque can be calculated using the wheel speed control deviation. The axle speed deviation and wheel speed deviation can be calibrated using parameters.

[0071] Drive shaft speed control deviation: e_d=ω_d_ref-ω_d Where, e_d: drive shaft speed control error; ω_d_ref: target speed of drive shaft; ω_d: actual speed of drive shaft.

[0072] Drive wheel speed control deviation: e_w=ω_w_ref-ω_w Where, e_w: drive wheel speed control error; ω_w_ref: target speed of drive wheel; ω_w: actual speed of drive wheel.

[0073] 3.1) Drive torque estimation: Sliding surface design: s_d=e_d+λ_d·∫e_d·dt Where, s_d: the sliding surface function of the drive control; λ_d: the sliding surface parameter of the drive control [1 / s] (the weighting coefficient of the integral term of the shaft speed error, which determines the dynamic characteristics of the sliding surface. The larger the value, the stronger the integral effect, λ_d>0); s represents the frequency unit; ∫e_d·dt: the integral term of the drive shaft speed error.

[0074] Sliding surface derivative: ds_d / dt=de_d / dt+λ_d·e_d =-dω_d / dt+λ_d·e_d =-(T_d-T_l-B_d·ω_d-T_fd) / J_d+λ_d·e_d Approach law design: ds_d / dt = -k_d·sgn(s_d) - ε_d·s_d in, k_d: Drive control switching gain [1 / s²] (k_d>0); ε_d: Drive control exponential approach coefficient [1 / s] (ε_d>0); sgn(s_d): Sign function, sgn(s_d)=1ifs_d>0, 0ifs_d=0, -1ifs_d<0.

[0075] The other two derivatives are equal: -(T_d-T_l-B_d·ω_d-T_fd) / J_d+λ_d·e_d=-k_d·sgn(s_d)-ε_d·s_d Using a saturation function, the target driving torque is: T_d=T_l+B_d·ω_d+T_fd+J_d·[λ_d·e_d+k_d·sat(s_d / Φ_d)+ε_d·s_d] Where, sat(s_d / Φ_d): saturation function, sat(x)=x if|x|≤1, sgn(x)if|x|>1; Φ_d: Drive control boundary layer thickness.

[0076] 3.2) Braking torque calculation: Sliding surface design: s_w=e_w+λ_w·∫e_w·dt Where, s_w: the sliding surface function of braking control; λ_w: the sliding surface parameter of braking control [1 / s] (the weighting coefficient of the wheel speed error integral term, which determines the dynamic characteristics of the sliding surface. The larger the value, the stronger the integral effect, λ_w>0); ∫e_w·dt: the integral term of the driving wheel speed error.

[0077] Sliding surface derivative: ds_w / dt=de_w / dt+λ_w·e_w =-dω_w / dt+λ_w·e_w =-(T_l-T_b-F_x·R-B_w·ω_w-T_fw) / J_w+λ_w·e_w Using the exponential reaching law: ds_w / dt=-k_w·sgn(s_w)-ε_w·s_w Where, k_w: braking control switching gain [1 / s²] (k_w>0); ε_w: braking control exponential approach coefficient [1 / s] (ε_w>0).

[0078] Let the two derivatives be equal: -(T_l-T_b-F_x·R-B_w·ω_w-T_fw) / J_w+λ_w·e_w=-k_w·sgn(s_w)-ε_w·s_w Continuous processing is used; target braking torque: T_b=T_l-F_x·R-B_w·ω_w-T_fw+J_w·[λ_w·e_w+k_w·sat(s_w / Φ_w)+ε_w·s_w] Wherein, Φ_w: the thickness of the braking control boundary layer.

[0079] Through the above steps, when the vehicle's drive wheels reach the junction of the high and low adhesion surfaces, the target drive shaft torque and drive wheel braking torque calculated by the system will ensure that the vehicle will not continue to slip or that there will be no impact when accelerating.

[0080] The present invention has the following technical effects: This invention uses algorithms to predict relevant information about the vehicle's contact with the road surface, providing early warnings to the traction control system and offering better estimates of drive and braking torque, thereby improving the accuracy and performance of traction control at the contact surface. This invention can accurately identify information at the contact point between high and low adhesion surfaces, enabling precise control of the vehicle's drive and braking torque at the contact surface. It eliminates the problems of continuous slippage when a vehicle with high adhesion enters a low adhesion surface, and poor acceleration and excessive impact when a vehicle with low adhesion enters a high adhesion surface.

[0081] As described above, the present invention can be implemented well.

[0082] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0083] In the description of this invention, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0084] In the description of this invention, although embodiments of the invention have been shown and described, it is understood that the above embodiments are exemplary and should not be construed as limiting the invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this invention.

[0085] In the description of this invention, all features disclosed in all embodiments of this specification, or steps in all methods or processes implied in the disclosure, may be combined and / or extended or replaced in any way, except for mutually exclusive features and / or steps.

[0086] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Based on the technical essence of the present invention, any simple modifications, equivalent substitutions, and improvements made to the above embodiments within the spirit and principles of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A traction control method for high- and low-adhesion road surfaces, characterized in that, Includes the following steps: The system collects vehicle dynamic information and then filters it. The vehicle dynamic information includes one or more of the following: vehicle speed, wheel speed, lateral and longitudinal acceleration, yaw rate, steering angle, tire force, drive status information, and braking status information. Calculate the initial road surface adhesion coefficient, compensate the initial road surface adhesion coefficient using filtered vehicle dynamic information, and obtain the compensated road surface adhesion coefficient. Estimate the reference vehicle speed based on the filtered vehicle dynamic information and the compensated road surface adhesion coefficient. The drive axle speed control error is calculated based on the reference vehicle speed, and the target drive torque is calculated based on the drive axle speed control error; and the drive wheel speed control deviation is calculated based on the reference vehicle speed, and the target braking torque is calculated based on the drive wheel speed control deviation.

2. The traction control method for high- and low-adhesion road surfaces according to claim 1, characterized in that, The calculation of the initial adhesion coefficient of the road surface includes the following steps: Acquire road surface images; The road surface image is input into the Swin Transformer V2 model and the ConvNeXt V2 model for feature extraction; The extracted features are fused; The initial adhesion coefficient of the docking surface is calculated based on the characteristics after fusion.

3. The traction control method for high- and low-adhesion road surfaces according to claim 2, characterized in that, The formula for calculating the initial adhesion coefficient of the road surface is: ； Where μ represents the initial adhesion coefficient of the docking surface, and μ_0 represents the base friction coefficient. T_roughness represents the first weight, and T_roughness represents the road surface roughness feature. This indicates the second weight, and T_contrast represents the texture contrast feature. This indicates the third weight, and M_type indicates the road surface material type. This indicates the fourth weight, and W_level represents the estimated level of slipperiness. This indicates the fifth weight, and Slip_ratio represents the estimated slip ratio.

4. The traction control method for high- and low-adhesion road surfaces according to claim 1, characterized in that, The initial road surface adhesion coefficient is compensated using filtered vehicle dynamic information to obtain the compensated road surface adhesion coefficient, including the following steps: A vehicle dynamics model is constructed based on the filtered vehicle dynamics information, and the longitudinal adhesion coefficient and lateral adhesion coefficient are calculated respectively. By combining the longitudinal adhesion coefficient and the transverse adhesion coefficient, the compensated road surface adhesion coefficient is obtained, as shown in the formula: μ_fused=w_long·μ_long+w_lat·μ_lat; Where μ_fused represents the compensated road surface adhesion coefficient, w_long represents the longitudinal weight, μ_long represents the longitudinal adhesion coefficient, w_lat represents the lateral weight, and μ_lat represents the lateral adhesion coefficient.

5. A traction control method for high- and low-adhesion road surfaces according to claim 4, characterized in that, The formula for calculating the longitudinal adhesion coefficient is: μ_long=(m·a_x+F_aero+F_roll) / (F_z_drive); Where μ_long represents the longitudinal adhesion coefficient, m represents the vehicle mass, a_x represents the longitudinal acceleration, F_aero represents air resistance, F_roll represents rolling resistance, and F_z_drive represents the vertical load on the drive shaft.

6. A traction control method for high- and low-adhesion road surfaces according to claim 4, characterized in that, The formula for calculating the lateral adhesion coefficient is: μ_lat=(v² / R) / [g·cos(θ)-(h_cg / t_w)·(v² / R)·sin(φ)]; Where μ_lat represents the lateral adhesion coefficient, v represents the actual vehicle speed, R represents the turning radius, g represents the gravitational acceleration, θ represents the road slope, h_cg represents the center of gravity height, t_w represents the wheel track, and φ represents the roll angle.

7. A traction control method for high- and low-adhesion road surfaces according to claim 4, characterized in that, Estimating the reference vehicle speed based on filtered vehicle dynamics information and compensated road adhesion coefficient includes the following steps: Construct a system state model; In the system state model, the state variables are: X=[v_x,v_y,γ,ω_fl,ω_fr,ω_rl,ω_rr,κ_fl,κ_fr,κ_rl,κ_rr]^T; Where v_x represents the longitudinal reference vehicle speed, v_y represents the lateral vehicle speed, γ represents the yaw rate, ω_fl represents the angular velocity of the left front wheel, ω_fr represents the angular velocity of the right front wheel, ω_rl represents the angular velocity of the left rear wheel, ω_rr represents the angular velocity of the right rear wheel, κ_fl represents the longitudinal slip ratio of the left front wheel, κ_fr represents the longitudinal slip ratio of the right front wheel, κ_rl represents the longitudinal slip ratio of the left rear wheel, κ_rr represents the longitudinal slip ratio of the right rear wheel, and T represents transpose; The input variable is: u=[T_fl,T_fr,T_rl,T_rr,δ_f]^T; Where T_fl represents the driving torque of the left front wheel, T_fr represents the driving torque of the right front wheel, T_rl represents the driving torque of the left rear wheel, T_rr represents the driving torque of the right rear wheel, and δ_f represents the front wheel steering angle; The output variable is: y=[ω_fl_meas,ω_fr_meas,ω_rl_meas,ω_rr_meas,a_x_meas,a_y_meas,γ_meas]^T; Wherein, ω_fl_meas represents the wheel speed measurement value of the left front wheel, ω_fr_meas represents the wheel speed measurement value of the right front wheel, ω_rl_meas represents the wheel speed measurement value of the left rear wheel, ω_rr_meas represents the wheel speed measurement value of the right rear wheel, a_x_meas represents the longitudinal acceleration measurement value, a_y_meas represents the lateral acceleration measurement value, and γ_meas represents the yaw rate measurement value. The constraints include: F_x,fl=μ_fused·F_z,fl·C_x(κ_fl,α_fl), F_x,fr=μ_fused·F_z,fr·C_x(κ_fr,α_fr), F_x,rl=μ_fused·F_z,rl·C_x(κ_rl,α_rl), F_x,rr=μ_fused·F_z,rr·C_x(κ_rr,α_rr), F_y,fl=μ_fused·F_z,fl·C_y(κ_fl,α_fl), F_y,fr=μ_fused·F_z,fr·C_y(κ_fr,α_fr), F_y,rl=μ_fused·F_z,rl·C_y(κ_rl,α_rl), F_y,rr=μ_fused·F_z,rr·C_y(κ_rr,α_rr); Where F_x,fl represents the longitudinal force of the left front wheel, F_x,fr represents the longitudinal force of the right front wheel, F_x,rl represents the longitudinal force of the left rear wheel, F_x,rr represents the longitudinal force of the right rear wheel, F_y,fl represents the lateral force of the left front wheel, F_y,fr represents the lateral force of the right front wheel, F_y,rl represents the lateral force of the left rear wheel, F_y,rr represents the lateral force of the right rear wheel, F_z,fl represents the vertical load of the left front wheel, F_z,fr represents the vertical load of the right front wheel, F_z,rl represents the vertical load of the left rear wheel, F_z,rr represents the vertical load of the right rear wheel, C_x represents the longitudinal force coefficient function, C_y represents the lateral force coefficient function, α_fl represents the sideslip angle of the left front wheel, α_fr represents the sideslip angle of the right front wheel, α_rl represents the sideslip angle of the left rear wheel, and α_rr represents the sideslip angle of the right rear wheel. The optimal state sequence is calculated using the formula: X*=[x*(k|k),x*(k+1|k),...,x*(k+N-1|k)]; Where X* represents the optimal state sequence, k represents the sampling time number, N represents the total number of sampling times, x*(k|k) represents the optimal estimate of the state at sampling time k, and x*(k+N-1|k) represents the optimal estimate of the state at sampling time k+N-1. The estimated reference vehicle speed is calculated using the formula: v_x,est(k)=[1,0,0,...,0]·x*(k|k) Where v_x,est(k) represents the estimated reference vehicle speed at sampling time k.

8. A traction control method for high- and low-adhesion road surfaces according to claim 1, characterized in that, The formula for calculating the target driving torque is: T_d=T_l+B_d·ω_d+T_fd+J_d·[λ_d·e_d+k_d·sat(s_d / Φ_d)+ε_d·s_d]; Where T_d represents the target driving torque, T_l represents the load torque, B_d represents the driving shaft damping coefficient, ω_d represents the actual speed of the driving shaft, T_fd represents the driving shaft friction torque, J_d represents the equivalent moment of inertia of the driving shaft, λ_d represents the driving control sliding surface parameter, e_d represents the driving shaft speed control error, k_d represents the driving control switching gain, sat represents the saturation function, ε_d represents the driving control exponential approach coefficient, s_d represents the driving control sliding surface function, and Φ_d represents the driving control boundary layer thickness.

9. A traction control method for high- and low-adhesion road surfaces according to any one of claims 1 to 8, characterized in that, The formula for calculating the target braking torque is: T_b=T_l-F_x·R-B_w·ω_w-T_fw+J_w·[λ_w·e_w+k_w·sat(s_w / Φ_w)+ε_w·s_w]; Where T_b represents the target braking torque, F_x represents the tire longitudinal force, B_w represents the wheel rotational damping coefficient, ω_w represents the actual speed of the drive wheel, T_fw represents the wheel bearing friction torque, J_w represents the wheel rotational inertia, λ_w represents the braking control sliding surface parameters, e_w represents the drive wheel speed control error, k_w represents the braking control switching gain, Φ_w represents the braking control boundary layer thickness, ε_w represents the braking control exponential approach coefficient, and s_w represents the braking control sliding surface function.

10. A traction control system for high- and low-adhesion road surfaces, characterized in that, The device includes a processor, a memory, and a computer program stored in the memory. When the computer program is executed by the processor, it implements the steps of the traction control method for high and low adhesion road surfaces as described in any one of claims 1 to 9.

Images